Philosophy and Climate Change (Engaging Philosophy) 9780198796282, 0198796285

Table of contents :
Cover
Philosophy and Climate Change
Copyright
Contents
Acknowledgments
List of Figures
List of Tables
List of Contributors
Abstracts of Chapters
Section I. Valuing Climate Change Impacts
1 A Convenient Truth? Climate Change and Quality of Life
2 Animals and Climate Change
3 Discounting under Risk: Utilitarianism vs. Prioritarianism
4 A Philosopher’s Guide to Discounting
5 Does Climate Change Policy Depend Importantly on Population Ethics? Deflationary Responses to the Challenges of Population Ethics for Public Policy
Section II. Cognition, Emotions, and Climate Change
7 The Wages of Fear? Toward Fearing Well About Climate Change
8 Climate Change and Cultural Cognition
Section III. Climate Change and Individual Ethics
9 Climate Change and Individual Obligations: A Dilemma for the Expected Utility Approach, and the Need for an Imperfect View
10 The Puzzle of Inefficacy
11 On Individual and Shared Obligations: In Defense of the Activist’s Perspective
12 How Much Harm Does Each of Us Do?
Section IV. Climate Change and Politics
13 How Quickly Should the World Reduce its Greenhouse Gas Emissions? Climate Change and the Structure of Intergenerational Justice
14 Political Realism, Feasibility Wedges, and Opportunities for Collective Action on Climate Change
15 Pareto Improvements and Feasible Climate Solutions
16 Climate Change, Liberalism, and the Public/Private Distinction
Introduction
Section I: Valuing Climate Change Impacts
Chapter 1: A Convenient Truth?: Climate Change and Quality of Life
1. Introduction
2. The Nature and Measure of Subjective Well-Being
3. Subjective Well-Beingand Its Correlates
4. Affect as Information and Guidance
5. Affect and Subjective Well-Being
6. A Recent Critique
7. Subjective Well-Beingand Climate
References
Chapter 2: Animals and Climate Change
1. Introduction
2. Farmed Animals, Climate Change, and a Duty to Resist
3. Wild Animals, Climate Change, and a Duty to Assist
4. Animals, Climate Change, and a Life Worth Living
5. Animals, Climate Change, and a Life Worth Creating
6. Conclusion
References
Chapter 3: Discounting under Risk: Utilitarianism vs. Prioritarianism
1. Introduction
2. Choice of the Social Welfare Framework: Utilitarianism vs. Prioritarianism
2.1 Risk and Equity: Aggregation Issues
2.2 Social Welfare Function
3. Implications for Discounting
3.1 Preliminaries
3.2 Utilitarianism and the Ramsey Rule
3.3 Utilitarian Discounting and the Precautionary Effect
3.4 Discounting under Alternative Welfare Frameworks
4. Conclusions
References
Chapter 4: A Philosopher’s Guide to Discounting
1. Introducing and Defending the Ramsey Rule
2. Why the Terms of Measurement Matter
3. Descriptivism and Prescriptivism in Discounting Methodology
4. The Role of Moral Experts in Parameter Assignments
5. Conclusion
References
Chapter 5: Does Climate Change Policy Depend Importantly on Population Ethics?: Deflationary Responses to the Challenges of Population Ethics for Public Policy
1. Introduction
2. Population Axiology and the Repugnant Conclusion
3. First Deflationary Response: Axiologies May Agree about Climate Change
4. Second Deflationary Response: Bounded Population Principles
4.1 Axiology with Population Size Bounds
4.2 Possibility Proof for Escaping the Repugnant Conclusion while Satisfying Bounded Versions of Population Ethics Desiderata
5. Conclusion
Appendix: A Smoothness Axiom and a New Argument for Total Utilitarianism
References
Section II: Cognition, Emotions, and Climate Change
Chapter 6: Way to Go, Me
1. Introduction
2. Climate Change as a Creeping Environmental Problem
3. Different Orientations
4. Switching Between Orientations and Mindset M
5. Seeking Self-Praiseversus Avoiding Self-Blame
6. Conclusion
References
Chapter 7: The Wages of Fear?: Toward Fearing Well About Climate Change
1. The Promise of Fear
2. The Wages of Fear
3. The Possibility of Hope
4. The Perils of Hope
5. Civic Fear
6. Conclusion
References
Chapter 8: Climate Change and Cultural Cognition
1. Introduction
2. Cultural Cognition
3. Values or Beliefs?
4. Cultural Cognition and Coincidence
5. Geoengineering
6. Conclusion
References
Section III: Climate Change and Individual Ethics
Chapter 9: Climate Change and Individual Obligations: A Dilemma for the Expected Utility Approach, and the Need for an Imperfect View
1. Introduction
2. A Dilemma for the Expected Utility Approach
3. Diagnosis
4. The Imperfect Approach
5. Extension to Other Cases
References
Chapter 10: The Puzzle of Inefficacy
1. Introduction
2. Introducing the Puzzle of Inefficacy
3. Ethical Structure and Social Structure
4. Contribution Ethics: A Sketch
5. Negligibility and Interaction
6. Conclusions
References
Chapter 11: On Individual and Shared Obligations: In Defense of the Activist’s Perspective
1. The Activist’s Perspective
2. Two Problems of Individual Incapability and Group Agency
3. Obligations
4. Irreducibly Shared Obligations
5. Collective Obligations in Spite of Individual Incapability
6. Remaining Problems of Collective Capability and Individual Incapability
References
Chapter 12: How Much Harm Does Each of Us Do?
1. Sorts of Harm and Their Quantity
2. New Data and Estimates
3. Lives for Money
4. The Consequences of Discounting
5. Conclusion and Why It Matters
References
Section IV: Climate Change and Politics
Chapter 13: How Quickly Should the World Reduce its Greenhouse Gas Emissions?: Climate Change and the Structure of Intergenerational Justice
1. Introduction
2. Reflections on the Dominant Approach
2.1 Basic Notions of Welfare Economics
2.2 Our Runaway Emissions Are Not in Fact Pareto-Inefficient
2.3 Kaldor-Hicks Efficiency Likewise Cannot Recommend Significantly Cutting Our Emissions
2.4 The Collapse of Climate Economics into a Narrow Form of Total Utilitarianism
2.5 Economic Analysis of Climate Policy Is Subject to the Repugnant Conclusion
2.6 The Need for Discounting Consumption Possibilities in the Future
2.7 Welfarism Is At Best a Class of Pro Tanto Reasons
2.8 We Do Not Owe It to Future People to Bring Them into Existence
3. Outline of a More Adequate Approach
3.1 The Real Puzzle of Intergenerational Justice
3.2 Understanding the Full Measure of the Problem
3.3 Elements of a Solution: Two Supporting Considerations
3.4 Elements of a Solution: Two Ideas Needed to Side-step the Non-Identity Problem
3.5 Extending the Reasoning Indefinitely Backwards and Forwards in Time
3.6 The Duty to Avoid Allowing Catastrophes to Take Place in the Distant Future
4. Conclusion
References
Chapter 14: Political Realism, Feasibility Wedges, and Opportunities for Collective Action on Climate Change
1. Overview
2. The Realist Argument and Efficiency Without Sacrifice
3. Posner, Weisbach, and Sunstein’s Version of Efficiency Without Sacrifice: Intratemporal Transfers
4. Broome’s Version of Efficiency Without Sacrifice: Intergenerational Transfers
5. Is Broome’s Proposal Feasible? How Realistic Is It?
6. Why the Realist Argument for Efficiency Without Sacrifice Is Invalid
7. Feasibility Wedges and a Meta-Architecture for Global Agreement
References
Chapter 15: Pareto Improvements and Feasible Climate Solutions
1. Introduction
2. The Concept of ‘Political Feasibility’
2.1 A Working Definition of ‘Political Feasibility’
2.2 Feasibility and Self-Interest
3. International Paretianism and Climate Change
3.1 Prospects for IP Climate Treaties
3.1.1 Climate Change as a Coordination Game
3.1.2 Climate Change as a Prisoners’ Dilemma
3.2 Does ‘Self-Interest’Suffice?
4. The Feasibility of IP Climate Deals
References
Chapter 16: Climate Change, Liberalism, and the Public/Private Distinction
1. Climate Change and the Anthropocene
2. Liberalism and the Public/Private Distinction
3. The Distinction Under Pressure
4. Pressure Drop?
5. Concluding Remarks
References
Index

Citation preview

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Philosophy and Climate Change

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

E N G AG I N G P H I L O S O P H Y This series is a new forum for collective philosophical engagement with controversial issues in contemporary society. Disability in Practice Attitudes, Policies, and Relationships Edited by Adam Cureton and Thomas E. Hill, Jr. Taxation Philosophical Perspectives Edited by Martin O’Neill and Shepley Orr Bad Words Philosophical Perspectives on Slurs Edited by David Sosa Academic Freedom Edited by Jennifer Lackey Lying Language, Knowledge, Ethics, Politics Edited by Eliot Michaelson and Andreas Stokke Treatment for Crime Philosophical Essays on Neurointerventions in Criminal Justice Edited by David Birks and Thomas Douglas Games, Sport, and Play Philosophical Essays Edited by Thomas Hurka Effective Altruism Philosophical Issues Edited by Hilary Greaves and Theron Pummer

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Philosophy and Climate Change Edited by M A R K BU D O L F S O N , T R I S T R A M Mc P H E R S O N , A N D DAV I D P LU N K E T T

1

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

1 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © the several contributors 2021 The moral rights of the authors have been asserted First Edition published in 2021 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2020949088 ISBN 978–0–19–879628–2 DOI: 10.1093/oso/9780198796282.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Contents List of Figures List of Tables List of Contributors Abstracts of Chapters

ix xi xiii xv

Introduction1 Mark Budolfson, Tristram McPherson, and David Plunkett SE C T IO N  I .   VA LU I N G C L I M AT E C HA N G E I M PAC T S 1. A Convenient Truth? Climate Change and Quality of Life Peter Railton

9

2. Animals and Climate Change Jeff Sebo

42

3. Discounting under Risk: Utilitarianism vs. Prioritarianism Maddalena Ferranna

67

4. A Philosopher’s Guide to Discounting Kian Mintz-Woo

90

5. Does Climate Change Policy Depend Importantly on Population Ethics? Deflationary Responses to the Challenges of Population Ethics for Public Policy Gustaf Arrhenius, Mark Budolfson, and Dean Spears

111

SE C T IO N I I . C O G N I T IO N , E M O T IO N S , A N D C L I M AT E C HA N G E 6. Way to Go, Me Chrisoula Andreou

139

7. The Wages of Fear? Toward Fearing Well About Climate Change Alison McQueen

152

8. Climate Change and Cultural Cognition Daniel Greco

178

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

vi Contents

SE C T IO N I I I .   C L I M AT E C HA N G E A N D I N D I V I D UA L E T H IC S 9. Climate Change and Individual Obligations: A Dilemma for the Expected Utility Approach, and the Need for an Imperfect View Julia Nefsky 10. The Puzzle of Inefficacy Tristram McPherson

201 222

11. On Individual and Shared Obligations: In Defense of the Activist’s Perspective252 Gunnar Björnsson 12. How Much Harm Does Each of Us Do? John Broome

281

SE C T IO N I V.   C L I M AT E C HA N G E A N D P O L I T IC S 13. How Quickly Should the World Reduce its Greenhouse Gas Emissions? Climate Change and the Structure of Intergenerational Justice Lucas Stanczyk 14. Political Realism, Feasibility Wedges, and Opportunities for Collective Action on Climate Change Mark Budolfson

295

323

15. Pareto Improvements and Feasible Climate Solutions Katie Steele

346

16. Climate Change, Liberalism, and the Public/Private Distinction Dale Jamieson and Marcello Di Paola

370

Index

397

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Acknowledgments Thanks to the Climate Futures Initiative, the University Center for Human Values, and the Department of Philosophy at Princeton University for support of this project, including financial support for a conference on these issues at Princeton University in May 2016. Thanks to everyone who participated in that workshop for the helpful feedback and discussion on the papers that were presented. Special thanks to Chuck Beitz, Marc Fleurbaey, Bert Kerstetter, Melissa Lane, Rob Socolow, and Michael Smith. Thanks to Coby Gibson, Max Frye, Daniel Gun Lim, Michael Morck, Joshua Petersen, Ira Richardson, Adrian Russian, Victoria Xiao, and Alice Zhang for their work as research assistants on this project. Thanks to Nithya Kasarla, Daniel Gun Lim, and Evan Woods for their help with the index. And thanks to David Braddon-Mitchell for providing the photo for the cover of this volume.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

List of Figures 1.1 Average reported subjective well-being index vs. per capita GDP, multinational sample, 2005–2008

24

1.2 Average reported subjective well-being index vs. per capita GDP, US sample, 1972–2006

25

1.3 Average reported subjective well-being index vs. household income, Australia, 2006

25

1.4 Average reported subjective well-being vs. income quintile, Switzerland, 1995

26

3.1 The current value of $1 received in t years depending on the social discount rate

68

5.1 The Repugnant Conclusion

116

5.2 The Sadistic Conclusion

117

5.3 Two population axiologies recommend the same “corner solution” to optimal decarbonization

120

5.4 Families of social evaluations that cohere with totalist axioms on the bounded set

126

13.1a Total welfare in time in time under business as usual

304

13.1b Total welfare in time under serious cuts

304

13.1c

304

Total welfare in time under enormous cuts

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

List of Tables 1.1 Per capita carbon footprints by nation and year (metric tons)

32

3.1 Summary of the main properties of the three approaches under analysis

77

3.2 List of the main studies that try to determine the size of the precautionary effect

84

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

List of Contributors Chrisoula Andreou, Professor of Philosophy, University of Utah. Gustaf Arrhenius, Director of the Institute for Futures Studies and Professor of Practical Philosophy, Stockholm University. Gunnar Björnsson, Professor of Practical Philosophy, Stockholm University. John Broome, Emeritus White’s Professor of Moral Philosophy, University of Oxford, and Honorary Professor, Australian National University. Mark Budolfson,  Assistant Professor in the Center for Population-Level Bioethics, the Department of Environmental and Occupational Health and Justice, and the Department of Philosophy, Rutgers University. Marcello Di Paola, Research Fellow in Philosophy, University of Palermo, and Lecturer in Environmental Studies, Loyola University Chicago. Maddalena Ferranna,  Research Associate at the T.  H.  Chan School of Public Health, Global Health and Population Department, Harvard University. Daniel Greco, Assistant Professor of Philosophy, Yale University. Dale Jamieson, Professor of Environmental Studies and Philosophy, Affiliated Professor of Law, and Director of the Center for Environmental and Animal Protection, New York University. Tristram McPherson, Professor of Philosophy, The Ohio State University. Alison McQueen, Associate Professor of Political Science, Stanford University. Kian Mintz-Woo, Lecturer in Philosophy, University College Cork. Julia Nefsky, Assistant Professor of Philosophy, University of Toronto. David Plunkett, Associate Professor of Philosophy, Dartmouth College. Peter Railton,  Gregory  S.  Kavka Distinguished University Professor, Arthur  F.  Thurnau Professor, and John Stephenson Perrin Professor of Philosophy, University of Michigan, Ann Arbor. Jeff Sebo, Clinical Associate Professor of Environmental Studies and Affiliated Professor of Bioethics, Medical Ethics, and Philosophy, New York University. Dean Spears, Assistant Professor of Economics, University of Texas at Austin. Lucas Stanczyk, Assistant Professor of Philosophy, Harvard University. Katie Steele, Associate Professor of Philosophy, Australian National University.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Abstracts of Chapters Section I. Valuing Climate Change Impacts 1  A Convenient Truth? Climate Change and Quality of Life By Peter Railton Justice would appear to require that those who are the principal beneficiaries of a history of economic and political behavior that has produced dramatic climate change bear a correspondingly large share of the costs of getting it under control. Yet a widespread material ideology of happiness holds that this would require sacrificing “quality of life” in the most-developed countries—hardly a popular program. However, an empirically-grounded understanding of the nature and function of “subjective well-being”, and of the factors that most influence it, challenges this ideology and suggests that well-being in more-developed as well as less-developed societies could be improved consistently with more sustainable resource utilization. If right, this would refocus debates over climate change from the sacrifice of “quality of life” to the enhancement and more equitable distribution of well-being within a framework of sustainable relations with one another and with the rest of nature.

2  Animals and Climate Change By Jeff Sebo This chapter argues that animals matter for climate change and that climate change matters for animals. In particular, animal agriculture will have a significant impact on the climate, and climate change will have a significant impact on wild animals. As a result, we morally ought to resist animal agriculture as part of our mitigation efforts and assist wild animals as part of our adaptation efforts. The chapter also evaluates different strategies for accomplishing these aims, and considers connections with debates about well-being, population ethics and duties to future generations, and the nature and limits of moral and pol­it­ ical theory.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

xvi  Abstracts of Chapters

3  Discounting under Risk: Utilitarianism vs. Prioritarianism By Maddalena Ferranna The debate on the economics of climate change has focused primarily on the choice of the social discount rate, which plays a key role in determining the desirability of climate policies given the long-term impacts of climate damages. Discounted utilitarianism and the Ramsey rule dominate the debate on discounting. The chapter examines the appropriateness of the utilitarian framework for evaluating public policies. More specifically, it focuses on the risky dimension of climate change, and on the failure of utilitarianism in expressing both concerns for the distribution of risks across the population and concerns for the occurrence of catastrophic outcomes. The chapter shows how a shift to the prioritarian paradigm is able to capture those types of concerns, and briefly sketches the main implications for the choice of the social discount rate.

4  A Philosopher’s Guide to Discounting By Kian Mintz-Woo This chapter introduces several distinctions relevant to what is called the “discounting problem”, since the issue is how (future) costs and benefits are discounted to make them comparable in present terms. The chapter defends the claim that there are good reasons to adopt Ramsey-style discounting in the context of climate change; the Ramsey Rule is robust, flexible, and well-understood. An important distinction involved in discounting—“descriptivism” and “prescriptivism”—is discussed. It is argued that, even if we adopt prescriptivism, and accept that this means there is a need for moral experts in parameter assignments, there is a significant issue. The type of moral expertise required for the discounting problem will not involve knowledge of moral theory—thus making moral philosophy unhelpful in terms of making particular parameter assignments, despite these being substantive moral judgments.

5  Does Climate Change Policy Depend Importantly on Population Ethics? Deflationary Responses to the Challenges of Population Ethics for Public Policy By Gustaf Arrhenius, Mark Budolfson, and Dean Spears Choosing a policy response to climate change seems to demand a population axiology. A formal literature involving impossibility theorems has demonstrated that

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Abstracts of Chapters  xvii all possible approaches to population axiology have one or more seemingly ­counterintuitive implications. This leads to the worry that because axiological theory is radically unresolved, this theoretical ignorance implies serious practical ig­ nor­ ance about what climate policies to pursue. This chapter offers two ­deflationary responses to this worry. First, it may be that given the actual facts of climate change, all axiologies agree on a particular policy response. In this case, there would be a clear dominance conclusion, and the puzzles of axiology would be practically irrelevant (albeit still theoretically challenging). Second, despite the impossibility results, the chapter proves the possibility of axiologies that satisfy bounded versions of all of the desiderata from the population axiology literature, which may be all that is needed for policy evaluation.

Section II. Cognition, Emotions, and Climate Change 6  Way to Go, Me By Chrisoula Andreou This chapter considers an interesting possibility related to human psychology and explores its relevance with respect to understanding and impacting behavior in the face of climate change understood as a “creeping environmental problem.” The possibility is that one’s best bet in terms of predicting and understanding someone’s take on her individually trivial contributions to positive or negative outcomes is not to look for clues about whether she is individualistic or group-oriented, but instead to figure out whether an individualistic take or a group-oriented take will facilitate her seeing herself as praiseworthy, or at least not blameworthy: insofar as one take is better suited to facilitating self-praise, or avoiding self-blame, with respect to the behavior at hand, it will, other things equal, be adopted. If this is right, it has important implications with respect to attempts at promoting environmentally friendly behavior by activating mo­tiv­ ations associated with the prospect of self-praise or self-blame.

7  The Wages of Fear? Toward Fearing Well About Climate Change By Alison McQueen What role, if any, should appeals to fear play in climate change communication? Moral and practical worries about fear appeals in the climate change debate have caused some to turn toward hope appeals. This chapter argues that fear can be a rational and motivationally powerful response to climate change. While there are good reasons to worry about the use of fear in politics, climate change fear appeals

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

xviii  Abstracts of Chapters can be protected against the standard criticisms of political fear. Hope appeals, by contrast, seem vulnerable to serious motivational d ­ rawbacks in the case of climate change. We should not therefore abandon fear appeals in favor of hope appeals. Instead, we should take our bearings from Aristotle in an effort to cultivate fear more responsibly. Aristotle offers an appealing model of “civic fear” that preserves the best aspects of hope, elicits rather than extinguishes our sense of agency, and invites rather than forecloses deliberation.

8  Climate Change and Cultural Cognition By Daniel Greco How should we form beliefs concerning global climate change? For most of us, directly evaluating the evidence isn’t feasible; we lack expertise. So, any rational beliefs we form will have to be based in part on deference to those who have it. But in this domain, questions about how to identify experts can be fraught. This chapter discusses a partial answer to the question of how we in fact identify experts: Dan Kahan’s cultural cognition thesis, according to which we treat experts on factual questions of political import only insofar as they share our moral and cultural values. The chapter then poses some normative questions about cultural cognition: is it a species of irrationality that must be overcome if we are to communicate scientific results effectively, or is it instead an inescapable part of rational belief management? Ultimately, it is argued that cultural cognition is substantively unreasonable, though not formally irrational.

Section III. Climate Change and Individual Ethics 9  Climate Change and Individual Obligations: A Dilemma for the Expected Utility Approach, and the Need for an Imperfect View By Julia Nefsky This chapter concerns the nature of our obligations as individuals when it comes to our emissions-producing activities and climate change. The first half of the chapter argues that the popular ‘expected utility’ approach to this question faces a problematic dilemma: either it gives skeptical verdicts, saying that there are no such obligations, or it yields implausibly strong verdicts. The second half of the chapter diagnoses the problem. It is argued that the dilemma arises from a very

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Abstracts of Chapters  xix general feature of the view, and thus is shared by other views as well. The chapter then discusses what an account of our individual obligations needs to look like if it is to avoid the dilemma. Finally, the discussion is extended beyond climate change to other collective impact contexts.

10  The Puzzle of Inefficacy By Tristram McPherson We appear to have reasons to act in light of the relationship between our choices and the horrors of factory farming or the escalating bad effects of climate change, even if we are unable to mitigate those bad effects through our individual choices. This idea can seem puzzling in two ways. First, it can seem puzzling how to explain these reasons, given our inefficacy. Second, it can seem that these reasons, even if they existed, would have to be vanishingly weak. This chapter develops a solution to this puzzle that appeals to a novel explanation of why a feature counts as a focal point in the explanation of ethical properties. This solution is applied to show how our relationship to certain social patterns can explain our reasons to respond to facts about factory farming and climate change, mentioned above.

11  On Individual and Shared Obligations: In Defense of the Activist’s Perspective By Gunnar Björnsson People who make substantial efforts to help resolve collective practical problems such as that of catastrophic climate change often think that: (1) Together, we can resolve the problem. (2) In virtue of this, we have an obligation to resolve it. (3) In virtue of the importance of the problem and our capacity to resolve it together, we have individual obligations to help resolve it. This “activist perspective” faces philosophical problems: How can the groups that are to solve the problems have obligations given that these groups are not themselves agents? How can members of such groups have obligations to help, given that they have no individual control over whether the problem is resolved? And how can the collective ability to solve a problem be relevant for individual obligations and individual moral de­lib­er­ ation? This chapter develops solutions to each of these problems based on an analysis of individual and shared obligations.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

xx  Abstracts of Chapters

12  How Much Harm Does Each of Us Do? By John Broome This chapter attempts to estimate the amount of harm an average American does by her emissions of greenhouse gas, on the basis of recent very detailed statistical analysis being done by a group of economists. It concentrates on the particular harm of shortening people’s lives. The estimate is very tentative, and it varies greatly according to how effectively the world responds to climate change. If the response is very weak, the chapter estimates that an average American’s emissions shorten lives by six or seven years in total. If the response is moderately strong, the figure is about half a year.

Section IV. Climate Change and Politics 13  How Quickly Should the World Reduce its Greenhouse Gas Emissions? Climate Change and the Structure of Intergenerational Justice By Lucas Stanczyk Given the accompanying sacrifices, how quickly should the present generation reduce its greenhouse gas emissions? The dominant framework for thinking about this question continues to be normative welfare economics. This chapter explains why the dominant approach should be rejected, and outlines the structure of what the author has come to think is the correct approach. On this approach, requirements of intergenerational justice are understood, not as the means to, but as the most important constraints on maximizing intertemporal welfare. The chapter explains why the main content of these constraints can be given by the theories of social and international justice. Finally, the chapter explains why the non-identity problem does not undermine the recommended way of thinking about intergenerational justice. Even if the business-as-usual baseline in greenhouse gas emissions will never harm any unborn future people, we can still say that humanity is forever subject to a suitably high environmental conservation standard.

14  Political Realism, Feasibility Wedges, and Opportunities for Collective Action on Climate Change By Mark Budolfson This chapter raises objections to the argument that a highly unjust response to the problem of climate change is the best that we can currently hope for and is thus

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Abstracts of Chapters  xxi the solution that we should actively promote even from an ethical point of view. Such an argument has been put forward by a wide range of commentators in phil­ oso­phy, economics, law, and international affairs including John Broome, Cass Sunstein, Eric Posner, and David Weisbach. Among other things, this chapter argues that the way in which this argument fails is both ethically and practically instructive, as its failure reveals how a realist approach to climate policy is consistent with a more equity-focused approach than is commonly appreciated. As a concrete illustration, it is explained how the lessons could be incorporated into a more ethical climate treaty architecture that shares structural features with proposals from William Nordhaus, Joseph Stiglitz, and others.

15  Pareto Improvements and Feasible Climate Solutions By Katie Steele Proponents of International Paretianism (IP)—the principle that international agreements should not make any state worse-off and should make some at least better off—argue that it is the only feasible approach to reducing the harms of climate change. They draw on some key assumptions regarding the meaning of ‘feasibility’ and the nature of the Pareto improvements associated with co­ord­in­ ated action on climate change. This chapter challenges these assumptions, in effect weakening the case for IP and allowing for broader thinking about what counts as a ‘feasible’ climate solution.

16  Climate Change, Liberalism, and the Public/Private Distinction By Dale Jamieson and Marcello Di Paola Climate change puts pressure on a distinction that is at the heart of liberal theory: that between the public and the private. Many of the GHGs-emitting behaviors that contribute to the disruption of the climate system—such as using computers, taking hot showers, eating this or that, driving cars, investing here or there, and having children—are traditionally regarded as private. Yet today, through climate change, these apparently private behaviors can have very public consequences, however indirect, across spatial, temporal, and genetic boundaries. The chapter introduces the public/private distinction and discusses the various ways in which it has figured in liberal theory. It goes on to show how climate change threatens the viability of the distinction, both by intensifying old tensions and by bringing new pressures to bear. It then considers some options for relieving the pressure, none of which seems particularly promising by liberal lights.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 13/02/21, SPi

Introduction Mark Budolfson, Tristram McPherson, and David Plunkett

This volume is guided by two thoughts. First, philosophers have much to contribute to the discussion of climate change. Second, reflection on climate change can ­contribute to our thinking about a range of general topics that are of independent interest to philosophers. This volume will be of interest both to philosophers working on climate change as well as those working in a range of other fields, ranging from public policy to economics to law to empirical disciplines including psychology, the science of climate adaptation, mitigation, and beyond. Part of what we aim to establish in this volume is that philosophers are in a strong pos­ ition to collaborate in the kind of interdisciplinary conversations needed to tackle pressing challenges for the world such as climate change. In this short introduction, we explain the guiding thoughts behind this vol­ ume, and provide a broad overview of some of the key themes that connect the chapters here. We also situate the chapters here within the broader inter­dis­cip­lin­ ary discussion of climate change. (A detailed abstract for each chapter precedes this introduction.) The first guiding thought behind this volume is that philosophers have much to contribute to the discussion of climate change. This is because philosophers have developed important intellectual tools and ideas that can help everyone think more clearly and carefully about central issues raised by climate change. For example, consider that the best scientific evidence suggests that climate change will have profound effects on a global scale, and that what we do in the near future can alter those effects for better or worse. In light of this, climate change poses some of the most profound ethical challenges of our time. Regardless of how we respond to climate change, our actions (or lack thereof) will have good effects on some and bad effects on others. To decide how to respond collectively, we need to think about who and what matters, about what sorts of effects on them matter, and about what would be a just or equitable way of ar­ran­ging those effects. For example, should our aim in responding to climate change be merely to maxi­ mize the total economic output of the world, as many influential economic models assume? Or should we also value the health and well-­being of all humans equally regardless of whether those individual people are rich or poor, and thus regardless of their contributions to global economic output? And what about future Mark Budolfson, Tristram McPherson, and David Plunkett, Introduction In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Mark Budolfson, Tristram McPherson, and David Plunkett. DOI: 10.1093/oso/9780198796282.003.0001

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

2  Mark Budolfson, Tristram M c Pherson, and David Plunkett humans, who do not yet exist, and who may or may not exist at all depending on what policies we choose? Does the health and well-­being of the environment and non-­human animals matter as well even beyond its value to humans? How should the burdens of adaptation and mitigation be distributed given the nature of the climate change problem and its causes? The first section of essays—Valuing Climate Change Impacts—addresses these questions of value theory about what ultimately makes outcomes better or worse, in connection with responses to climate change and beyond. At the same time, even supposing we have answered the preceding questions and thus are confident which responses to climate change are better and worse, we must still grapple with further political questions: which of these responses are politically infeasible, and how should feasibility affect which response we decide to aim for with policy and political advocacy? More generally, to what extent are classic goals of political philosophy—such as liberalism, protection of basic rights, justifiability to each person of the structure of society, and so on—helpful in iden­ tifying the best way to restructure society in response to climate change? And might the problem of climate change itself provide a challenge to some of these classic goals of political philosophy? The fourth section of essays—Climate Change and Politics—addresses these questions of political philosophy. Even if we were to settle all of the preceding questions about what we col­lect­ ive­ly should do about climate change both nationally and internationally, there is a gap between those facts and what conclusions you and I should draw for our own individual behavior. In particular, we need to better understand the connec­ tion between the best collective responses to a problem such as climate change, and what actions individual people should take. For example, is climate change essentially a global collective action problem that requires a global carbon pricing response in a way that makes no special action required by individuals, beyond merely supporting and complying with such a collective-­level scheme? Or instead are individual people required to take costly actions even beyond that, given that climate change is literally killing more and more people all the time as a result of our collective emissions? The third section of essays—Climate Change and Individual Ethics—addresses these questions, and provides a toolbox of important resources for thinking well about these (and other) pressing ethical questions that face us as individuals in contemporary society. Another set of issues arises from the notorious mismatch between (a) the actual beliefs and motivations of people in the world right now concerning ­climate change, which are demonstrably inadequate to cope with the challenges we face, and (b) what beliefs and motivations would be needed for us to effectively respond to climate change—and how to get from (a) to (b). The second section of essays—Cognition, Emotions, and Climate Change—demonstrates how philo­ sophers can help us to understand and evaluate the nature of the relevant motives and beliefs, and contribute to the effort to improve the beliefs and motivations of

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Introduction  3 people in order to create the right conditions for an adequate response to climate change to emerge. In addition to the issues above that structure this volume, there are many other noteworthy issues that appear throughout the volume. To take one example, philo­sophers can help illuminate and critically examine the often obscure meta­ phys­ic­al and ethical assumptions about the future that underlie influential con­ temporary policy discussions about climate change. This is an especially pressing topic because experts often agree that the best response to climate change depends heavily on the correct approach to specific questions about the future, namely intertemporal discounting and population ethics. The chapters in this volume demonstrate how philosophy can help us to make further progress toward shared understanding, and perhaps even narrowing our uncertainty about these issues. In addition, other issues that appear throughout the volume include: • The ethical and political significance of uncertainty, and risk, and decision-­theoretic principles • The significance of feasibility considerations and ideal vs. non-­ideal theory in political philosophy • The ethical importance (or unimportance) of making a difference in one’s actions as a single individual person • The epistemology and normative psychology of risk, beliefs, and emotions regarding oneself and others, including updating beliefs in light of the assessment of external experts The sixteen chapters collected in this volume contribute to these important topics, and give readers important entry points into the burgeoning philosophical literature on themes relevant to climate change. Our goal was to include philo­ sophers working in different subareas of philosophy, ranging from ethics to political philosophy to epistemology to the philosophy of science. We have also sought to include both philosophers who have already made important contributions to philosophical discussion of climate change, as well as those who had not written directly on this topic prior to this volume. The second thought that animates this volume is that thinking about climate change can be illuminating for a range of topics of abiding interest to philo­sophers. For example, philosophers aim to understand the nature of political norms, distributive justice, how to respond to uncertainty, our obligations to non-­human animals, the relationship between collective and individual responsibility, the ethical significance of currently non-­existent future persons, and when the state is justified in using coercion to promote its policies. The threats posed by climate change not only make these issues more pressing; they can also shed new light on them, giving philosophers important reasons to revisit assumptions that have structured much of our thinking about these questions. In addition to the issues

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

4  Mark Budolfson, Tristram M c Pherson, and David Plunkett directly addressed in this volume, other important philosophical questions about climate change that are not addressed here but that promise to lead to more gen­ eral philosophical insight include:1 • Individual and collective responsibility for climate change, and the relation­ ship between the two, including issues regarding historical responsibility, benefitting from injustice, and corrective justice • Institutional responsibility, for example the responsibilities of universities to invest in research and/or divest from fossil fuels and animal agriculture • Responsibility for climate-­based immigration and refugees • Climate change and the application of distributive justice to the global sphere • The potential ethical significance of intentions and/or the distinction between doing vs. allowing harm to responsibility, or to reasons for action • The epistemology of scientific risk assessment: empirical evidence, scientific models, and subjective: assessing the empirical evidence for anthropogenic climate change • The role of scientific expertise in public discourse and political decision-­making • Civil disobedience and political action • The ethics of communication and public persuasion • The ethics of agriculture, including issues about the emissions footprint of animal agriculture • The ethics of procreation, including issues about the emissions footprint of having children • The ethics of species loss and wilderness loss • The ethics and epistemology of risks related to potential technological solu­ tions to climate change, including geoengineering • The ethics of difficult tradeoffs, including between meeting the current urgent needs of the desperately poor for cheap energy vs. the needs of future people • Economic growth, climate change, and the correct objective/axiology for evaluating social policy • The importance of public justification and the legitimacy of coercion and international institutions, especially in the face of looming catastrophe and unsolved global collective action problems, and general ethical and political principles that apply in emergencies and global collective action situations • Technocracy and the importance of public justification, especially given empirical issues that are opaque to citizens, and the time inconsistency of the costs to citizens now vs. benefits to others (mostly non-­citizens) in the future • Religious toleration and religion-­based denial of climate change 1  Many of these topics are addressed in other work that has appeared in climate change and philosophy.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Introduction  5 In light of the above, we believe that philosophical work on climate change— including the work showcased in this volume—can make an important contribu­ tion to the broader interdisciplinary conversation about this pressing topic. We hope the volume also spurs others (both from within philosophy as well as from other fields) to enter these conversations, and collaborate together for the good of society and future generations. In our view, understanding climate change, and discussing how to address it, should be at the very center of public conversation for the foreseeable future. We think that philosophy can make an enormous contribution to that conversation, but will do so only if both philosophers and non-­philosophers understand what philosophy can contribute. We hope that this volume contributes to that understanding.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

SECTION I

VALU ING C L IMAT E C HA NG E IMPACT S

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

1

A Convenient Truth? Climate Change and Quality of Life Peter Railton

1. Introduction One of the chief obstacles in the developed world to undertaking the measures needed to contend with global climate change is the widespread belief that these measures would inevitably come at the expense of the standard of living now enjoyed by a significant portion of the population in these countries. This in turn raises concerns about justice and fairness. Could it be just or fair to take steps that would impose upon future generations in the US and elsewhere a lower standard of living than our own? Such worries are especially acute since, arguably, those currently enjoying a high standard of living in the developed world are the ones who have inherited most of the benefit from the long history of exploitation of the environment. Moreover, economic growth in developing countries over recent decades has been credited with raising millions from poverty (Deaton 2013), and this trend could be dramatically affected by slowed growth in developed and developing countries. Yet if developed and developing countries continue to enlarge their burden on the environment at current rates, the problem of global climate change is likely only to worsen, with pervasive long-­run negative effects on well-­being worldwide—effects that can be expected to be borne most heavily by the world’s least advantaged populations. It seems we are in a double-­bind, and while technological innovations or economic changes might occur that would permit continuing current rates of economic growth indefinitely, to plan on this would be taking a tremendous risk with the lives and well-­being of others, including other species.1 Additionally, there is risk to the social and political environment to consider. Limited economic growth could also mean that politics and economics come to approximate a zero-­sum game, intensifying social rivalry. In such competition, those already possessing greater material resources are likely to be able to use their influence to consolidate their position, entrenching inequality and undermining 1  In this chapter our primary focus will be upon human well-­being, not because that should be our sole concern, but because it is certainly concern enough to raise these issues acutely. Peter Railton, A Convenient Truth? Climate Change and Quality of Life In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Peter Railton. DOI: 10.1093/oso/9780198796282.003.0002

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

10  Peter Railton democratic processes. Indeed, we have already seen increasing inequality and ­rising illiberal and non-­democratic political tendencies in well-­developed countries where large sectors of the workforce have experienced an erosion of their relative economic standing. How to mobilize sufficient support for more sustainable policies with respect to the natural environment even as a deteriorating social and political environment seems to be undermining public institutions and the sense of shared purpose? But what if this apparent double-­bind involves a mistaken view of the relation of well-­being to “standard of living”? And if this mistaken view has in turn misled us in how we frame the worry that future generations will “live less well than we do”? While part of the argument of this chapter will be that an analysis of people’s actual standard of living must go beyond the orthodox criterion of real per capita gross domestic product (GDP), this criterion is the most widely studied indicator of “standard of living” across countries and across time, and so will be our starting point.2 And I will mostly focus on “well-­being” in the sense of “subjective well-­being” typical of contemporary social science: how individuals report experiencing their daily lives, express overall happiness or satisfaction with that life, and evaluate their lives against their ideas of what makes a life better or worse. As  philosophers, we might have more elaborate or demanding conceptions of ­well-­being, but subjective well-­being, too, is the most heavily studied indicator across countries and time. Moreover, the experiential and self-­evaluative elements it involves are likely to be part of any plausible theory about what it is to be living well, and to figure heavily in shaping people’s social attitudes and political behavior. So the literature on standard of living and subjective well-­being will afford us some empirical traction in thinking about how the changes needed in order to contend with global climate change might actually affect standard of living or well-­being in a fuller and more normative sense, or how this might interact with questions about fairness or justice. However, we cannot simply take the subjective well-­being literature at face value—to understand its potential significance for the questions we are examining here we will need to look for underlying explanations of the phenomena this literature records. For example, I will be arguing that understanding the sources and components of subjective well-­being can help us understand some of the seemingly puzzling behavior of measurements of subjective well-­being. And this will enable us to see more sustainable alternatives to current patterns of economic growth, especially if our aim is to achieve greater and more equal subjective well-­being worldwide. While we cannot ignore the fact that very substantial material development will be needed to bring more of the world’s population to a level of material sufficiency and security, we must also take into account that the way this growth is achieved 2  GDP includes personal consumption (68% in the US in 2019, Q1), private investment (17.5%), net exports (−3%), and government spending (17.5%) (Federal Reserve 2020).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  11 can have effects on subjective well-­being that, taken together, are of greater importance than the absolute level of real per capita GDP. Societies in the developed and developing world that give greater emphasis to these other factors— which include levels of social support, public health, civil and personal freedom, and effectiveness and responsiveness of governance—already provide evidence that our children could live better than we do, even at a lower level of real per capita GDP than the contemporary US. Decades of study indicate that increasing levels of real per capita GDP can be generated in already-­prosperous countries, at tremendous cost to the environment, without noticeable gains in average reported subjective well-­being. By contrast, high levels of social support, public health, civil and personal freedom, and effectiveness of governance, which need not impose comparable burdens upon the environment, are reliably associated with increased subjective well-­being across a wide range of levels of real per capita GDP. Indeed, the social, political, and economic structures that tend to generate the greatest increases in per capita income and wealth may also tend to undermine levels of social support, public health, civil and personal freedom, and effectiveness and responsiveness of governance. We must of course be realistic as we think about mobilizing responses to global climate change. But it seems to me unrealistic to think that continuation of existing patterns of economic growth will best serve the realization of human well-­being. Current economic structures and practices will be hard to change in the direction of greater sustainability, but not because moving them toward greater sustainability would necessarily conflict with effective pursuit of widespread human ­happiness—that, I will argue, is happiness ideology, not happiness science. It is the same ideology that transformed the jointly-­realizable “American Dream” as first articulated by historian James Truslow Adams—a “dream of a land in which life should be better and richer and fuller for every man, with opportunity for each according to his ability or achievement” (Adams  1931)—into the very thing he was warning against: a competitive, individualistic dream of happiness in life as gaining in wealth and social position. It is possible to imagine a more realistic picture of human happiness, as we will see below, but not a picture more convenient for channeling people’s legitimate desire for a better life for themselves and their children into a drive for material consumption and for consumption-­fueled economic growth, a drive that may well be a risk factor with respect to attaining human happiness—even for those most consumed by this drive.3 Global climate change has been said to be an “inconvenient truth”, since it conflicts with maintaining cherished ways of living and won’t go away by being 3  Kasser (2003) and Nickerson et al. (2003) have found evidence that having a high level of personal concern with materialistic values is associated with lower levels of subjective well-­being, though this association is, as one might expect, mediated somewhat by income level. And Van Boven (2005) and others have found that experiential and shared goods tend to be more rewarding, at the moment and in retrospect, than material goods.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

12  Peter Railton ignored.4 I will be arguing that, if we are genuinely concerned with human ­well-­being and adopt a more plausible theory of its nature and sources, existing ways of living need not be so cherished. That is a “convenient truth”, both from the moral standpoint of commitment to well-­being and justice and from the practical standpoint of mobilizing support to contend with potentially devastating climate change. Since I have no recipe for translating this truth into actual social practice, I may justly be accused of being too optimistic on the strength of it. With Kant, though, I believe that a bias for hope, even in the face of a seemingly dark future, is one characteristic lapse of sober human reason that we should not attempt to correct.

2.  The Nature and Measure of Subjective Well-­Being To make the case for this “convenient truth”5 we will have to devote much of this chapter’s length to looking into research on subjective well-­being, so we had better say more to adumbrate that notion. When speaking strictly, psychologists will emphasize that, as an empirical construct, subjective well-­being is a collection of answers to standard arrays of questions, not a theory of well-­being or happiness. These answers are self-­reports, not judgments by trained researchers, and thus they rely upon how people understand the questions they are asked and the answers they choose among. The standard questions divide into two broad categories, those concerning the hedonic tone of one’s life and those concerning one’s satisfaction with one’s life overall. While the answers given in these two categories correlate with each other, they also behave differently in a number of important ways, as we’ll see. And some researchers have argued that it is misleading to treat these two elements as forming together a unified construct. The first, hedonic category has generally taken one of two forms. In the ­longest-­running surveys, for example the World Values Survey (Inglehart et al. 2014), individuals are asked about their happiness, e.g., “Taking all things together, would you say you are (i) Very happy, (ii) Rather happy, (iii) Not very happy, or (iv) Not at all happy?” This kind of response range is sometimes called a Likert Scale. More recent surveys may divide the range into five, or seven, or ten steps. The other, also more recent, form of question within the hedonic category asks individuals to report, not their happiness overall, but the frequency of their current or recent experience of a range of more specific or episodic emotions or moods, such as joy, laughter, sadness, stress, anger, or worry. For example, in the 4  These are, of course, only two of the many ways in which global climate change is an “in­con­veni­ ent truth”. 5  Potentially convenient, more accurately—it will matter how this truth is appreciated and used.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  13 massive Gallup World Poll subjects are asked to focus on the previous day, from morning until night, and answer whether they experienced “a lot” of anger, or depression, or enjoyment, etc. (Gallup World Poll). Other variants deploy a scale for these answers, ranging from “Not at all” to “All the time” for specified emotions or moods over the specified period of time. Questions in the second main category, often called “life satisfaction”, ask in­di­ vid­uals to evaluate, or indicate a degree of satisfaction with, their lives overall. Here subjects are typically given a Cantril Ladder for reporting their answers, where the top rung is something like “Best possible life” or “Best possible life for me” and the bottom rung is “Worst possible life” or “Worst possible life for me”. Respondents are invited to indicate “where they stand” on the ladder. Since such self-­ranking constitutes a judgment rather than a feeling of satisfaction as such, this component is sometimes called “cognitive life satisfaction”. To obtain subjective well-­being, the values for the “hedonic” answers and the “life satisfaction” answers are combined into an overall score or “index”, usually giving each category equal weight. In addition to using large-­scale surveys, measures of subjective well-­being are also obtained in laboratory settings, face-­to-­face interviews, physiological measures, and “experience sampling”, in which infor­ mants are contacted at random times during their daily lives and asked to report what they are currently doing and to answer some form of the usual questions (Eid and Diener 2004). Since the answers given are self-­reports of individuals following their own interpretation of the questions, one might expect significant, arbitrary variability in the results of these surveys. And they do in fact show some variability. But what is more striking is how much consistency is shown in the results of these surveys, even cross-­culturally, and how much the variability that is found is systematic rather than arbitrary—e.g., linked in predictable ways to context, mood, etc., which then can be controlled for (for discussion, see Eid and Diener 2004). And when psychometric techniques are deployed on the data generated by sampling subjective well-­being, or latent variable models are developed to explain these data, the results vary for different measures and models, but a number show reasonable statistical behavior and such features as normalcy in distribution, discriminant power, and internal consistency (see Lucas et al. 1996; Eid 2008; Diener et al. 2013). Subjective well-­being has come to be of interest as an indicator in part because it can add predictive power to psychological and economic models based upon other, more “objective” measures (Frey and Stutzer 2002). For example, knowing someone’s score on subjective well-­being can improve accuracy in predicting health outcomes, performance in school or at work, success or failure in relationships, and more (Luhmann et al. 2010). Despite its seeming “softness” or arbitrariness, then, measurements of subjective well-­being appear to be correlated with something genuine, and as a result an active industry has grown up around measuring and utilizing subjective well-­being in research, clinical practice, public policy, and commerce.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

14  Peter Railton Philosophers have tended to be skeptical about subjective well-­being, especially given its reliance upon self-­report surveys and the tendency of the press and some psychologists to ignore the warnings and frame their presentation of data on sub­ ject­ive well-­being in terms of happiness or well-­being, full stop. However, it seems to me that philosophical accounts of personal well-­being cannot afford to be so far alienated from the perspective of actual individuals as to treat as irrelevant what people themselves will say when asked to reflect on how well their lives are going, or how they feel day-­in and day-­out. And attention to the literature on subjective well-­being may help philosophers keep some aspects of normative theorizing in proportion. For example, it is not uncommon for philosophers to treat “full realization of characteristic human potential” as central to well-­being. Yet in most countries average reported sub­ ject­ive well-­being tends to rise in later decades of life—after having fallen during the middle years of life—even as older individuals are undergoing changes that reduce a number of their abilities. Could this simply be “adaptive preference”, in which people settle for getting increasingly less from life as their capacities diminish? It is, of course, rational to adjust one’s goals in light of changes in resources or potentials (Rothermund and Brandstadter  2003), but this need not result in a diminished life and would not explain why, by contrast, average reported sub­ject­ ive well-­being falls dramatically in the later decades of life in some countries. For example, a comparison of prosperous Anglophone countries with former Soviet bloc countries (Deaton 2007) found that, in both groups, average individual reported satisfaction with health declined in the final decades of life, indicating that health preferences do not simply adapt to “like what one gets”. By contrast, average individual reported satisfaction with life in one’s later years either rose or stayed constant in the wealthy Anglophone countries, while it declined markedly in the former Soviet bloc countries that had experienced a collapse of the systems for socially-­subsidized health care and support of the aged. It therefore is not inherent in the human condition that diminished health and capacities lessen one’s experienced quality of life—societies and social relations have the power to affect this significantly. Disabled individuals and disability advocates have argued strongly that it is too easy to assume that well-­being requires full development of typical human capacities, or that those who experience disability must somehow be settling for less rewarding lives. Not only does this misrepresent the lived experience of many disabled individuals, it may be another manifestation of the material ideology of happiness, ignoring evidence that material resources need not dominate in one’s quality of life (cf., for example, Solomon 2012).

3.  Subjective Well-­Being and Its Correlates Standardized measures of subjective well-­being began to be used systematically in World War II, trying to understand how well individuals bear up in, or after,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  15 combat situations. Once the war was over, the measures began to be used to observe trends in the population more broadly. What was found? Focusing first on the highly-­developed and highly-­studied societies of North America and Western Europe, the great bulk of the population reported subjective well-­being crowded into the top half of the scales, with an average of 2.3 on a three-­point scale or 3.8 on a five-­point scale not uncommon. So the average reported subjective well-­being in these societies isn’t “neutral” (mid-­way on the scale, say), but distinctly positive. Moreover, distribution around this average was statistically “normal”, with most scores clustered within a band around the median value, and frequency falling off fairly uniformly on either side of the central mass of scores. And by the time the curve reaches the bottom of its range (the top being truncated somewhat by the “ceiling” of the maximal value), the tail has become very thin. Moreover, if we look at individuals’ reported scores over time, there is a tendency to stay within a range centered on a fairly constant modal value, spending some time above, and some time below, but in a fashion that is fairly balanced to either side. Individual differences in this average value have often been attributed to genetic or personality factors. Strikingly, within a country over significant stretches of time, the average value for reported subjective well-­being also tended to remain fairly stable, despite changes in other social variables, including real per capita GDP. For example, for the US, real per capita GDP tripled between 1947 and 2004. Yet during this same period, average reported subjective well-­being remained fairly constant, even falling slightly (World Database of Happiness; Clark et al. 2008). One might fish for explanations in the peculiar political and cultural trajectory of the US, except that a similar phenomenon has occurred in the large Western European countries, where the real GDP per capita has also risen significantly since the 1960s, while average reported satisfaction with life as a whole has, with fairly limited variation, remained essentially constant (Eurobarometer; Clark et al. 2008). A potential implication for the most-­developed countries of the US and European data is striking. A doubling or even tripling of real GDP per capita did not seem to have a significant effect on the average reported level of subjective well-­being, so that the creation of a staggering additional amount of material wealth—and consumption of an equally staggering amount of material resources, with attendant environment effects—might not have made a significant difference to the experienced quality of life.on the whole.6 As we look more closely into

6  Of course, many other factors were at work in these societies during the same period—for example, income inequality in the US and Europe was initially fairly low by historical standards from the 1940s to the 1980s, and rose markedly afterwards (Piketty and Saez 2013). Still, the striking pattern of rising real per capita GDP with relatively flat average reported subjective well-­being held during both the periods of lower inequality and those of higher inequality. Throughout, unless otherwise noted, I will follow the standard usage in the social science literature of speaking of “statistical effects”, such as regression coefficients, as “effects” tout court. This, I grant, is tendentious and potentially misleading, since “effect” is also a causal term, and less is known about

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

16  Peter Railton the sources and dynamics of subjective well-­being, this fundamental fact should not be lost sight of. In response, someone might argue this pattern shows only that people’s reported subjective well-­being is not a sensitive indicator of real changes in their conditions of life. Perhaps people’s reported subjective well-­being is driven by relatively constant factors of individual psychology, for example, and so tells us little about how people’s objective quality of life is changing over time? However, we will be reviewing a considerable body of evidence that suggests that individuals’ reported subjective well-­being is often quite sensitive to changes in features of their existence that we would normally think of as important to quality of life, including variation in economic conditions. For example, the “output gap” is the difference within an economy between available productive cap­ acity and actual output, expressed as a percentage of GDP. Mapping the output gap over time serves as one indicator of changes in overall economic activity. Over the three-­decade period 1972–2006 in the US, fluctuations in the output gap on the order of a 1–2 percent increase or decrease were reflected, with a brief time lag, in similar changes in expressed subjective well-­being, varying on either side of a fairly constant mean (General Social Survey). Output gap is not a widely followed statistic, so this effect is unlikely to have arisen simply from responses to news reports. Instead, the changes taking place in actual lives reflected in this statistic—e.g., unemployment vs. reemployment, reduced vs. increased ability to cover basic household expenses, etc.—appear to have been reflected in turn in average reported subjective well-­being. The question is less, “Is subjective well-­being an informative social measure?”, than “Which aspects of life is subjective well-­being sensitive to, and why?” Here is a first hint: both individual and social reported subjective well-­being generally seem to fluctuate around a mean, being especially responsive to changes in condition—to trajectories in key indicators. Thus average reported subjective well-­being fell when the Global Financial Crisis hit in 2008, though it rose subsequently as the economy improved (Helliwell et al. 2019). In Russia in the aftermath of the collapse of the Soviet Union, average reported subjective well-­being fell sharply and deeply, only returning toward its former level as conditions slowly began to normalize (Ingelhart et al. 2013). In Venezuela for the period 2010-­2016, during which it experienced a series of major social and economic disruptions, average reported life-­satisfaction fell from over 7 to just above 4 (on a 10-­point scale, Rojas 2018). And for individuals, average reported subjective well-­being, as we’ll see, also seems significantly tied to changes in life-­condition.

causal effects upon, or of, subjective well-­being. So claims about “effects” must be understood as qualified in this way.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  17

4.  Affect as Information and Guidance That subjective well-­being would in this way be especially responsive to change can be linked to quite general features of affect in the human psychology. ‘Affect’ is a generic term in psychology, used to encompass both episodic emotions and moods. Affect enters into both components of subjective well-­being, though in different ways. The experiential component of subjective well-­being is a direct self-­report of felt “happiness” or of recent positive vs. negative felt emotions (e.g., joy, sadness, anger, worry, interest, etc.); the life-­satisfaction component typically combines an evaluative dimension (e.g., where one takes oneself to stand relative to the “best possible” or “worst possible” lives for oneself) with a more affective dimension (e.g., “Taking all things together, how satisfied are you with your life?”), and evidence suggests that people often use recent or prospective felt affect as one component in determining how well their lives are going. What do we know about the nature and function of the affective system, then, that might help us understand the behavior of subjective well-­being as an indicator? In contemporary psychology, a conception of the affective system has emerged in which its principal functions are to inform the individual and help guide subsequent behavior (Schwartz and Clore 2003). As Phoebe Ellsworth, a cognitive social psychologist, and Randolf Nesse, an evolutionary psychologist and clinical psychiatrist, write: Emotions are modes of functioning, shaped by natural selection, that coordinate physiological, cognitive, motivational, behavioral, and subjective responses in patterns that increase the ability to meet the adaptive challenges of situations that have recurred over evolutionary time. (Nesse and Ellsworth 2009: 129, original italics)

Fear, for example, is a response to evidence of threat, danger, or loss of standing, and it works to rapidly mobilize in a coordinated way a wide range of physical and mental responses appropriate to contending with threating circumstances— reorienting attention, increasing vigilance, priming memory and inference, and intensifying motivation and bodily action-­readiness. In the brain, the affective system comes into play very early in perceptual processing (for some dimensions of emotion, within 100 milliseconds of the onset of a sensory stimulus). Fear thus can be engaged and begin to operate prior to any self-­conscious judgment about risk—one can be uneasy about a situation without knowing why, and this emotional state can prime subsequent thought and begin to affect behavior (Diano et al. 2017). The affective system is also a primary nexus for reward and learning, using on-­going feedback from experience to update our expectations and thereby shape our choices and actions. Consistent with this central informing and guiding role, the affective system recruits acquired information widely and projects to

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

18  Peter Railton virtually every region of the brain (Pessoa 2008). Equally important for informing and guiding, the affective system plays a critical role in mapping the physical environment, representing others’ states of mind, and simulating and evaluating alternative courses of action (for more extended discussion, see Duncan and Barrett 2007; Seligman et al. 2013). While fear is a paradigmatic emotion, we should not think of the evolution and working of the affective system solely in terms of episodic arousal states. Although some have defended the idea of evolutionarily-­selected “basic emotions” with particular behavioral manifestations, it is increasingly common to see the af­fect­ ive system as a general-­purpose capacity for the appraisal of situations and prospects, and the adjustment of mental responses and behaviors (Duncan and Barrett  2007; Moors et al.  2013). It has become clear that affect is continuously operative, not only in such acute states as fear, joy, anger, or surprise, but equally in such states as satisfaction, trust, interest, and affection. It is not surprising, given this continuous informing and guiding function, that people are generally ready with answers when happiness researchers approach them with questions about how their life is going, or what positive or negative feelings they have had recently. For example, the anterior insula, a central component of the affective system, projects to brain regions involved in conscious ex­peri­ence, and appears to carry out a real-­time synthesis of information collected from internal and external sensation, bodily requirements, goals, and memory to provide a unified experiential sense of how well one is doing from moment to moment in meeting one’s needs or realizing one’s aims (Craig 2009). The affective elements of subjective well-­being thus can be seen as part of an evolved system for informing individuals in real time about how well they are doing in many dimensions of life, whether significant or minor, short term or long, or personal, familial, economic, or social. What might we predict about sub­ ject­ive well-­being, given this picture of its affective elements?

5.  Affect and Subjective Well-­Being Consider another system evolved to inform and guide: perception. Upon entering a room with a ticking clock or buzzing light fixture, the noise will distract at first, but if the unvarying sound continues, its impact on perceptual experience tends to attenuate to inaudibility, and only a deliberate shift of attention, or a variation in the noise, will reinstate the sound in perceptual experience (Elliott et al. 2011). This attenuation of experiential impact has been characterized as perceptual “habituation” or “adaptation”,7 which is thought to free the perceptual system to attend to other, more informative features of one’s situations. 7  These terms are used somewhat interchangeably, and there is no clearly recognized distinction or  definitive account of underlying mechanisms. It seems to me important not to confuse

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  19 One of the earliest findings in the subjective well-­being literature was that various important life events, which initially have a strong positive or negative effect on individual reported subjective well-­being—marriage, the loss of a spouse, winning a lottery, or serious disability, for example—often have an attenuated effect on reported subjective well-­being over time, as an individual tends to return a personal “set point” (Brickman et al. 1978; Diener et al.  2006; Oswald and Powdthavee  2008). This process is often spoken of as “hedonic adaptation” or “habituation”, but we should ask whether habituation is in fact the right model.8 Even in the case of perception, “habituation” can be a bit of a misnomer. The perceptual system appears to draw upon the mind’s predictive modeling of its environment, so that sensory attention can be allocated preferentially to the most informative inputs. This predictive modeling of the environment is complex, and can lead to the shaping and production of conscious experience, e.g., strengthening the signal for spoken words against background noise, generating constant color for objects despite changes in illumination, filling in the “blind spot”, and coloring the full visual field even though color is only directly perceived in a small region of the retina. That is, instead of simple, passive “habituation”, the system is intelligent, active, and sensitive to information value on a continuing basis (for discussion, see Teufel and Fletcher 2020). Should the continuous buzzing of a fluorescent light fixture abruptly stop, we may immediately respond to the change, suggesting that the aural environment was not simply being “tuned out”, but rather was being continuously monitored for information that does not fit prediction. For example, the small, involuntary eye-­movements being made approximately five times a second to take in the visual environment—saccades—are regulated by a learning-­based system that shifts gaze in the direction of the most informative incoming signals, and at a rate proportional to information value (Morris et al. 2016). So what we might call “intelligent perception” of the environment is a continuous process involving the coordinated action of many elements: statistical and causal modeling of the environment, sensitivity to one’s own in­tern­al state or goals, and detection of discrepancies with expectation, all of which go into shaping momentary conscious awareness. Should we expect the affective system—which receives information directly from sensation, internal states, and goals, models the physical and social environment, and spontaneously allocates or shapes such processes as attention, cognition, conscious experience, motivation, and action—to be any less actively intelligent adaptation—e.g., the adaptation of the concentration of hemoglobin in one’s bloodstream to moving to higher altitudes—with habit—e.g., relatively stereotyped and inflexible behaviors arising when reward from a response to a stimulus is unchanging.. 8  “Set point” theory has been subject to increasing criticism. As we saw in the case of age, and will see in a wider array of circumstances, below, there can be medium or long-­term changes in an individual’s level of subjective well-­being. One important source of evidence is the study of “life tra­jec­tor­ ies” using longer-­term data sets (see Heady and Muffels 2017).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

20  Peter Railton in monitoring the environment we inhabit? Consider a study of the nucleus accumbens, a part of the affective system which plays a critical role in reward and behavior regulation in animals and humans, and is highly conserved evolutionarily. Reynolds and Berridge (2008) found that, as the background level of risk in a mouse’s environment changes, the portion of the nucleus accumbens shell that is responsive to risk also varies, permitting finer-­grained discrimination of levels of risk. The allocation of the shell to positive, assuring stimuli returns to normal when background risk decreases and greater sensitivity is needed across the full range. This constitutes re-­calibration in response to risk, to render the system more attentive to gradations of risk when potential danger is predominant. Such re-­calibration for the purpose of maximizing sensitivity to the most important new information is not “habituation” or return to a “set point”—it can, for example, result in chronically heightened vigilance, risk-­sensitization, and action-­readiness when high risk levels are persistent. Reynolds and Berridge refer to this as “re-­tuning the affective keyboard” in response to changes in the environment. One way to see how this might be reflected in subjective well-­being is to think of the pattern of various important life events. Marriage, for example, is ac­com­ pan­ied by an anticipatory gain in subjective well-­being, a peak of excitement around the time of the wedding, and then a return to a level essentially the same as it was prior to the build-­up to marriage. For many, life is not so very different after marriage, and one must not re-­calibrate permanently upward lest one lose sensitivity to the variations one must be attentive to in other areas of one’s life. But the presence of a spouse is rarely “tuned out” or ignored, and threats to, or changes in, one’s marital condition, including most dramatically separation, divorce, or the death of a spouse, retain the capacity to deeply affect subjective well-­being (Diener et al. 2006). By contrast, life is seldom “essentially the same” when one has children—children need continuing attention and care in ways that change regularly over time and pose significant adaptive challenges for parents, in which a sense of having mastery over, or even control of, one’s life situation is likely to be lower. Having a child thus is somewhat correlated with higher levels of momentary negative affect and lower levels of overall satisfaction, but strongly correlated with persistently higher levels of subjective stress, especially among women--even if overall mood tends to return closer to previous levels (Kahneman and Deaton 2010). Such re-­calibration also is not “adaptive preference” in the usual sense. After childbirth, one does not simply “like what one gets” when unable to sleep through the night or calm a baby’s distress, or contend with a distraught and difficult teenager. What one “likes getting” has probably changed—one’s preferences are now likely to include, and give considerable weight to, the well-­being and interests of the child, usually to an extent one could not have anticipated—but these changes can make it more difficult, rather than less, to “get what one likes”. One isn’t “settling for less” from life as a way of making things easier and less stressful—one is likely trying to achieve more, with higher levels of effort and stress.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  21 The need to struggle in order to manage is not, however, incompatible with finding rewards, enjoyment, and satisfaction in life. If the struggle is a continuing, required condition, then the affective system must calibrate so that needing to struggle on behalf of oneself or one’s family is not in itself taken as a sign of failure, or as evidence that one is not meeting reasonable personal or social ex­pect­ations—and successful struggling is rewarded. Adaptive expectation and evaluation thus must be distinguished from adaptive preferences—less struggle for oneself and one’s family can remain preferred. But one’s expectations need to be calibrated to the actual prospects of one’s environment if one is to retain adequate sensitivity to the actual, incremental gains and losses needed to guide daily action and planning intelligently. Consider in this light the fact that, in surveys during the 2010s (Gallup World Poll; World Bank 2005–2011), the percentage of the population of Thailand and Bangladesh expressing that they are satisfied with their “standard of living” (74% and 78%, respectively) was only somewhat lower than in the US (83%), even though the per capita GDP of Thailand and Bangladesh was a small fraction of that in the US (approximately one-­sixth and one-­twentieth, respectively). Does this mean that people in Thailand and Bangladesh have adapted their preferences so that they no longer see their condition as lacking? No—in Thailand 60% and in Bangladesh 76% described themselves as “struggling”, and only 37% and 13%, respectively, described themselves as “thriving”. By contrast, in the US at the time, 38% described themselves as “struggling” and 58% described themselves as “thriving”. In less-­developed countries, where the only reasonable personal and social expectation for many in the population is that one will need to struggle in order to make ends meet, and where “achieving a high income” would be an uninformative standard for when one has succeeded in making ends meet, the daily accomplishment making ends meet through struggle should be felt as the success it is, and reward and motivation should continue to encourage this, not discourage it as continuing failure. Compare this to the reported experience of those individuals living in ­highly-­developed countries who are unemployed (though not retired). A study of cor­re­la­tions between demographic factors and reported subjective well-­being in Switzerland (Leu et al. 1997) found that, of a wide array of possible social conditions, unemployment had the most severe negative effect. For example, while being widowed without a new partner was associated with a reported average subjective well-­being of 8.16 (on a 10-­point scale; the average for Swiss citizens at the time was 8.30), being unemployed was associated with a reported average of 6.56 (see also Frey and Stutzer 2002). That a deeply negative departure from the social average can be found among the unemployed is not idiosyncratic to Switzerland, and has been across a range of highly-­developed countries. Why, even when countries like Switzerland provide a system of social support for the unemployed, would there be such a precipitous fall in subjective well-­being?

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

22  Peter Railton One explanation is that, in these countries, active employment is a strong social norm, and failure to succeed at being employed is a failure to meet social ex­pect­ations as well as a personal difficulty in managing. Frey and Stutzer (2002) found, for example, that unemployed workers in Switzerland suffer social exclusion, stigmatization, anxiety, and depression—with these negative effects particularly pronounced for those unemployed for a long period. A study in the US found that worry, sadness, and stress increased linearly with length of unemployment, while happiness, enjoyment, and even laughter fell (Kahneman and Deaton  2010). When struggling-­to-­manage-­materially is the norm, managing-­ materially-­ through-struggle is success. But when active employment is the norm, unemployment, even if one is not struggling materially, is failure. Lucas et al. (2004) found in a 15-­year longitudinal study of Germany that individuals who have lived through an extended period of unemployment do not simply return to their pre-­unemployment level of reported subjective well-­being once they return to work, and tend to react more strongly to ­unemployment if it occurs again—signs of psychological sen­si­tiza­tion rather than habituation (Janicki-­Deverts et al. 2008). Contrast this with the effectiveness of social support in securing a relatively high level of reported subjective well-­being in another “persistent” condition—aging and age-­related declines in health and ability. Aging and conditions typically arising from aging, while they can challenge one’s capacity to manage, are not departures from normative expectations—these are, instead, conditions we all expect to encounter in the natural course of events. When adequate social support is provided, then even though ill health takes some toll on subjective well-­being, the overall well-­being of the older population can be comparable to that of the total population—whereas, as we saw, the absence of such support is correlated with an older population suffering a decline in reported subjective well-­being year-­by-­year (Deaton 2007). What we find, then, are not forms of simple “habituation” or “liking what one gets”, but affective responses that re-­calibrate to circumstances to enable in­di­vid­ uals to stay sensitive to what constitutes, in their circumstances, success or failure— whether in material terms or in terms of meeting personal or social norms. And social expectations and norms play an especially important role. Across the globe, marginalized and stigmatized groups are found to have lower levels of average reported subjective well-­being and higher levels of stress, with associated poorer health outcomes (Bollini et al 2009; Pascoe and Smart Richman, 2009; Paradies et al. 2015; Sattler and Lemke 2019). This is cruel, but it is also rooted in function: being less-­than-­readily-­accepted in the larger society requires one to have persistently higher vigilance in one’s social interactions with that society. By contrast, the residual long-­term impact upon average reported subjective well-­being from some negative life events that are part of a socially-­typical life—for example, losing a spouse—may be quite small.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  23 If we wish for our children to “live better than we do”, then providing greater security of employment and other forms of social support (Easterlin and Switek 2014), and lessening stigmatization and marginalization of the unemployed, should be high on our list—as should reducing social stigmatization and marginalization generally. This is not the same as producing more economic growth or more jobs as such, since gains from economic growth often do not extend equally across the whole population, and the jobs created may be of a kind that are themselves stigmatized and those occupying them may come overwhelmingly from marginalized groups.9 Even relatively high rates of employment across society are compatible with a mix of forms of employment in which an increasing number of jobs lead to economic precariousness and inability to escape personal or family poverty—as we have seen in recent years in the US, where reported levels of ­life-­satisfaction on average fell between 2015 and 2018, despite growth in GDP and declining unemployment (Kalleberg 2018; Helliwell et al. 2019). Focusing on economic growth or employment but ignoring these forms of stigmatization, marginalization, unequal distribution, and precariousness is likely to produce at best transient gains in average reported subjective well-­being. Indeed, growing evidence suggests that our children are already “living less well” than we do (Newport 2015; Helliwell et al. 2019). Understanding the relation of economic growth to average reported subjective well-­being thus requires a complex analysis. For example, while China and India both had fairly high rates of economic growth from 2005–2008 to 2015–2018, reported levels of life-­satisfaction rose in China but fell by a significantly larger amount in India (Helliwell et al.  2019; see also Otis 2017). If one plots average reported subjective well-­being against per capita GDP across nations, one sees a scattered distribution of nations, but the trend line shows diminishing marginal return from per capita GDP, with gains to indexed subjective well-­being becoming increasingly flat above a level of reasonable sufficiency, about $15,000 at the time (2005–2008), a level about equal to Taiwan or the Czech Republic, or one-­third of the US per capita GDP (Figure 1.1).10 It appears from Figure 1.1 that the nineteenth-­century Utilitarians were right that money has diminishing marginal utility and that the actual priority should be improving the income or wealth of those who are less well-­off, where marginal gains are much greater, rather than increasing economic growth or wealth overall. Unfortunately, the opposite can be observed in the US and a number of European countries, where income and wealth growth in recent decades has 9  For example, foreigners living in Switzerland are heavily over-­represented in less socially valorized jobs, and have an average reported level of subjective well-­being of 7.62, as opposed to 8.30 for Swiss citizens (Leu et al. 1997). 10  Well-­being indices are constructed from individuals’ expressed satisfaction with various features of their lives, which can include standard of living, health, social connectedness, and future security.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

24  Peter Railton 3.5 3 2.5 2 1.5 1 0.5 0 –0.5

0

5,000

10,000

15,000

20,000

25,000

30,000

35,000

40,000

Index of subjective well-being combines reported hedonic state and satisfaction with various aspects of life, using mean results from a large number of surveys.

Figure 1.1  Average reported subjective well-­being index vs. per capita GDP, multinational sample, 2005–2008 Source: Data from Ingelhart et al. (2014).

disproportionately gone to households in the top percentages. Though inequality had begun to decline recently (Horowitz et al. 2020), the current pandemic may well erase these gains. If we look at correlations between income and reported subjective well-­being within a country, we see a very similar curve. Figure 1.2 suggests that the large gains in real income of the upper income groups in the US in the period 1976–2016, while middle- and lower-­income groups saw little or no gain, could be expected to have contributed little to increase overall well-­being. And, indeed, overall average reported subjective well-­being in the US showed no net upward trend during these years. If we are concerned with the subjective component of well-­being, it would seem that channeling more income toward those who are already above a level of material comfort squanders material gains that could instead secure a much greater effect if channeled to lower-­income groups. If we compare levels of income equality among highly-­developed societies— e.g., in 2014–2016 the US had a Gini coefficient of 41.5, while Australia stood at 35.8 and Switzerland at 32.3—we can ask how the overall trend of reported sub­ ject­ive well-­being with income varies from Figure 1.2 in the more egalitarian societies. Although variations in the components of subjective well-­being measured do not make for strict comparability, it is notable that the trend of reported subjective well-­being with respect to income is much “flatter” in Australia and Switzerland than in the US, in that those in the lower income groups express an average level of subjective well-­being much closer to the levels reported by middleor upper-­income groups (Figures 1.3 and 1.4). Figures 1.3 and 1.4 suggest that, in prosperous countries like Australia and Switzerland, the social policies and political and economic conditions that are

A Convenient Truth?  25 0.6 0.4 0.2 0 –0.2 –0.4 –0.6 –0.8 –1 –1.2

0

20,000

40,000

60,000

80,000

100,000 120,000 140,000

160,000

Index of subjective well-being combines reported hedonic value and satisfaction with various aspects of life, using mean results from a large number of surveys.

Figure 1.2  Average reported subjective well-being index vs. household income, US sample, 1972–2006 Source: Data from General Social Survey

78 77 76 75 74 73 72 71 70 69 68

150,000

Average reported subjective well-being (indexed) vs. household income, Australia, 2006

Figure 1.3  Average reported subjective well-­being index vs. household income, Australia, 2006 Source: Data from Household, Income and Labor Dynamics in Australia Survey.

associated with a more egalitarian income distribution also are associated with better life-­conditions for those in lower-­income groups, perhaps because such societies often provide more adequate health care, family support, schools, or other benefits that, in more unequal countries, one needs income to obtain.11 11  Effects of income inequality on subjective well-­being in developing countries, where resources may not exist for the provision of adequate social services and support, are more complex (Reyes-­ Garcia et al. 2019).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

26  Peter Railton 8.6 8.5 8.4 8.3 8.2 8.1 8 7.9 7.8 7.7

Lowest 1/5th

Next 1/5th

Next 1/5th

Next 1/5th

Highest 1/5th

Average reported subjective well-being vs. income quintile, Switzerland, 1995

Figure 1.4  Average reported subjective well-­being vs. income quintile, Switzerland, 1995 Source: Data from Frey and Stutzer (2002)

Flattening need not come at the expense of overall subjective well-­being, and ­levels of average reported subjective well-­being in Australia and Switzerland are very similar to the US. Moreover, the prosperous countries that have moved furthest in the direction of income equality and provision of health, family support, and other social services, the Scandinavian countries, have average reported sub­ ject­ive well-­being consistently higher than the US, despite lower levels of GDP per capita.12 And in all of these countries, the top income groups experience very high average reported subjective well-­being. There is not, then, the supposedly stark trade-­off between, on the one hand, overall social subjective well-­being or the subjective well-­being of more prosperous groups, and, on the other, more egalitarian income distribution and higher subjective well-­being for lower income groups. “Leveling down” need not be the inevitable consequence of more egalitarian policies and conditions. Indeed, a large-­scale panel data study of Germany from 2005 to 2013 found that the average reported subjective well-­being of regions with higher rates of poverty was lower overall, and lower for upper- as well as lower-­income individuals. Welsch and Biermann (2019) conclude that reducing poverty is a problem not only of social justice, but of allocative efficiency in raising well-­being society-­wide, now and in the future. For example, comparing regions within the US, Pickett and Wilkinson found a negative association between regional income inequality and UNICEF measures of the well-­being of children, even controlling for the average income in the regions. They conclude that improvements in child well-­being in general could be enhanced more effectively by reducing income inequality or reducing poverty than by increased average 12  Gini coefficient during the same years: Sweden, 29.2; Denmark, 28.2; Iceland, 27.8; Norway, 27.5; Finland, 27.1.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  27 income (Pickett and Wilkinson 2007). And a wide range of research indicates that improving the early childhood environment is a highly efficient social investment for increasing quality of life (Knudsen et al. 2006). A number of personal, social, and political factors other than level of economic development can have associations with overall average reported subjective well-­being at least as important as money income (Helliwell et al. 2017). Returning to cross-­national comparisons, Figure 1.1, which shows only the general trend line, masks a wide variation in GDP per capita among countries with similar levels of reported subjective well-­being. For example, during the time period of Figure 1.1 (2005–2008), the Lain American countries of Columbia, Mexico, Costa Rica, and Venezuela had average levels of reported subjective well-­being comparable to Finland, the UK, Italy, Australia, and Belgium, despite having per capita GDPs less than one-­half of these nations. And at the same time, the Eastern European countries of Hungary, Slovenia, Poland, East Germany, and the Czech Republic, had average levels of reported subjective well-­being much lower than these Latin American countries, despite their greater per capita GDP. The relatively high levels of average reported subjective well-­ ­ being in Latin America are often explained in terms of the denser social and family connections and support re­la­ tionships and lower levels of social isolation—none of which mere income can procure, and all of which are independently known to make considerable contributions to sub­ject­ive well-­being. In a host of measures of involvement with family and ability to rely upon family resources in times of need or difficulty, Latin American countries dominate the world rankings (Rojas 2018). Even as level of per capita GDP does not tell us which individuals or households are able to achieve a reasonable sufficiency in income, it may tell us even less about which individuals or households are able to enjoy a reasonable level of support and security in such central areas of life as housing, food, care for children, care for the ill or elderly, or protection against economic precariousness. Social support, security, and protection from precariousness may also play a primary role in explaining why certain Scandinavian countries have long had distinctively higher average reported subjective well-­being than other highly-­developed countries, including those with higher GDP per capita and higher rates of growth of GDP, such as the US (Kalleberg 2018; World Bank). A recent study found, for example, that Scandinavian countries may owe their higher average reported subjective well-­being primarily to the higher reported average subjective well-­being of the least prosperous (cf. Biswas-­Diener et al. 2010). We see this reflected as well in the proportion of a nation’s population reporting themselves as “struggling” or “thriving”. In the US in the 2010s, we saw, 38% report themselves as “struggling” and 58% as “thriving”. In Denmark at the same time, 24% reported themselves as “struggling” and 74% as “thriving” (Gallup World Poll). We can also look at associations of subjective well-­being with varied personal, social, and political conditions within a country. In the US, for example, if we

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

28  Peter Railton compare those with low as opposed to middling or high income, we find higher levels of reported stress as well as higher levels of blood cortisol, along with poorer long-­term health outcomes (Cohen et al. 2006; Lupien et al. 2009; Kim et al. 2013). Kahneman and Deaton (2010) used a large US sample to compare the effects of various life conditions on elements of subjective well-­being in comparison to income. They found that such factors as being the care-­giver for an elderly or disabled family member, suffering a chronic health condition, and being alone, had stronger associations with negative affect, stress, and reduced life-­satisfaction than having an income of $48,000 in 2008–2009. Because formal and informal social institutions can have a significant effect upon whether individuals end up being solely responsible for the care of an elderly person, or lacking adequate health care, or living cut off from social contact, changes in social support can make differences in people’s lives comparable to, or greater than, significant increases in income. People do not simply “habituate” to these life-­conditions, or adapt their preferences to “like what they get”. If we wish to reduce these serious sources of subjective ill-­being we would do well to invest existing economic resources more heavily in increasing formal and informal social support than in generating economic growth without reference to such distributive and social concerns. A final, and potentially important set of within-­country factors concern the political and civil life of the society. On a worldwide basis, there is a remarkably robust, linearly increasing relationship between a nation’s civil rights, political democracy, social freedoms, and absence of corruption or rule of law, and its average reported subjective well-­being (Inglehart and Klingemann 2000; Frey and Stutzer 2002). This contrasts sharply with the diminishing marginal increase of average reported subjective well-­being with respect to per capita GDP (Diener et al. 1995). The causes of this relationship are difficult to disentangle. The capacity of a society to provide such rights, liberties, and effective institutional practices is an important social accomplishment, and will reflect many contributory features. Among these features would appear to be the level of subjective well-­being itself, since there is evidence that rising popular discontent is associated with increasing authoritarian tendencies. For example, the period 2006–2020—which includes the Global Financial Crisis and its aftermath—has seen decline in civil liberties and political democracy in most regions of the world (Freedom House 2020). Over the same years the population-­weighted averages for reported positive affect and life-­satisfaction have declined, while those for negative affect have risen sharply (Helliwell et al. 2019). A dynamic correlation is not, of course, causation, and multivariate modeling suggests that responsiveness and effectiveness of political institutions, rather than the bare existence of democratic civil rights, may be the primary factors in the association (Ingelhart and Klingemann 2000). In an effort to achieve comparability, Frey and Stutzer (2002) studied the various cantons within Switzerland, with their different constitutional forms of direct popular governance, and found a positive association of democratic forms with

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  29 subjective well-­being; Stadelmann-­Steffen and Vatter (2012) looked more closely at the Swiss cantons and concluded that rates of actual participation in these democratic forms, rather than their constitutional presence as such, accounted for most of the variation. And cross-­national studies in Europe and Latin America have found that confidence in national institutions, economic as well as political, is positively correlated with subjective well-­being (Hudson 2006; Machia and Plagnol 2019). In any given country and in any given region within a country, there will be many factors influencing both subjective well-­being and civil and political life, and these variables will have complex interrelations among themselves. One important change in the US in recent decades, for example, has been the rapid rise of economic inequality. Some analysts have suggested that the growth of in­equal­ity in the US has affected both levels of perceived well-­being and the ef­fect­ ive­ ness and freedom from corruption of US political institutions. For ex­ample, in the US since World War II, there has been a tight correlation between in­equal­ity, as measured by the Gini coefficient, and degree of political partisanship (Voteview 2016), which in turn can reduce the responsiveness, effectiveness, and fairness of governance institutions in relation to widespread public needs and aspirations. Such effects are of particular importance in the context of thinking about climate change, since they can dramatically reduce the capacity of polities to take appropriate actions in response to the challenges of global climate change. In sum, we have seen a range of factors that are significantly correlated with higher reported subjective well-­being, beyond per capita GDP—factors that can help explain the significant variation in average reported subjective well-­being across countries with essentially similar levels of per capita GDP, or essentially similar levels of average reported subjective well-­being across countries with significantly different per capita GDP. Attention to these factors could help our children to “live better than we do” even if we do not bequeath to them a higher material “standard of living”, with its attendant burden upon the environment. And even in much less well-­developed countries, these factors could help make it the case that children around the globe “live better than their parents” without explosive material growth that would make contending with global climate change yet more difficult. We might not, then, face quite the hard choice often imagined between promoting lives people find more rewarding and curbing the destruction of the environment.

6.  A Recent Critique Recently, the picture presented above of the relationship between income and reported subjective well-­being—which is sometimes known as the “Easterlin Paradox” in honor of the economist who gave it prominence (see Easterlin

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

30  Peter Railton 1974)—has been vigorously challenged. The economists Betsy Stevenson and Justin Wolfers (2008, 2013) point out that decreasing marginal utility gains as a function of absolute money income is compatible with continuing increase in utility as a function of log(absolute money income), and large-­scale Gallup polling in the US and worldwide appears to bear this out for the life-­satisfaction component of subjective well-­being (Kahneman and Deaton 2010). It has now become something of a new orthodoxy that “money does buy happiness”. How significant is this argument, and how does it bear on our discussion here? Recall that the life-­satisfaction portion of subjective well-­being is measured by responses to a “Cantril Ladder” poll. Typical instructions for such a poll are that subjects are asked to place themselves on a ladder in which the lowest step is “Worst possible life for me” and the highest step is “Best possible life for me” (Gallup World Poll). This kind of question, which frames the individual’s answer in terms of a representation that emphasizes a hierarchy of lives, tends to elicit answers that are sensitive to the individual’s estimate of where they stand in the social hierarchy, even if they are being explicitly asked how they themselves “feel” about where they stand (Diener et al. 2010). It is perhaps unsurprising, then, that reported life-­satisfaction as measured in this way continues to rise with income, even beyond the point at which individuals report, on average, any improved quality of affective experience. In the most widely studied large-­scale database for the US, Kahneman and Deaton (2010) found that, as household income rose from $0 to $80,000, respondents reported higher levels of positive affective ex­peri­ence and lower levels of negative affective experience, though at a diminishing rate of increase such that, above $80,000, reported positive affect and negative affect remain virtually constant. Thus, while respondents continued to place themselves higher on the rungs on the ladder of life-­satisfaction as household income moves upward beyond $80,000, they do not register this in any difference in the positive or negative experiential quality of their affective lives. It would seem that there is an upper limit on how much money can do to enhance the experiential quality of one’s life, but not because one hits a ceiling of constant positive affect and freedom from negative affect—life always remains a mixture of good and bad experiences. Of life’s problems and challenges, only so many can be solved with money, and among those that can’t are some of the most important sources of positive or negative affect in life: developing loving relations with one’s partner or children, making and keeping lasting friendships, earning the respect of one’s co-­workers or community, gaining a sense of meaning from one’s life, resolving conflicts with one’s parents, avoiding substance or mental ­illness, the loss of someone close, and so on. Indeed, Quoidbach et al. (2010) find evidence that possessing large amounts of money can impair enjoyment of life’s non-­monetary pleasures. To say that reported ladder-­ranking correlates with log(money income), even at higher income levels, is really just another way of saying that ladder-­ranking

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  31 shows a diminishing marginal return from money income. For example, in the US sample Kahneman and Deaton (2010) studied, a household income of approximately $50,000 was associated with an average reported ladder position one rung higher than the average reported position of those with a household income of $15,000—so $35,000 made “one rung” of difference, which could be expected to be a meaningful difference in life satisfaction and subjective well-­being. But the average reported life satisfaction of someone with a household income of $160,000 was only half a rung higher than that for $50,000—so an income gain three times larger made half as much difference to average reported life satisfaction. Because it is a logarithmic scale, continuing to move upward will require income to double, then redouble, then redouble, etc. for each fractional gain. This is a spectacularly expensive way to raise the life-­satisfaction component of subjective well-­being by diminishing increments—while having no measurable effect on positive or negative experience. One can ask what sort of social policy would place serious weight on such gains, when much larger gains can be made at much lower cost for those with more modest incomes—especially since ladder rankings appear to be so heavily influenced by the individual’s assessment of where he or she stands in the economic or social hierarchy, which of course will tend to go up with income, even at the highest levels. And we’ve seen that exacerbating income inequality may have negative effects on subjective well-­being more generally through a number of personal, social, and political pathways, some of which seem to lower well-­being even for those with higher incomes (Alesina et al. 2004; Welsch and Biermann  2019). So policies that channel an exponentially larger share of the nation’s income to those already at the highest levels of income— policies akin to those that have prevailed in the US over recent decades—might actually reduce reported subjective well-­being quite widely in society. Policies favoring equality are sometimes accused of “leveling down”, but policies favoring increasing inequality in the name of growth can bring about a general lowering of levels of well-­being. From the standpoint of social policy, the implication would seem to be that, if one is trying to increase overall subjective well-­being, it is increasingly expensive to have even a small impact on the subjective well-­being of those already well above the level of reasonable sufficiency through income alone. Yet, as we have seen, there are other factors that likely could make a significant difference to individual lives, including ways of reducing precariousness and unemployment, increasing social connectedness, providing better-­targeted social support, improving political institutions, and reducing social stigmatization and corruption.13 13 See also Easterlin (2010, 2013) on the import of data obtained in recent, large-scale global s­ amples. Measures and methods in this area are fraught with controversy, but in his studies he finds no positive association between rate of economic growth without regard to distribution and rate of change in average reported life satisfaction, even in developing countries. Overall, he concludes that if society has the goal of improving the felt quality of life for individuals, full employment and an

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

32  Peter Railton There are already countries—not only highly-­developed countries, but countries at lower overall levels of economic development—whose populations report ­levels of subjective well-­being that would seem to confirm this. If we will but pay attention.

7.  Subjective Well-­Being and Climate Let’s look a bit closer at the possible relationship between thriving in life and burden upon climate. Consider that among the countries where individuals report the highest levels of thriving are the Scandinavian nations, along with Costa Rica (World Database of Happiness). At the same time, notice that the per capita carbon footprints of these countries are markedly lower than the US (Table 1.1). Clearly there are far too many variables in play to compare the US directly with, say, Denmark, Sweden, or Costa Rica—if for no other reasons than questions of size and diversity of population. However, consider that such differentiating factors as size and composition of population were factors already in place in 1980, and this may help explain why per capita carbon emissions in these three countries were much lower than the US at that time. But notice that Denmark and Sweden also effected greater percentage reductions in carbon footprint in the ensuing decades, while retaining higher levels of thriving and lower levels of per capita GDP than the US. It appears to be possible to substitute higher levels of   “production” and “consumption” of non-material goods for higher levels of production and consumption of material goods while retaining high levels of average reported subjective well-­being. Table 1.1.  Per capita carbon footprints by nation and year (metric tons)  

1980

2018

United States Denmark Finland Iceland Norway Sweden Costa Rica

20.8 11.8 12.2 8.2 9.3 8.6 1.0

16.1 5.8 8.8 12.1 9.4 4.5 1.8

Source: World Bank; Crippa et al. 2019.

effective set of social supports will do more than economic growth as such, and that these measures may be practicable and affordable through good institutional design and practice, even in less well-­ developed societies.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  33 Positive psychology is the study of the nature and sources of human well-­being, and one widely-­used model is Seligman’s PERMA model (2018), according to which the most significant components of well-­being are Positive affective ex­peri­ ences, Engagement in work and social relations, a network of well-­developed Relationships, activities or work that are experienced as Meaningful, and success in Accomplishment of one’s goals. The elements of PERMA are not inherently expensive, demanding of carbon-­based resources, or exclusionary (i.e., restricted by their nature to a small portion of the population, as occupying a high relative position in society would be). Neither are they beyond the capacity of education, employment policies, social support structures, and effective governance to affect. While material goods can contribute to the realization of these elements of a life, the level of material goods required in any given case need not be high for the realization of the component element to be high. Indeed, pursuing the elements of PERMA in one’s life may often mean de-­emphasizing the importance of income or wealth in deciding what one will seek or do. Of course, to the extent that individuals have materialist or positional values and goals (e.g., possession-­defined success, happiness as acquisition, a focus on relative economic standing as a measure of life-­success), then material success or social position will play a significant role in what they take to have meaning or constitute accomplishment or important relationships with others. But there is evidence that giving priority to material values is associated with lower levels of subjective well-­being at all levels of income (cf. Kasser 2003), and that individual consumption of material goods, as opposed to experiential and shared goods, is associated with lower levels of satisfaction both at the time of consumption and in retrospect (Van Boven  2005; Gilovich and Kumar  2015; Brown et al. 2016). Accepting a value framework that assesses “standard of living” in material terms is not only a risk factor for the environment, it appears to be a risk factor with regard for individual and social well-­being. The functioning of the affective system in informing and guiding can help us understand why the spectacular creation of material wealth in the US and Western Europe in the post-­World War II period seems to have done so little to improve the experienced quality of life in these countries—not because subjective well-­being is an insensitive indicator of how well people’s lives are going, but because how well people’s lives are going, once they have reached a reasonable level of economic sufficiency and security, is more a matter of PERMA and social and political institutions than the accumulation of possessions. Indeed, we are beginning to observe a decline in subjective well-­being in the US even during recent periods of economic growth (Helliwell et al. 2019). While post-­World War II economic growth in the US was fairly widely distributed across society for the first three decades, economic growth since the 1980s has led to greatly increased economic inequality, a change that has been associated with higher levels of economic and social precariousness, falling levels of trust in

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

34  Peter Railton government, institutions and one another, political pessimism, and, as we saw earlier, markedly increased political partisanship, contributing to a decline in the responsiveness and efficiacy of governance institutions (Voteview 2016; Rainie and Perrin 2019; Ward 2019). US economic growth in recent decades, then, may have exerted a negative effect on its capacity to respond to environmental challenges, even as other social and cultural changes have promoted awareness of these challenges. We should understand growth that takes this form not only as threatening to the economic security and prospects of individuals and families, but also as threatening to such core elements of subjective well-­being as social connectedness, engagement, accomplishment, and meaningfulness. As Durkheim (1857/1951) argued, we should expect under such circumstances to see higher rates of alienation and suicide. And suicide rates have risen steadily in the US over recent decades in comparison with Western Europe (National Center for Health Statistics 2017), particularly within the white, non-­Hispanic population in the US. If we group suicide together with other “deaths of despair” associated with social and economic marginalization (Case and Deaton  2017)—e.g., drug overdose and alcohol toxicity—the upward trend of rates among middle-­aged white, non-­Hispanic men and women in the US is especially striking, more than doubling between1990 and 2015. At the same time, incidence of stress-­related psychological difficulties such as anxiety, and of despair-­related psychological difficulties such as depression, has risen among the youth in the US (Case and Deaton 2017). The overall rate of anxiety disorder among college and university students in the US rose by 48 percent between 2008 and 2014 (Scheffler et al. 2018), and the precollege population experienced an increase in anxiety and suicide, and a sharp decline in subjective well-­being, from 2006 to 2017 (Helliwell et al. 2019). So we have a second “environmental problem” to solve as part of trying to solve the first: the existence in countries making a massive contribution to global climate change, and the US especially, of economic and social inequality and precariousness that tend to make materialist concerns an urgent matter for increasingly many individuals and families, to erode public institutions, and to undermine the realization of important social goods and intrinsic values.14 We of course should not wait for improvement in this second environmental problem before tackling the problem of global climate impact, but it seems likely that, without some progress on this second environmental problem, the conditions for broadly-­based and politically-­effective participation in a movement for environmental change will continue to deteriorate. Combatting social and economic inequality and precariousness in the developed as well as developing world, and lessening the threat 14  For example, an experiment in guaranteeing minimum incomes in the US, while controversial to interpret (Hum and Simpson 1993), was seen by some interpreters as indicating that individuals on average were still eager to work, and were able to search longer for, and find, work that better suited their skills and interests. A recent Finnish experiment showed, as a preliminary result: increased trust in others and in the political and legal system; gains in confidence about the future as well as sub­ject­ ive well-­being and the reduction of stress; and less tendency to lose interest in various things previously considered enjoyable (Kangas et al. 2019).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  35 this poses to meaningful democratic political participation and social solidarity, may be vital for sustaining the long-­term commitment that working together for solutions to the global climate change will require. Importantly, what we have observed about subjective well-­being suggests that widespread participation in an effective common project to benefit the environment and future generations could itself make a contribution to subjective well-­being. The elements of PERMA are not new to thinking about what objective well-­being consists in, neither is the way they connect individual and social goods. They echo eudaimonistic ideas handed down from the ancient Greeks about what kinds of lives and polities are suitable to the distinctive nature of humans. What’s new is that social psychology and large-­scale survey research have provided evidence that these are not just philosophical ideals, but key contributors to how people actually everyday experience life. By contrast, preserving the existing way of life and social and political tendencies in the developed countries, and in the US in particular, will not only further imperil the world environment, but will likely fail to ensure that “our children will live better than we do”—at least, if we are concerned with the lived quality of their experience. And while it would be absurd and cruel to deny the importance of economic growth for improving the lives of disadvantaged populations the world over, it would also be absurd and cruel to ignore evidence that higher per capita GDP and economic growth may also come at the expense of “living better” and may lead to living worse—even while other measures, not demanding high levels of economic growth or increased environmental burdens, could lead to substantial gains in subjective well-­being. That is the “convenient truth” about subjective well-­being. It is a material ideology of happiness, not a matter of hard-­headed analysis, to claim that there is a necessary trade-­off between promoting a better environment and promoting the quality of life for the great majority of the population. Unbalanced development that prioritizes growth over achieving widespread sufficiency and security may be an unhealthy diet for people and polities as well as the planet. So, while there are strong interests arrayed on behalf of fostering inequality as an engine of economic growth and encouraging a material conception of the “standard of living”, “living better”, or “the American Dream”, we need not concede to these interests the idea that this is a “realistic” picture of what makes people happy. A commitment to enhancing the subjective well-­being of one’s own children and the world’s children may be an ally, not an obstacle, in making the political, social, and economic changes necessary to reduce the harms and injustices of climate change. But keeping this in focus will require the evidence, understanding, and imagination to countervail against ideologically powerful conceptions of the nature and sources of well-­being itself.15

15  I would like to thank the Editors of this volume and David Brink, Cody Gibson, Michael Morck, and Ira Richardson for very helpful comments on previous drafts. I am also indebted to Roy Baumeister, Kent Berridge, Daniel Haybron, Martin Seligman, Chandra Sripada, and Valerie Tiberius for conversations about subjective well-­being and happiness over the years.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

36  Peter Railton

References Adams, W. T. (1931). The American Epic. Boston: Little, Brown, and Company. Alesina, A., R. D. Tella, and R. MacCulloch (2004). “Inequality and Happiness: Are Europeans and Americans Different?” Journal of Public Economics 88: 2009–2042. Biswas-Diener, R., J. Vitterson, and E. Diener (2010). “The Danish Effect: Beginning to Explain High Well-Being in Denmark.” Social Indicators Research 97: 229–246. Bollini, P., S. Pampallona, P. Wanner, and B. Kupelnick (2009). “Pregnancy Outcome of Migrant Women and Integration Policy: A Systematic Overview of the International Literature.” Social Science and Medicine 68: 452–461. Brickman, P., D. Coates, and R. Janoff-Bulman (1978). “Lottery Winners and Accident Victims: Is Happiness Relative?” Journal of Personality and Social Psychology 36: 917–927. Brown, K. W., T. Kasser, R. M. Ryan, and J. Konow (2016). “Materialism, Spending, and Affect: An Event-Sampling Study of Marketplace Behavior and Its Affective Costs.” Journal of Happiness Studies 17: 2277–2292. Case, A. and A.  Deaton (2017). “Mortality and Morbidity in the 21st Century.” Brookings Papers on Economic Activity, Spring: 397–476. Clark, A. E., P. Frijters, and M. A. Shields (2008). “Relative Income, Happiness, and Utility: An Explanation for the Easterlin Paradox and Other Puzzles.” Journal of Economic Literature 46: 95–144. Cohen, S., W. J. Doyle, and A. Baum (2006). “Socioeconomic Status Is Associated with Stress Hormones.” Psychosomatic Medicine 68: 414–420. Craig, A.  D. (2009). “How Do You Feel—Now? The Anterior Insula and Human Awareness.” Nature Reviews: Neuroscience 10: 59–70. Crippa, M., G. Oreggioni, D. Guizzardi, M. Muntean, E. Schaaf, E. Lo Vullo, E. Solazzo, F. Monforti-Ferrario, J. G. J. Olivier, and E. Vignati (2019). “Fossil CO2 and GHG Emissions of All World Countries.” JCR Science for Policy Report, European Commission. Deaton, A. (2007). “Income, Aging, Health and Wellbeing around the World.” Center for Health and Wellbeing, Princeton, and National Bureau of Economic Research. Deaton, A. (2013). The Great Escape: Health, Wealth, and the Origins of Inequality. Princeton: Princeton University Press. Diano, M., A.  Celeghin, A.  Bagnis, and M.  Tamietto (2017). “Amygdala Response to  Emotional Stimuli without Awareness: Facts and Interpretations.” Frontiers in Psychology 7: article 2029. Diener, E., M.  Diener, and C.  Diener (1995). “Factors Predicting Subjective WellBeing in Nations.” Journal of Personality and Social Psychology 69: 851–864. Diener, E., R. E. Lucas, and C. N. Scollon (2006). “Beyond the Hedonic Treadmill: Revising the Adaptation Theory of Well-Being.” American Psychologist 61: 305–314.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  37 Diener, E., W.  Ng, J.  Harter, and R.  Arora (2010). “Wealth and Happiness across the World: Material Prosperity Predicts Life Evaluation, Whereas Psycho-Social Prosperity Predicts Positive Feeling.” Interpersonal Relations and Group Processes 99: 52–61. Diener, E., R. Ingelhart, and L. Tay (2013). “Theory and Validity of Life Satisfaction Scales.” Social Indicators Research 112: 497–527. Duncan, S. and L. F. Barrett (2007). “Affect Is a Form of Cognition: A Neurobiological Analysis.” Cognition and Emotion 21: 1184–1211. Durkheim, E. (1857/1951). Suicide: A Sociological Study, trans. J.  A.  Spaulding and G. Simpson, ed. G. Simpson. London: Routledge & Kegan Paul. Easterlin, R.  A. (1974). “Does Economic Growth Improve the Human Lot?,” in Nations and Households in Economic Growth: Essays in Honor of Moses Abramovitz, ed. P. A. David and M. S. Reder. New York: Academic Press, 89–125. Easterlin, R.  A. (2010). Happiness, Growth, and the Life Cycle. New York: Oxford University Press. Easterlin, R. A. (2013). “Happiness, Growth, and Public Policy.” Economic Inquiry 51: 1–15. Easterlin, R. A. and M. Switek (2014). “Set Point Theory and Public Policy,” in Stability of Happiness: Theories and Evidence on Whether Happiness Can Change, ed. K. M. Sheldon and R. E. Lucas. Amsterdam: Elsevier, 201–217. Eid, M. (2008). “Measuring the Immeasurable: Psychometric Modeling of Subjective Well-Being Data,” in The Science of Subjective Well-Being, ed. M. Eid and R. J. Larsen. New York: Guilford Press, 141–167. Eid, M. and E. Diener (2004). “Global Judgments of Subjective Well-Being: Situational Variation and Long-Term Stability.” Social Indicators Research 65: 245–277. Elliott, T.  M., J.  Christensen-Dalsgaard, and D.  B.  Kelley (2011). “Temporally Selective Processing of Communication Signals by Midbrain Neurons.” Journal of Neurophysiology 105: 1620–1632. Eurobarometer. https://ec.europa.eu/commfrontoffice/publicopinion/index.cfm Federal Reserve 2020. https://fred.stlouisfed.org/release/tables?rid=53&eid=13146#s nid=13148. Freedom House (2020). Freedom in the World: 2020 Report. https://freedomhouse. org/report/freedom-world/2020/leaderless-struggle-democracy Frey, B.  S. and A.  Stutzer (2002). Happiness and Economics. Princeton: Princeton University Press. Gallup World Poll. https://www.gallup.com/analytics/318875/global-research.aspx General Social Survey. https://gss.norc.org/ Gilovich, T. and A. Kumar (2015). “We’ll Always Have Paris: The Hedonic Payoff from Experiential and Material Investments.” Advances in Experimental Social Psychology 51: 147–187.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

38  Peter Railton Heady, B. and R. Muffels (2017). “Towards a Theory of Medium Term Life Satisfaction: Similar Results for Australia, Britain and Germany.” Social Indicators Research 134: 359–384. Helliwell, J. F., R. Layard, and J. D. Sachs (2017). World Happiness Report 2017. https:// worldhappiness.report/ed/2017/. Helliwell, J. F., R. Layard, and J. D. Sachs (2019). World Happiness Report 2019. https:// worldhappiness.report/ed/2019/. Horowitz, J. M., R. Ingielnik, and R. Kochhar (2020). “Trends in Income and Wealth Inequality.” Pew Research Center Report, January 9. https://www.pewsocialtrends. org/2020/01/09/trends-in-income-and-wealth-inequality/#upper-income-householdshave-seen-more-rapid-growth-in-income-in-recent-decades. Household, Income and Labor Dynamics in Australia Survey: https://melbourneinstitute.unimelb.edu.au/hilda Hudson, J. (2006). “Institutional Trust and Subjective Well-Being across the EU.” Kylos 59: 43–62. Hum, D. and W.  Simpson (1993). “Economic Response to a Guaranteed Annual Income: Experience from Canada and the United States.” Journal of Labor Economics 11: S263–S296. Inglehart, R. and H.-D.  Klingemann (2000). “Genes, Culture, Democracy, and Happiness,” in Culture and Subjective Well-Being, ed. E.  Diener and E.  M.  Suh. Cambridge, MA: MIT Press, 165–183. Inglehart, R., R. Foa, E. Ponarin, and C. Welzel (2013). “Understanding the Russian Malaise: The Collapse and Recovery of Subjective Well-Being in Post-Communist Russia.” National Research University Higher School of Economics, Working Paper BRP 32/SOC/2013. Inglehart, R., C.  Haerpfer, A.  Moreno, C.  Welzel, K.  Kizilova, J.  Diez-Medrano, M. Lagos, P. Norris, E. Ponarin, B. Puranen et al. (eds.) (2014). World Values Survey: All Rounds—Country-Pooled Datafile Version. http://www.worldvaluessurvey.org/ WVSDocumentationWVL.jsp. Madrid: JD Systems Institute. Janicki-Deverts, D., S. Cohen, K. A. Matthews, and M. R. Cullen (2008). “A History of Unemployment Predicts Future Elevations in C-Reactive Protein among Male Participants in the Coronary Artery Risk Development in Young Adults (CARDIA) Study.” Annals of Behavioral Medicine 36: 176–185. Kahneman, D. and A. Deaton (2010). “High Income Improves Evaluation of Life but Not Emotional Well-Being.” Proceedings of the National Academy of Sciences of the United States of America 107: 16489–16493. Kalleberg, A.  L. (2018). Precarious Lives: Job Insecurity and Well-Being in Rich Democracies. Medford, MA: Polity Press. Kangas, O., S. Jauhiainen, M. Simanainen, and M. Ylikanno (eds.) (2019). “The Basic Income Experiment 2017–2018 in Finland: Preliminary Results.” Reports and Memorandums of the Ministry of Social Affairs and Health 2019–9 [Finland].

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  39 Kasser, T. (2003). The High Price of Materialism. Cambridge, MA: MIT Press. Kim, P., G. W. Evans, M. Angstadt, S. S. Ho, C. S. Sripada, J. E. Swain, I. Liberzon, and K. L. Phan (2013). “Effects of Childhood Poverty and Chronic Stress on Emotion Regulatory Brain Function in Adulthood.” Proceedings of the National Academy of Science 110: 18442–18447. Knudsen, E. I., J. J. Heckman, J. L. Cameron, and J. P. Shonkoff (2006). “Economic, Neurobiological, and Behavioral Perspectives on Building America’s Future Workforce.” Proceedings of the National Academy of Science 103: 10155–10162. Leu, R. E., S. Burri, and T. Priester (1997). Lebensqualität und Armut in der Schweiz. Bern: Haupt. Lucas, R.  E., A.  E.  Clark, Y.  Gerogellis, and E.  E.  Diener (2004). “Unemployment Alters the Set Point for Life Satisfaction.” Psychological Science 15: 8–13. Lucas, R. E., E. Diener, and S. Eunkook (1996). “Discriminant Validity of Well-Being Measures.” Journal of Personality and Social Psychology 71: 616–628. Luhmann, M., R. E. Lucas, M. Eid, and E. Diener (2010). “The Prospective Effect of Life Satisfaction on Life Events.” Social Psychological and Personality Science 4: 39–45. Lupien, S.  J., B.  S.  McEwen, M.  R.  Gunnar, and C.  Heim (2009). “Effects of Stress throughout Lifespan on the Brain, Behavior, and Cognition.” Nature Reviews: Neuroscience 10: 434–445. Macchia, L. and A. C. Plagnol (2019). “Life Satisfaction and Confidence in National Institutions: Evidence from South America.” Applied Research in the Quality of Life 14: 721–736. Moors, A., P. C. Ellsworth, K. R. Scherer, and N. H. Frijda (2013). “Appraisal Theories of Emotion: State of the Art and Future Development.” Emotion Review 5: 119–124. Morris, A. P., F. Bremmer, and B. Krekelberg (2016). “The Dorsal Visual System Predicts Future and Remembers Past Eye Position.” Frontiers in Systems Neuroscience 10: Article 9. National Center for Health Statistics (2017). https://www.cdc.gov/nchs/index.htm. Nesse, R. M. and P. C. Ellsworth (2009). “Evolution, Emotions, and Emotional Disorders.” American Psychologist 64: 129–139. Newport, F. (2015). “Fewer Americans Identify as Middle Class in Recent Years.” Gallup Research Report, April 28. Nickerson, C., N. Schwartz, E. Diener, and D. Kahneman (2003). “Zeroing in on the Dark Side of the American Dream: A Closer Look at the Negative Consequences of the Goal of Financial Success.” Psychological Science 14: 531–536. Oswald, A.  J. and N.  Powdthavee (2008). “Does Happiness Adapt? A Longitudinal Study of Disability with Implications for Economists and Judges.” Journal of Public Economics 92: 1061–1077. Otis, N. (2017). “Subjective Well-Being in China: Associations with Absolute, Relative, and Perceived Economic Circumstances.” Social Indicators Research 132: 885–905.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

40  Peter Railton Paradies, Y., J.  Ben, N.  Denson, A.  Elias, N.  Priest, et al. (2015). “Racism as a Determinant of Health: A Systematic Review and Meta-Analysis.” PLOS One 10: e0138511. Pascoe, E. A. and L. Smart Richman (2009). “Perceived Discrimination and Health: A Meta-Analytic Review.” Psychological Bulletin 135: 531–554. Pessoa, L. (2008). “On the Relationship between Emotion and Cognition.” Nature Reviews: Neuroscience 9: 148–158. Pickett, K. E. and R. G. Wilkinson (2007). “Child Well-Being in Rich Societies: Ecological Cross-Sectional Study.” BMJ Online First. doi:10.1136/bmj.39377.580162.55. Piketty, T. and E.  Saez (2013). “Top Incomes and the Great Recession: Recent Evolutions and Policy Implications.” IMF Economic Review 6: 456–478. Quoidbach, J., E. W. Dunn, K. V. Petrides, and M. Mikolajczak (2010). “Money Giveth, Money Taketh Away: The Dual Effect of Wealth on Happiness.” Psychological Science 21: 759–763. Rainie, L. and A. Perrin (2019). “Key Findings about Americans’ Declining Trust in Government and Each Other.” Pew Research Center. https://www.pewresearch.org/ fact-tank/2019/07/22/key-findings-about-americans-declining-trust-in-governmentand-each-other/. Reyes-Garcia, V., A.  Angelson, G.  E.  Shively, and D.  Minkin (2019). “Does Income Inequality Influence Subjective Well-Being? Evidence from 21 Developing Countries.” Journal of Happiness Studies 20: 1197–1215. Reynolds, S.  M. and K.  C.  Berridge (2008). “Emotional Environments Retune the Valence of Appetitive versus Fearful Functions in Nucleus Accumbens.” Nature Neuroscience 11: 423–425. Rojas, M. (2018). “Happiness in Latin America Has Social Foundations,” in World Happiness Report 2018, Chapter 8. https://worldhappiness.report/ed/2018/. Rothermund, K. and J. Brandstadter (2003). “Coping with Deficits and Losses in Later Life: From Compensatory Action to Accommodation.” Psychology and Aging 18: 896–905. Sattler, F.  A. and R.  Lemke (2019). “Testing the Cross-Cultural Robustness of the Minority Stress Model in Gay and Bisexual Men.” Journal of Homosexuality 66: 189–208. Scheffler, R., H. Qaazi, and G. Dimick (2018). “The Anxious Generation: Causes and Consequences of Anxiety Disorder among Young Americans. Preliminary Findings.” Berkeley Institute for the Future of Young Americans. Schwartz, N. and G.  L.  Clore (2003). “Mood as Information: 20 Years Later.” Psychological Inquiry 14: 296–303. Seligman, M. E. P. (2018). “PERMA and the Building Blocks of Well-Being.” Journal of Positive Psychology 13: 333–335. Seligman, M. E. P., P. Railton, R. F. Baumeister, and C. Sripada (2013). “Navigating into the Future or Driven by the Past.” Perspectives on Psychological Science 8: 119–141.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

A Convenient Truth?  41 Solomon, A. (2012). Far from the Tree: Parents, Children and the Search for Identity. New York: Scribner. Stadelmann-Steffen, I. and A.  Vatter (2012). “Does Satisfaction with Democracy Really Increase Happiness? Direct Democracy and Individual Satisfaction in Switzerland.” Political Behavior 34: 535–559. Stevenson, B. and J. Wolfers (2008). “Economic Growth and Subjective Well-Being: Reassessing the Easterlin Paradox.” Brookings Papers on Economic Activity 2008 (Spring): 1–102. Stevenson, B. and J. Wolfers (2013). “Subjective Well-Being and Income: Is There any Evidence of Satiation?” American Economic Review 103: 598–604. Teufel, C. and P. C.  Fletcher (2020). “Forms of Prediction in the Nervous System.” Nature Reviews: Neuroscience 21: 231–242. Van Boven, L. (2005). “Experientialism, Materialism, and the Pursuit of Happiness.” Review of General Psychology 9: 132–142. Van Campen, C. and M. van Sandvoort (2013). “Explaining Low Subjective Sell-Being of Persons with Disabilities in Europe: The Impact of Disability, Personal Resources, Participation, and Socio-Economic Status.” Social Indicators Research 111: 839–854. Ward, G. (2019). “Happiness and Voting Behavior.” In J. Helliwell et al. (2019): 47–66. Voteview (2016). https://legacy.voteview.com/Politics_of_Income_Inequality_2016.htm. Welsch, H. and P.  Biermann (2019). “Poverty Is a Public Bad: Panel Evidence from Subjective Well-Being Data.” Review of Income and Wealth 65: 187–200. World Bank data. https://data.worldbank.org/. World Database of Happiness. https://worlddatabaseofhappiness.eur.nl/.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

2

Animals and Climate Change Jeff Sebo

1. Introduction The standard story about climate change is that humans are the main causes of climate change and will also be the main victims. But nonhuman animals are central to climate change too. In particular, animal agriculture is a major contributor to climate change, and climate change will be a major contributor to wild animal suffering. As a result, I believe, resistance to animal agriculture should be a major part of climate change mitigation, and assistance for wild animals should be a major part of climate change adaptation. Both of these topics—animal agriculture as a cause of climate change, and wild animals as victims of climate change—are neglected, but the latter is especially neglected.1 I think that there are at least two reasons for this. First, whereas humans have self-­interested reason to care about animal agriculture as a cause, we do not have self-­interested reason to care about wild animals as victims (except insofar as we benefit from the existence of wild animals). Second, whereas it is relatively clear how we should address the problem of animal agriculture and ­climate change, it is not at all clear how we should address the problem of wild animals and climate change. Indeed, the problem of wild animals and climate change sits at the intersection of some of the hardest problems that we face in science and philosophy, and it is therefore not only especially important but also especially daunting. My aim in this chapter is to examine these connections between animals and climate change from a primarily philosophical perspective. I will proceed as ­follows. In §2, I will examine the impacts that animal agriculture will have on ­climate change, and I will argue that, in light of these impacts, we morally ought to resist animal agriculture as part of our mitigation efforts. In §3, I will examine the impacts that climate change will have on wild animal populations, and I will argue that, in light of these impacts, we morally ought to assist wild animals as part of our adaptation efforts. Then, in §§4–5, I will discuss questions about well-­being, creation ethics, population ethics, duties to future generations, and more that we 1  See Hsiung and Sunstein (2007) and Palmer (2011) for representative exceptions.

Jeff Sebo, Animals and Climate Change In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Jeff Sebo. DOI: 10.1093/oso/9780198796282.003.0003

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  43 need to ask before we can make these duties more concrete. I will not attempt to answer these questions here. Instead, my aim is to start to unpack the relevant issues and argue for two first steps that will make sense no matter how we answer many of them. First, we should research alternatives to animal agriculture and interventions in wild animal suffering. Second, we should extend moral and political standing to all animals. Before I begin, a quick note about the scope of this topic. While my focus here will be on animal agriculture as a cause of climate change and on wild animals as victims, we can, and should, also talk about farmed animals as victims of climate change and about wild animals as causes. For an example of the former, in fall 2016 Hurricane Matthew resulted in flooding that killed millions of farmed ­animals in North Carolina alone (Philpott 2016). And for an example of the latter, aquatic animals such as whales function as carbon sinks if they die of natural causes (since they sink to the bottom of the ocean, bringing stored carbon with them) and as carbon sources if we kill them (since we bring them to the surface, releasing stored carbon into the atmosphere) (Pershing et al. 2010). There is much more to say about these topics (as well as about the relevance of other domesticated and wild animals), but since the most important topics in this area are ­animal agriculture as a cause of climate change and wild animals as victims, these two will be my focus here.

2.  Farmed Animals, Climate Change, and a Duty to Resist We know that the climate is changing, and that human activity is changing it. We will likely see a 2–4 degree Celsius increase in global average temperatures above pre-­industrial levels by the end of the century (IPCC 2014). There is also a small chance that a runaway effect will occur, causing as much as a 10 degree Celsius increase. Climate experts generally agree that a 5–10 degree increase would be catastrophic. But even if we limit the damage to 2–4 degrees, climate change will have a massive and pervasive impact on this planet. It will cause melting ice caps; rising sea levels; flooding coastal cities; an increase in the frequency and intensity of extreme weather events such as hurricanes and tsunamis; and regional conflicts over land, water, energy, and food (NASA  2016). As the Pentagon notes, ­climate change will also be a “threat multiplier” that makes existing problems worse (Bergengruen 2017). This raises questions about what if anything we can and should do to mitigate human-­caused climate change. These questions are hard to answer because the activities that contribute most to climate change are activities that humans really enjoy: reproduction and, especially, consumption. Consider reproduction first. In 1900, we had a global human population of about 1.5 billion (Kremer 1993). Now we have more than 7 billion

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

44  Jeff Sebo and counting, and population experts estimate that we will likely have 9 or 10 billion by 2050 and 10 or 11 billion by 2100 (Gerland et al. 2014). Now consider consumption. Not only do we have more people, but, much more importantly, we are consuming much more land, water, energy, and animals than ever before. In  particular, to feed even a fraction of the current human population, we are currently farming 20+ billion terrestrial animals at any given time, and using a substantial proportion of the land, water, and energy that we consume in order to sustain this population (FAO 2016; see also Tilman et al. 2011). Moreover, this farmed animal population is increasing with each passing year, since developed countries are expanding into aquaculture and developing countries are seeing increased demand for animal products and adoption of industrial animal agricultural methods. Needless to say, this global animal agricultural system causes unimaginable pain and suffering for nonhuman animals (Singer 2009). It also harms workers (Pachirat 2013) and leads to many public health and environmental risks, ranging from increased mental and physical health risks from animal waste seeping into the land and water near factory farms to increased risk of a global pandemic from antimicrobial use on factory farms (Pew Commission 2008). Thus, even if we set aside climate change, we have strong reason to advocate against animal agriculture. With that said, and most importantly for our purposes here, animal agriculture is also a leading contributor to human-­caused climate change. Specifically, animal agriculture is responsible for an estimated 9 percent of human-­caused carbon emissions, 37 percent of human-­caused methane emissions, and 65 percent of human-­caused nitrous oxide emissions (Steinfeld et al. 2006). When you put these figures together, the upshot is that industrial animal agriculture is responsible for anywhere from 14.5–51 percent of global greenhouse gas emissions. Why is this range so wide? The 14.5 percent estimate comes from a 2013 FAO report (Gerber et al. 2013, updated from an 18 percent estimate in Steinfeld et al. 2006) whereas the 51 percent estimate comes from a 2009 World Watch report (Goodland and Anhang  2009), which assesses the climate impacts of animal agriculture on a different timescale. In particular, the FAO report assesses the ­climate impacts of animal agriculture on a 100 year timescale (which amplifies the impacts of other industries, since animal agriculture is responsible for a relatively small share of global carbon emissions and carbon stays in the atmosphere longer than methane and nitrous oxide). In contrast, the World Watch report assesses the climate impacts of animal agriculture on a 20 year timescale (which amplifies the impacts of animal agriculture, since animal agriculture is responsible for a relatively large share of global methane and nitrous oxide emissions and methane and nitrous oxide trap heat more effectively than carbon while in the atmosphere). Insofar as the FAO and World Watch reports are assessing the climate impacts of animal agriculture on different timescales, then, they might not be disagreeing so much as reporting similar information in different ways.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  45 With that said, there are other, more substantive differences between these reports as well. For example, the World Watch report estimates that the global population of farmed animals is higher than the FAO report does, and it also considers a wider range of food-­related activities as part of animal agriculture than the FAO report does.2 Which report is more accurate in these respects? I am not sure. Fortunately, I do not think that we need to answer this question in order to make substantive moral and political progress on this issue. In particular, even if we assume for the sake of argument that both of these reports are overestimating the real climate impacts of animal agriculture, and that the real climate impacts of animal agriculture are only, say, 10 percent of all human-­caused greenhouse gas emissions (on a 100 year timescale), this is still a massive contribution, and it will only increase as the human population increases and a higher proportion of the global population produces and consumes industrial animal products. Thus, even though many questions remain about the relative climate impacts of animal agriculture, I think that we at least have enough information to know that we need to take action on this issue. What action should we take? I think we can say at least this much with confidence: Given that animal agriculture is a major contributor to human-­caused climate change, we should resist animal agriculture as part of our mitigation efforts. In particular, we should resist animal agriculture individually through activism, advocacy, and philanthropy, and we should resist it collectively through public policy. Unfortunately, many environmentalists have been slow to accept these conclusions. Why is that? There are many sources of resistance, ranging from historical tensions between the animal and environmental movements to personal reluctance to accept that we might have a moral duty to give up food that we like. Some people also have reasonable concerns about resistance of animal agriculture, though these concerns do not, I think, provide us with sufficient reason not to engage in resistance. I will focus on two reasonable concerns here. The first concern one might have is that individual activism and advocacy around animal agriculture will alienate people from the environmental movement, thereby limiting the growth and overall impact of this movement. Why should we focus on a relatively controversial topic such as food (which would push people away) when we could instead focus on relatively uncontroversial t­opics such as deforestation and natural resource extraction (which would bring people in)? I  think that this concern is reasonable, especially since many people in social movements, including the animal and environmental movements, do sometimes engage in unnecessarily alienating activism and advocacy. However, I also think that this concern does not provide us with sufficient reason not to engage in individual activism and advocacy around this issue, since (a) these efforts might

2  Thanks to Christopher Schlottmann for helpful conversation on this point.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

46  Jeff Sebo not alienate people from the environmental movement, and (b) even if they do have this impact, they might still be warranted. First, individual activism and advocacy around this issue might not alienate people from the environmental movement. There are many ways of addressing controversial issues without pushing people away. For example, to the degree that activists frame animal agriculture as a structural social, political, and economic problem that demands a structural social, political, and economic solution rather than as an individual behavioral problem that demands an individual behavioral solution, they can address this issue without shaming individual producers and consumers. Moreover, to the degree that activists advocate for individual behavioral change in moderate, conciliatory ways rather than in radical, confrontational ways (for example, to the degree that they invite people to reduce meat consumption rather than insist on veganism as a moral baseline), they can address this issue without shaming individual producers and consumers, even when they do take an individual approach. These structural and/or conciliatory approaches might not be a full solution, as I will emphasize in a moment. But they are part of the solution, and insofar as activists worry about alienating p ­ eople, they can mitigate this risk by taking these approaches. Second, even if individual activism and advocacy around this issue do alienate people from the environmental movement, they might still be warranted. After all, a smaller and/or slower growing movement that focuses on the right issues in the right ways can sometimes do more good overall than a larger and/or faster growing movement that focuses on the wrong issues in the wrong ways. This is part of why people in many social movements, including the animal and en­vir­on­ men­tal movements, agree that social movements should address at least some controversial issues and make at least some controversial claims about those issues. Yes, this approach risks alienating people in the short term. But it also raises awareness about important issues and shifts the center of debate about them, thereby paving the way for long-­term change in spite of these short-­term risks. And in this case, we have an industry that produces at least 10 percent of all human-­caused greenhouse gas emissions. We have reason to believe that this industry is neglected in both the animal and environmental movements (ACE 2016). And we have reason to believe that activism that addresses this industry can mitigate greenhouse gas emissions about as cost-­effectively as carbon offsets can (ACE 2012). In light of these considerations, I find it plausible that we should resist animal agriculture as part of our mitigation efforts whether or not doing so will alienate people in the short term (which, again, is a risk that we can mitigate if we like). The second concern one might have is that policy change around this issue will restrict individual liberty, thereby violating individual rights. Why should we attempt to restrict the kinds of options that people have access to (which would restrict individual liberty) rather than attempt to persuade people to make good

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  47 choices given existing options (which would not restrict individual liberty)? I  think that this concern is reasonable too, especially since many movements, including the animal and environmental movements, do sometimes promote good causes in coercive and manipulative ways. However, as with the previous concern, I do not think that this concern provides us with sufficient reason not to pursue policy change, since (a) policy change might not restrict individual liberty, and (b) even if it does, this impact might be warranted. First, policy change might not restrict individual liberty. There are many ways of addressing controversial issues through policy without restricting individual liberty. Many countries currently support animal agriculture more than they support alternatives through taxes, subsidies, and deregulation. For example, part of why animal products seem to be so accessible and affordable is that many states do very little by way of regulating how animal products are made, very little by way of enforcing the regulations that they do have, and very little by way of punishing food corporations in cases of enforcement. This allows food corporations to “externalize” public health and environmental costs of animal agriculture by relying on taxpayers to foot the bill. So, if states were to tax animal agriculture more (relative to alternatives), subsidize animal agriculture less (relative to alternatives), and regulate animal agriculture more (relative to alternatives), they could motivate people to support alternatives without restricting individual liberty in the relevant sense at all. Second, even if public policy does restrict individual liberty, that might be ­warranted. After all, many people believe that states are morally permitted, if not morally required, to restrict liberty in cases where exercising liberty foreseeably and avoidably causes harm. And of course, animal agriculture foreseeably and avoidably causes a lot of harm. Thus, we might think, states are morally permitted, if not morally required, to restrict individual liberty in order to prevent the harms that animal agriculture causes. However, we need to say more before we can draw this conclusion. For example, we have to ask: Do individuals cause harm when they eat animal products? Also, is the state morally permitted to restrict individual liberty only in order to prevent individuals from causing harm, or is it also morally permitted to restrict individual liberty in order to prevent collectives from causing harm? These questions matter because if we accept that both (a) individuals do not cause harm when they eat animal products and (b) the state is permitted to restrict individual liberty only in order to prevent individuals from causing harm, then we might deny that the state is permitted to restrict individual liberty in this case. However, if we accept (as I do) that either (a) individuals do cause harm when they eat animal products or (b) the state is permitted to restrict individual liberty in order to prevent collectives from causing harm, then we might accept that the state is permitted to restrict individual liberty in this case. Of course, to say that we should resist animal agriculture as part of our mitigation efforts is not to say exactly how much or in what ways we should do that. The

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

48  Jeff Sebo causes of climate change are many and varied, and our approaches to mitigation should be many and varied as well. But I do think that we should be spending much more time, energy, and money on this issue than we currently are on this issue. Otherwise, even if we (miraculously) bring about a decrease rather than increase in human population, and even if we (miraculously) bring about a decrease rather than increase in consumption in general, we will neither be addressing the main population crisis nor be addressing the main consumption crisis.3 To do that, we have to either abolish animal agriculture or regulate it so substantially that animal products become a rare luxury item from a global perspective, so that we can substantially reduce the animal welfare, public health, and environmental impacts of this industry.

3.  Wild Animals, Climate Change, and a Duty to Assist We are now in the Anthropocene/Capitalocene, a geological era defined by the impacts of humanity/capitalism on the planet. The destruction of nonhuman spaces, construction of human cities, industrialization of human economies, and other such activities are already having a systematic impact on the planet. The prospect of climate change will only amplify these impacts. This raises, or at least amplifies, questions about duties of assistance for wild animals whose lives our activities will impact. These questions are complicated, in part because climate change might have very different kinds of impacts for individuals, species, and ecosystems. At the level of species and ecosystems, it is relatively easy (while still objectively hard) to say what the overall impacts of climate change will be. In particular, climate change will bring about mass extinction, biodiversity loss, and ecosystem collapse, “with the worst-­case scenarios leading to . . . the sixth mass extinction in the ­history of the earth” (Bellard et al. 2012; see also Kolbert 2014).4 Granted, it can be difficult to assess whether and to what degree certain kinds of change are good or bad for species or ecosystems. But we can grant for the sake of discussion that, if anything is bad for species or ecosystems, then the systematically destructive effects of human activity in general, and of human-­caused climate change in particular, are bad for them. In contrast, at the level of individual wild animals, it is much harder to say what the overall impacts of climate change will be. It might be tempting to think that if climate change is bad for species and ecosystems, then it will also be bad for individuals. But of course, this might not be true, since for every species that 3  For more, see Garnett (2016), Ranganathan et al. (2016), and Stehfest et al. (2009). 4  Though see Palmer (2011) for an argument that, even if we think that ecosystems have intrinsic moral value, it is hard to tell what counts as good or bad for them.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  49 goes extinct, another might expand to take its place. Thus, if we want to assess how climate change will impact individual wild animals, we will have to ask questions such as: Which kinds of animals will likely do well in a world reshaped by climate change, and which kinds of animals will likely do badly? And how good or bad is life, on average or in total, for the kinds of animals who will likely do well and for the kinds of animals who will likely do badly? Unfortunately, there are currently many uncertainties about which kinds of animals will likely do well and badly in a world reshaped by climate change. This issue will depend on the complex interplay of impacts on land, water, primary productivity, pathogens and parasites, and more (Tomasik 2008). With that said, consider for the sake of discussion two possible trends that we might see in a world reshaped by climate change. First, we might discover that climate change favors adaptive generalists, i.e., animals who can survive in a wide range of en­vir­ on­men­tal conditions, over niche specialists, i.e., animals who can survive only in a narrow range of environmental conditions.5 Second (this is more speculative), we might discover that climate change favors r-­strategists, i.e., animals with shorter lifespans and higher reproduction rates, over K-­strategists, i.e., animals with longer lifespans and lower reproduction rates (Tomasik 2008). The reason that climate change might have these impacts is that it will radically and rapidly alter natural habitats, and so animals who are (a) able to survive in a wider range of environmental conditions and/or (b) have shorter lifespans and higher reproduction rates are more likely to survive to sexual maturity in these circumstances than animals who are not all else equal.6 As Díaz et al. put the point: On average, the organisms that are losing out [as a consequence of global change drivers such as climate change] have longer lifespans, bigger bodies, poorer dispersal capacities, more specialized resource use, lower reproductive rates, and other traits that make them more susceptible to human activities such as nutrient loading, harvesting, and biomass removal by burning, livestock grazing, ploughing, clear-­felling, etc.  (Díaz et al. 2006: 1301)7

As I said, we will have to ask many other questions (including the many moral and political questions that we will consider below) before we can say what if anything follows from these possibilities. However, I think that we can say this much now: Insofar as human-­caused climate change will harm wild animals, we 5  Thanks to Sue Donaldson for helpful discussion on this point. 6  Of course, many r-­strategists will also be niche specialists and many K-­strategists will also be adaptive generalists, so these two considerations might push against each other to a degree. But they might not always do so, and in any case these possible impacts will be useful to consider for our purposes here. 7  One complication is that insect populations may be decreasing for other reasons (Vogel 2017). If they are, then even if climate change has a positive impact on insect populations, human activity may still have a negative impact on insect populations overall.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

50  Jeff Sebo should assist wild animals as part of our adaptation efforts, assuming we can do so effectively. In particular, and as with mitigation, we should assist wild animals individually through activism, advocacy, and philanthropy, and we should assist them collectively through public policy. Of course, the idea that we should help wild animals is controversial. However, I think that our complicity in climate-­related wild animal suffering might change that for some people. In what follows I will focus on the utilitarian and rights theoretic cases for intervention in climate-­related wild animal suffering, though I should say that I think similar con­sid­er­ations apply for other moral theories as well.8 With that in mind, consider utilitarianism first. Utilitarians think that we have a moral duty to maximize well-­being in the world. Thus, for utilitarians the main questions will be: How important is climate-­related wild animal suffering? How neglected is this problem? And how tractable is this problem? At present, the first two questions are easier to answer than the third. With respect to importance, wild animal suffering is a high priority cause area for utilitarians with or without climate change, since quadrillions if not quintillions of sentient lives are at stake. With respect to neglectedness, wild animal suffering is a high priority cause area for utilitarians with or without climate change, since hardly anybody is working on it at all right now (even in the animal and environmental movements). And with respect to tractability, it remains to be seen. We can currently help wild ­animals in small-­scale ways, for example through small-­scale feeding, vaccination, or contraception programs. But we are not yet capable of helping wild animals in large-­scale ways. Thus, utilitarians should say that our focus in the short term should be on researching interventions in wild animal suffering (so that we know how to help them) and advocating for reducing wild animal suffering (so that we have the power to act on this knowledge). And then, if and when we have a clear sense of the tractability of this problem, we can decide what if anything we should to do at scale. Are there limits to the ways in which we can be permitted or required to help wild animals according to utilitarianism? In principle, no. For example, if we determine that we can promote average or total well-­being among wild animals by increasing or decreasing their populations, then a utilitarian might say that we have a moral duty to do exactly that. Similarly, if we determine that we can promote average or total well-­being among wild animals by placing them in captivity (for example by turning all of nature into an animal sanctuary), then a utilitarian might say that we have a moral duty to do exactly that as well. (More on these possibilities below.) Of course, in practice, a lot depends on how confident we are about these judgments. Again, given how difficult these issues are and how poor 8  For example, see Gruen (2014), Hursthouse (2011), and Palmer (2010) for alternative moral theories that imply that we have moral duties to wild animals (or at least that our treatment of wild animals is morally evaluable in some way, shape, or form).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  51 our track record is on these issues, caution is warranted. Still, the mere fact that utilitarians are open to these options in principle (whether or not they are also open to these options in practice) is noteworthy. Now consider rights theory. Rights theorists deny that we have a moral duty to maximize well-­being. Instead, many rights theorists think that (a) we have a moral duty to respect autonomy, and that (b) whereas we have a perfect duty of non-­maleficence (i.e., we always have a moral duty not to harm others), we have only, at best, an imperfect duty of beneficence (i.e., we only, at best, sometimes have a moral duty to help others).9 As a result, the main questions for rights the­ or­ists will be: If we intervene in climate-­related wild animal suffering, will we be interfering in wild animal autonomy, and will we be engaging in beneficence or non-­maleficence? This is why, as I indicated above, many rights theorists deny that we have a moral duty to intervene in wild animal suffering in general, since, they think, non-­interference is morally permissible whereas beneficent interference may not be. However, they might accept that we have a moral duty to intervene in climate-­related wild animal suffering. Why? Because insofar as humans are already interfering with wild animals through climate change and other such impacts, the choice is not between non-­interference and beneficent interference. Instead, the choice is between relatively maleficent interference and relatively beneficent interference. In this case, rights theorists might support intervening in wild animal suffering after all, in order to reduce and repair the harms that human activity is causing. Are there limits to the ways in which we can be permitted or required to help wild animals according to rights theory? In principle as well as in practice, the answer is very likely yes. For example, even if we determine that we can promote average or total well-­being among wild animals by increasing their populations, decreasing their populations, or placing them in captivity, a rights theorist might oppose these interventions if they believe that these interventions violate wild animal autonomy or harm some wild animals as a means to helping others. However, a lot will depend on the details. For example, rights theorists will have to ask which cases of wild animal suffering we should take responsibility for in a world reshaped by climate change, as well as what it means to respect wild animal autonomy and treat wild animals as ends in themselves in a world reshaped by climate change. Still, given our complicity in climate-­related wild animal suffering, rights theorists should accept that we have a moral duty to assist wild animals in at least some cases, either as a matter of non-­maleficence

9  To be clear, I am focusing here on rights theories according to which wild animals have moral status. So I will be drawing less from rationalists like Immanuel Kant and more from sentientists, in a broad sense of the term, like Tom Regan (2004), Christine Korsgaard (2011), and Sue Donaldson and Will Kymlicka 2011.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

52  Jeff Sebo (insofar as assistance aims at mitigation) or as a matter of reparation (insofar as assistance aims at adaptation). As I said above, I think that if these arguments are correct, then we morally ought to assist wild animals, assuming we can do so effectively, individually as well as collectively, through public policy. Of course, the idea that we should assist wild animals through public policy is controversial, because it raises challenging questions about the political standing of nonhuman animals. And the traditional view about this issue is that nonhuman animals cannot have political standing, because only political agents can have political standing, and nonhuman animals are not political agents (for discussion, see Cochrane 2010). However, in recent years many theorists have started to challenge this analysis. Why? Because as Donaldson and Kymlicka (2011) note, many aspects of political standing do not, in fact, require the kind of political agency that nonhumans seem to lack. Granted, you might need this kind of political agency in order to have responsibilities, as well as in order to have at least some rights, such as the right to hold public office or vote in an election. But you do not need this kind of political agency in order to have other rights, such as the right to reside in your territory, the right to return to your territory, or the right to have your interests represented in the political process. This development creates space for nonhumans to have at least partial political standing whether or not they have political agency in the relevant sense. How if at all might this development create space for the state to have duties of assistance to wild animals in light of climate change? Donaldson and Kymlicka (2011) have developed what is currently the most prominent account of the political standing for nonhuman animals. With many caveats, they argue that we should extend our conception of citizenship (i.e., full membership in our political ­community) to domesticated animals, our conception of denizenship (i.e., partial membership in our political community and partial membership in their own political community) to liminal animals, and our conception of sovereignty (i.e., full membership in their own political community) to wild animals. On this view, we should think about our political duties to wild animals on the model of our political duties to other nations. So what political duties do we have to other nations? The standard view is that primarily, we should let other nations be. However, in cases of emergency, and especially in cases of emergency for which we are responsible, we should provide assistance. Thus, for example, Henry Shue argues that if developed nations are disproportionately responsible for climate change and if developing nations will be disproportionately impacted by climate change, then developed nations have a duty to help developing nations adapt to climate change (for example, see Shue 1999). If we apply this model to wild animals, then, we might reach a similar conclusion: Primarily, we should let them be. But if human communities are disproportionately responsible for climate change and if nonhuman communities will be disproportionately harmed by climate change, then human communities have a duty to help nonhuman communities adapt to climate change.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  53

4.  Animals, Climate Change, and a Life Worth Living Suppose, then, that we accept that we morally ought to resist animal agriculture as part of our mitigation efforts and assist wild animals as part of our adaptation efforts. We have many other questions that we need to ask as well before we can make our duties regarding animals and climate change more concrete. This is especially true for wild animals and climate change (which will be my focus here), since this topic requires us to think about many challenging questions holistically. In what follows I will consider how questions about well-­being, creation ethics, population ethics, duties to future generations, and more all bear on this topic. As I said above, my aim is not to answer all these questions here but rather to explore them, show how they interact, and argue for two first steps that will make sense no matter how we answer many of them. With that in mind, consider first several questions about well-­being that we need to answer in order to make our duties regarding animals and climate change more concrete. First, we need to ask: What is the nature of well-­being? This question breaks into several further questions; we will here consider two. The first is: Does the intrinsic value of a life for its subject depend only on subjective facts such as how much pleasure or pain it contains, or does it also depend on objective facts such as how many meaningful projects, relationships, and other such goods it contains? The second issue is: What marks the difference between a good life, i.e., a life worth living, and a bad life, i.e., a life not worth living? Should we set the baseline for a life worth living at zero, such that a life is worth living if it contains more good things than bad and not worth living if it contains more bad things than good (perhaps taking into account the order of good and bad things in life)? Alternatively, should we set the baseline higher than zero, such that a life can be not worth living even if it contains more good things than bad, or lower than zero, such that a life can be worth living even if it contains more bad things than good? (See Bradley 2015 for discussion.) What should we think about these issues? With respect to the nature of ­well-­being, subjectivist views are appealing because they provide a simple, intuitively plausible explanation of why good things are good and bad things are bad. Meaningful projects, relationships, and so on are good to the degree that they promote pleasure and the absence of pain, period. In contrast, objectivist views are appealing because they capture the messiness, complexity, and variety that many people take the intrinsic value of a life to have (especially in a world where so many individuals lead such different lives). With respect to the second issue, this is a very challenging question. It feels tempting to set the baseline for a life worth living at zero, since it makes sense that a life with positive net value would be worth living and that a life with negative net value would be not worth living, all else equal. However, it also feels tempting to set the baseline higher than zero,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

54  Jeff Sebo since some people would rather not live at all than live a life that contains only a bit more good than bad. And it also feels tempting to set the baseline lower than zero, since other people would rather live a life that contains a bit more bad than good than not live at all (and in both cases, who are we to dispute that?). What depends on these issues? How we answer these questions will affect our thinking about which wild animals have lives worth living. For example, if well-­being depends entirely on subjective facts, then we might feel relatively ­pessimistic about how many wild animals have lives worth living, since we might feel relatively pessimistic about how many wild animals have pleasant lives. In contrast, if well-­being depends at least partly on objective facts (for example facts about whether or not a particular animal has a species-­typical life), then we might feel relatively optimistic about how many wild animals have lives worth living, since we might feel relatively optimistic about how many wild animals satisfy the relevant criteria. As for the baseline of a life worth living: To the degree that we set a high baseline, relatively few wild animals will count as having lives worth living, and therefore relatively few wild animals will have the kinds of lives that we think we should bring about (assuming that we think we have stronger reason to bring about lives worth living than lives not worth living). Whereas, to the degree that we set a low baseline, relatively many wild animals will count as having lives worth living, and therefore relatively many wild animals will have the kinds of lives that we think we should bring about (again, assuming we think we have stronger reason to bring about lives worth living than lives not worth living). The second question that we need to ask is: Which wild animals have the capacity for well-­being? In other words, which animals have lives that are evaluable as worth living or not worth living? Of course, this depends on which theory of well-­being we accept. But suppose for the sake of discussion that we accept that well-­being depends on sentience, i.e., pleasure and pain experience. In this case, our question becomes: Which wild animals are sentient? What should we think about this issue? Unfortunately, this question is difficult if not impossible to answer in any precise way. As long as we reject radical ­skepticism about other minds in general, we can be relatively confident that other vertebrates are sentient, since other vertebrates are behaviorally, psychologically, and evolutionarily continuous with humans in respects that appear relevant to sentience (DeGrazia 2002). However, we can be less confident about invertebrates, especially insects. After all, invertebrates such as insects are similar to us in many ways that appear relevant to sentience as well as different from us in many ways that appear relevant to sentience (see Mendl and Paul 2016 for discussion of the current science on this topic in the case of bumblebees). Ultimately, we might never know what if anything it is like to be an insect, and therefore we might always be at least somewhat uncertain about this issue (though of course, we might always be somewhat uncertain about other human and nonhuman animals as well; see Hyslop 2014 for discussion).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  55 What depends on this issue? A lot. For example, scientists estimate that there are about 10 quintillion insects in the world at any given time (Encyclopedia Smithsonian 2016). It follows that if there is even a 1/1000 chance that the average insect has even 1/1000 the amount of well-­being that the average primate does at any given time (which, in my view, is a conservative pair of assumptions), then the total amount of expected insect well-­being in the world at any given moment is equal to that of 10 trillion primates. Thus, the question whether and to what degree we regard insects as having well-­being may affect our moral calculations considerably. For example, if we accept utilitarianism, then we may or may not think that the impacts of climate change on insect populations are morally de­cisive, because of how much total well-­being insects have overall. Similarly, if we accept rights theory, then we may or may not think that the impacts of climate change on insect populations are morally decisive as well, depending on how we flesh out the details of the theory (because of how many insects we could be harming unnecessarily as means to our ends). The third question that we need to ask is: Which animals have lives worth ­living? Of course, this depends on what theory of well-­being we accept too. But suppose for the sake of discussion that we accept that well-­being depends on sentience and that the baseline for a life worth living is zero, i.e., that a life is worth living if it contains more pleasure than pain and not worth living if it contains more pain than pleasure. In this case, our question becomes: Which animals have net pleasurable lives, and which animals have net painful lives? What should we think about this issue? Unfortunately, this question is difficult if not impossible to answer in any precise way as well. Many people think that most wild animals have net pleasurable lives because, when they think about wild animals, they think about charismatic megafauna such as elephants, lions, or tigers roaming the countryside (and they tend to assume that natural processes are good). However, some people such as Oscar Horta (2010) and Brian Tomasik (2015) think that most animals have net painful lives.10 Why? Because they think that, when we consider the full story about what life in nature is like (and correct for our tendency to assume that natural processes are good), we see that most wild animals experience lives full of pain as a result of hunger, thirst, disease, weather, predation, and more. Horta and Tomasik think that this is especially true of r-­strategists (especially insects), since r-­strategists (especially insects) aim for quantity over quality in reproduction, which means that a relatively high proportion of these individuals experience a painful death as soon as they become cap­ able of experiencing anything at all (assuming that they do experience anything at all of course).

10  See also Ng 1995 and Groff and Ng 2019.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

56  Jeff Sebo What depends on this issue? Again, a lot. Consider for the sake of simplicity two possible scenarios (though please keep in mind that the actual state of affairs is likely to be much more complicated). First, suppose that no wild animals have lives worth living, and that K-­strategists have worse lives on average (because they have longer lives and more intensely negative experiences) and that r-­strategists have worse lives in total (because there are so many more of them). In that case, if human-­caused climate change increases the ratio of r- to K-­strategists, then it will make quality of life for wild animals better on average but worse in total, all else equal. Conversely, suppose that all wild animals have lives worth living, and that K-­strategists have better lives on average (because they have longer lives and more intensely positive experiences) and r-­strategists have better lives in total (because there are so many more of them). In that case, if human-­caused climate change increases the ratio of r- to K-­strategists, then it will make quality of life for wild animals worse on average but better in total, all else equal. Thus, these questions could determine whether we think the impact of climate change on wild animals is net positive or negative, which, in turn, could affect what we think we owe ­certain wild animal populations in light of climate change.11 As we have seen, these questions about well-­being are both deeply important and deeply challenging, given that we may never know what if anything it is like to be members of other species. This raises more general methodological questions about how to treat other animals in cases of uncertainty about whether they have lives worth living. One possibility, which I discuss in more detail elsewhere (Sebo 2018), is that we can use principles of risk and uncertainty to assess how to treat other animals in these cases. In this case we would have two main options to consider. First, we can use a precautionary principle that tells us not to risk bringing about an unacceptable level of harm. Second, we can use an expected harm principle that tells us to (a) multiply the probability that our action will cause harm by the level of harm it would cause if it did and (b) treat the product of this equation as the amount of harm that our action will actually cause, for purposes of deciding what to do. What would these principles imply about the questions that we have been ­considering here? In cases of uncertainty about whether or not animals have the capacity for well-­being in general, both of these principles would tell us to treat them as though they do, at least to a degree. In particular, the precautionary ­principle would tell us to treat other animals as though they have the capacity for well-­being, and the expected value principle would tell us to multiply our credence that they have this capacity by the amount of well-­being they would have if 11  These questions could also end up determining whether animal agriculture has a net positive or negative impact on farmed animals, which will be relevant for some moral theories. It is interesting to ask what theory of well-­being, if any, implies the common view that wild animals have good lives overall but that farmed animals do not. It is also interesting to ask how if at all we should change our thinking about animal ethics if we are not able to vindicate either or both parts of this common view.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  57 they did, and to treat the product of that equation as the amount of well-­being that they actually have. However, in cases of uncertainty about whether or not animals have lives worth living, it is less clear what these principles would tell us to do, since there are real risks either way. For example, if we decide to say that animals have lives worth living in cases of uncertainty, then we risk false positives, i.e., we risk accidentally bringing about more animals with lives not worth living (insofar as we think we should bring about animals with lives worth living). Whereas if we decide to say that animals have lives not worth living in cases of uncertainty, then we risk false negatives, i.e., we risk accidentally bringing about fewer animals with lives worth living (insofar as we think we should not bring about animals with lives not worth living). Thus a lot will depend on which of these risks we think is greater, as well as on what we think about the relative value of bringing about lives worth living and not bringing about lives not worth living. We will consider this issue below.

5.  Animals, Climate Change, and a Life Worth Creating Next, consider several questions about creation ethics, population ethics, and duties to future generations that we need to ask in order to make our duties regarding animals and climate change more concrete. First, do we have an equally strong duty to bring about lives worth living and not bring about lives not worth living, or do we have a stronger duty to not bring about lives not worth living than we have to bring about lives worth living? This question concerns creation ethics, and it actually divides into multiple questions. One question is about the lives that we cause to exist. Do we have an equally strong duty to create lives worth living and not create lives not worth living? Or do we have a stronger duty to not create lives not worth living than we have to create lives worth living? Another question is about the lives that we allow to exist. Do we have an equally strong duty to prevent lives not worth living and not prevent lives worth living? Or do we have a stronger duty to prevent lives not worth living than we have to not prevent lives worth living? What should we think about this issue? That depends in part on what moral theory we accept. For example, since (many) utilitarians think that (a) benefits and harms are equally morally important and (b) actions and omissions are equally morally important, they would likely say that we have an equally strong duty to create lives worth living, not create lives not worth living, prevent lives not worth living, and not prevent lives worth living (at least in theory).12 In contrast, since (many) rights theorists think that (a) harms are more important than benefits and 12  Though see Benatar (2006) for an argument that whereas creating lives not worth living counts as a harm, creating lives worth living does not count as a benefit.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

58  Jeff Sebo (b) actions are more important than omissions, they would likely say that we have a stronger duty to not create lives not worth living than we have to create lives worth living, prevent lives not worth living, or not prevent lives not worth ­living (in theory as well as in practice). However, there are exceptions to these generalizations. For example, negative utilitarians think that (a) harms are more important than benefits and (b) actions and omissions are equally important. Thus, they should say that we have a stronger duty to prevent, and not create, lives not worth living than we have to create, and not prevent, lives worth living. At the limit, this kind of view seems to support mass extinction for all sentient life.13 How much depends on this issue? A lot. For example, if we think that benefits and harms are equally important, then we might think that we should be neutral about whether to treat wild animals as having lives worth living in cases of uncertainty all else equal (because false positives and false negatives are equally risky all else equal). Whereas if we think that harms are more important than benefits, then we might think that we should treat wild animals as not having lives worth living in cases of uncertainty all else equal (because false positives are riskier than false negatives all else equal). Similarly, if we think that actions and omissions are equally important, then we might think that we have moral duties to future wild animals whether or not we are responsible for their existence. Whereas if we think that actions are more important than omissions, then we might think that we have moral duties to future wild animals only if we are responsible for their existence (in which case we will have to decide whether the Anthropocene/ Capitalocene/climate change make it the case that we are responsible for the existence of future wild animals). Of course, there are further questions that we have to ask here too. For example, a rights theorist will want to know whether certain population control measures would interfere with wild animal autonomy or harm certain wild animals as a means to helping others. Still, what we think about creation ethics will have a major effect on what we take our moral duties to future wild animals to be. The second question is: Do we have an equally strong duty to create all lives worth living, or do we have a stronger duty to create better lives worth living than worse lives worth living? This question raises what Derek Parfit calls the non-­identity problem (Parfit  1984: 351–377). Roughly speaking, the non-­identity problem works as follows. Many of our choices will affect not only the quality but also the identities of future lives. Thus, for example, suppose that we have two policy options: A and B. If we select option A, that will bring about one population of individuals (population A) who have lives very much worth living. Whereas if we select option B, that will bring about another population of individuals (population B) who have lives at least somewhat worth living. Intuitively, if we bring about

13  See Ord (2013) for critical discussion of negative utilitarianism.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  59 population B instead of population A, we cause harm, and therefore we act wrongly, all else equal. But whom are we harming? Not population B, since this is the only future in which they can exist at all. So, our challenge is to either show how option B causes harm and is therefore wrong all else equal, or accept that it does not cause harm and is therefore not wrong all else equal. What should we think about this issue? Some philosophers think that we do not, in fact, cause harm in this kind of situation. As long as everyone in the future has lives worth living, it is not worse to bring about individuals with worse lives than individuals with better lives. Other philosophers think that we do, in fact, cause harm in this kind of situation. On impersonal views (the kinds of views that utilitarians tend to accept), the solution is that option A will result in higher levels of well-­being than option B, independently of who experiences that well-­being. On person-­affecting views (the kinds of views that rights theorists tend to accept), the solution is that certain individuals in population A function as counterparts to certain individuals in population B, and so we can compare lives worth living across possible futures for purposes of assessing harms and benefits (see, for example, Višak 2016). How much depends on this issue? That depends on how we answer some of the other questions that we have asked here. For example, to the degree that climate change will produce more wild animals with lives not worth living, we can say that climate change will harm wild animals whether or not we think the non-­identity problem has a solution. In contrast, to the degree that climate change will not produce more wild animals with lives not worth living, we can say that climate change will harm wild animals only if we think that the non-­identity problem has a solution. Of course, we will not be able to resolve these issues here. But note that, if we want to claim that climate change will harm future humans by causing them more pain and suffering even if they still have lives worth living, then we will have to find a solution to the non-­identity problem, in which case this solution will likely apply in the nonhuman case as well. The final question that we will consider here (though there are many others as well) is: Does average or total well-­being matter more? This question raises what Derek Parfit calls the repugnant conclusion (Parfit 1984: 381–90). Roughly speaking, the problem here is that we run into seemingly implausible results no matter what we say. On one hand, if we say that total well-­being matters more than average well-­being, then we seem to be committed to the idea that a world with, say, 10 quintillion “happy insects” (each of whom, we can stipulate, has a life containing one unit of well-­being) is better than a world with, say, 10 billion happy, flourishing humans (each of whom, we can stipulate, has a life containing one million units of well-­being). Why? Because the insect world would contain 10 quintillion (1e+19) units of well-­being overall, whereas the human would contain only 10 quadrillion (1e+16) units of well-­being overall. On the other hand, if we say that average well-­being matters more than total well-­being, then we seem to be

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

60  Jeff Sebo committed to the idea that a world with, say, 10 billion happy, flourishing humans is better than a world with 10 billion happy, flourishing humans and 10 quintillion happy insects. Why? Because the average quality of life in the human-­only world would be higher than the average quality of life in the human and nonhuman world. Intuitively, many people want to reject both of these conclusions. This raises the question: Which, if either, should we accept, and why? What should we think about this issue? For some philosophers, the solution is to simply accept one of these conclusions, no matter how implausible it may appear to be. For example, total utilitarians (the more common type) think that we should accept that total well-­being matters more than average well-­being (and therefore that a world with 10 quintillion happy insects is better than a world with 10 billion happy humans all else equal), and average utilitarians (the less common type) think that we should accept that average well-­being matters more than total well-­being (and therefore that a world with 10 billion happy humans alone is better than a world with 10 billion happy humans and 10 quintillion happy insects all else equal). For other philosophers, the solution is to reject the reasoning that leads to both conclusions, sometimes in ways that concern other issues that we have discussed here. For example, if we raise the baseline for a life worth living high enough, then we can say that the insects in question are “miserable” rather than “happy,” and that their lives are adding negative rather than positive value to the world. Of course, in this case we might still face a version of this problem with other kinds of animal. For example, even if we are no longer committed to the idea that a world with 10 quintillion insects is better than a world with 10 billion humans, we might still be committed to the idea that a world with, say, 10 quadrillion squirrels is better than a world with 10 billion humans. But maybe this bullet is easier to bite. How much depends on this issue? Again, it depends on how we answer some of the other questions that we have considered here. For example, if we think that climate change will increase or decrease both average and total well-­being in the world, then we will not encounter this problem. Whereas if we think that climate change will increase average well-­being but decrease total well-­being in the world (or vice versa), then we will encounter this problem, and it will matter whether or not we have a solution. Thus, for example, suppose that all wild animals have lives worth living, that r-­strategists have better lives in total and K-­strategists have ­better lives on average, and that climate change will increase the ratio of r- to K-­strategists in the world. In this case, climate change will increase total well-­being and decrease average well-­being in the world all else equal. Now suppose, in contrast, that all wild animals have lives not worth living, that r-­strategists have worse lives in total and K-­strategists have worse lives on average, and that climate change will increase the ratio of r- to K-­strategists in the world. In this case, climate change will decrease total well-­being and increase average well-­being in the world all else equal.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  61 There are many other questions that we need to consider in this area as well, for example about whether or not we have equally strong duties to nearby and distant future individuals (see Chapter 4 by Mintz-­Woo in this volume for discussion), and about whether or not we are epistemically and practically capable of  predicting what future individuals will need and providing it for them (see Beckerman  1999). Without going into too much detail about these questions here, we can note two related points of contact with the questions that we have considered here. First, if we accept an impartial moral theory, we will accept that we have strong and direct duties to distant future animals (at least in theory), whereas if we accept a partial moral theory, we may or may not, depending on the details. Second, insofar as we make progress researching large-­scale interventions in wild animal suffering, we might accept that we have strong and direct duties to distant future animals in practice as well (i.e., we might accept that we morally ought to engage in bioengineering or geoengineering programs designed to promote the interests of distant future human and nonhuman animals). Whereas insofar as we do not make progress researching these interventions, we might accept that we have only weak and/or indirect duties to distant future animals in practice (i.e., we might think that we morally ought to attempt to bring about a more just future society in which people can treat human and nonhuman animals well, rather than attempting to treat them well directly). Some of these views will be more revisionary than others. With that said, I think that any reasonable view about our duties to future animals will be at least somewhat revisionary, since there are potentially so many more future animals than present animals that we would have to set an implausibly steep discount rate, in theory or in practice, in order to vindicate anything like our current moral and political priorities.

6. Conclusion My aim in this chapter has been to survey some, but not remotely all, of the connections between animals and climate change, as well as some, but not remotely all, of the scientific and philosophical questions that we must ask in light of these impacts. I have not, of course, attempted to answer all of these questions here, though I did come closer to doing that in the case of farmed animals than in the case of wild animals. In the case of farmed animals: Since animal agriculture is a leading cause of climate change, we should resist animal agriculture in our efforts to mitigate climate change. The question is whether and how we can do that in an ethical and effective way. I have indicated why I think we can move in this direction individually as well as collectively without doing more harm than good or violating individual rights, but much more needs to be said before we can fully evaluate this approach to environmental advocacy.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

62  Jeff Sebo In the case of wild animals: If climate change will be a major contributor to wild animal suffering in the future, then we should assist wild animals in our efforts to adapt to climate change. How should we do so in particular? We will need to make more progress on the many issues that we have considered here before we can say for sure. However, I think that we can make two general, provisional observations for now. First, we cannot wait for consensus about the many issues that we have con­ sidered here in order to take action. We may never achieve consensus about these issues, yet the fact remains that climate change will have an increasingly massive and pervasive impact on all sentient life on this planet with each passing decade. We need to confront that reality now and see what, if anything, we can do now that will make sense no matter how we answer many of the questions that we have asked here. Otherwise we will be no better with respect to animals and climate change in particular than the many politicians who say, “We need another decade of research before we can take action” are with respect to climate change in general. Second, as we have seen, there are at least two general courses of action that we can take now that will make sense no matter how we answer many of the questions that we have asked here. First, we can make epistemic progress by researching what we should aim for and how we should aim for it. Should we resist or embrace the impact that climate change will have on wild animal populations, and, either way, how can we help surviving animals to flourish as much as pos­ sible? And, second, we can make practical progress by advocating for moral and political standing for wild animals, so that we will have the social and political capital necessary for taking the interests of wild animals into account as we recreate human societies in the face of climate change. If we take these steps now, then we will be more likely to know what to do and be able to do it when the time comes—no matter what that happens to be (see Horta 2010 for similar recommendations). Granted, the idea that we might have a moral or political duty to resist animal agriculture or assist wild animals raises further questions about how restrictive and demanding such policies would be. This is especially true of the latter idea. Insofar as we extend moral and political standing to nonhuman animals, the class of moral and political agents, i.e., those of us who have moral and political duties, will be much smaller than the class of moral and political patients, i.e., those of us who have moral and political rights. So, we will have to decide how willing we are to accept a conception of private and public morality according to which humans must place many new demands and restrictions on our behavior in order to reduce and repair the harms that human activity is causing to nonhuman animals. Indeed, we might even wonder whether such a situation is compatible with what Hume (1998), and then Rawls (1999), called “the circumstances of justice,” i.e., the circumstances in which justice is possible and necessary to try to achieve (see Plunkett 2016 for more on this issue).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  63 These concerns are reasonable. However, I think that it would be premature to dismiss the idea that we morally ought to resist animal agriculture and/or assist wild animals on the basis of these concerns at this point. After all, even if we think that there are limits to how restrictive or demanding morality can be, we might also think that substantial self-­sacrifice is called for in a situation in which potentially quintillions of lives are at stake as a result of our activity. Also, even if it would be excessively restrictive or demanding for us to fully address these issues now, it might not be quite as restrictive or demanding for us to do so in the future, especially once we have new alternatives to animal agriculture, new interventions in wild animal suffering, and new social, political, and economic systems for implementing them. Thus, for example, if we focus on researching interventions in wild animal suffering and extending moral and political standing to wild animals right now, we might find that we can help them in part simply by considering and representing their interests as we work to recreate our own communities to cope with climate change later on.14

References Animal Charity Evaluators (2012). “Top Animal Charities and Climate Change.” https:// animalcharityevaluators.org/blog/top-animal-charities-and-climate-change/. Animal Charity Evaluators (2016). “Where Donations Go.” https://animalcharityevaluators.org/research/donation-impact/where-donations-go/. Beckerman, W. (1999). “Sustainable Development and Our Obligations to Future Generations,” in Fairness and Futurity: Essays on Environmental Sustainability and Social Justice, ed. A. Dobson. Oxford: Oxford University Press, 71–92. Bellard, C., C. Bertelsmeier, P. Leadley, W. Thuiller, and F. Courchamp (2012). “Impacts of Climate Change on the Future of Biodiversity.” Ecology Letters 15: 365–377. Benatar, D. (2006). Better Never to Have Been. Oxford: Oxford University Press. Bergengruen, V. (2017). “Trump May Doubt Climate Change, but Pentagon Sees it as a ‘Threat Multiplier’.” McClatchy. http://www.mcclatchydc.com/news/nation-world/ national/national-security/article153908824.html. Bradley, B. (2015). Well-Being. Cambridge: Polity Press. 14  Thanks to Sue Donaldson, Maxwell Frye, Coby Gibson, Dale Jamieson, Tyler John, Michael Morck, Josh Millburn, Kian Mintz-­Woo, Ira Richardson, Adrian Russian, Christopher Schlottmann, Brian Tomasik, the Editors of this volume, and the organizers and participants at the 2016 Philosophy and Climate Change Conference at Princeton University, the 2016 Value Theory Workshop at the University of North Carolina at Chapel Hill, and the 2017 Conference on Convergences and Divergences between Animal and Environmental Ethics at McGill University for helpful feedback and discussion. Finally, thanks to Peter Ohlin and Walter Sinnott-­Armstrong at OUP for supporting me in expanding this chapter into a book, and thanks to the Editors of this volume for setting this expanded project in motion.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

64  Jeff Sebo Cochrane, A. (2010). An Introduction to Animals and Political Theory. New York: Palgrave Macmillan. DeGrazia, D. (2002). Animal Rights. Oxford: Oxford University Press. Díaz, S., J.  Fargione, F.  S.  Chapin III, and D.  Tilman (2006). “Biodiversity Loss Threatens Human Well-Being.” PLoS Biology 4(8): e277. https://doi.org/10.1371/ journal.pbio.0040277. Donaldson, S. and W. Kymlicka (2011). Zoopolis: A Political Theory of Animal Rights. Oxford: Oxford University Press. Encyclopedia Smithsonian (2016). “Numbers of Insects (Species and Individuals).” https://www.si.edu/Encyclopedia_SI/nmnh/buginfo/bugnos.htm. FAO (2016). “FAOSTAT Data.” http://www.fao.org/faostat/en/#data. Garnett, T. (2016). “Plating up Solutions.” Science 353: 1202–1204. Gerber, P. J., H. Steinfeld, B. Henderson, A. Mottet, C. Opio, J. Dijkman, A. Falcucci, and G.  Tempio (2013). Tackling Climate Change through Livestock: A Global Assessment of Emissions and Mitigation Opportunities. Rome: Food and Agriculture Organization of the United Nations (FAO). Gerland, P., A.  E.  Raftery, H.  Ševcíková, N.  Li, D.  Gu, T.  Spoorenberg, L.  Alkema, B. K. Fosdick, J. Chunn, N. Lalic, G. Bay, T. Buettner, G. K. Heilig, and J. Wilmoth (2014). “World Population Stabilization Unlikely This Century,” Science 346: 234–237. Goodland, R. and J. Anhang (2009). “Livestock and Climate Change.” World Watch (November/December): 10–19. Groff, Z. and Y.-K. Ng (2019). “Does Suffering Dominate Enjoyment in the Animal Kingdom? An Update to Welfare Biology.” Biology and Philosophy 34(4): 40. Gruen, L. (2014). “Dignity, Captivity, and an Ethics of Sight,” in The Ethics of Captivity, ed. L. Gruen. Oxford: Oxford University Press, 241–247. Horta, O. (2010). “Debunking the Idyllic View of Natural Processes: Population Dynamics and Suffering in the Wild.” Télos 17: 73–88. Hsiung, W. and C.  Sunstein (2007). “Climate Change and Animals.” University of Pennsylvania Law Review 155: 1695–1740. Hume, D. (1998). An Enquiry Concerning the Principles of Morals. Oxford: Oxford University Press. Hursthouse, R. (2011). “Virtue Ethics and the Treatment of Animals,” in The Oxford Handbook of Animal Ethics, ed. T. L. Beauchamp and R. G. Frey. Oxford: Oxford University Press, 119–143. Hyslop, A. (2014). “Other Minds.” The Stanford Encyclopedia of Philosophy, ed. E. N. Zalta. https://plato.stanford.edu/archives/spr2016/entries/other-minds/. IPCC (2014). Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Core Writing Team: R. K. Pachauri and L. A. Meyer. Geneva: IPCC. Kolbert, E. (2014). The Sixth Extinction: An Unnatural History. New York: Henry Holt.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Animals and Climate Change  65 Korsgaard, C. (2011). “Interacting with Animals: A Kantian Account,” in The Oxford Handbook of Animal Ethics, ed. T. L. Beauchamp and R. G. Frey. Oxford: Oxford University Press, 91–118. Kremer, M. (1993). “Population Growth and Technological Change: One Million B.C. to 1990.” The Quarterly Journal of Economics 108: 681–716. Mendl, M. and E. Paul (2016). “Bee happy.” Science 353: 1499–1500. NASA (2016). “The Consequences of Climate Change”. http://climate.nasa.gov/effects/. Ord, T. (2013). “Why I’m Not a Negative Utilitarian.” http://www.amirrorclear.net/ academic/ideas/negative-utilitarianism/. Ng, Y.-K. (1995). “Towards Welfare Biology: Evolutionary Economics of Animal Consciousness and Suffering.” Biol Philos 10(3): 255–285. https://doi.org/10.1007/ bf00852469 Pachirat, T. (2013). Every Twelve Seconds. New Haven: Yale University Press. Palmer, C. (2010). Animal Ethics in Context. New York: Columbia University Press. Palmer, C. (2011). “Does Nature Matter? The Place of the Nonhuman in the Ethics of Climate Change,” in The Ethics of Global Climate Change, ed. D.  G.  Arnold. Cambridge: Cambridge University Press, 272–291. Parfit, D. (1984). Reasons and Persons. Oxford: Clarendon Press. Pershing, A.  J., L.  B.  Christensen, N.  R.  Record, G.  D.  Sherwood, and P.  B.  Stetson (2010). “The Impact of Whaling on the Ocean Carbon Cycle: Why Bigger Was Better.” PLoS ONE 5(8): e12444. doi:10.1371/journal.pone.0012444. Pew Commission on Industrial Farm Animal Production (2008). Putting Meat on the Table. http://www.pewtrusts.org/en/research-and-analysis/reports/2008/04/29/ putting-meat-on-the-table-industrial-farm-animal-production-in-america. Philpott, T. (2016). “Hurricane Matthew Killed Millions of Farm Animals in North Carolina.” Mother Jones, October 14. http://www.motherjones.com/environment/ 2016/10/hurricane-matthew-killed-animals-hog-poop. Plunkett, D. (2016). “Justice, Non-Human Animals, and the Methodology of Political Philosophy.” Jurisprudence 7: 1–29. Ranganathan, J., D.  Vennard, R.  Waite, B.  Lipinski, T.  Searchinger et al. (2016). “Shifting Diets for a Sustainable Food Future.” Working Paper, Installment 11 of Creating a Sustainable Food Future. World Resources Institute. doi:10.13140/ RG.2.1.3808.2961. Rawls, J. (1999). A Theory of Justice. Revised Edition. Cambridge, MA: Harvard University Press. Regan, T. (2004). The Case for Animal Rights. Berkeley: University of California Press. Sebo, J. (2018). “The Moral Problem of Other Minds.” The Harvard Review of Philosophy 25: 51–70. Shue, H. (1999). “Global Environment and International Inequality.” International Affairs 75: 531–545.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

66  Jeff Sebo Singer, P. (2009). Animal Liberation. New York: Harper Perennial Modern Classics. Stehfest, E., L.  Bouwman, D.  P.  van Vuuren, M.  G.  J.  den Elzen, B.  Eickhout, and P. Kabat (2009). “Climate Benefits of Changing Diet.” Climatic Change 95: 83–102. Steinfeld, H., P. Gerber, T. Wassenaar, V. Castel, M. Rosales, and C. de Haan (2006). Livestock’s Long Shadow. Rome: Food and Agriculture Organization of the United Nations. http://www.fao.org/docrep/010/a0701e/a0701e00.htm. Tilman, D., C.  Balzer, J.  Hill, and B.  Befort (2011). “Global Food Demand and the Sustainable Intensification of Agriculture.” Proceedings of the National Academy of Sciences of the United States of America 108: 20260–20264. Tomasik, B. (2008). “Climate Change and Wild Animals.” Last modified December 13, 2016. http://reducing-suffering.org/climate-change-and-wild-animals/. Tomasik, B. (2015). “The Importance of Wild Animal Suffering.” Relations 3: 133–152. Višak, T. (2016). “Do Utilitarians Need to Accept the Replaceability Argument?” in The Ethics of Killing Animals, ed. T. Višak and R. Garner. Oxford: Oxford University Press, 117–135. Vogel, G. (2017). “Where Have all the Insects Gone?” Science 356: 576–579.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

3

Discounting under Risk Utilitarianism vs. Prioritarianism Maddalena Ferranna

1. Introduction The debate on the economics of climate change has focused primarily on the choice of the social discount rate, which plays a key role in determining the desirability of climate policies given the long-­ term impacts of climate damages. Discounted utilitarianism and the Ramsey Rule dominate the debate on discounting. The chapter examines the appropriateness of the utilitarian framework for evaluating public policies. More specifically, it focuses on the risky dimension of climate change, and on the failure of utilitarianism in expressing both concerns for the distribution of risks across the population and concerns for the occurrence of catastrophic outcomes. I show how a shift to the prioritarian paradigm is able to capture those types of concerns, and briefly sketch the main implications for the choice of the social discount rate. The social discount rate represents the weight attached to the future monetary consequences of a public policy. Suppose that the policy maker has to decide about the introduction of a carbon tax. The tax will likely produce some negative consequences in the short term, such as a temporary increase in energy bills and the shut-­down of a few coal power plants, with a resulting rise in unemployment. The tax will also accelerate the transition to clean energy production, which will curb the generation of carbon emissions and limit the increase in atmospheric temperature, thereby reducing the damages caused by a hotter climate. Assume that the policy will cost $5 billion today, while the future benefits can be estimated at $100 billion in 100 years (where the future $ are inflation-­adjusted).1 Let us further assume that those figures completely capture all the effects of the policy.2 1  The tax will likely create benefits also in the short term, e.g. the creation of new jobs in the renewable energy sector, or improvements in air quality, and associated health improvements, due to the reduced use of fossil fuels. However, for the sake of simplicity, I will be very conservative and assume that in the short term there are only costs, while all the benefits accrue in the long term. 2  In other words, I am introducing the simplifying assumption that all the relevant impacts of the policy are fully encapsulated in those monetary amounts. This is a standard assumption in the

Maddalena Ferranna, Discounting under Risk: Utilitarianism vs. Prioritarianism In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Maddalena Ferranna. DOI: 10.1093/oso/9780198796282.003.0004

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

68  Maddalena Ferranna Present value of $1 in the future 1.2 1

$ value

0.8 0.6 0.4 0.2 0

0

10

20

30

40

50

60

70

80

90

100

time (t) 1.4%

3%

5.5%

7%

Figure 3.1  The current value of $1 received in t years depending on the social discount rate. $1 where The present value (PV) of $1 in the future is computed as: PV  (1  SDR)t SDR represents the social discount rate

Ought the policy to be introduced? The answer depends on how to compare future benefits to current costs. If $1 in 100 years is equivalent to $1 today, then the policy ought to be financed because the estimated benefits are far larger than the costs. But if $1 in 100 years is valued less than $1 today, it might be better not to finance the policy. Figure 3.1 depicts the dramatic consequences of choosing a  large social discount rate. For example, if the social discount rate is equal to 1.4  percent (the rate proposed by Stern 2006), $1 in 100 years is equivalent to $0.25 today, and the policy will be financed.3 If the social discount rate is 5.5 percent (the rate proposed by Nordhaus 2007), $1 in 100 years is equivalent to just $0.005 today: $100 billion in 100 years are valued only $500 million today; as a consequence, the policy will not be introduced.4 The theory of discounting analyzes to economic analysis of the goodness of a climate intervention – the type of analysis I am referring to in the chapter. The use of a single monetary metric presupposes our ability to accurately quantify all impacts of climate change in monetary terms. There is a lively debate on whether it is appropriate to monetize all impacts, and how to accomplish that (e.g. should we put a $ value to saving a life? how much is a life worth?). Since the chapter focuses on the limits of utilitarianism and the consequences for the discount rate, I will overlook this debate, and assume all monetary amounts fully capture the positive and negative impacts of climate change. 3  With a social discount rate of 1.4%, $100 billion in 100 years are equivalent to $25 billion today (0.25*100), larger than the cost of $5 billion. 4  A further increase in the discount rate will also undermine policies with medium-­term effects: for example, if the discount rate is 7% (the rate suggested by OMB in 2003 and recently re-­instated by the Trump administration for all types of cost–benefit analysis), $1 in 50 years is valued only $0.034

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Discounting under Risk  69 which extent the passing of time reduces the value of future benefits for ethical purposes. The question of discounting reflects a consequentialist approach to public decision-­making. The desirability of a public policy is assessed based on the outcomes that are produced. The policy maker should choose the policy that brings about the best state of affairs as possible, or, equivalently, that contributes the most to overall good. If the future consequences of a policy are found to contribute less to overall good than otherwise-­similar hypothetical consequences occurring today, then we are introducing discounting. Formally, each future policy impact is multiplied by a weight (the so-­called social discount factor) that reflects the importance of that particular impact for the maximization of overall good; if the weight is lower than 1, then there is discounting.5 The social discount rate is a number that represents the normative importance of the future impacts of a policy with respect to the current ones. Clearly, its choice depends on how we define overall good (usually represented through a ‘social welfare function’ in economic theory). In the economics of climate change, most of the literature adopts a utilitarian approach, where overall good is represented by summing well-­being across individuals. Moreover, the literature makes the additional simplifying assumption that all relevant dimensions of well-­being are embodied in a single ‘consumption’ index, expressed in monetary values.6 For example, well-­being could positively depend on receiving a sufficiently high income, having good health, living in a non-­polluted environment, benefitting from basic rights, and so on. The climate change literature makes the assumption that all non-­monetary ‘goods’ (health, nature . . .) can be expressed in monetary terms. The main consequence is that now individual well-­being depends on a single good, usually referred as ‘consumption’. Therefore, the benefits of a climate policy are measured in monetary amounts (adjusted by inflation), and, loosely speaking, the social discount rate defines how much less the policy maker prefers to give money to future generations compared to the current generation, on the grounds that overall good would increase to a minor extent.

today (EPA 2018). The high discount rate contributes to explain the very low current social cost of carbon adopted by the U.S.  ($1-$7 for the year 2020); the low value is due also to the decision of excluding the climate damages attributable to U.S. economic activitiy but suffered by the rest of the world. To draw a comparison, the central value of the social cost of carbon computed by the US ­government during the Obama administration used a 3% discount rate, resulting in a social cost of carbon for the year 2020 of about $40 (IWG 2016). 5  The social discount factor (SDF) for an impact occurring in t years is inversely related to the 1 social discount rate (SDR): = , exactly as described in the caption of Figure 3.1. (1 + SDR)t 6  See Greaves (2017) for a discussion of the main assumptions made in the economics literature of climate change, and the consequences for discounting. As mentioned in note 2, the assumption that all impacts can (and should) be properly monetized is highly debatable, but we will take it for granted here.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

70  Maddalena Ferranna In a deterministic context (i.e., one where there is no risk), the utilitarian social welfare function seems a suitable framework for the evaluation of public policies. Under the common assumption of decreasing marginal utility with respect to consumption, utilitarianism is sensitive to the distribution of consumption across the population, by giving more weight to the poor than to the rich. According to this view, the future returns of a climate policy should be discounted if they will benefit individuals that are richer than the current individuals sustaining the climate policy costs.7 Moreover, since well-­being depends only on a consumption index, changing the sensitivity of welfare to consumption level can easily accommodate different ethical intuitions, from egalitarianism (where only the policy impact to the poorest individual gets a positive weight) to prioritarianism (where impact to the poor gets larger weight than under utilitarianism).8 In contrast, once we introduce risk over future consumption levels9 (where consumption is a proxy for well-­being), utilitarianism imposes some very re­strict­ ive assumptions on what a policy maker should do. Indeed, by definition, the form of the utility function represents both concerns for inequality in consumption levels and concerns for risk in consumption.10 Risk aversion is usually considered a matter of population preference, while inequality aversion is an ethical choice. As a consequence, by assuming that the two attitudes coincide, we are suggesting that the policy maker ought to be as much inequality averse as individuals are risk averse; for example, if the population is risk neutral, then the policy maker should be inequality neutral, and attach the same weight to rich and poor.11 Moreover, as will be more clear from the discussion in the following

7  The climate change economics literature usually makes the additional assumption of a positive utility discount rate, or pure rate of time preference, according to which future generations should have a lower weight than current ones because they do not exist today. There is a vast literature on the suitability of discounting. I neglect this issue here, and I focus on other characteristics of the welfare approach that matter for the policy evaluation. 8  In other words, it is virtually impossible to distinguish the behavior of a policy maker that gives priority to the worst off (a prioritarian one) from the behavior of a utilitarian policy maker who is very sensitive to low consumption levels, as long as the worst off coincides with the poor. See also Fleurbaey et al. (2019) and Ferranna and Fleurbaey (forthcoming) for a discussion of this point. 9  With “risk” I mean that the evaluator is uncertain about which outcome will occur in the future among a set of possible outcomes. This uncertainty is represented by a probability distribution over the set of outcomes, i.e., a probability for each possible outcome. In the framework considered in the chapter, the only outcome is the consumption index, as a proxy for well-­being. Thus, “risk” or “consumption risk” means that we do not know how much consumption future generations will enjoy (or, equivalently, what will be the well-­being level of future generations). The risk may be associated to uncertainty about economic growth, the occurrence of international conflicts, climate disasters, etc. 10  See also Fleurbaey et al. (2019) for a discussion about the disentanglement between risk attitudes and inequality attitudes. 11  In the utilitatian context, individual risk aversion dictates also how much the policy maker should be risk averse. In particular, since most climate-­economy models assume that all individuals are equally risk averse, the policy maker preference for risk is exactly the same as the population’s one. The coincidence between the risk attitudes of the population and the risk attitudes that the policy maker should display is less problematic, as it is usually argued (especially in economic theory) that the policy maker should respect individuals’ preferences, without being paternalistic. See

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Discounting under Risk  71 s­ ection, if the policy maker cares equally about risk and about inequality, then she is indifferent to both the distribution of individual risks (i.e., if some individuals or generations are more at risk than others) and to the occurrence of catastrophic outcomes (i.e., if it is more likely that individuals’ fates are independent or positively correlated—and potentially exposed to a catastrophic outcome where every­ one suffers a loss). Indeed, by definition, utilitarianism sums individual well-­beings without attaching a particular weight to some individuals. As a consequence, an individual at high risk has the same priority as an individual at low risk, and the distribution of individual risks (i.e., who is more at risk) does not matter. Similarly, since utilitarianism simply sums individual well-­beings, it does not attach special consideration to interdependence across individual risks. For example, if we have two individuals exposed to the same risk of suffering damages, the utilitarian policy maker does not care if both of them will actually suffer the loss (i.e., the risks are interdependent) or if it is more likely that only one of them will; the utilitarian policy maker cares only about the fact that both individuals are at risk.12 This chapter discusses how utilitarianism fails to properly consider distributional issues in the presence of risk, and shows how a prioritarian approach (or, at least, some forms of prioritarianism) could overcome those limits, by giving priority to either the individual who faces the worst risk (thereby expressing concerns for the distribution of individual risks), or by giving priority to the state of affairs with the highest realized inequality (thereby expressing concerns for how individual outcomes should correlate). The shift from utilitarianism to prioritarianism has important consequences for discounting. In particular, if the policy maker cares about the distribution of individual risks, and if future individuals, even though at risk, are expected to have better chances than current ones, then the policy maker should discount more heavily than a utilitarian one. In contrast, if the policy maker dislikes ex-­post inequality, i.e., she has a preference for realized correlated outcomes, then we would soften the controversial Weitzman result, according to which the risk of large consumption loss due to climate change should lead to infinite precautionary motives (Weitzman 2009). The chapter is organized as follows. Section 2 examines three different social criteria for aggregating concerns for risk and concerns for inequality: the standard utilitarian framework (Harsanyi 1955); a framework that focuses on the distribution of individual risks (advocated, e.g., by Diamond 1967 and Epstein and

Rabinowicz (2002) for an argument supporting extra moral risk aversion on behalf of the policy evaluator, even if that imposes a paternalistic precautionary attitude on the population. 12  For example, Posner (2004) suggests the idea that we should add a ‘premium’, or additional value, to policies that prevent catastrophic outcomes, such as many deaths at the same time. Such premium would capture the horror of extinction, which reflects the feeling that the loss in welfare derived by many simultaneous fatalities is more than the mere sum of individual losses. A utilitarian approach would care only about the latter, while neglecting the correlation among individuals’ well-­beings.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

72  Maddalena Ferranna Segal 1992); a framework that focuses on collective risks and the distribution of realized outcomes (advocated, e.g., by Hammond 1983, Broome 1984, and Fleurbaey 2010). I will then illustrate how a prioritarian social welfare function could represent the previous preferences by disentangling risk and equity concerns. In particular, ex-­ante prioritarianism can display preferences for equity in individual risks, while ex-­post prioritarianism in the form of an expected equally distributed equivalent as proposed by Fleurbaey (2010) can display preferences for the actual distribution of outcomes and the form of the collective risk, while neatly disentangling risk and equity preferences.13 Section 3 introduces the notion of discounting under utilitarianism, examines the famous Ramsey Rule and then discusses the implications of prioritarianism for discounting. Finally, Section 4 concludes.

2.  Choice of the Social Welfare Framework: Utilitarianism vs. Prioritarianism 2.1  Risk and Equity: Aggregation Issues A well-­known problem in social choice is how to determine a procedure for collective decision-­ making in the presence of risk. In a famous contribution, Harsanyi (1955) argued that we should use the utilitarian criterion, where the decision maker operates behind “the veil of ignorance” (i.e., without knowing individuals’ identity), and measures total welfare by summing individual expected utilities, or, equivalently, the expected value of the sum of individual utilities. Mathematically, let ui , s denote the utility of individual i in state of nature s, and ps the probability of state s; let us assume that there are N individuals and S states of nature. Then, the expected utilitarian social welfare function is equal to:

N

S

W U = ∑∑ ps ui ,s =i 1 =s 1

Harsanyi defended utilitarianism on the ground that it is the only aggregation method able to both respect individual preferences for risk and lead to collective rational choices. Indeed, Harsanyi (1955) proved that the social welfare function is a linear combination of individuals’ utility functions if: (i) society maximizes expected social welfare; (ii) individuals maximize expected utility; (iii) society is indifferent between two probability distributions whenever all individuals are. The result depends on the fact that, behind the “veil of ignorance”, the decision 13  As I will argue later, the most common ex-­post prioritarian approach (Adler and Sanchirico 2006) does not disentangle risk and equity concerns.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Discounting under Risk  73 maker does not know which type of individual she will be, i.e., which individual risk she will face. As a consequence, inequality across individuals is identical to inequality (in the sense of risk) over states of nature. However, the criterion has been recurrently criticized for being unable to ad­equate­ly evaluate inequality in the presence of risk.14 In particular, it fails to capture two types of concerns: (a) concerns for the distribution of individual risks, i.e., whether the risk is equally shared among people or concentrated on a specific group (concern for ex-­ante equity); (b) concerns for the collective consequences of the risk, i.e., whether the ­realizations are likely to be equally distributed across the population, or whether the negative effects of the risk are borne only by a fraction of it (concern for ex-­post equity). Note that the first situation could lead also to catastrophic outcomes, where everyone in the population suffers a loss. The ex-­ante approach to equity postulates that the decision maker should be sensitive to the distribution of individual risks, and describes whether the decision maker prefers to grant the same chances to everyone or to concentrate the risk on few individuals. Consider, for instance, the rise of an epidemic, and the choice between an imperfect vaccine and quarantine. Assume that the two options are mutually exclusive and equally effective, and that it is not possible to perfectly recognize the symptoms of the disease (e.g., if someone is fevered, he might have contracted the disease, but it could also be a heatstroke). If the policy maker opts for the vaccine, then she will prefer equally shared risks to concentrated risks: thanks to vaccination, the probability that an individual will contract the disease has decreased, but since the vaccine is imperfect, there is still a chance that the disease will spread, thereby putting everyone at risk. If the policy maker chooses the quarantine, she will prefer to put a small part of the population at a large risk rather than spread a small risk to the entire population: by separating those that are at higher risk and/or those that seem to have contracted the disease, the policy maker is protecting the majority of the population while endangering a few. In contrast, the ex-­post approach to equity focuses on the realizations of the individual risks, and on how they are correlated. In particular, it describes whether the decision maker prefers fair situations where every individual bears a  loss (a ‘misery loves company’ result) over unfair situations where only few ­individuals sustain a loss. Note that, if the policy maker prefers to have states with no ex-­post inequality, she will also be ready to accept states with potential 14  Prominent criticism includes Diamond (1967), Sen (1973), Myerson (1981), Broome (1984), Epstein and Segal (1992), Ben-­Porath et al. (1997), and Fleurbaey (2010).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

74  Maddalena Ferranna catastrophic outcomes, where, for instance, the majority of the population suffers a loss. Consider, for instance, urban planning strategies. To which extent should the policy maker encourage concentration of people and assets on productive areas? Concentration will likely reduce initial inequalities, but, at the same time, induce a catastrophic loss in case the areas are struck by a natural disaster. To understand the limits of utilitarianism, and its inadequacy in addressing concerns for either ex-­ante and/or ex-­post equity, let us consider the following three lotteries (which constitute a canonical example of Diamond 1967). Assume that there are only two individuals in the economy, and two possible states of nature, where a state of nature describes the level of utility (or well-­being) enjoyed by each individual. In each of the following matrices, rows stand for individuals, columns for states with equal probabilities, and cells for utility levels. 1 1 A:   0 0

1 0  B:   0 1 

1 0  C:   1 0 

Assume that there is a benevolent policy maker,15 who has to choose between the three lotteries. Under utilitarianism, the three lotteries are equivalent, as the decision maker would simply compute the utility level associated to each cell, and then sum them. As a consequence, she acts as if each cell were a potential alternative for a hypothetical representative agent. For example, in lottery A, the utilitarian policy evaluator cares only about the fact that with probability 0.5 one unit of utility will be enjoyed, and with probability 0.5 no unit will be enojoyed. Similarly for the other two lotteries. Therefore, we can say that the utilitarian decision maker cares only about the average individual utility risk, or, equivalently, about the expected distribution of utilities across individuals. Clearly, though, the three lotteries differ in terms of fairness. For example, in lottery B and C the two individuals face the same risk: each will lose with prob­ abil­ity 0.5. In lottery A, instead, only one agent will face a loss for sure. Hence, the policy maker may prefer lotteries B and C over lottery A because in the latter individual utility risks are unequally distributed between the two individuals. In that case, we say that the policy maker has ex-­ante equity concerns, i.e., concerns for the distribution of individual risks. Moreover, if we look at the collective consequences of the risk, in lottery A and B only one agent suffers a loss in each state of nature, while in lottery C either both agents bear a loss or neither of them. Therefore, in lottery C there is a more equitable distribution of utility in each state of nature with respect to lottery A and B. The downside is the presence of a collective risk: in the first two lotteries 15  With benevolent policy maker, I mean an external agent who has no private interest in the lottery choice, and who cares only about maximizing total goodness (or welfare) in the economy. The lottery choice will depend on how total welfare is characterized.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Discounting under Risk  75 we know with certainty the distribution of realized well-­being (one agent enjoys 1 unit of utility and the other 0), while in lottery C there is uncertainty about the final distribution of well-­being. If the policy maker has preferences about how individual outcomes should correlate (as opposed to risks), we say that she has ex-­post equity concerns. The difference among the evaluation of the three lotteries can be modeled also in terms of the order in which risk and inequality are processed. If the policy maker first looks at risk and then at inequality, she may argue that in lottery A individuals have different chances, so there is inequality; in lotteries B and C both individuals face the same chances, so there is no inequality. On the other hand, if the policy evaluator first processes inequality and then risk, she may argue that in lottery C there will always be equality after the resolution of uncertainty, while in lotteries A and B there will be inequality ex-­post. Therefore, the order in which the two types of variability are treated matters for defining social preferences. Two aspects are worth highlighting. First of all, the previous examples show that it is possible to neatly separate risk and equity concerns by focusing on one dimension at a time. However, that will give rise to two different ethical theories on the interaction between fairness and risk. One theory looks at individual risks and inequality in distribution of individual chances. The second theory, instead, looks at collective risks, and whether they lead to correlated outcomes or unequal outcomes. The choice between them responds purely to our normative intuitions.16 Second, there is a tension between equity and avoidance of catastrophic outcomes, which has long been recognized also in the literature on public risks:17 the best way to avoid catastrophic outcomes in terms of widespread losses is to concentrate the risk on few individuals, although that may raise equity concerns.

2.2  Social Welfare Function A number of alternative social welfare functions have been proposed to capture the type of concerns raised by the previous examples.18 Here, I will focus only on two social welfare functions belonging to the prioritarian family, the first one ­representing concerns for ex-­ante equity and the second one concerns for ex-­post 16  Although the two approaches are mutually exclusive (one takes a purely ex-­ante perspective while the other a purely ex-­post one), we could imagine a policy evaluator who considers a com­bin­ ation of the two, since she is sometimes interested in ex-­ante risks and sometimes in ex-­post outcomes. This ‘mixed’ approach presents additional challenges to the two extreme ones, due to the need of specifying which intuitions should lead us to the definition of a particular combination (how much weight should be given to ex-­ante concerns with respect to ex-­post ones)? Therefore, in this chapter I will focus exclusively on the two ‘pure’ approaches. See Gajdos and Maurin (2004) as one of the (very) few studies that look at the combination of ex-­ante and ex-­post evaluations. 17  See, for example, Fishburn and Sarin (1991); Keeney (1980). 18  See, for example, Adler (2012) for a review.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

76  Maddalena Ferranna equity. Let j be an increasing function representing aversion to inequality in (expected and/or realized) utility. 1. Ex-­ante prioritarianism: the policy evaluator first consider individual expected utilities and then aggregates across individuals:

N  S  W EAP = ∑j  ∑pS ui , S  =i 1 = S 1 

2. Expected equally distributed equivalent approach, or ex-­post transformed ­prioritarianism (Fleurbaey 2010): the policy evaluator first computes a statedependent social welfare function, and then aggregates over states of the world:

S  N j (ui , s )  W EEDE = ∑pSj −1  ∑  S 1= = i 1 N 

Concavity of j signals that the policy maker does not like unequal distributions of individual risks (ex-­ante case), and unequal distributions of outcomes (ex-­post case). Convexity of j, instead, denotes preferences for concentrated risks (ex-­ante case), and preferences for non-­correlated outcomes which will avoid potential catastrophic scenarios (ex-­post case). To better understand the proposed welfare frameworks, let us try to apply them to the previous example: EAP EAP • Ex-­ ante prioritarianism (EAP): W = ( B) W = (C ) 2j (0.5) > W EAP ( A) = j (1) + j (0) if and only if j is concave. For example, if j = u (concave 2 function): W EAP (B) = W EAP (C ) => 1.41 W EAP ( A) = 1. If j = u (convex fun­ EAP EAP EAP W (C ) = 0.5 < W ( A) = 1. ction): W (B) = EEDE ( A) = W EEDE (B) • Expected equally distributed equivalent (EEDE): W  j (1) + j (0)  For example, if j = u (concave func0.5 > W EEDE (C ) = = j −1  . 2   EEdE EEDE tion): W ( A) = W ( B) = 0.5 > W EEDE (C ) = 0.25 If j = u2 (convex funcEEDE EEDE EEDE tion): W ( A) = W ( B) = 0.5 < W (C ) = 0.7.

Since the two welfare functions separate risk and inequality concerns, in applications there will be two different parameters to calibrate, one that represents aversion to inequality (captured by the form of the function j) and one that represents aversion to risk (captured by the form of the function u). Note that the function j represents well-­being inequality aversion (or propensity for well-­being inequality, if j is convex instead of concave). However, in applications we are mainly interested in consumption inequality aversion; for example, in the discounting problem, the policy evaluator cares about differences in consumption across states of nature and differences in consumption across generations. Nevertheless, it can be shown that the calibration of well-­being inequality aversion will allow us to calibrate also consumption inequality aversion (Ferranna

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Discounting under Risk  77 Table 3.1.  Summary of the main properties of the three approaches under analysis  

Stochastic Sure Time Ex-­ante dominance thing consistency Pareto principle

Util. Yes EAP No EEDE Yes

Yes No Yes

Yes No Yes

Separability Disentanglement of risk and equity concerns

Yes Yes Yes Yes Yes, if no No inequality

No Yes Yes

2019).19 Without entering into details, the two proposed welfare functions allow one to perfectly disentangle risk and equity preferences, and introduce new types of concerns that seem to be quite relevant in evaluating public policies. The choice between an ex-­ ante or ex-­ post approach is an ethical one.20 However, the two approaches differ in the underlying properties, and that difference may induce us to choose one approach over the other. Table 3.1 summarizes the main properties of the two approaches, and compares them to the utilitarian one. Take, for instance, the ex-­ante approach. In that case, the focus of inequality is individual risks, represented by individuals’ expected utilities (ex-­ante Pareto). As a consequence, the ex-­ante approach respects individuals’ preferences over risk, but, unlike utilitarianism, it does not have the form of an expected social utility. The consequence is the violation of principles of stochastic dominance,21 and, in particular, of the sure thing principle,22 which is generally taken to be an 19  In particular, in the ex-­ante case, the concavity of u determines risk aversion, and the shape of j consumption inequality aversion. In the ex-­post case, the combination of j and u determines consumption inequality aversion, and the shape of j −1 risk aversion. 20  As previously noted, I am implicitly assuming an on/off choice between the two approaches (see note 13). More generally, we could combine the two approaches. The choice of how much weight to attach to ex-­ante considerations versus ex-­post ones would still be an ethical choice, which could be governed by looking at the properties of the two approaches and weighing those that seem more ­ethic­al­ly relevant. 21  Consider the following example, drawn by Adler (2012): 9 1 A:   1 9

5 − e B:  5 − e

5−e   5−e 

where rows represent individuals, columns aggregate states, and each entry is a utility level. If we fix the state, the distribution of individual utilities in B is fairer than the one in A: for e sufficiently small, the vectors (9,1) and (1,9) can be ranked worse than the vector (5 − e , 5 − e ). Thus, all outcomes of lottery B are ranked higher than all outcomes of lottery A. Therefore, lottery B stochastically dom­in­ ates lottery A from a collective point of view. However, the ex-­ante approach will prefer A, as each individual expected utility is larger. 22  The sure thing principle (Savage 1954) says that outcomes that occur regardless of the actions that are chosen should not affect the agent’s preferences. Consider again the following two lotteries: 1 1 1 0 A:  B:    0 0 0 1 If state of nature 1 occurs, the results of lottery A and B are the same, as only one agent gets 1. Therefore, state 1 should not affect the final decision. If state of nature 2 occurs, the results of A and B

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

78  Maddalena Ferranna essential requirement of rationality (Broome 1984). Moreover, the ex-­ ante approach is time inconsistent, i.e., optimal choice changes over time even if no information arrives. In contrast, the ex-­post approach retains the form of an expected social utility (thus dominance principles are fulfilled), but does not respect individuals’ preferences, except in the case in which no inequality arises.23 Fleurbaey (2010) argues that respecting individuals’ preferences is much less compelling in an uncertain environment with respect to a risk-­free one, since, in the context of risk, individuals are not perfectly informed about the consequences of their decisions. In particular, when taking a decision under risk, each individual focuses on his own utility, and ignores the correlation with other individuals. However, those correlations matter as they define how bad a collective risk is, and whether society as a whole can recover from it. The drawback, of course, is that there is no concern for individual chances: it matters only who and/or how many individuals face a loss, not if some individuals are more at risk than others. Another drawback of this approach is the fact that it is not separable across individuals: the evaluation of a policy depends also on the well-­being of the individuals that are not directly affected by it. The two social welfare functions previously discussed have a prioritarian-­like form. In the prioritarian world, a common approach to represent concerns for the ex-­post distribution of individual utilities is to use an ex-­post prioritarian welfare function (Adler and Sanchirico 2006), which takes the expectation of the sum of transformed utilities:

N

S

W EPP = ∑∑psj (ui , s ) =i 1 = S 1

However, this approach is not able to disentangle risk and equity preferences, which was one of the reasons for using alternatives to utilitarianism. In particular, the ex-­post prioritarian policy evaluator will be indifferent between the lotteries A, B, and C just like a utilitarian one. Therefore, the combination of the functions j and u will denote both consumption inequality aversion and risk aversion. The difference with respect to utilitarianism is in the level of risk/inequality aversion: if j is concave, the policy evaluator gives more priority to the worse-­off, which is

are symmetric, and therefore indifferent. By the sure thing principle, the policy evaluator should be indifferent between lottery A and lottery B. Instead, a policy evaluator averse to inequality in the distribution of individual risks is found to prefer lottery B. 23  In that case, the policy maker would simply consider the expected utility of a representative agent, which is exactly what is assumed in the standard Ramsey equation.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Discounting under Risk  79 equivalent to having a larger degree of consumption inequality/risk aversion; if j is convex, the opposite holds.24

3.  Implications for Discounting This section discusses the implications in terms of discounting if we do not adopt the utilitarian framework, but one of the alternatives previously described. First of all, I will briefly introduce the theory of discounting and define the discount rate I am going to analyze: the discount rate for safe projects. Then, I will describe the famous Ramsey Rule, and how it changes in the presence of risk by intro­du­ cing the so-­called precautionary effect. Finally, I will explain how the prioritarian approach will affect this precautionary effect.

3.1 Preliminaries Cost–benefit analysis constitutes the common approach to evaluate the desirability of a public policy. Once all the costs and benefits of the policy have been detected, cost–benefit analysis recommends the implementation of the policy only if the temporal distribution of its impacts leads to a ‘social improvement’, which is usually represented as an increase in intertemporal welfare. Given the difficulty in measuring and interpreting changes in welfare, a cost–benefit ana­ lysis exercise usually employs two shortcuts. First of all, it focuses only on the policy impacts that can be monetized and it reduces the policy to a temporal sequence of cash flows. Second, it assumes that the increase in welfare in a given period is simply equivalent to the cash flow received in that period scaled by an exogenous factor (the discount factor) which describes how much we value having a monetary flow in that particular period.25 The discount factor is inversely related to the discount rate. A positive discount rate means that future monetary values matter less than current ones: the larger the discount rate, the lower the weight attached to future monetary benefits/costs. Conversely, a negative discount rate means that future valuations are weighed more than current ones.

24 See also Bovens (2015) on the difference between what he calls the ‘Prioritarian View’ (expressed in ex-­post prioritarianism) and the ‘Distribution View’ (expressed in ex-­post transformed prioritarianism). 25  The weight is exogenous because it does not depend on the introduction of the policy itself, i.e., the implementation of the policy does not alter the temporal distribution of weights. In other words, there is the presumption that the policy is only “marginal”, i.e., it does not have any big impact on the growth process and equilibrium prices. Arguably, climate change policies do not constitute “marginal” projects. See Gollier (2012) and Greaves (2017) for reasons why the discussion keeps focusing on the discount rate for marginal projects.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

80  Maddalena Ferranna Therefore, the larger the discount rate, the lower the likelihood that a project with future benefits will be implemented. Even though the appropriate discount rate is set exogenously, it does not mean that all projects should have the same discount rate. Rather, the appropriate discount rate should be tailored to the characteristics of the policy. In particular, we can distinguish between two types of projects according to the riskiness of their returns: safe projects and risky projects. Safe projects are those whose returns are known, i.e., they can, to some extent, be easily anticipated. Just to be clear, the widely discussed Ramsey Rule provides a formula to compute the discount rate of safe projects. Risky projects are those whose returns are uncertain: for example, climate change mitigation is a risky project because the uncertainty surrounding the phenomenon makes it hard to anticipate what will be the effects of the policy (if we cut our emissions by 20 percent today, how much can we reduce future damages?). Even though climate change policy is a risky project, the debate usually focuses on the discount rate for sure projects, since there is empirical uncertainty on how to measure the riskiness of climate policies and its interaction with the macroeconomic (non-­climate related) risk.26 As a consequence, the following discussion focuses on discount rates for safe projects.27

3.2  Utilitarianism and the Ramsey Rule The discussion on discounting is usually based on the so-­called Ramsey Rule (Ramsey 1928), which is derived from a discounted utilitarian social welfare function, and takes the following form:

SDR= d + h g

The parameter SDR represents the social discount rate, d the utility discount rate (or pure rate of time preference), h the coefficient of inequality aversion, and g  the average growth rate of the economy. Thus, the Ramsey Rule posits that we should discount future monetary values for two reasons: a strict preference for the present over the future (represented by d ), and the prediction of positive economic growth, which would transform the climate policy into a transfer from the poor current generation to rich future generations, thereby raising inequality 26  See, e.g., Dietz et al. (2018) on the correlation between aggregate risk and climate risk (the ­so-­called “beta” of a project). 27  A second salient feature is the timing of the returns of the policy. Indeed, the discount rate does not need to be fixed over time: cash flows occurring at different time periods may be discounted at different rates (the following sections will explain why that could be the case). For example, looking again at Figure  3.1, we may find it appropriate to discount cash flows in 50 years at the 5.5% rate (meaning that $1 in 50 years is equivalent to $0.07 today), and cash flows in 100 years at the 1.4% rate (resulting in $0.25 today for any future $).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Discounting under Risk  81 concerns. The coefficient of inequality aversion h is related to the form of the ­utility function. In particular, the larger h is, the more the policy maker ought to be sensitive to differences in consumption levels across individuals. The term h g summarizes the so-­called wealth effect of discounting: since future generations are expected to be wealthier than the current one, the policy maker should be conservative in investing in climate policy so as not to increase the inequality between the present and the future. Following the publication of the Stern Review (Stern 2006), most of the debate has focused on the appropriateness of introducing a preference for the present in the evaluation of public policies. Two contrasting positions have emerged: those who believe that d should reflect ethical principles (Arrow 1999; Stern 2006), and those who believe that it should express individuals’ preferences as elicited from market transactions (Nordhaus 2007; Weitzman 2007b).28 A few papers have focused also on the choice of h, and the role of inter-­generational inequality in shaping the optimal policy (Anthoff et al. 2009; Dasgupta 2008).29 Thinking in terms of the Ramsey Rule has helped to frame and organize the analysis on discounting, but it has also concentrated most of the discussion on one specific aspect of the climate change problem: its long-­term nature, and the resulting inter-­generational conflicts. There is at least one other feature of the climate change problem that deserves particular attention, and which could be even more important than the inter-­generational equity problem in driving the choice of the social discount rate: the high uncertainty in our predictions about long-­term economic growth, and the impact of climate change on that.30 The Ramsey Rule problematically assumes a steady and continuous economic growth due to persistent technological progress. The problem is that we are not sure if future generations will be richer than us. After all, we have experienced destruction of civilizations in the past, and nothing guarantees that it will not happen also in the future. Indeed, many believe that depletion of natural resources, inefficient allocation of capital, and high population growth will induce low, if not negative, growth. Climate change will contribute to that: large increase

28  Ramsey (1928) himself suggests that, when considering inter-­generational projects, partiality about time is a “practice which is ethically indefensible and arises merely from the weakness of im­agin­ation”. If one treats different generations equally, the only argument for a positive rate of time preference is the risk of extinction: the policy evaluator should be impatient because the future generation may never exist; as a consequence, policies that favor mainly the future should have a lower value since their benefits may never be enjoyed. For example, this is the stand taken by Stern (2006) in assessing the value of climate mitigation policies. See also Kelleher (2017) for a discussion of prescriptive vs. descriptive approaches to discounting. 29  See also Mintz-­Woo (Chapter  4, this volume) on the topic of how to elicit the coefficient of in­equal­ity aversion. 30 Other important aspects related to discounting for climate change are: intra-­ generational in­equal­ity, and how the impacts of climate change are distributed across the world population; population risks, and whether climate change creates a risk of extinction (Fleurbaey and Zuber  2015; Bommier and Zuber 2008).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

82  Maddalena Ferranna in temperature might trigger catastrophic effects and destroy the world as we know it now. When the future prospects of the economy are uncertain, the Ramsey Rule is unable to guide our choice of the socially efficient discount rate simply because we do not know which growth rate to use. Weitzman (2007a, 2007b, 2009) is likely the most known advocate of this idea. By adopting a discounted expected utilitarian welfare framework, Weitzman argues that we should use a low discount rate not because we should treat the future equally to us (low d and/or low h), but because we might face catastrophic impacts that will shrink the future economy to survival levels; as a consequence, we should favor policies that tend to rise future consumption by choosing a low discount rate. Most im­port­ ant­ly, Weitzman’s “dismal theorem” (2009) shows that, if there are fat-­tailed risks,31 the social discount rate should be infinitely negative independently of the value of the parameters of the Ramsey equation (or, alternatively, any investment for the future should be undertaken today, independently of its costs). Weitzman’s result shows the importance of considering risk in the picture, but it also calls into question the appropriateness of the utilitarian approach. Millner (2013) points out that Weitzman’s extreme conclusion is driven mainly by his particular choice of the social welfare function. Independently of Weitzman’s work, that observation raises the more general question of which welfare framework should be used to determine the social discount rate, once we have more than one dimension of variability: risk (variability across states of nature) and inequality (variability across individuals, either living in the same generation or in different generations). As I have argued in the previous sections, the utilitarian framework has one main shortcoming when evaluating situations with multiple dimensions of vari­ abil­ity (such as risk and inequality): it does not disentangle concerns towards risk from concerns towards inequality.32 Looking at the Ramsey equation, the coefficient of inequality aversion h now measures also aversion to risk. That has two main problems, a practical one and a theoretical one. The practical problem is that we do not know how to calibrate the parameter h: should it represent individuals’ aversion to risk, or an ethical choice about the degree of inequality aversion? Experiments that try to elicit the value of h usually end up with different numbers according to the particular definition of h that the builders of the ex­peri­ ment have in mind (Carlsson et al. 2002; Atkinson et al. 2009). The theoretical problem concerns the relationship between risk and inequality, and the inability of utilitarianism to capture concerns for both ex-­ante and ex-­post equity.

31  A fat-­tailed risk denotes a probability distribution that has small, although significant, prob­abil­ ity mass on very low outcome levels. 32  Amother shortcoming of utilitarianism concerns population ethics, and the treatment of populations of different sizes. See, for example, Fleurbaey and Zuber (2015) for a discussion of population principles and discounting for climate change.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Discounting under Risk  83

3.3  Utilitarian Discounting and the Precautionary Effect The Ramsey Rule is based on a world without risk, and it is derived from a utilitarian social objective; as a consequence, it focuses exclusively on inequality and economic growth. However, it can be easily extended to the presence of risk. In particular, it can be shown that a third term should be added to the Ramsey Rule, which reflects the so-­called precautionary effect (Gollier 2012). When the future becomes uncertain, individuals tend to sacrifice more in the present for the future. The same reasoning applies for public projects. The wealth effect relies on the assumption that future generations will be richer than the present one for sure. If the rate of economic growth is uncertain, the current generation should transfer resources from the present to the future so as to protect the well-­being of future populations, i.e., the policy maker should act in a more precautious way. While the wealth effect induces high discount rates, the precautionary effect operates in the opposite direction, by increasing the value of $1 in the future with respect to $1 today. Mathematically, the precautionary effect depends on the size of the risk in the economic growth rate (basically a measure of the degree of uncertainty, often summarized by the variance) and on the coefficient of risk aversion.33 Risk aversion describes how much the policy maker dislikes het­ero­gen­eity in the distribution of outcomes across different states of nature. The more risk averse the policy maker is, the more willing she is to accumulate resources to protect the future generations. In the utilitarian framework, the concavity of the utility function defines both risk aversion and inequality aversion. As a consequence, the choice of the level of risk aversion is constrained by the normative choice of the level of inequality aversion, and vice versa. Therefore, we may end up either underestimating or overestimating the precautionary effect with respect to the wealth one. As long as we limit our analysis to ‘safe projects’ (i.e., projects whose returns are known), most of the debate on the impact of risk on discounting concerns the actual size of the precautionary effect. For a given coefficient of risk aversion, the precautionary effect has a relevant impact only if the risk concerning the future is sufficiently large. Here is where opinions differ, mainly because there is disagreement on how to predict the future. Some papers calibrate an almost zero precautionary effect (e.g., Kocherlakota 1996), while others find an infinitely negative precautionary effect (e.g., Weitzman 2009). Table  3.2 lists the main papers on this subject, and their conclusions in terms of precautionary effect and discount rate.

33  To be precise, the precautionary effect depends on the coefficient of prudence (Gollier 2012), which measures the willingness to accumulate resources to protect future generations. However, the form of the utility function that is commonly used in the climate policy literature is such that the coefficient of risk aversion unequivocally determines also the coefficient of prudence.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

84  Maddalena Ferranna Table 3.2.  List of the main studies that try to determine the size of the precautionary effect. The main difference concerns the assumptions about the characteristics of the growth process. In all cases, the discount rate is computed under the same baseline assumptions:= d 0;= h 2;= g 2%  

Assumptions

Precautionary Discount rate effect

Ramsey Rule Kocherlakota (1996) Bansal and Yaron (2004)

No risk, steady positive growth Steady positive growth with some volatility (standard deviation = 3.6%) Economic cycles: positive growth may be followed by recession periods (or even negative growth); shocks are persistent Extreme events: small annual probability (≅ 2%) of large and permanent drop in consumption Model uncertainty (we do not know the right growth process), which will lead to fat tails, i.e., high weight on small probability events with catastrophic impacts

0 0.4%

4% 3.6%

≅ 1%

3%

Barro (2006) Weitzman (2007a, 2009)

Large

Negative, if drop in consumption > 40% Infinite, Infinitely independently negative of value judgments

3.4  Discounting under Alternative Welfare Frameworks What happens to the discount rate if we replace the utilitarian criterion with  an  ex-­ante/ex-­post approach to equity? For the sake of simplicity, let us ­consider again a very simple inter-­generational model where there is just one representative individual per generation, and future generations face an income growth risk. In the ex-­ante approach, the policy evaluator is interested in the distribution of generational risks, i.e., whether the future generation has better or worse chances than the current one. If she does not like unequal distributions of risks, then she is more concerned about inequality than about risk. Indeed, when talking about “risk”, we have to think of a combination of expected outcome and variability in the possible outcomes. For example, a risk with high volatility, but also high mean, may be preferred over a risk with low (or zero) volatility, but also low mean. As a consequence, a policy maker interested in equalizing chances may be unwilling to transfer from an individual with a low mean/low volatility risk to an individual with a high mean/high volatility risk, if, from an ex-­ante point of view, the latter has better chances than the former. The same reasoning can be applied to discounting. A policy evaluator concerned with ex-­ante equity realizes that, under the common assumption of positive growth, the future generation has overall better chances than the current one. Therefore, she will have less

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Discounting under Risk  85 incentives to invest for the future than a utilitarian one: even though there is the possibility of very bad consequences for the future generation, the future is expected to have better prospects than the present, thereby softening the precautionary motive for investment (Adler and Treich 2017). Notice that this effect is distinct from the wealth effect previously described. The policy maker has low willingness to invest for the future not because the future is richer for sure, but because the future faces better prospects than the present. If the policy maker does not like inequality in chances, she will have lower incentives than the utilitarian one to transferring resources for the future. Note that this effect due to concerns for the distribution in risks is comparable to a pure rate of time preference, or utility discounting. As a consequence, the policy maker should have a preference for the present not because future generations are not yet born (encapsulated in the pure rate of time preference), but because the future has better prospects than the present. With an ex-­post approach to equity, instead, we do not care if the current and the future generations face the same prospects, but if one generation stands out for being particularly lucky or unlucky. In other words, the correlation of risks across generations plays a key role in deriving the discount rate. For example, if the risk affecting a generation has only short-­term effects and is independent of the risks falling on subsequent generations, the incentives to invest for the future will likely be similar to the ones under utilitarianism. If anything, the incentives to invest will be larger if that future generation has a relatively large probability of facing a catastrophic outcome. Note, however, that this result is not driven by the presence of the catastrophe per se, but by the fact that the future generation will be unluckily treated with respect to the others. In contrast, if the risk affecting a generation has persistent effects, the incentives to invest for the future will be lower than the utilitarian ones. Indeed, the occurrence of correlated positive outcomes has the perverse effect of realizing ex-­post equality (a misery loves company effect); as a consequence, if the policy maker is adverse to ex-­post inequality, the discount rate will tend to increase. Perhaps the most interesting consequence is the weakening of Weitzman’s dismal result. Indeed, given that catastrophic impacts bring equality, the willingness to invest for the future under uncertainty and fat tails will no longer be infinitely positive: if we are concerned about ex-­post equity, we will not be ready to invest in any policy as the Dismal Theorem suggests (Fleurbaey and Zuber 2015). Nevertheless, we would still have a precautionary motive to invest, which will increase with the size of the risk: as much as we like ex-­post equality created by catastrophic impacts, we would still prefer to have equality and non-­catastrophic impacts, which explains the incentives to invest to protect against the negative outcomes. Therefore, by using an ex-­post approach to equity, we would be able to disentangle risk and equity concerns, and we would still get a Weitzman type of result without its unrealistic side (i.e., the infinitely negative discount rate).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

86  Maddalena Ferranna

4. Conclusions The chapter has highlighted the main limits of utilitarianism in evaluating public risks with unequal consequences across the population, and has described two possible alternative social welfare functions based on the disentanglement of risk and equity preferences: ex-­ ante prioritarianism (ex-­ ante equity) and equally ­distributed equivalent approach (ex-­post equity). For each alternative approach, I have sketched the main implications in terms of discounting. I do not claim that these social welfare functions are the best solution to the risk/equity issue, nor that they are strictly better than utilitarianism in all respects. However, given the limits of utilitarianism, I think that it is worth discussing its potential alternatives and their implications in the evaluation of public policies. A minor goal would be to understand whether the results are very much different from the utilitarian framework, or whether the utilitarian framework would still be an acceptable approximation. For example, I have discussed how ex-­ante prioritarianism induces a preference for the present that is similar to the presence of a utility discount rate, or pure rate of time preference. As a consequence, ex-­ante prioritarianism is approximately equal to discounted utilitarianism, with the difference that, in the first case, utility discounting is endogenous. Similarly, in the ex-­post (transformed) prioritarian case, concerns for ex-­post equity tend to raise the discount rate if risks are positively correlated across generations (and due to the persistence in growth, they probably are). So, again, ex-­post prioritarianism will lead to similar results as discounted utilitarianism. In contrast, the choice between utilitarianism and the ex-­post approach to equity would matter in the presence of fat tails. Indeed, a utilitarian policy maker would have an infinite willingness to invest for the future, while an evaluator interested in ex-­post equity would not call for the implementation of any type of investment (the willingness to pay would be finite), although she would adopt a very low (perhaps negative) discount rate. The main advantage of prioritarianism (or at least some versions of it) over utilitarianism is the possibility to disentangle risk and equity preferences, thereby improving the calibration of the associated parameters. Because of the high sensitivity of climate policy to the specification of preferences, and, in particular, to the degrees of inequality and risk aversion, selecting the ‘right numbers’ is not just a theoretical exercise. More research on the definition of the most suitable welfare approach and on how to actually measure preferences is more than welcome. The chapter has to be interpreted as a small contribution in that direction. Note, however, that the discussion rested on two important assumptions: risk refers only to consumption and there is no inequality inside a given generation.34 Additionally, 34 See, for instance, Dennig et al. (2015) on the dramatic increase in the carbon tax if intra-­ generational inequality was taken into account, in particular the disproportionate burden of climate damages falling on the future poor.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Discounting under Risk  87 the chapter has focused on situations of risk but not of uncertainty (in the sense of Knightian uncertainty, i.e., cases of unknown unknowable).35 Extensions to other types of risk (e.g., population or mortality risks), other forms of inequality (intra- vs. inter-­generational inequality), or ambiguity over the likelihoods of different events could change the conclusions or the choice of the most appropriate social welfare framework.

References Adler, M. (2012). Well-Being and Fair Distribution. Oxford: Oxford University Press. Adler, M. and C. Sanchirico (2006). “Inequality and Uncertainty: Theory and Legal Applications.” University of Pennsylvania Law Review 155: 279–377. Adler, M. and N. Treich (2017). “Utilitarianism, Prioritarianism and Intergenerational Equity: A Simple Cake Eating Model.” Mathematical Social Sciences 87: 94–102. Anthoff, D., R. S. J. Tol, and G. W. Yohe (2009). “Risk Aversion, Time Preference and the Social Cost of Carbon.” Environmental Research Letters 4: 1–7. Arrow, K.  J. (1999). “Discounting, Morality, and Gaming,” in Discounting and Intergenerational Equity, ed. P.  R.  Portney and J.  P.  Weyant. Washington, DC: Resources for the Future, 13–21. Atkinson, G., S.  Dietz, J.  Helgeson, C.  Hepburn, and H.  Saelen (2009). “Siblings, Not Triplets: Social Preferences for Risk, Inequality and Time in Discounting Climate Change.” Economics: The Open-Access, Open-Assessment E-Journal 3: 2009–2026. Bansal, R. and A. Yaron (2004). “Risks for the Long-Run: A Potential Resolution of Asset Pricing Puzzles.” Journal of Finance 59: 1481–1509. Barro, R. (2006). “Rare Disasters and Asset Markets in the Twentieth Century.” Quarterly Journal of Economics 121: 823–866. Ben-Porath, E., I.  Gilboa, and D.  Schmeidler (1997). “On the Measurement of Inequality under Uncertainty.” Journal of Economic Theory 75: 443–467. Bommier, A. and S.  Zuber (2008). “Can Preferences for Catastrophe Avoidance Reconcile Social Discounting with Intergenerational Equity?” Social Choice and Welfare 31: 415–434. Bovens, L. (2015). “Evaluating Risky Prospects: The Distribution View.” Analysis 75: 243–253. Broome, J. (1984). “Uncertainty and Fairness.” The Economic Journal 94: 624–632. Carlsson, F., D.  Daruvala, and O.  Johansson-Stenman (2002). “Measuring Future Grandparents’ Preferences for Equality and Relative Standing.” Economic Journal 112: 362–383.

35  Knight (1921).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

88  Maddalena Ferranna Dasgupta, P. S. (2008). “Discounting Climate Change.” Journal of Risk and Uncertainty 37: 141–169. Dennig, F., M.  B.  Budolfson, M.  Fleurbaey, A.  Siebert, and R.  H.  Socolow (2015). “Inequality, Climate Impacts on the Future Poor, and Carbon Prices”. Proceedings of the National Academy of Sciences 112(52): 15827–158232. Diamond, P. (1967). “Cardinal Welfare, Individualistic Ethics and Interpersonal Comparisons of Utility: Comment.” Journal of Political Economy 75: 765–766. Dietz, S., C. Gollier, and L. Kessler (2018). “The Climate Beta.” Journal of Environmental Economics and Management 87: 258–274. EPA (2018). Regulatory Impact Analysis for the Proposed Emission Guidelines for Greenhouse Gas Emissions from Existing Electric Utility Generating Units; Revisions to Emission Guideline Implementing Regulations; Revisions to New Source Review Program. U.S.  Environmental Protection Agency. https://www.epa.gov/sites/ production/files/2018-08/documents/utilities_ria_proposed_ace_2018-08.pdf Epstein, L. and U.  Segal (1992). “Quadratic Social Welfare Function.” Journal of Political Economy 100: 691–712. Ferranna, M. (2019). “Fairness in Cost Benefit Analysis: Equity-Enhanced Mean Variance Rules.” Unpublished manuscript. Ferranna, M. and M. Fleurbaey (forthcoming). “Climate Change.” In Prioritarianism in Practice, ed. M. Adler and O. Norheim. Cambridge: Cambridge University Press. Fishburn, P. and R.  Sarin (1991). “Dispersive Equity and Social Risk.” Management Science 40: 1174–1188. Fleurbaey, M. (2010). “Assessing Risky Social Situations.” Journal of Political Economy 118: 649–680. Fleurbaey, M., M. Ferranna, M. B. Budolfson, F. Dennig, K. Mintz-Woo, R. Socolow, D.  Spears, and S.  Zuber (2019). “The Social Cost of Carbon: Valuing Inequality, Risk, and Population for Climate Policy.” The Monist 102(1): 84–109. Fleurbaey, M. and S.  Zuber (2015). “Discounting, Risk and Inequality: A General Approach.” Journal of Public Economics 128: 34–49. Gajdos, T. and E. Maurin (2004). “Unequal Uncertainties and Uncertain Inequalities: An Axiomatic Approach.” Journal of Economic Theory 116: 93–118. Gollier, C. (2012). Pricing the Planet’s Future. Princeton: Princeton University Press. Greaves, H. (2017). “Discounting and Public Policy: A Survey.” Economics and Philosophy 33: 391–439. Hammond, P. (1983). “Ex-Post Optimality as a Dynamically Consistent Objective for Collective Choice under Uncertainty,” in Social Choice and Welfare, ed. P. Pattanaik and M. Salles. Amsterdam: North-Holland, 175–205. Harsanyi, J. (1955). “Cardinal Welfare, Individualistic Ethics and Interpersonal Comparison of Utilities.” Journal of Political Economy 63: 309–321. IWG (2016). Technical Update of the Social Cost of Carbon for Regulatory Impact Analysis-Under Executive Order 12866. Report prepared by the Interagency

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Discounting under Risk  89 Working Group on Social Cost of Greenhouse Gases, United States Government. https://www.epa.gov/sites/production/files/2016-12/documents/sc_co 2_tsd_ august_2016.pdf Keeney, R. (1980). “Equity and Public Risk.” Operations Research 28: 527–534. Kelleher, J. P. (2017). “Descriptive versus Prescriptive Discounting in Climate Change Policy Analysis.” Georgetown Journal of Law and Public Policy 15: 957–977. Knight, F. H. (1921). Risk, Uncertainty, and Profit. Boston: Hart, Schaffner and Marx; Houghton Mifflin Company. Kocherlakota, N.  R. (1996). “The Equity Premium: It’s Still a Puzzle.” Journal of Economic Literature 34: 42–71. Millner, A. (2013). “On Welfare Frameworks and Catastrophic Climate Risks.” Journal of Environmental Economics and Management 65: 310–325. Myerson, R. (1981). “Utilitarianism, Egalitarianism, and the Timing Effect in Social Choice Problems.” Econometrica 49: 883–897. Nordhaus, W.  D. (2007). “The Stern Review of the Economics of Climate Change.” Journal of Economic Literature 45: 686–702. Posner, R. A. (2004). Catastrophe: Risk and Response. New York: Oxford University Press. Rabinowicz, W. (2002). “Prioritarianism for Prospects”. Utilitas 14(1): 2–21. Ramsey, F. P. (1928). “A Mathematical Theory of Savings.” Economic Journal 38: 543–559. Savage, L. (1954). The Foundation of Statistics. New York: John Wiley & Sons. Sen, A. (1973). On Economic Inequality. Oxford: Oxford University Press. Stern, N. (2006). The Economics of Climate Change: The Stern Review. Cambridge: Cambridge University Press. Weitzman, M.  L. (2007a). “Subjective Expectations and Asset-Return Puzzles.” American Economic Review 97: 1102–1130. Weitzman, M. L. (2007b). “The Stern Review of the Economics of Climate Change.” Journal of Economic Literature 45: 703–724. Weitzman, M.  L. (2009). “On Modeling and Interpreting the Economics of Catastrophic Climate Change.” Review of Economics and Statistics 91: 1–19.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

4

A Philosopher’s Guide to Discounting Kian Mintz-­Woo

One of the challenges of climate change lies in the temporal distances between policies and effects. This is not only because of the lag between emissions and effects of greenhouse gases on the climate system but also because some of these gases are very long-­lived and continue to have effects over long timespans. For these reasons, one prominent focus of climate ethics has been on various problems in intergenerational ethics.1 My discussion will focus on the distinctions relevant to what is called the discounting problem, since the issue is how (future) costs and benefits are “discounted” to make them comparable in present terms.2 More precisely, the discounting problem is about which weightings to place on those costs and benefits as a function of time.3 An important moment in the history of discounting was the release of the so-­called “Stern Review” of climate change, where Nicholas Stern (2006) argued that moral or normative justifications needed to be brought into long-­term cost– bene­fit climate analyses. In particular, he held that a form of utilitarian impartiality implied that the pure rate of time preference should be (near-)zero.4 This led to a spate of objections (Dasgupta 2008; Nordhaus 2007a, 2007b; Weitzman 2007), many of which criticized Stern’s introduction of moral considerations. However, there are some important subtleties that are often mistaken about this debate. The main intention of this chapter is to introduce and explain the most important distinctions with respect to the discounting problem for a 1  Sometimes, we frame these problems in terms of models with overlapping generations, but here I am less concerned with the generational aspects and more concerned with temporal ones. 2  Helpful recent surveys of this question can be found in Davidson (2015), Greaves (2017), Heal (2007), and Kelleher (2017a). 3  Note that discounting in this context is a very technical economic term. It is, for instance, completely unrelated to the everyday usage of the term “discounting”, which is about lowering prices (such as in retail). Also, note that although the weightings are a function of time in the technical sense— they are a function of a variable t for time—the justifications for different weightings may be about things which are only correlated with time (see Parfit 1984: Appendix F). 4  In particular, he said that the pure rate of time preference should be positive not because future outcomes are weighted less heavily, but for probabilistic reasons having to do with the uncertainty of exogenous human extinction. However, properly speaking, human extinction is just a sub-­class of a broader class of exogenous events where the action does not generate the intended outcome, so Stern’s reasoning should actually generate a higher pure rate of time preference than whatever is implied merely by the probability of human extinction (Mintz-­Woo 2019).

Kian Mintz-­Woo, A Philosopher’s Guide to Discounting In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Kian Mintz-­Woo. DOI: 10.1093/oso/9780198796282.003.0005

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

A Philosopher ’ s Guide to Discounting  91 philosophical audience. A secondary intention is to defend the theory underlying this problem and to point out limitations of moral theory when addressing this problem. In Section 1, I defend the claim that there are good reasons to adopt Ramsey­style discounting in the context of climate change. I emphasize that Ramsey-­style discounting need not be intrinsically wedded to utilitarianism. In Section  2, I introduce two important distinctions involved in discounting. The first is between pure discounting5 and what I call impure discounting 6 (i.e. between discounting in utility/welfare terms and discounting in monetary terms). The second is between consumption-­based and investment-­based discounting (i.e. depending on whether we are measuring the costs and benefits in terms of displaced social value or in terms of financial flows). In Section  3, I distinguish between descriptivism and prescriptivism. The descriptivist about discounting assigns values to the relevant parameters on the basis of observed data grounded in stated or revealed individual preferences; the prescriptivist assigns values to the parameters on the basis of factors beyond stated or revealed individual preferences, including social preferences and considerations from moral theory. In Section 4, I argue that, even if we adopt prescriptivism, and accept that this means there is a need for moral experts in parameter assignments, there is a significant problem. The type of moral ex­pert­ise required for the discounting problem, I argue, will not involve know­ ledge of moral theory—thus making moral philosophy unhelpful in terms of making particular parameter assignments, despite these being substantive moral judgments. In Section  5, I conclude with a few thoughts about which kind of moral expertise might be relevant to the discounting problem.

1.  Introducing and Defending the Ramsey Rule Integrated assessment models (IAMs) perform an important role in climate economics; they allow us to do things like assess the value of particular policies by combining (very) basic climate models with various macroeconomic measurements. As mentioned before, various aspects of the discounting problem are of particular importance in climate policy evaluations as the costs and benefits of 5  Pure discounting is also known as “pure rate of time preference” or “utility discounting”. 6  This is a term which I am introducing. Usually we distinguish pure discounting with discounting in consumption terms. However, since I also want to draw attention to the distinction between investment-­based and consumption-­based discounting, I suspect that using standard locutions like “discounting in consumption terms” and “consumption-­based discounting” will be unnecessarily confusing, since these are entirely different things. Since “impure” discounting is the obvious opposite of “pure” discounting, I expect it to be an intuitive contrast. For prioritarians, there will be another level for something like “super-­pure discounting” which involves the translation between utility/welfare and social value, an even purer concept.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

92  Kian Mintz-Woo climate action are temporally dispersed. The results of IAMs are thus quite sensitive to the parameters which govern time preference. The key to time preference on standard models is the Ramsey Rule.7 The Ramsey Rule involves discounting in terms of the consumption of individuals. It governs the optimal consumption path with a particular social welfare function.8 It can be stated in a very compact form which relates the social discount rate r on the left-­hand side to what is sometimes called the social rate of time preference on the right-­hand side. The social rate of time preference is the sum of the pure rate of time preference (the Greek letter δ , “delta”) and a term which is the product of the growth rate of consumption g and the elasticity of marginal utility of consumption (the Greek letter η , “eta”):

r = δ + (η × g )

Intuitively, discounting of consumption depends on both pure discounting, that is, factors which are not particular to consumption but time itself (or things correlated with time) (δ ) and the diminishing marginal value of consumption as consumption grows (η × g ).9 The elasticity η measures the relative change in ­marginal utility, or how quickly marginal utility decreases as consumption rises. Higher values indicate that marginal utility decreases more quickly and η = 1 meaning that marginal utility decreases proportionately to consumption increases. Choosing η = 0 indicates no concern about distribution at all (i.e. the only concern is increasing total consumption with no regard to who consumes), while choosing η = ∞ indicates that only benefits to the very worst off improve society. In practice, the higher η is, the more important it is to distribute consumption to those with lower initial consumption. It may also be helpful to think of pure discounting as discounting applied in the pure units of utility or welfare, i.e. as a utility discount rate. To give a sense of potential values, δ is usually taken to be between 0 and 0.015 (i.e. 0 to 1.5% per year);

7  Derivations can be found in Dasgupta (2008) and Goulder and Williams (2012). Note that the Ramsey Rule is a conclusion following from various assumptions. 8  Beyond the consumption path being optimal, the Ramsey Rule is optimal also in a second sense that it does not describe actual—rather, idealized—economies, due to inter alia market imperfections and distortions (Groom et al. 2005). Many of the issues, such as the divergence between descriptivism and prescriptivism, depend on the assumption we do not live in an optimal economy. Since we do not, I take these issues to be live. 9  As Kelleher (2017a) points out, it may be “highly misleading” to take δ > 0 to be due simply to the fact that some cost or benefit is in the future. Beyond being justified because there is uncertainty about the future, it may also be justified by considering the demandingness of δ = 0 for current people. Some philosophers take discounting which is particular to time itself to be the only real types of discounting. However, this is incorrect; discounting is simply differential weighting. In this context, most discounting is indexed by time, but this need not be the case. Caney (2009) and Mintz-­Woo (2017), for instance, point out that discounting for uncertainty is also a form of discounting, a form which inherently has nothing to do with time.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

A Philosopher ’ s Guide to Discounting  93

η is usually taken to be anywhere between 0.5 and 4; and g is usually taken to be

between 0.01 and 0.03 (i.e. 1 to 3% per year).10 The normative parameters are taken to be δ and η , with r being of—so to speak—derivative normative importance.11 The former factor, δ , governing pure discounting, is much more intuitive and has held the bulk of philosophical attention (Broome  1994; Caney  2014; Parfit  1984; Sidgwick  1907/1981). However, given a growth rate g of more than 0.01 (or 1% growth per year), changes in η affect r more than changes in δ , since η is multiplied by g. So η is in general a greater driver of the social discount rate than δ . It deserves philosophical attention to a greater extent than it has received. While uncertainty about future growth (i.e. appropriate values for g) is also of importance, I take it to be of less distinctly ethical12 importance; at least in prin­ ciple, uncertainty about consumption growth can be eventually resolved given sufficient empirical (macro)economic data, which distinguishes this parameter from η and δ in social contexts, a point I will return to below.13 The parameter η governs utility functions given that they increase in a particular predictable manner with respect to increases in consumption.14 By “pre­dict­ able”, I mean that η is constant throughout the utility function as we change consumption levels; a 1% increase in consumption is worth the same relative utility gain regardless of whether you started with a lot of consumption or a little. The main alternative in discounting to the Ramsey Rule is setting the social discount rate to equal the (highest) rate of return on investment, i, focusing on the left-­hand equation:

i = r = δ + (η × g )

Sometimes, these alternatives are contrasted by calling the Ramsey approach “consumption-­ based discounting” as opposed to this “investment-­ based 10  The growth rate is. the change in consumption given as a percentage of current consumption, c(t ) that is, g= (t ) g= (t ) , where the dot indicates the first derivative, but if we take some long-­term c(t ) ­stable growth rate, then we can drop the argument of g (i.e. t). 11  In other words, we can say the social discount rate r is not itself of direct ethical importance, because any value of r underdetermines what the normative parameters are. 12  While it is certainly of ethical relevance how we plan or execute growth, g is not a moral parameter since the Ramsey Rule just depends on how much consumption actually grows (or is expected to occur). While one might think that decreasing consumption growth might be a good way to lower the discount rate, this would be a very costly way to affect the discount rate from the point of view of society, since decreasing consumption for members of society could involve failing to improve people’s lives. 13  For more on risk and uncertainty, see Chapter 3 by Maddalena Ferranna. 14  Technically, that utility functions are written as a function of consumption in a special isoelastic c1−η or constant relative risk aversion (CRRA) form that looks like U (c) = (with U (c ) = log c , for η = 1), 1 −η which has the pragmatic implication that derivatives of the utility function are easy to work with.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

94  Kian Mintz-Woo discounting” (or “opportunity cost discounting”).15 In an ideal economy, both of these approaches will converge because each of these three terms converge. However, in a non-­ideal economy (e.g. with distortions due to incomplete markets and taxes), we can either take the left-­hand equation (i = r) or the right-­hand equation (r = η × g ). The Ramsey Rule, or consumption-­based, discounting adopts the right-­hand side whereas investment-­based discounting depends on the left­hand side equation. I expand on this distinction in Section 3. However, before explaining how we can assign values to the normative parameters in the Ramsey Rule in the following section, I will close this section by discussing three reasons to work with the right-­hand equation in the context of (climate) public policy evaluation. In other words, here are three reasons for adopting consumption-­based discounting. First, it is robust in that it can be ­generated under different ethical approaches, such as intuitionism, not just discounted utilitarianism (Dasgupta  2005). It is also flexible; for instance, the Ramsey Rule can be “extended” or modified to include terms for certain types of risk or uncertainty (e.g. Fleurbaey et al. 2019; Groom et al. 2005) and can model situations where consumption growth is negative. Finally, after having been through several waves of theoretical work, the framework is well-­understood and is required in many sources of public policy. For instance, in the United States, regulations have to be subject to cost–benefit analyses, and these require some type of discounting, at least some justifications of which are grounded in a version of the Ramsey Rule. So three reasons that we should work within this framework are that it is robust, flexible, and well-­understood. I present these three reasons just in order to make the case; I lack the space here to fully address competing reasons for investment-­based discounting although I do not think that similarly convincing reasons can be presented in turn.16 Let me expand on these three points in turn. The Ramsey Rule is robust in that it arises from different moral theories. In particular, there are both, in the Rawlsian parlance, teleological and intuitionist ways of generating the Ramsey Rule (Rawls  1971). For Rawls, teleological moral theories took the good to be conceptually prior to the right, with the right defined in terms of increasing the good in some manner. Intuitionists rely upon putatively self-­evident or prima facie principles. As Ramsey was of the traditional British utilitarian tradition, it is unsurprising that the Ramsey Rule accords with a teleological approach. However, as Dasgupta (2005) points out, the Ramsey Rule also applies under intuitionist 15  The Ramsey approach is called “social welfare equivalent” discounting, as opposed to “finance equivalent” discounting, by Goulder and Williams (2012). 16  Interesting questions arise when moving beyond discounted utilitarianism. For instance, if you disentangle risk aversion and inequality aversion and introduce variable populations, then the results can diverge quite dramatically. I lack the space to explore these non-standard options but refer the interested reader to Fleurbaey and Zuber (2015) and Fleurbaey et al. (2019) for more. Another alternative is restricting the scope of the evaluation over space, by limiting impacts nationally instead of globally. Interesting moral issues arise here as well (Mintz-Woo 2018b).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

A Philosopher ’ s Guide to Discounting  95 approaches to moral theory. As well as deriving this framework from an approach which is concerned with maximizing utility, we can also develop it from an approach which is concerned with putatively self-­evident axioms used for comparing consumption streams. In particular, Koopmans (1960) showed that certain axioms imply a social welfare function that looks like Ramsey’s (pure) discounted utilitarianism—and implies discounting which complies with something like a Ramsey Rule. Koopmans was interested in ranking infinite streams. The axioms used on the rankings of those streams are part of an intuitionist approach to ­ethics if we consider those axioms to be subject to evaluation by ethical intuitions, whereupon we work out the implications of adopting those particular axioms.17 If we interpret the streams as over consumption, then the functional form implied by Koopmans’ axioms has some monotonically increasing function of consumption. (If we interpret the streams as over utility, then Koopmans’ axioms yield some type of prioritarianism!) It is worth noting that this differs from the project of many intuitionists, who were concerned both with self-­evident or prima facie duties or rules—not axioms—and with the ways these could conflict—not the implications that could be drawn from them. However, if we take axioms to be subject to the same kind of tests of self-­ evidentness as the intuitionists employed, this method also generates similar functional forms to a teleological approach, making the results more robust to moral methodology than is commonly recognized. It is also appropriate for theories which include non-­welfarist con­sid­er­ations, depending on what we take to be what Kelleher (2017b) calls the “evaluative scope”. For instance, we can adopt this framework and modify recommendations with additional side-­constraints from the point of view of justice or deontology. For these reasons, Ramsey-­style discounting is robust with respect to a variety of potential moral commitments. The second reason to adopt Ramsey-­type discounting is its flexibility. The basic Ramsey Rule applies under deterministic conditions and without distinctions for different individuals. These are simplifications that can reasonably be criticized as incorrect in reality. However, they are not intrinsic to the general framework and we can extend the framework to relax these conditions (obviously, with costs incurred for simplicity and computability) (Fleurbaey et al. 2019). To give a sense of potential extensions, I describe a couple here. Extensions to Ramsey discounting have been made to accommodate risk over consumption growth (e.g. Groom et al. 2005: formula (28)) as well as having differential intratemporal and intertemporal discounting (e.g. Dasgupta 2019). The former introduces a term which lowers the social discount rate r by appealing to risk over discount rates; the latter distinguishes between intragenerational discounting (which is subject to gen­er­ ation­al preference and is not ruled out) and intergenerational discounting (which 17  The scientific interest of Koopmans’ results is that the axioms required to generate these social welfare functions are surprisingly weak.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

96  Kian Mintz-Woo is thought to be morally impermissible and set to 0). While the resultant ­formulations are, unsurprisingly, ungainly, they are still within the same general framework, showing how flexible it can be. Besides extending the Ramsey Rule, we can generate very different requirements simply by varying the values of the parameters in the Ramsey Rule itself. For instance, undiscounted utilitarianism can be modeled as a special case of discounted utilitarianism simply by setting δ = 0. We can also model different intertemporal approaches by considering extreme values of η : as noted above, we can demand that consumption streams must only contain maximized consumption (η = 0) or must be arranged such that we maximize minimum consumption (η = ∞).18 It is also worth pointing out that the Ramsey Rule applies even in cases where expected consumption growth is negative (i.e. that there is expected ­consumption decline and g < 0 ). In such cases, the marginal value of current consumption is less than that of future time periods (where consumption is lower), which would be reflected in η × g < 0, decreasing the social discount rate (r ). Intuitively, if future generations are poorer, then utilitarianism tells us that we should value their consumption more highly than our own given that they will be poorer. This can lead to negative social discount rates, indicating a bias towards the future and away from the present. Because of the range of views that can be formulated in this framework, the framework’s flexibility allows for a straightforward conversation while avoiding the challenges inherent in translating positions originating in different frameworks. The final point is that the Ramsey Rule is well-­understood. This applies both to the assumptions that are required to generate it, as well as economic methods for gathering data that might be relevant to assigning values to the normative parameters within it. Unlike the other two reasons, this is meant to be a pragmatic reason; it is easier to act within a well-­understood framework which also forms the basis of climate policy analysis. Because we know the implications of changing different parameters, our moral judgments can be more informed. Simple in­tu­ itions about distributions, whether over risks or equity, lack this coordinated aspect. By considering risk or equity in isolation from the intertemporal trade­offs, such connections can be obscured. For instance, we are used to thinking of distribution as zero-­sum; with a particular quantity of goods, we can distribute them more or less equally (for instance). When considering goods over time, however, we have to take into account potential growth of invested capital. This is considerably less intuitive and, while not the only way, Ramsey-style discounting can help structure these judgments.

18  Solow (1974) was among the first to point out that Rawlsian distribution can be given as a limit case of utilitarianism where η = ∞.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

A Philosopher ’ s Guide to Discounting  97

2.  Why the Terms of Measurement Matter When considering the Stern Review (2006), people often point to his moral claim: that future people (more specifically, pure benefits accruing to the future) should be counted equally to our own. This implies that pure discounting (δ ) should be near-­zero.19 The pure rate of time preference is a kind of discount rate; a positive value for it means that future benefits (respectively, costs) are to be counted less than the present, and even less the further in the future they are. It is a pure discount rate. However, it is not the only kind of discount rate. In contrast, William Nordhaus, an influential economic modeler who rejected this introduction of moral judgments, suggested that the social discount rate (r) should reflect (risk-­free, real) market interest rates. He straightforwardly adopted the investment-­based discounting approach since he thought of climate policies as investments and, as such, they needed to be competitive with other available investments. Since there are often well-­paying investment opportunities in growing economies, his values for the social discount rate (r) were relatively high (and certainly much higher than near-­zero rates), in the range of 4% to 6%. From this discussion, non-­experts may draw the conclusion that moral judgments (of Stern’s type) lead to social discount rates that are near-­zero and market appeals (of Nordhaus’ type) lead to rates that are relatively high. But this is a conflation, since the rates that Stern was discussing using these moral judgments are not the same kind of rates that Nordhaus is considering; Stern was talking directly about δ whereas Nordhaus about r. Although δ and r are both discount rates, the former applies to (pure) costs and benefits (i.e. as measured in utility/welfare) whereas the latter applies to consumption equivalents. For this reason, the relatively high values of the social discount rate that Nordhaus advocates are not directly comparable to the near-­zero values of the pure rate of time preference that Stern advocates. To make comparable claims, Stern advocated r = 1.4 for δ = 0.001, η = 1 and g = 0.013 .20 This is, of course, lower than Nordhaus’ r values, but Stern’s near-­zero δ value is only a component of Stern’s r, as is clear from the Ramsey Rule. The difference in application is what is meant to be suggested by the term “pure” as opposed to what I am calling “impure”. Unfortunately, however, the economic terminology is rather opaque, so what the “purity” in “pure rate of time preference” refers to may not be obvious. Furthermore, this opacity is compounded by the fact that the term “social discount rate” is not helpful as both rates may be social. The names are more reflective of historical contingencies than transparent labeling. 19  The value actually suggested was δ = 0.001, for an assumed 0.1% annual risk of extinction. For our purposes, we can consider this zero or near-­zero. 20  To contrast, Nordhaus (1994) takes δ = 0.03 and η = 1 and Cline (1992) takes δ = 0 and η = 1.5.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

98  Kian Mintz-Woo Here is another way of making this point: if we believe that the marginal value of consumption gains decline the more initial consumption one has (that consumption generates “decreasing marginal utility”) and that there will be consumption growth, then one should be in favor of a positive (impure) social discount rate even if one thinks there should be no pure discounting. Why is this? Formally, for r = δ + η × g , if η > 0 and g > 0, then r > 0, even given δ = 0. Intuitively, if we consider how to distribute extra consumption between (by assumption) on average richer future people and poorer present people, we should prefer distributing it to the present. Due to the relative poverty of the present, the marginal utility generated by additional consumption would be greater for present people than future people. In other words, this argument shows we should discount consumption, or have positive impure discount rates even if we do not accept any pure discounting at all.21 So what is the important difference? Moral philosophers will be more familiar and used to working in what I’m calling pure units, such as utility/welfare, whereas economists tend to work with impure units. There are important reasons for these differences; pure units are of immediate moral importance, whereas impure units are more readily observable and therefore more amenable to em­pir­ ic­al treatment. Furthermore, moral philosophers usually assume that we can talk about pure units in cardinal terms; that is, we can make well-­defined comparisons between ratios of interpersonal utility levels. This is a significant assumption; if we do not make it, then it makes more sense to talk in impure units, since these are observable and comparable. This is one of the reasons that modern economics tends towards discussion only in terms of increasing returns in financial terms. Since the advent of positive economics, economists have tended to stay away from such assumptions (Dennig  2018). The relevance for this difference in the context of climate economics is that philosophers focus on pure discounting (which naturally applies to pure units) while neglecting the importance of impure rates which apply to, for instance, consumption discounting. Much of this focus by economists is due to the fact that we cannot empirically determine the car­din­ al­ity of pure units (only their individual ordinality). In contrast, philosophers are not required to restrict themselves to the empirically verifiable, so are free to talk in pure units. This practice is justifiable in moral philosophy, but it is important to remember that it requires a contentious assumption. However, keeping the philosophical discussion attuned only to pure units has an important cost. As discussed in the previous section, δ is only one component in r, the social discount rate. Moreover, the pure rate of time preference is a

21  Recall that the opposite result, negative discounting (weighting future consumption more heavily than present consumption), occurs if we expect consumption levels to decline, i.e. if g < 0, given η > 0, g ×η < 0.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

A Philosopher ’ s Guide to Discounting  99 relatively small component, since g > 0.01 implies that changes in the elasticity of marginal utility of consumption, η , are more important. In other words, r is driven more by changes in η than in changes to δ . For these reasons, it is im­port­ ant to pay attention to the units that a given discount rate is being applied to, since the meaning changes dramatically and pure discount rates are (less im­port­ ant) drivers of impure social discount rates. One might worry that η is merely an empirical unit that does not lend itself to theory but I argue below both that there are reasons to take it to be moral and that there are interesting things that can be said about η if one takes it to be moral. The next point to make is about the role of “consumption”. Since there are a variety of goods that one can consume in order to satisfy one’s preferences, a complete model would incorporate heterogeneity in goods. However, for tractability in modeling contexts, we often assume that any goods have equivalents in terms of a single homogenous good called “consumption” (conventionally represented by the variable c). There are important difficulties here, since climate change may differentially affect various categories of goods—in particular, natural goods might not grow as quickly under climate change as artificial goods, so we might need to take the relative prices of natural goods into account (Sterner and Persson 2008). This debate is sometimes conducted in terms of “strong” versus “weak” sustainability, depending on whether one thinks that there are some goods that cannot be (or should not be) converted into consumption equivalents, or whether one thinks that such goods are worth large amounts of consumption equivalents. A more philosophical objection is that goods might not be tradable against each other in terms of a single homogenous good, since they might have values which are agent-­relative or intrinsic. So, for instance, a given object’s particular history might make it matter in agent-­relative terms (e.g. this cutlery was passed down from my parents) or it could be that some goods are valuable independent of anyone’s preferences (e.g. an ecosystem’s dynamics being maintained might matter even if no one were aware of them). One version of such claims is that one cannot replace the good without losing that historical or personal connection, which is true, but it does not follow that there are no replacements which would be of equivalent value to the person. I would argue that one can develop connections or history with replacement goods as well, although of course this may cost time and involve regret, things that should be factored in. These are important and complicated discussions, but I wish to set them aside here, since they are not directly linked to the discounting problem. Goulder and Williams (2012: fn 9), for instance, point out that, in theory, introduction of different goods which are imperfect substitutes can be incorporated into what is gained through mitigation using a single discount rate. Intuitively, we can get equivalent results by assuming that the (relative) value of mitigation benefits rise over time, showing that the problem of multiple goods can be isolated from the discounting problem.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

100  Kian Mintz-Woo In the previous section, in the context of public policy evaluation, I advocated consumption-­based discounting. This is because the purpose of public decision­making is to increase social welfare and not to maximize social value in monetary terms. However, this can also be bolstered by the descriptive claim that, since national income primarily comes from tax revenue, and that revenue primarily displaces consumption, consumption-­based discounting is appropriate (Arrow 1999b). However, if one wishes to adopt a(n impure) social discount rate which is a weighted average of consumption-­based and investment-­based discounting, this discussion can be framed as discussing the consumption-­based component of the weighted average. After all, the investment-­based discounting component is of considerably less philosophical interest and should be approximated by market rates of return in well-­functioning markets.22 Since the social welfare function is an obvious candidate for moral discussion, it is worth considering how moral considerations can be adopted in consumption-­based discounting. In the next section, I argue that there are surprising difficulties in doing so.

3.  Descriptivism and Prescriptivism in Discounting Methodology An important distinction between families of discounting methodologies is between descriptive and prescriptive methodologies (or descriptivism and prescriptivism, sometimes also referred to as “positive” and “normative” economics, respectively). In the context of discounting and climate economics, this distinction comes from Arrow et al. (1996), following Manne (1995). Descriptivism has been the dominant historical consensus in economics from the late 1970s (Dennig 2018). However, the theoretical differences potentially encompass welfare economics whenever there is the possibility of utility being construed in either normative or positive fashions. One of the important claims in this section is that both are motivated by some form of impartiality; descriptivism about impartiality between revealed (or stated) preferences of different individuals and prescriptivism about impartiality between temporally dispersed pure costs and benefits. Note that, in this context, the term “preference” is very broad, and need not be confined to what philosophers sometimes call mere preferences; in other words, preferences may be informed by moral values, be other-­regarding, and/or be highly consequential. The first thing to recognize is that the distinction between descriptivism and prescriptivism is kept even if we accept the distinction between investment- and consumption-­based discounting (although the conjunction of prescriptivism 22  I say “approximated” because the market rates of discount need not converge to observed returns on investment.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

A Philosopher ’ s Guide to Discounting  101 and investment-­based discounting would be hard to defend). As I have argued for consumption-­based discounting here, it is worth considering how the meth­od­olo­gies would differ between descriptivism and prescriptivism.23 The claim I argue for is that, no matter which way we go, there will be an important role for moral experts. If considering prescriptivism in consumption-­based discounting, the relevant questions are to what extent society should discount future consumption. Stern (2006) and Cline (1992) take moral considerations to imply that δ ≈ 0, based on temporally impartial principles possibly stemming from classical utilitarian assumptions. Insofar as these are moral considerations, these judgments clearly require some type of (moral) expertise. If considering descriptivism in consumption-­based discounting, the relevant question is to what extent society does value greater over lesser consumption. Sometimes this is phrased as “inequality aversion”, but more properly we should think of it just as the generated utility. If the actual utility generated by those with higher income from a given marginal increase in consumption is lower than those with a lower income—i.e. if consumption has decreasing marginal value, then we do want to allocate it in such a manner that it does generate more utility. With the advent of positive economics in the twentieth century, it became clear that there was no empirically objective manner with which to evaluate such interpersonal comparisons of utility (cf. Dennig 2018 for a discussion of this historical debate). In the context of standard IAMs, the problem is even more acute since any decreasing marginal value is captured by a single (constant) parameter, η .24 So there is a very blunt parameter which governs this phenomenon which in practice could be very complicated. However, there are very good pragmatic reasons to take this parameter to be constant: keeping it constant makes the calculations tractable and making it a function would generate too many degrees of freedom for the modeler without clear benefits in accuracy. For instance, if the function was non-­differentiable every­where, that would make analyses far more complicated. Also, it would require determination of which kind of function to adopt (e.g. square root, cube root) as well as a variety of data points to fit curves to. In this constant, but tractable, form, η is not directly connected to observation. It is highly unlikely that society has preferences over consumption that can be modeled by a constant elasticity of marginal utility of consumption. On the one 23  Kelleher (2017b) points out that we can distinguish between descriptive approaches which are concerned with ranking potential outcomes in terms of a social welfare function (which is what I have been calling consumption-­based discounting as opposed to investment-­based) with those that are simply concerned with searching for potential Pareto improvements over the status quo. While this is in fact a distinction between outputs, I do not see the need here to be so fine-­grained about distinguishing methodologies. 24  By constant, I mean that it does not vary depending on the income level; it is not expressed as a function of consumption, which it could theoretically be.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

102  Kian Mintz-Woo hand, we can take the modeling as tractable approximations of individual ­preferences. On the other hand, by adopting these simplified forms, we divorce parameter assignments to η from straightforward observed preferences. In other words, in theory the parameter assignments could be directly responsive to empirical results but in practice they are approximations—justifiable because computationally tractable, but approximations nonetheless. So my claim is that, even under descriptivism, we require (expert!) judgment in order to amalgamate and ap­proxi­mate the empirical information. This means that descriptivist mo­tiv­ ations also justify, albeit in a more roundabout way, a demand for expertise. Thus, the first parallel between these methodological families is both require expertise. There is another interesting parallel between the families. If we are being ­char­it­able, both sides are appealing to impartiality, but for the descriptivist it is impartiality over (existing) people’s preferences and for the prescriptivist it is impartiality over (pure) costs and benefits.25 Since, in fact, people’s preferences are not temporally impartial, we have a conflict between the two. Many people’s preferences exhibit pure discounting (and often even hyperbolic discounting, which means significant discounting near the present and diminishing or even no discounting in the future (Frederick et al.  2002)), which can be exemplified by temporal myopia in the individual case. In other words, many people prefer that pleasant experiences occur closer to the present, even when this comes at a cost. To illustrate, a famous example of temporal myopia is the marshmallow experiment for children (Mischel et al.  1972). In this experiment, preschool children (aged three to five) were asked to delay a less preferred reward (such as eating a pretzel) in order to get a more desired reward (such as eating a marshmallow). The experiment was about what strategies children used to delay gratification, such as distracting themselves with the toys which were provided to some of them. The difficulty in delaying gratification, even when doing so yields a preferred outcome, is a well-­known way in which temporal myopia can be instantiated. We now turn to explaining the appeal of descriptivism.26 Why should we think that existing people’s preferences are normatively important? The standard economic response is that policy choices are about getting people what they want, and there are at least three reasons for trying to do so (I do not endorse all of these points).27 First, because introducing evaluation of people’s extant preferences is thought of as just applying one’s own preferences and there is no general way to show that one’s own preferences are superior. Second, because consumer 25  I think in the most charitable interpretation of the descriptivist position, it is impartiality over preferences at all times (and places) but since we cannot observe non-­current preferences, we make do with what is observable. 26  I thank Tristram McPherson for pushing for this explanation. 27  Dennig (2018: 44) notes that the standard economic view has “widespread agreement in the profession”.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

A Philosopher ’ s Guide to Discounting  103 preferences are subject to consumer sovereignty, and should be accorded priority due to democratic concerns. So, for instance, we have Lerner (1972: 258) writing that “This view I find very close to the idea of democracy or freedom—the idea of normally letting each member of society decide what is good for himself, rather than have someone else play a paternal role”. Third, because, unlike some alternatives, it is empirically applicable. A rough caricature of the position is: preferences are measurable, and the only preferences we have access to are the preferences of extant people. So, for instance, philosophers prefer to appeal to reasons or arguments, but these may not be the right kind of tool for such quantitative questions, or, to the extent they are, they may just feed into the preferences of those who endorse those reasons or arguments. These claims lead to the discussions in the final sections. While I think the first two reasons can be answered, my own view is that this third point is important and not sufficiently recognized. I think that a practical methodology has to recognize that there are only extant preferences available to consider. However, in the context of social policy, I deny that this implies that all preferences should count equally. My claim is that some social preferences can justifiably be weighted more highly than others. So although we have only preferences to contend with in setting these parameters, some of them may be better-­informed in a morally relevant manner.

4.  The Role of Moral Experts in Parameter Assignments My contention in this section is that, if one is being prescriptivist about discounting, a position which has been appealing to philosophers (e.g. Caney  2008; Moellendorf 2014), there are unrecognized challenges in appealing to moral theory in this context.28 I think there is a deep reason for it not being moral philo­ sophers who should decide the values of the normative parameters η and δ . By and large, moral philosophers do not have the training to recognize the implications of changing these parameter values. While they have deep familiarity with the moral theories that make these parameter values important in the evaluation of policy, they are not trained in how to assign values to these parameters or which data in the real world might suggest certain values (or ranges of values). More contentiously, I am going to argue that moral theory does not help make up for this lack. To be precise, I am conceiving of moral theory in a narrow manner, where it is the construction of valid and invalid arguments based on moral prin­ ciples. While this is far from all that moral philosophers do, it is a large and

28 Prescriptivism is not only of appeal to philosophers; for instance, Arrow (1999a) explicitly adopts prescriptivism.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

104  Kian Mintz-Woo important part.29 Furthermore, it is obviously the case that some moral ­philo­sophers are well-­ informed about the welfare assumptions in modern ­economics and, indeed, about the relevant empirical work. But, insofar as there is a trade-­off between knowledge of moral theory and the relevant domains, my view will be that the latter dominates in assigning values. In my view, applied ethics does require combining moral theory with empirical information. Note that these claims may not be true in general with respect to applied ethics. In most contexts, unlike the discounting problem, one is not working within a systematic framework but trying to balance intuitions about cases with principles. However, the point is that this will be challenging in this context where cost– bene­fit analysis already has significant presumptions about the normatively relevant factors—namely, costs and benefits. Furthermore, there is a dearth of principles that might point to specific values for the normative parameters. This is especially clear in the case of η . Utilitarianism, for instance, tells us to maximize utility but provides no guidance on how society gains utility with changes in consumption. But, again, the empirical data tell us that different approaches to measuring those changes provide divergent estimates for this parameter. So it is not a straightforwardly empirical question that can be simply considered the role of economists or other social scientists to solve.30 Prescriptivism is the view that the assignment of parameter values should involve moral considerations. Introducing moral considerations into discounting is the role of (some type of) moral expert. It would be reasonable to assume that these will be moral philosophers, but I will argue that this is not the case in this context. Here, I provide the argument in brief. Let us define moral philosophers as those who are knowledgeable about, and typically trained in the use of, moral theories.31 In other words, they know the implications of various moral principles or judgments and which moral theories cohere with different principles or judgments. So my contention is that prescriptivism requires a moral expert. However, I will argue that the first approximation—that the relevant moral experts will be moral philosophers—is problematic in this context, since the parameter value assignments can both have infinite ranges and take what I call “non-­critical values”. While I endorse this strong claim, it is worth noting that it is not necessarily needed here. I could point to the contingent point that philosophical training is unlikely to provide the mathematical acuity or the familiarity with the relevant 29  In Mintz-­Woo (forthcoming), I argue that it is the key role of moral philosophers; the evalu­ ation—and ultimately acceptance or rejection—of theories is plausibly the role of the public instead. 30  I will suggest that the preferences of relevant domain experts should have their preferences priv­ il­eged but not because it is a merely empirical question; rather that it is a moral question which is difficult to answer using moral argumentation. 31  In the limit case, those theories may not be systematic; for my purposes, a non-­systematic theory like moral particularism counts as a theory—just one with no theoretical implications for particular judgments. In other words, it generates no principled arguments which can be used to assess domains which satisfy the two criteria mentioned below.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

A Philosopher ’ s Guide to Discounting  105 empirical facts which are clearly relevant to having informed views on appropriate values for these normative parameters. While I think the conceptual point is ­correct, the contingent point is enough to suggest the limitations of moral philo­ sophers playing the role of moral experts. This is not to gainsay the role that moral philosophers can and should play in addressing intergenerational issues. Although I deny that philosophical argumentation can be adduced directly for particular parameter assignments, we can consider distinct roles for philosophers. A first is a standard, but more modest, role of evaluating or considering scientific methodologies and justifications. In other words, this role is not in discussing any particular parameter assignments, but instead discussing potential methodologies for generating parameter assignments. I think this is a helpful role for philosophers to play (indeed, I view the contributions of this chapter to be of this type). A second role concerns substantive moral presuppositions. So, for instance, the kinds of social welfare functions at issue seem to presuppose claims including, but not limited to, aggregationism, welfarism, and consequentialism. Moral theory and argumentation have an important role to play in determining whether to adopt these controversial moral claims. However, even considering this more limited second role, I think appearances could be misleading. The first point to consider is that it is possible that these controversial moral claims are not presupposed in standard economic analyses. As Kelleher (2017b) points out, we can imagine that the discounting problem has limited scope. For instance, the resultant judgments about which policies to pursue might only be value-­concerning, and not be all-­things-­considered judgments. There might be, for instance, countervailing justice-­related or deontological concerns. So, regardless of the theory adopted, insofar as we are concerned with avoiding negative consequences—a consideration that most theories have even if they take value to be one of multiple aspects of morality—the problem of discounting is relevant for the axiological questions which could arise.32 The second, and more important, point is that we can ask these questions internal to the moral theory while bracketing the external question of whether the theory is true. Considering these points conditional on the truth of the appropriate value claims, we see that the moral theory still only supports particular parameter assignments in a very few critical cases. My remaining worry is that, since moral philosophy is conducive to arguing for critical parameter values, one will take those to be the correct values. The worry is a version of the law of the instrument, usually credited to Abraham Kaplan. He wrote that “It comes as no particular surprise to discover that [one] formulates problems in a way which requires for their solution just those 32  There are of course moral theories in which consequences have no weight whatsoever. I lack space here to discuss theories of this type.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

106  Kian Mintz-Woo techniques in which [one] is especially skilled” (Kaplan 1964: 28).33 Philosophical argumentation is too coarse-­grained to assign parameter values outside of these critical values, since there are not philosophical arguments for δ = 1 or η = 2. The worry is that this could lead to endorsement of critical values simply because they are amenable to philosophical argumentation.34 So, to be clear, while prescriptivism requires appeal to moral theory and ­argumentation, it is not so clear that moral philosophy can actually fulfill this role outside of a few critical values. My claim is not that these critical values are ­necessarily false, but that focus upon them could lead us to neglect the rest of the possibility space and that simply because these critical values are subject to philosophical argumentation, they might appear more plausible to philosophers. An objection to this position might be that it is in light of the possibility of applying moral argumentation to critical values that we know they are the correct ones. So, for instance, many philosophers think that pure discounting is impermissible and that δ = 0 is required by morality,35 either by consequentialist impartiality (Broome 1994; Gardiner 2011; Jamieson 2014; Parfit 1984) or by broadly deontological concerns (Caney  2008,  2009; Davidson  2006).36 However, this is sometimes accompanied by the claim that η should be revealed from consumer behaviour—or, more often, ignoring the role of η altogether, despite its relative importance. I will argue that this position is more difficult to defend than is usually recognized. Notice that this is the position taken by both Stern (2006) and Cline (1992), who argue that δ ≈ 0 is implied by utilitarianism but that η should be revealed through observed choices.37 As Dasgupta (2008: 159) argues: Cline and Stern would appear not to have bothered at all about consistency. They chose η on the basis of estimates obtained from consumer behaviour, but ignore consumer behaviour entirely when it came to the choice of δ and sought the advice of moral philosophers instead.

33  The first recorded use of the law was a spoken report from a medical conference, where Kaplan said, “Give a boy a hammer and everything he meets has to be pounded” (Horowitz 1962: 637). 34  I may be granting too much to the zero pure discounters; Heath (2017: 453) suggests that most philosophers don’t argue for zero pure discounting but instead act as if “the commitment to a zero rate of pure time preference serv[es] as some sort of framework principle”. 35  Sometimes, such as with Parfit (1984) or Kelleher (2012), there is imprecision about whether philosophers argue that δ = 0 or that r = 0, but I think it more charitable to read them all as arguing for the former (and see Heath 2017 for similar concerns). 36  However, rejecting probability discounting, as Caney (2009) does, becomes problematic because one’s theory risks becoming too coarse-­grained to address small changes in risk (Mintz-­Woo 2017). 37  The motivation is that η is dependent on how actual people smooth consumption, revealing how their utility is affected by temporally unequal consumption. However, in the context of social policies, the parameter η doesn’t just govern intrapersonal utility, but interpersonal social utility. Since we cannot directly measure interpersonal utility, η is not simply a descriptive parameter in the intergen­er­ation­al context.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

A Philosopher ’ s Guide to Discounting  107 When arguing for a consistent methodology, we can view this as a reductio.38 Suppose that these moral philosophers should determine the value of the normative parameter δ . Then, to maintain methodological consistency, they should determine the value of the normative parameter η . It is implausible that moral philosophers should be the source for parameter assignments for η . So, assuming we want to treat different normative parameters consistently, this could be seen as a reductio on moral philosophers determining the value of δ .39 There is a second reason to take η and δ to be both treated as normative parameters (or neither). If we adopt the Ramsey Rule and fix the social discount rate r (for instance, by appeal to observed rates of return) and the growth rate of consumption g (for instance, keeping this low to keep intergenerational consumption similar), then η and δ end up being ethical duals of each other (Dasgupta 2005). In other words, the Ramsey Rule is observed using infinitely many pairs of η and δ , and by reducing one and increasing the other, we could respect the rule. So, in practice, we have two moral judgments we can adjust as pairs of parameter values. Since we can manipulate one normative parameter by changing the value of the other, this places them in some sense morally on a par. The upshot is that these parameter value assignments are normative, broadly speaking, but, I claim, not the especial realm of moral theory or moral philo­ sophers. This point has gone unrecognized by many working in this area.

5. Conclusion While I have argued that consumption-­based discounting is appropriate for social projects, and that prescriptivism about the discounting problem is appropriate, the major claim of this chapter has been negative: moral philosophy is of limited use in the context of quantitative issues such as these. However, this can also be seen as an interesting theoretical claim: practical ethical judgments may require normative considerations that are not a part of traditional philosophical training. Although I do think that moral experts are needed to inform prescriptivism about the discounting problem, I have left it open which moral experts they are—i.e. if they are not moral philosophers, but we might think of policy makers, economists, or statisticians. In other work, I discuss the role that those not especially

38  Note that we are assuming that the normative parameters, η and δ to be subject to prescriptivism and moral considerations. My point is that doing so in a parallel manner for both is, at the very least, challenging. 39  Frej Klem Thomsen suggests that we can run this as a ponens rather than a tollens: if one thinks non-­moral philosophers should have no role in determining the appropriate value of δ then, by parity, they should have no role in determining values for η . I am sympathetic insofar as these parameters should be taken as ethical duals and, while I think the tollens is more plausible, am happy to grant that the ponens would be more plausible for some.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

108  Kian Mintz-Woo trained in moral or ethical issues can bring to moral questions.40 In particular, I would suggest that those who know about the theory and practice of the implications of changing the moral parameter assignment values are better placed to be experts in the context of the discounting problem. While those who have not worked directly with the parameter values will have little intuitive feel for differences between parameter value assignments, those who do will. My view is that means that their intuitions should be weighted more heavily, perhaps even by taking them to be the moral experts in this context.41

References Arrow, K. (1999a). “Discounting, Morality, and Gaming,” in Discounting and Intergenerational Equity, ed. P.  R.  Portney and J.  P.  Weyant. Washington, DC: Resources for the Future, 13–22. Arrow, K. (1999b). “Inter-Generational Equity and the Rate of Discount in LongTerm Social Investment,” in Contemporary Economic Issues, Volume 4, ed. M. R. Sertel. Basingstoke: Macmillan, 69–88. Arrow, K., W.  R.  Cline, K.-G.  Mäler, M.  Munasinghe, R.  Squitieri, and J.  E.  Stiglitz (1996). “Intertemporal Equity, Discounting, and Economic Efficiency,” in Climate Change 1995—Economic and Social Dimensions of Climate Change, ed. J.  Bruce, H. Lee, and E. Haites. Cambridge: Cambridge University Press, 125–144. Broome, J. (1994). “Discounting the Future.” Philosophy and Public Affairs 23(2): 128–156. Caney, S. (2008). “Human Rights, Climate Change, and Discounting.” Environmental Politics 17(4): 536–555. Caney, S. (2009). “Climate Change and the Future: Discounting for Time, Wealth, and Risk.” Journal of Social Philosophy 40(2): 163–186. Caney, S. (2014). “Climate Change, Intergenerational Equity and the Social Discount Rate.” Politics, Philosophy & Economics 13(4): 320–342. Cline, W. R. (1992). The Economics of Global Warming. Washington, DC: Institute for International Economics. Dasgupta, P. (2005). “Three Conceptions of Intergenerational Justice,” in Ramsey’s Legacy, ed. H. Lillehammer and D. H. Mellor. Oxford: Clarendon Press, 149–169. Dasgupta, P. (2008). “Discounting Climate Change.” Journal of Risk and Uncertainty 37: 141–169. 40  This thesis work is summarized in Mintz-­Woo (2018a). 41 Thanks to Francis Dennig, Eike Düvel, Maddalena Ferrana, Marc Fleurbaey, Paul Kelleher, Lukas Meyer, Jeff Sebo, Karl Steininger, Frej Klem Thomsen, and to the editors of this volume for their extensive and helpful commentary. I also appreciate comments from audiences at the Mercator Research Institute on Global Commons and Climate Change, especially commentators Martin Kowarsch and Linus Mattauch, and at the University of Graz.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

A Philosopher ’ s Guide to Discounting  109 Dasgupta, P. (2019). Time and the Generation: Population Ethics for a Diminishing Planet. New York: Columbia University Press. Davidson, M.  D. (2006). “A Social Discount Rate for Climate Damage to Future Generations Based on Regulatory Law.” Climatic Change 76(1): 55–72. Davidson, M.  D. (2015). “Climate Change and the Ethics of Discounting.” Wiley Interdisciplinary Reviews: Climate Change 6(4): 401–412. Dennig, F. (2018). “Climate Change and the Re-evaluation of Cost-Benefit Analysis.” Climatic Change 151(1): 43–54. Fleurbaey, M., M.  Ferranna, M.  Budolfson, F.  Dennig, K.  Mintz-Woo, R.  Socolow, D.  Spears, and S.  Zuber (2019). “The Social Cost of Carbon: Valuing Inequality, Risk, and Population for Climate Policy.” The Monist 102(1): 84–109. Fleurbaey, M. and S.  Zuber (2015). “Discounting, Beyond Utilitarianism.” Economics 9(12). Frederick, S., G.  Loewenstein, and T.  O’Donoghue (2002). “Time Discounting and Time Preference: A Critical Review.” Journal of Economic Literature 40(2): 351–401. Gardiner, S. (2011). A Perfect Moral Storm: The Ethical Tragedy of Climate Change. Oxford: Oxford University Press. Goulder, L. H. and R. C. Williams (2012). “The Choice of Discount Rate for Climate Change Policy Evaluation.” Climate Change Economics 03(04): 1250024. Greaves, H. (2017). “Discounting for Public Policy: A Survey.” Economics and Philosophy 33(3): 391–439. Groom, B., C.  Hepburn, P.  Koundouri, and D.  Pearce (2005). “Declining Discount Rates: The Long and the Short of It.” Environmental and Resource Economics 32(4): 445–493. Heal, G. M. (2007). “Discounting: A Review of the Basic Economics.” The University of Chicago Law Review 74(1): 59–77. Heath, J. (2017). “Climate Ethics: Justifying a Positive Social Time Preference.” Journal of Moral Philosophy 14(4): 435–462. Horowitz, M.  J. (1962). “Trends in Education.” Journal of Medical Education 37(6): 634–637. Jamieson, D. (2014). Reason in a Dark Time: Why the Struggle to Stop Climate Change Failed—and What It Means for Our Future. Oxford: Oxford University Press. Kaplan, A. (1964). The Conduct of Inquiry. San Francisco: Chandler. Kelleher, J. P. (2012). “Energy Policy and the Social Discount Rate.” Ethics, Policy & Environment 15(1): 45–50. Kelleher, J. P. (2017a). “Descriptive versus Prescriptive Discounting in Climate Change Policy Analysis.” Georgetown Journal of Law & Public Policy 15: 957–977. Kelleher, J. P. (2017b). “Pure Time Preference in Intertemporal Welfare Economics.” Economics and Philosophy, 33(3): 441–473. Koopmans, T. C. (1960). “Stationary Ordinal Utility and Impatience.” Econometrica: Journal of The Econometric Society, 28(2): 287–309.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

110  Kian Mintz-Woo Lerner, A.  P. (1972). “The Economics and Politics of Consumer Sovereignty.” The American Economic Review 62(1/2): 258–266. Manne, A. S. (1995). “The Rate of Time Preference.” Energy Policy 23(4/5): 391–394. Mintz-Woo, K. (2017). “A New Defence of Probability Discounting,” in The Ethical Underpinnings of Climate Economics, ed. A.  Walsh, S.  Hormio, and D.  Purves. Oxford: Routledge, 87–102. Mintz-Woo, K. (2018a). “Moral Uncertainty Over Policy Evaluation.” Erasmus Journal for Philosophy and Economics 11(2): 291–294. Mintz-Woo, K. (2018b). “Two Moral Arguments for a Global Social Cost of Carbon.” Ethics, Policy & Environment 21(1): 60–63. Mintz-Woo, K. (2019). “Principled Utility Discounting Under Risk.” Moral Philosophy and Politics 6(1): 89–112. Mintz-Woo, K. (Forthcoming). “Public Values in the Right Context.” Australasian Philosophical Review. Mischel, W., E. B. Ebbesen, and A. Raskoff Zeiss (1972). “Cognitive and Attentional Mechanisms in Delay of Gratification.” Journal of Personality and Social Psychology 21(2): 204–218. Moellendorf, D. (2014). The Moral Challenge of Dangerous Climate Change. Cambridge: Cambridge University Press. Nordhaus, W.  D. (1994). Managing the Global Commons: The Economics of Climate Change. Cambridge, MA: MIT Press. Nordhaus, W. D. (2007a). “A Review of the ‘Stern Review on the Economics of Climate Change.’ ” Journal of Economic Literature 45(3): 686–702. Nordhaus, W.  D. (2007b). “Critical Assumptions in the Stern Review on Climate Change.” Science 317(5835): 201–202. Parfit, D. (1984). Reasons and Persons. Oxford: Clarendon Press. Rawls, J. (1971). A Theory of Justice. Cambridge, MA: Belknap Press. Sidgwick, H. (1907/1981). Methods of Ethics (7th edition). Indianapolis: Hackett Publishing. Solow, R.  M. (1974). “Intergenerational Equity and Exhaustible Resources.” The Review of Economics and Statistics 41: 29–45. Stern, N. (2006). The Economics of Climate Change: The Stern Review. Cambridge: Cambridge University Press. Sterner, T. and U. M. Persson (2008). “An Even Sterner Review: Introducing Relative Prices into the Discounting Debate.” Review of Environmental Economics and Policy 2(1): 61–76. Weitzman, M. L. (2007). “A Review of the ‘Stern Review on the Economics of Climate Change.’ ” Journal of Economic Literature 45(3): 703–724.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

5

Does Climate Change Policy Depend Importantly on Population Ethics? Deflationary Responses to the Challenges of Population Ethics for Public Policy Gustaf Arrhenius, Mark Budolfson, and Dean Spears

To plan an appropriate response to climate change, it is important to evaluate each of the alternative responses that are available. How can we take into account changes in the world’s population? Should society aim to promote the total of people’s wellbeing in the world, or their average wellbeing, or something else? The answer to this question will make a great difference to the conclusions we reach. Intergovernmental Panel on Climate Change (Pachauri et al. (2015))

1. Introduction The Intergovernmental Panel on Climate Change (IPCC) and some leading philo­sophers and economists have expressed unease about the implications of population change for evaluating responses to climate change and other intergen­ erational policy challenges. Their unease derives from a common view among those who investigate the questions of population ethics, that is, theories about the value of outcomes where the number of people, the quality of their lives, and their identities may vary. The view is that we do not know what to do about inter­ generational policy until we know what to do about population ethics. John Broome, in particular, has prominently voiced the concern that climate policy could turn critically on unresolved questions in population ethics.1 The worry expressed by Broome and reflected in the quote from the IPCC above might be stated as follows: 1  See, e.g., Broome (1992), (2004: ch. 1), and (2012).

Gustaf Arrhenius, Mark Budolfson, and Dean Spears, Does Climate Change Policy Depend Importantly on Population Ethics? Deflationary Responses to the Challenges of Population Ethics for Public Policy In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Gustaf Arrhenius, Mark Budolfson, and Dean Spears. DOI: 10.1093/oso/9780198796282.003.0006

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

112  Gustaf Arrhenius, Mark Budolfson, and Dean Spears Worry:  Because climate change, climate policy, the size of the population, and population policy all may have effects on one another, and because population ethics is so theoretically unresolved as to permit a wide range of reasonable dis­ agree­ment about social evaluation, our ignorance of the correct population ethic implies serious practical ignorance about what climate policies to pursue.2 In this chapter, we argue that the Worry is not obviously well-­founded: we may already know enough to make good choices about climate policy even without further progress in population ethics, and further progress might not make much difference to the conclusions that are ultimately correct. More generally, we high­ light some reasons—some philosophical, some empirical—why intergenerational policymaking might not be very sensitive to classic arguments from population ethics in the way that have often been assumed. To understand why the IPCC and many others share the Worry, we must begin by noting that intergenerational policymaking seems to require a concept of goodness that aggregates consequences for many different people (perhaps even non-­humans), with different properties, living at different times. Most of these people are not yet alive. Most of them will only ever be born depending on which particular climate policy is chosen. But any response to climate change requires integrating over the consequences for all of them. For example, consider the Integrated Assessment Models (IAMs) of climate policy constructed by economists and other researchers. In 2018, William Nordhaus was awarded the Economics prize to the memory of Alfred Nobel, partly for his family of climate policy IAMs. IAMs like Nordhaus’ choose an optimal carbon tax policy, balancing the disadvantages of more expensive energy with the advantages of reduced global warming. More broadly, reducing fossil fuel consumption could increase present-­day economic costs for both poor ­people and rich people; could slow economic growth and poverty alleviation in the developing world; and could prevent future harm from temperature increases—increases which will help some people, but hurt many more people, and have consequences for inequality. The socially optimal carbon tax or fossil fuel policy depends on taking all of these and other relevant factors into proper account—which seems to require weighing the aggregate of these consequences conditional on different policy options. So, choosing a policy response to climate change seems to demand an aggrega­ tive concept of goodness—an axiology. Those who study axiology have devoted 2  “We do not know what value to set on changes in the world’s population. If the population shrinks as a result of climate change, we do not know how to evaluate that change. Yet we have reason to think that changes in population may be one of the most morally significant effects of climate change. The small chance of catastrophe may be a major component in the expected value of harm caused by climate change, and the loss of population may be a major component of the badness of catastrophe. . . . So we face a particularly intractable problem of uncertainty, which prevents us from working out what we should do. Yet we have to act; climate change will not wait while we sort our­ selves out” (Broome 2012: 183–185).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Responses to the Challenges of Population Ethics  113 considerable theoretical attention to population ethics: to the questions of how rankings of aggregate social goodness extend to ranking outcomes in which ­different people and different numbers of people exist. Parfit (1984) identified many of the core questions of population ethics, which are widely regarded to remain open. A number of candidate resolutions have been offered in the literature, but a formal literature involving impossibility theorems—led by Arrhenius (2000a, 2000b) and subsequent work—has demonstrated that each approach (and all pos­ sible approaches) has one or more seemingly counterintuitive implication. These theorems appear to show that our considered moral beliefs are mutually incon­ sistent, that is, that necessarily at least one of our considered moral beliefs is false. Since consistency is, arguably, a necessary condition for moral justification, it may appear that we are forced to conclude that there is no moral theory which can be justified. Moreover, we would then lack the theoretical tools needed to evaluate climate options in which the number of people, the quality of their lives, and their identities will differ. In Section  2 we introduce in more detail these paradoxes and the related popu­ la­ tion axiology literature, with special focus on Parfit’s well-­ known Repugnant Conclusion. With this introduction in hand, Section  3 offers the first and simplest of two deflationary responses to the Worry: it may be, given the actual facts of climate change, that all axiologies agree on a particular policy response. In this case, there would be a clear dominance conclusion, and the puzzles of population ethics would be practically irrelevant (albeit still theoretically challenging). Section 4 offers the second more complex deflationary response: despite the impossibility results from Arrhenius, it is nonetheless possible to prove the possibility of axiologies that satisfy bounded versions of all of the desiderata from the population ethics literature that Arrhenius’ proofs marshal. In this way, an incomplete population axiology that is defined over the practically relevant bounded space can avoid the Repugnant Conclusion and satisfy other relevant bounded versions of the adequacy conditions in population ethics. Assuming that we only need to consider the bounded versions of the adequacy conditions when we consider policy issues, and that analogous impossibility theorems cannot be proved in the bounded domain, we can for practical purposes put the impossibility the­orems that have haunted population ethics to the side. These deflationary responses do not show that theoretical progress towards population axiology should not continue. Indeed, as we shall show below, an important consequence of the second deflationary response is that it shows the need of more scrutiny of what the core intuitions behind the adequacy conditions in population ethics really are, and further investigation of axiologies on bounded domains. The upshot of this chapter is that responding to climate change, and policy analysis more generally, may not need to wait for greater consensus in population ethics on unbounded domains, and that the possibility of deflationary responses to the impossibility theorems deserves further attention.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

114  Gustaf Arrhenius, Mark Budolfson, and Dean Spears

2.  Population Axiology and the Repugnant Conclusion Population axiology concerns how to evaluate populations of different sizes in regard to their goodness: how to assign a value to increases and decreases in population size. The first few papers in this field were not published until the late 1960s and it did not become a significant field until Derek Parfit’s famous book Reasons and Persons, published in 1984. It is now a very lively field of inquiry. As John Broome has noted, policymakers seem to almost universally ignore the effects of policy on population size. Why do they ignore it? One possible explanation is that many people have what Broome calls the Intuition of Neutrality, which holds that adding a person to the world’s population makes the world neither better nor worse.3 Hence, effects on population size is something that we do not need to think about, or if we do need to think about it, it is because it makes people’s lives better or worse; other than that, having a bigger or smaller population does not make any difference to the value of outcomes. There are likely to be limits to Neutrality. For example, most people would probably agree that if population growth leads to having many people with very bad lives, then that would make the world worse. In light of this, we think that among those people who have intuitions in this neighborhood, it is more likely that they endorse the more limited Asymmetry Intuition (which also appeared earlier in the literature):4 We have no moral reasons for or against creating people with positive welfare stemming from the welfare these people would enjoy, but, on the other hand, we have moral reasons against creating people with negative welfare stemming from the negative welfare these people would suffer. Hence, those people are neutral only about adding people with positive welfare.5 However, assuming that future people have positive or neutral welfare, the idea is that population size is neutral in terms of value and that we can ignore this aspect when considering different policies. However, Neutrality and Asymmetry each on their own lead to inconsistency given some other beliefs that most of us share. Consider the following two possible additions to the present population A, each of which would be the result of an alternative climate policy: • Population B consists of a number of people with very low positive ­welfare, and 3  For a more detailed discussion of the neutrality intuition, see Broome (2004) and 2010). 4  How many people in fact endorse the Asymmetry is an empirical question; in one recent survey Spears (2019) finds that only a minority of respondents do. The study also provides suggestive evi­ dence for weaker versions of the Asymmetry focused on the weight of suffering and parental pro­cre­ ative autonomy, as discussed in Arrhenius (forthcoming: section 9.5). 5  This formulation is from Arrhenius (forthcoming) and (2000b). For earlier formulations, see McMahan (1981) and Parfit (1982).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Responses to the Challenges of Population Ethics  115 • Population C is a population of the same size as B but made up of people with very high welfare. According to Neutrality and Asymmetry, either adding B or adding C to A each would make the resulting populations equally good, given full com­par­abil­ity.6 But surely, when other things are equal, it must be better to create people with very high welfare rather than people with very low welfare. Hence, popu­la­tion A+C is better than population A+B, which contradicts Neutrality and Asymmetry. So they are false. And because they are false, climate policymaking must consider population size in its evaluation of outcomes. The opening quotation from the IPCC listed two alternative approaches to aggregating welfare. One approach is Total Utilitarianism: when we evaluate future populations in respect of population change, we look at the total welfare in the different possible outcomes and rank them by how much total welfare they contain. According to this view, we should maximize the total amount of welfare in the world. So if there are more people with lives worth living, then that is better. Now a problem with this view is that it has a number of very counterintuitive implications. Much theoretical attention in population ethics has focused on a particular implication of Total Utilitarianism. Total welfare can be increased in two ways when the size of the population is no longer fixed: by keeping the popu­ la­tion at a constant size and making people’s lives better, or by increasing the size of the population by adding new people with lives worth living. So, according to Total Utilitarianism, a future with an enormous population with lives barely worth living could be better than a future with a smaller population with very high individual quality of life. But the idea that it would be better to radically increase the world’s population at the expense of future people’s individual welfare seems repugnant to many, and rather a reason to reject Total Utilitarianism. It is an instance of Parfit’s infamous Repugnant Conclusion: Repugnant Conclusion:  For any population consisting of people with very high positive welfare, there is a better population in which everyone has a very low positive welfare, other things being equal.7 6  Giving up full comparability isn’t sufficient to save the Neutrality and Asymmetry Intuition, see Arrhenius (forthcoming) and Broome (2004). 7  Here’s how Parfit (1984: 388) formulates the conclusion: “For any possible population of at least ten billion people, all with a very high quality of life, there must be some much larger imaginable population whose existence, if other things are equal, would be better, even though its members have lives that are barely worth living.” Hence, our formulation from Arrhenius (2000b) is more general than his. The ceteris paribus clause in the formulation is meant to imply that the compared popu­la­ tions are roughly equal in all other putatively axiologically relevant aspects apart from individual wel­ fare levels. Although it is through Parfit’s writings that this implication of Total Utilitarianism has become widely discussed, it was already noted by Henry Sidgwick (1907: 415), before the turn of the century. For other early sources of the Repugnant Conclusion, see Broad (1979: 249–250), McTaggart (1927: 452–453), and Narveson (1967).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

116  Gustaf Arrhenius, Mark Budolfson, and Dean Spears In Figure 5.1, the width of each block represents the number of people; the height represents their lifetime welfare. Dashes indicate that the block in question should be much wider than shown, that is, the population size is much larger than shown. These populations could consist of all the past, present, and future lives, or all the present and future lives, or all the lives during some shorter time span in the future such as the next generation, or all the lives that are causally affected by, or consequences of a certain action or series of actions, and so forth. All the lives in the diagram have positive welfare, or, as we also could put it, all the people have lives worth living. The A-­people have very high welfare whereas the Z-­people have very low positive welfare. The reason for this could be that in the Z-­lives there are, to paraphrase Parfit, only enough ecstasies to just outweigh the agonies, or that the good things in those lives are of uniformly poor quality, e.g., eating potatoes and listening to Muzak.8 Or it could be that the Z-­people have quite short lives as compared to the A-­people. We could imagine that in A, the people live for, say, 80 years whereas in Z the average life expectancy is, say, 40 years, like in some developing countries in the 1970s. However, because there are many more people in Z, the total sum of welfare in Z is greater than in A. Hence, a theory like Total Utilitarianism, according to which we should maxi­ mize the welfare in the world, ranks Z as better than A—an instance of the Repugnant Conclusion. As the name indicates, many people find the Repugnant Conclusion a reason to reject Total Utilitarianism; to these, the idea that we can make the world better by expanding the population at the expense of future people’s individual quality of life seems very counterintuitive. The Repugnant Conclusion has sometimes been taken in the literature as the major objection to Total Utilitarianism that allegedly disqualifies it as a plausible axiology.9 The other approach mentioned by the IPCC is to maximize average welfare in the world. This is what Average Utilitarianism tells us to do. Returning to Figure 5.1, in the case of the A and Z populations the average principle recommends A,

Very low positive welfare Population Z is much larger than A

Very high positive welfare A

Z

Figure 5.1  The Repugnant Conclusion 8  See Parfit (1984: 388) and Parfit (1986: 148). 9  There are other implications of Total Utilitarianism in population ethics that arguably are even more counterintuitive than the Repugnant Conclusion, see e.g., Arrhenius (forthcoming), (2000b), and (2011). More on this below.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Responses to the Challenges of Population Ethics  117 because average welfare is much higher in A than in Z. Hence, Average Utilitarianism avoids Parfit’s Repugnant Conclusion, which may seem to count in its favor.10 Unfortunately, it has even worse problems. One problem with maximizing aver­ age welfare is that it implies that it can be better to add one group of people to the population rather than some other group, even if each person in the former group has a life that is not worth living and each person in the latter group has a life that is worth living. This is illustrated in Figure 5.2. Here, we have the A population where the x-­people’s quality of life is very high. Assume that we can either increase population either by adding the y-­people that have quite low but positive welfare—their lives are worth living—or by adding the z-­people, all of whom are suffering horribly—their lives are not worth living. Because adding a lot of people with very low but positive welfare can decrease the average welfare of the population more than adding fewer people suffering horribly, it might be better, according to Average Utilitarianism, to add the suffer­ ing lives (the z-­people) rather than the lives worth living (the y-­people). Again, we have a very counterintuitive conclusion on our hands. This is what Arrhenius called the Sadistic Conclusion: Sadistic Conclusion: It can be better to expand the population by adding ­people with negative welfare rather than adding people with positive welfare, other things being equal.11 The path away from the Repugnant Conclusion towards the Sadistic Conclusion illustrates the puzzles that motivate the Worry. There may be no principle for

x

x

x y

A

A+

z

A–

Figure 5.2  The Sadistic Conclusion 10  As explained below, Budolfson and Spears (2018c) have argued that Parfit’s initial illustration is only a subset of the classical Repugnant Conclusion, and that we should understand it to include a version (based on addition to a base population, explained in their paper) that is implied by Average Utilitarianism and other axiologies that are commonly taken to avoid the Repugnant Conclusion. Throughout this section, for clarity we maintain the standard terminology in the population literature, except where it is clear we are discussing the argument of Budolfson and Spears. Anglin (1977) and Arrhenius (2000b: chs. 3 and 10) note that Average Utilitarianism implies a version of the Repugnant Conclusion to the effect that for any population with very high welfare, it can be worse to add this population rather than a population with very low welfare. As Anglin summarized simply: “in some cases the average principle also leads to the Repugnant Conclusion” (1977: 746). 11  See e.g., Arrhenius (2000b) and (2000a).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

118  Gustaf Arrhenius, Mark Budolfson, and Dean Spears evaluating populations that is not in some way very counterintuitive. This ­possibility was originally raised informally by Parfit, who presented a number of paradoxes in population ethics. Much of the important theoretical progress since then has been in formalization of these conclusions and axiologies, as well as many others, and their integration into rigorous proofs. This literature has progressed, at first, through a dialogue in which researchers proposed and formalized alternative population axiologies (Greaves 2017). Each was specially formulated to avoid versions of the Repugnant Conclusion, and then further explored by researchers. So, Ng (1989) introduced a variable-­value axiology, in which the average utility of a population is inflated by a positively increasing, concave function of population size, such that social evaluation asymptotes from nearly-­Total Utilitarianism to nearly-­Average Utilitarianism as population size increases. Like Average Utilitarianism, Ng’s theory does not escape the Sadistic Conclusion. Blackorby and Donaldson (1984) and later Blackorby et al. (1995) propose Critical-­Level Generalized Utilitarianism; this approach also avoids the Repugnant Conclusion at the cost of implying the Sadistic Conclusion. Other approaches, such as Sider’s (1991) theoretical example of Geometrism, or Asheim and Zuber’s (2014) Rank-­Dependent Generalized Utilitarianism, attend to people’s rank within a population, like maximin does. These avoid the Repugnant Conclusion, but have other implausible properties, including in cases where population size does not change, such as recommending redistribution from the worst off to the best off in some cases.12 None of these proposals has resolved the paradoxes. Led by Arrhenius (2000b), the literature has now established a number of impossibility theorems that dem­ onstrate that no axiology can simultaneously satisfy various sets of very compel­ ling adequacy conditions or principles. Trying to satisfy all of them at the same time leads to contradiction. These conditions are of the type that we have been considering—for example, what Arrhenius calls the Egalitarian Dominance Condition, which states that one population A is better than another same-­sized population B if A is perfectly equal and every person in A is better off than every person in B. This condition is incompatible with several other compelling condi­ tions, including conditions that are formulated to rule out the Repugnant and the Sadistic Conclusions. The first and perhaps most well-­known of these impossibil­ ity theorems is the following: Impossibility Theorem (Arrhenius  2000a): There is no welfarist axiology that satisfies the Dominance, the Addition, and the Minimal Non-­Extreme Priority Principle and avoids the Repugnant, the Sadistic, and the Anti-­Egalitarian Conclusion.13 12  See Arrhenius (forthcoming) and (2000a), and Arrhenius et al. (2014). 13 For theorems with logically weaker and intuitively even more compelling conditions, see Arrhenius (forthcoming), (2000a), (2001), and (2011).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Responses to the Challenges of Population Ethics  119 Although we refer the reader to the formal statement by Arrhenius (2000a), we emphasize here that each of the conditions listed in the theorem is intuitively compelling. For example, the Dominance Condition is simply that if everyone in population A is better off than everyone in population B, then A is better than B. Moreover, as Arrhenius has shown, there are theorems with logically weaker and intuitively even more compelling conditions.14 Impossibilities such as these are the challenges that motivate the Worry. One type of response to this challenge that we will set aside here is to offer a purported philosophical resolution to the challenge of the Repugnant Conclusion. Most of these purported resolutions argue that the Repugnant Conclusion should simply be accepted as true. For example, Hare (1988), Huemer (2008) Mackie (1985), Tännsjö (2002), and Gustafsson (forthcoming) have all offered arguments in favor of endorsing the Repugnant Conclusion, because of various arguments that the apparent repugnance of the conclusion is illusory or based on misunderstand­ ing. One drawback with this resolution is that the theorems with logically weaker conditions are not based on avoidance of the Repugnant Conclusion but on the intuitively more compelling Very Repugnant Conclusion: For any perfectly equal population with very high positive welfare, and for any number of lives with very negative welfare, there is a population consisting of the lives with negative welfare and lives with very low positive welfare which is better than the high welfare population, other things being equal.15 More recently, Budolfson and Spears (2018c) have offered an alternative type of resolution of the Repugnant Conclusion. They argue that Parfit’s original ex­ample of the Repugnant Conclusion should be understood as describing only a proper subset of instances of the Repugnant Conclusion, and that the full set of instances of the Repugnant Conclusion should be understood to include a broader set, including cases in which there is a base population that is unaffected by the choice between a larger or a smaller population.16 Given their more gen­ eral characterization of the Repugnant Conclusion, they prove that all of the most commonly discussed aggregative welfarist population axiologies imply at least one instance of this unrestricted Repugnant Conclusion. They then argue that because the Repugnant Conclusion so understood is a problem for all of the most

14  See, e.g., Arrhenius (forthcoming), (2000a), (2001), and (2011). 15  See, e.g., Arrhenius (forthcoming), (2000b), and (2011). For a detailed discussion of other prob­ lems with debunking arguments with regard to the Repugnant Conclusion, including Hare et al.’s arguments, see Arrhenius (forthcoming: ch. 3) and (2000b). 16  Budolfson and Spears’ general characterization of the Repugnant Conclusion including instances with non-­zero base populations is comparable to Arrhenius’ Strong Quality Addition Principle (Arrhenius forthcoming,  2000b), which is violated by both Total and Average Utilitarianism (and some other population axiologies). Arrhenius draws, however, a different conclusion from this result, namely that the Strong Quality Addition Principle should be rejected as an adequacy condition since it rules out too many axiologies in one fell swoop and thus is in that sense too strong.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

120  Gustaf Arrhenius, Mark Budolfson, and Dean Spears commonly discussed welfarist axiologies, it can no longer be reasonable to assume that a plausible axiology must avoid it. We set aside these purported solutions in this chapter. The problem we focus on is what the upshot of the population ethics literature is for policy on the assumption that there is no resolution to the challenges of population axiology at hand.

3.  First Deflationary Response: Axiologies May Agree about Climate Change The open theoretical questions of population axiology only turn out to be a prac­ tical problem for a policy challenge if population axiologies sufficiently disagree about the best policy response to that challenge. To see how this could turn out not to be the case in connection with climate change, consider the toy illustrative example in Figure  5.3. The figure plots a stylized version of the sort of climate policy decision considered by Integrated Assessment Models discussed earlier, including those developed by Nordhaus. If Figure  5.3 correctly described the full climate policy problem, then the Worry could be false, even though the candidate population axiologies differ. In the figure, the ethical question under consideration is what future decarboniza­ tion rate should be achieved: 100%, 0%, or some other optimum in between? The recommendations of two population axiologies are considered. These give differ­ ent evaluations of different options. Total Utilitarianism rises convexly as the decarbonization rate increases; Average Utilitarianism rises only concavely. Thus, Overall goodness Total Utilitarianism

Average Utilitarianism

0%

100% Future decarbonization

Figure 5.3  Two population axiologies recommend the same “corner solution” to optimal decarbonization

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Responses to the Challenges of Population Ethics  121 Average Utilitarianism thinks that a decarbonization rate of 90% would be only slightly worse than 100%, but Total Utilitarianism thinks 90% would be much worse than 100%. Note that Average and Total Utilitarianism even have different scales for goodness: neither their lowest level of goodness nor their highest levels of goodness are the same number, and their evaluations cover ranges of different length. This is important because some responses to normative uncertainty—such as Expected Moral Value—recommend an average or expectation over alternative theories (Budolfson and Spears 2018a; Bykvist 2017; Bykvist et al. 2019; Greaves and Ord 2017; Hedden 2016). This moral-­expectation approach has found difficulty in the need to compare evaluation quantities across theories, but that problem is not relevant in the case of Figure 5.3, because the two axiologies agree on the optimum. The point of Figure 5.3 is that both Average and Total Utilitarianism recom­ mend the same corner solution. In optimization, a “corner solution” is when the optimal policy is equal to a boundary constraint. Because Average and Total Utilitarianism both recommend full decarbonization, in this example, there is no practical disagreement between them, only theoretical disagreement. Whether or not actual climate policy is well-­described by Figure 5.1 is substantially an empirical question (concerning economics, demography, climate science, etc.), although also a normative one (because different losses, such as of life and wealth, must be aggregated). However, it is not implausible that actual climate policy questions could be resolved by dominance—that is to say, by agreement across candidate axiologies. For example, if we are confident that a particular set of future lives would be full only of terrible suffering and thus not worth living, and if by preventing those lives from occurring we prevent some harmful carbon emissions, and if furthermore we know these are the only relevant considerations, then all plausible population axiologies recommend not creating those lives. Although that example was fanciful, another might be quite realistic (see Scovronick et al. 2017 for detailed evaluation of the following). Consider invest­ ments in human development in developing countries, with a special focus on women’s social status and the education and well-­being of young women. This would have a range of likely consequences, which we can assume for hypothesis that we know with certainty (a level of confidence beyond the actual reach of social science): • The women who receive the program and the lives lived by other people in their places and times would be better: an increase in the near-­term average. • Long-­ term average well-­ being would be improved by reduced climate change and by accelerated economic development. • Some twenty-­first century lives that would have been worth living would not be lived, because of empowered young women choosing to reduce their fertility. (Under Total Utilitarian-­like theories, this would be a social cost.)

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

122  Gustaf Arrhenius, Mark Budolfson, and Dean Spears • Because of the reduced threat of climate change, the expected number of future good lives lived increases by more than the number of twenty-­first century lives reduced. In this case, the total expected number of lives lived would increase, average well-­being would increase within every time period, and average across-­time well-­being would increase because the average human would live later in his­­toric­al time. Moreover, it is not implausible that the welfare of the worst-­off lives would be higher (a property that matters to some egalitarian views), although this was not specified above. So, according to every plausible axiology in the literature and more—including Average Utilitarianism and related views, Total Utilitarianism and related views, maximin, and others—implementing the human development policy is recommended, in expectation. The upshot is that we can know whether to implement the policy without knowing the correct population axiology, and also without a general solution to moral uncertainty. In this case, the Worry would be deflated. More generally, other practical policy questions that are commonly taken to hinge on the choice of population axiology may be resolved by similar dominance arguments or corner solutions.17 This would depend on social, economic, and scientific facts. For example, some have argued that an implication of Total Utilitarianism is that substantially more resources should be invested in prevent­ ing human extinction (Beckstead 2013; Bostrom 2013). However, it may be that commonly discussed policy options (such as asteroid deflection) offer a small marginal benefit of further investment as compared to merely pursuing standard economic growth, technological progress, and human development. The reason being that such standard policies would have large co-­benefits against existential risk, perhaps because war of mass destruction or resistant, pandemic infectious disease would be less likely, or because survival-­promoting technologies would be invented. If so, both Average and Total Utilitarianism would recommend serious investment in thoughtful, long-­term human development, economic growth, and technical progress: Average Utilitarianism because it increases average well-­being, and Total Utilitarianism because it does this while also offering the co-­benefit of promoting survival. To be sure, this would not be the set of policies that humanity is currently pursuing, but it would not be a major reallocation into activities that 17  One exception to this possibility is the welfare of non-­human animals. The number and well-­being of non-­human animals is generally governed by ecological forces such as natural selection, to a greater extent than the number and well-­being of humans, which is regulated, in part, through complex tech­ nology and culture. In many cases, the implication of this fact may be that the average well-­being of non-­human animal species is kept within a narrow species-­specific range, while adjustment to changing conditions occurs in population size (on the extensive rather than the intensive margin, in economists’ language). If so, Average and Total Utilitarianism, as extended to non-­human animals, may give very different recommendations. See Hsiung and Sunstein (2006), and Budolfson and Spears (2018b) for more on climate and non-­human animals.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Responses to the Challenges of Population Ethics  123 only have the benefit of reducing existential risk, and nor would it turn on the choice of population axiology. Of course, it may be that the climate policy menu under consideration does not yield one dominating option. Also, there could be additional considerations, such as bounded political capital. If political capital is scarce, a politician who needs to compromise across politically linked issues (such as climate policy and domestic health care or tax policy) may care about how much worse 95% would be than 100%, which cannot be settled by this sort of dominance-­identification procedure. Still, this is a promising avenue for further research that should be pursued in light of the impossibility theorems in population axiology.

4.  Second Deflationary Response: Bounded Population Principles The Repugnant Conclusion—and especially the search for a sensible population-­sensitive social welfare function that does not imply the Repugnant Conclusion—has been a central focus of the population ethics literature since Parfit (1984) introduced it. For example, Arrhenius et al. (2014) have called it “one of the car­dinal challenges of modern ethics” and Greaves (2017) introduces the Repugnant Conclusion as “the key objection” to Total Utilitarianism and related views. Because most of the literature on population axiology takes it as an adequacy condition that an acceptable social welfare function should not imply the Repugnant Conclusion, researchers have proven that many social welfare functions, in addition to Total Utilitarianism, imply the Repugnant Conclusion if  the populations being evaluated can be unboundedly large. As noted above, Arrhenius (2000a, 2000b) presents an impossibility theorem that proves that no social welfare function can escape implying the Repugnant Conclusion, if the function is defined for unboundedly large population and has desirable—and plausibly ethically necessary—properties. Such properties are formalized as ax­ioms for Arrhenius’ theorems. These are impressive and rigorous philosophical results. But what are the implications for policy analysis? Do these results show that the assumptions of many leading policy analyses are illegitimate, as suggested by the quotes above from IPCC and John Broome? More generally, how should policy analysis respond to these results? Arrhenius notes that one response could be a thorough­ going skepticism or paralysis. However, he is much more enthusiastic about the possibility of a deflationary response: namely, to “try to find a way to explain away the relevance of the [Repugnant Conclusion and associated impossibility] the­orem for moral justification.”18 18  Arrhenius (forthcoming: ch. 13) and (2000b: ch. 12).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

124  Gustaf Arrhenius, Mark Budolfson, and Dean Spears Our goal in this section is to articulate another deflationary response to the impossibility theorems to the effect that policy analysis can in some cases le­git­im­ ate­ly ignore them and the Repugnant Conclusion when that analysis applies to bounded problems, as Arrhenius’ impossibility theorems assume unboundedness. We show that unlike unbounded cases, in bounded cases that are relevant to ­policy analysis, it is indeed possible to identify an axiology that captures all of the intuitions that support Total Utilitarianism while also avoiding the Repugnant Conclusion. This shows that it may be possible to endorse both the intuitions that motivate Total Utilitarianism and the intuition that tells against accepting the Repugnant Conclusion. The idea is that there might be a mere appearance of ­conflict between these intuitions that arises from taking our intuitions about the realistic range of cases relevant to policy as also extending to cases in the unbounded penumbra. In other words, this second deflationary response to the Worry exploits the possibility of interpreting the intuitively compelling axioms of population ethics as restricted to a bounded domain.19 An adequacy condition to avoid the Repugnant Conclusion on unbounded space has no implications for such a family of bounded axiologies. As we detail below, in our formal argument, our approach is not to reject that populations can be unboundedly large; instead, we propose bounded axioms that, in some cases, apply to only some of the space of possible populations.

4.1  Axiology with Population Size Bounds The practically relevant set of policy options that humanity will ever face is a bounded set, along many dimensions. This is partly because the set of practically relevant population sizes is bounded. This is true even if the possible values of social welfare are unbounded, in part because policy choices could only have boundedly large effects on individual welfare. In making the empirical observa­ tion that the set of practically relevant population sizes is bounded, we have in mind a very large upper bound. The upper bound could be much larger than the largest set that an expert predicts could ever be relevant. It is sufficient for our purposes, for example, that the bound be 1080, which is an estimate of the number of atoms in the universe, or 1058, which is the estimate of Bostrom (2013) of the number of simulated human lives that a superintelligence could create with the 19  Shiell (2008) offers a formal proof of an intuition (related to a point made by Parfit 1984: 387), namely that within a truncated domain, Total Utilitarianism need not imply the Repugnant Conclusion within that domain. In this way, Shiell’s proof depends essentially on truncating the choice set. In contrast, our proof below does not truncate the choice set. Our axiological principles cover the entire choice set, fully specify how to rank all outcomes within a policy-­relevant range, but do not fully specify how rank all outcomes beyond that range. Moreover, the principles also satisfy certain bounded analogues of the central population ethics desiderata involved in the impossibility theorems in the area.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Responses to the Challenges of Population Ethics  125 available energy in the universe. The lower bound on the policy relevant set of population sizes is the number of humans who already have ever been born. In this vein, even outside of population ethics, practical policy analyses are untroubled by imaginable, unbounded marginal utilities or counts of small harms; in this section, we formalize that observation by weakening some axioms of popu­la­tion ethics to a bounded domain. We can consider axioms that only apply to a very large but bounded subset of the potentially unbounded complete, im­agin­able social choice set, and choose a family of axiologies that (a) satisfies attractive axioms defined over the bounded set and (b) has no implications about the Repugnant Conclusion. A requirement to avoid the Repugnant Conclusion has no implications for this bounded family of axiologies. The purpose of axiomatic representation theorems is to rule in and rule out sets of functional forms. In general, a representation theorem permits a family of function shapes that leaves certain features unspecified. For an example in the context of axiologies, critical level generalized utilitarianism is consistent with concave or affine transformations of utility and with positive or zero critical ­levels; each of these combinations would have different normative implications. Similarly, a family of population-­sensitive axiologies could leave unspecified how populations are evaluated outside of the bounded set. Such a family of axiologies would ignore the Repugnant Conclusion—while fully specifying the social evalu­ ation on the bounded set. The literature has identified the following very general characterization of the space of a number of important aggregative welfarist axiologies:20

(

)

W = g (n) h n−1 ∑ i f (xi ) − h ( f (a) )  ,20  

where:

• n is population size, • xi is the utility of person i, • a is 0 or positive and is a critical level for adding a life to be a social improvement. • The functions f, g, and h are all non-­decreasing. If f and h are both the identity function, then we have utilitarianism. If f is concave and h is the identity func­ tion, then we have additively separable prioritarianism. If f is concave and h = f−1, we have a type of non-­separable egalitarianism. This general functional form is intended to clarify that the shape of g could be chosen independently of any combination of otherwise permissible features for

20  Budolfson and Spears (2018c); compare Greaves and Ord (2017).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

126  Gustaf Arrhenius, Mark Budolfson, and Dean Spears the other elements of the function. It includes as special cases many axiologies in the literature, although not rank-­dependent axiologies such as maximin or Asheim and Zuber’s (2014) rank-­dependent generalized utilitarianism, nor so-­called person affecting theories.21 In Total Utilitarianism g is linear; in Average Utilitarianism g is constant; and in Ng’s Theory X’ g is concave. Below, we will use the term “totalist” to refer to the family of theories according to which g is linear. Figure 5.4 illustrates a possibility for g that is the focus of this section of the chapter: a family of functional forms for g could be chosen that fully specifies g on the bounded policy-­relevant set, while avoiding the Repugnant Conclusion and taking no stand on the shape of g outside the bounded set. Functional forms a, b, c, and d would rank policy options over the practically relevant set identically, for any given specification of f, h, and a. Form a matches Total Utilitarianism, if f and h are the identity function. Forms b, c, and d are blank at populations smaller that the bounded choice set, to illustrate that they do not make assumptions about how to rank populations this small. It is not essential to our argument that the bounded set have either a zero or a positive lower bound: the possibility of a lower bound greater than zero represents the minimum on policy-­relevant population sizes due to the fact that billions of humans have already been born. Forms a, b, and c have different implications for the Repugnant Conclusion, and may or may not invoke other undesirable properties outside of the practically relevant set. Form d is not a fully specified function form, but is merely a repre­ sentation of the possibility of a decision-­ maker remaining uncertain about options outside of the bounded set. The existence of functional forms a, b, and c and of the options in d tells us that a climate policymaker could say: Because over the practically relevant set of policy options I am both attracted to totalist intuitions (or axioms), and I am fully comfortable with a generalized total social welfare function; and because this practically relevant set is bounded, I should make policy according to any of a, b, c, or d. I remain troubled by the Repugnant Conclusion, but that can be a problem for future research, because it (a)

g(n)

(b)

n

g(n)

(c)

g(n)

n

(d)

g(n)

n

Figure 5.4  Families of social evaluations that cohere with totalist axioms on the bounded set Note: Curly braces on the horizontal axis note the finite bounded set. 21  For a general discussion of the latter, see Arrhenius (forthcoming).

n

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 18/02/21, SPi

Responses to the Challenges of Population Ethics  127 does not threaten my conviction about how policy options should be ordered in the practically relevant set of policy options.

Of course, someone with less totalist intuitions, for example someone who leans more toward Average Utilitarianism, wouldn’t be able to say this. Likewise for theories that do not fall under the general characterization above, such as rank-­order theories and person affecting theories.22 Still, it shows that restricting the applicability of the axioms to bounded sets opens up for convergence on policy recommendations for a number of different theories.

4.2  Possibility Proof for Escaping the Repugnant Conclusion while Satisfying Bounded Versions of Population Ethics Desiderata The graphical examples of the prior section suggest a route to avoiding the Repugnant Conclusion. In this section, we prove that this is possible by adopting a plausible set of axioms: namely, bounded versions of familiar axioms. For example, in one of his pioneering informal results, Parfit (1984) makes use of the controversial (since it makes it easy to derive the Repugnant Conclusion) Mere Addition Principle: Mere Addition:  An addition of people with positive welfare does not make a ­population worse, other things being equal.23 This axiom could be weakened to: Bounded Mere Addition:  An addition of people with positive welfare does not make a population worse, other things being equal, if each population (with and without the addition) is within the bounded domain. One could similarly modify other adequacy condition axioms such as Arrhenius’ Non-­Sadism Condition to a Bounded Non-­Sadism Condition, and the Egalitarian Dominance Condition to a Bounded Egalitarian Dominance Condition. In each case, the modified axiom would reflect an analogous axiological intuition as the original axiom, but with the restriction that it only applies to comparisons of populations within the bounded set. Such bounded axioms would simply make no claims about ranking populations outside of the bounded set. Relatedly, but 22  For a discussion of the latter family, see Arrhenius (forthcoming) and (2000b). 23  See also Blackorby et al. (2005), Arrhenius (forthcoming) and (2000b). Like many contributors to the debate, Arrhenius and Blackorby et al. rejects the Mere Addition Principle as an adequacy condition for a satisfactory population axiology.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

128  Gustaf Arrhenius, Mark Budolfson, and Dean Spears outside of an axiomatic framework, one could assess the constructive argument that Broome (2004) presents for generalized, Critical-­Level Total Utilitarianism, but—unlike Broome—only assess and apply the argument while considering populations within the bounded set.24 Would such bounded axioms be intuitively compelling? Because they are logic­ al­ly weaker than their unbounded counterparts, they must be at least as compel­ ling. The impossibilities of population ethics are only interesting because the original axioms are compelling. Anyone who agrees with the original axioms will also agree with these, which are weaker: they make the same claims about fewer cases. And they may attract the new support of cautious evaluators who are hesi­ tant to make axiomatic claims about unbounded populations. In particular, consider a social evaluator who accepts the axiom of a complete and transitive social order for all populations, and accepts anonymity and same-­number Pareto for all populations, but then accepts only the Bounded Mere Addition and similarly modified and bounded versions of Separability and the other axioms that Blackorby and Donaldson (1984) demonstrate entail general­ ized Critical Level Total Utilitarianism. Such a set of axioms would entail a family of social welfare functions—each same-­number utilitarian—where g is increasing and linear over the bounded set, and could have any shape outside of the bounded set (perhaps disciplined by further continuity axioms). In particular, the resulting axiologies need not be separable outside of the bounded set. Such bounded ax­ioms would also rule out a positive critical level within the bounded set, due to Bounded Mere Addition. The modified axioms would provide a principled mo­tiv­ ation for the social evaluator to use this family of social welfare functions. Such an axiology would be sufficient for a climate IAM and to answer any question posed by climate ethics, and the Repugnant Conclusion is not entailed. More broadly, we now prove: Possibility Theorem for Bounded Axiologies:  There exist complete welfarist axi­ ologies that satisfy the Bounded Dominance, the Bounded Addition, and the Bounded Minimal Non-­Extreme Priority Principles and avoid the Repugnant, the Bounded Sadistic, and the Bounded Anti-­Egalitarian Conclusion. The proof is by example. Forms b and c from Figure 5.4 satisfy the theorem, as does any form of W in which h and f are the identity functions, g is the identity function on the bounded set (as in Total Utilitarianism), and g is everywhere non-­decreasing and is bounded above outside the bounded set. At very large

24  Of course, a more substantive axiology such as Critical-­Level Total Utilitarianism could still have unintuitive violations of other bounded conditions; for example, Critical-­ Level Total Utilitarianism violates a Bounded Non-­Sadism that modifies the Non-­Sadism axiom to only apply to the bounded set.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Responses to the Challenges of Population Ethics  129 population sizes outside of the bounded set, this family of axiologies would imply the (unbounded) Sadistic Conclusion, just as Ng’s Theory X’ does—but that is no contradiction, because the Possibility Theorem only requires avoiding the Sadistic Conclusion in the bounded set. Note that bounded Average Utilitarianism (g is constant in the bounded set) is not an example consistent with the Possibility Theorem because it does not satisfy avoiding even the Bounded Sadistic Conclusion; nor does Theory X’, if g is concave within the bounded set. A worry, however, is that the impossibility theorems might reappear over a bounded domain by further reformulating the adequacy conditions to take into account that we are now dealing with a bounded domain. Such reformulations can be done in multiple ways; one straightforward example is as follows: Bounded Repugnant Conclusion I: In the bounded domain, for any popu­la­tion consisting of people with very high positive welfare, there is a better popu­la­tion in which everyone has a very low positive welfare, other things being equal. Rather trivially, this cannot be an implication of axiologies that verify the Possibility Theorem above. Consider, for example, the largest population size within the bounded domain, and assume each member of that population has a very high welfare. Because this involves the largest population size within the domain, there cannot be a population with much lower welfare that is better. However, there are other reformulations of the Repugnant Conclusion that are not as easily avoided in the bounded domain. Here is one example: Bounded Repugnant Conclusion II:  In the bounded domain, there are very large populations consisting of people each with very high positive welfare for which there are better populations in which everyone has a very low positive welfare, other things being equal. The idea behind the Bounded Repugnant Conclusion II is the intuition that if a population is sufficiently big and everyone enjoys very high welfare, then such a population is better than each of the populations with only very low positive welfare in the domain. This intuition is one candidate for being the main intuition behind the counterintuitiveness of the original Repugnant Conclusion (recall that Parfit formulated it in terms of “any possible population of at least ten ­billion people”25). Along this line, it could be further argued that what is fundamental to repug­ nance is the existence of a Large Quantity-­Quality Trade-­off—meaning, a case where a large increase in quantity is allowed to compensate for a large decrease in

25  Parfit (1984: 388, emphasis added).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

130  Gustaf Arrhenius, Mark Budolfson, and Dean Spears quantity, or the reverse. According to this take on the Repugnant Conclusion, unboundedness is not essential to repugnance. This raises the important question of what is essential to the repugnance of the Repugnant Conclusion, and how many versions or instances there may be. As it is sometimes expressed, there can be various instances of the Repugnant Conclusion (Parfit 2016). If so, perhaps a satisfactory population axiology should not imply any instances of it. Depending on the size of the domain, the size of the very large populations, and on what the difference is between lives with very high and very low welfare, Bounded Total Utilitarianism might imply Bounded Repugnant Conclusion II. For example, let’s assume that a life with very high welfare is at least 100 times better than a life with very low positive welfare and let’s use Bostrom’s estimate, mentioned above, of 1058 simulated human lives as an upper bound on the size of possible populations. It follows from Bounded Total Utilitarianism that there is a very high welfare level such that for any population up to size 1056 enjoying this level, there is a better very low welfare population in the domain. So, according to Bounded Total Utilitarianism, a population with lives barely worth living would be better than an enormous population with very high individual quality of life. And given that an intuitively sufficiently large population with very high welfare is smaller than 1056, which seems intuitively compelling (compare Parfit’s specifi­ cation of “at least 10 billion people”), Bounded Total Utilitarianism implies the Bounded Repugnant Conclusion II in this domain. One can, of course, argue for other smaller upper bounds on the size of pos­sible populations and for other differences between very high and very low positive welfare lives. However, what this shows is that the unbounded scope of the classical Repugnant Conclusion is not needed to produce extreme quantity-­quality trade-­offs. More importantly, it shows that there may be impossibility theorems looming even in the bounded domain with the adequacy conditions from the unrestricted domain appropriately adjusted. Of course, this has to be appropriately shown by proving such theorems. The mere fact that some set of axioms is impossible to combine is not sufficient, of course, for an important challenge to climate policymaking. The involved ­conditions also have to be intuitively compelling. As the example above hints at, these conditions might or might not be sufficiently compelling depending on what one takes to be the main intuition behind classical unbounded conditions. Hence, the results we get when restricting population ethics to a bounded domain raises new and important questions that need to be further investigated: Is the implication of Bounded Repugnant Conclusion II sufficiently counterintuitive to work as an adequacy condition for a satisfactory population ethics? Might it even capture the main intuition behind the counterintuitiveness of the original Repugnant Conclusion? Or is unboundedness an essential part of the counterintuitiveness of the Repugnant Conclusion? More broadly, this result suggests asking why exactly the Repugnant Conclusion is counterintuitive. Is the quantity-­quality trade-­off involved in the Bounded

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Responses to the Challenges of Population Ethics  131 Repugnant Conclusion II sufficiently similar to a general quality-­quantity trade-­off problem for every aggregative axiology (see Budolfson and Spears 2018c, discussed above) to make it unsuitable as a condition on theory choice with respect to aggregative axiologies? Ultimately, we need to scrutinize more carefully the source of the counterintui­ tiveness of the original Repugnant Conclusion to know whether it will carry over to the bounded domain. Moreover, could the force of bounded impossibility the­orems be weakened by finding good reasons to restrict the upper bound on the domain further? And will the further assumptions that seem to be needed for bounded theorems, such as assumptions regarding the possible size of the involved populations, the difference between very high and very low positive wel­ fare, and the measurement of welfare (in the above example we assumed a ratio scale which isn’t necessary for the unbounded theorems) open up for ways of escaping the theorems that are not available in the unbounded domain? This is an important but neglected area of research in population ethics which the second deflationary response puts focus on.

5. Conclusion Policy analysis requires an axiology, population dynamics are important to ­climate change, and there is radical disagreement among experts about population axiology (Arrhenius forthcoming, 2000a, 2000b, 2001, 2011). Does this state of affairs limit our ability to know how to respond to climate change? Although several prominent voices have voiced this Worry, we suggested that it is not obviously well-­ founded, and we have highlighted two possible deflationary responses. In the first, we noted that many important policy questions are likely to be subject to simple, cross-­theoretical dominance resolutions, as illustrated by a corner solution to an optimization problem. In the second deflationary response, we observed that the intuitions that support the axioms that lead to the Repugnant Conclusion also support the axioms in the bounded case while avoiding the Repugnant Conclusion. Because any real-­world policy question is a question about a bounded population domain (even if potentially very large in quantity), we can adopt these axioms for purposes of policy in their modified bounded form. We also noted some important limitations and possible problems for these deflationary strategies. Regarding the first deflationary response, we noted that the climate policy menu under consideration may not yield one dominating option. Moreover, there could be additional considerations, such as bounded political capital, which could complicate the issue such that it cannot be settled by the suggested dominance-­identification procedure, or could simplify the issue by further reducing the practical space of policy options to those in which many axiologies agree.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

132  Gustaf Arrhenius, Mark Budolfson, and Dean Spears Regarding the second deflationary response, there is the worry that the impossibility theorems might reappear over a bounded domain when the classical ad­equacy conditions are appropriately adjusted for the bounded domain. An important challenge highlighted by considering the Repugnant Conclusion on a bounded domain is the need to identify exactly what constitutes the main coun­ terintuitiveness of the Repugnant Conclusion and whether it carries over from the unbounded to the bounded domain (or, perhaps, to any other domains). This is a neglected but important area for further research in light of the impossibility theorems in population axiology on unbounded domains and the possibility the­orem above on bounded domains. In the meantime, we need not overstate the practical importance of the Repugnant Conclusion and other challenging problems in population ethics as we seek to cope with important challenges for the future of humanity. As we have shown, skepticism and paralysis are not yet warranted, as there are promising deflationary responses to the impossibility theorems and strategies for gaining consensus given disagreement for practical policymaking. Policy analysis may not need to wait for greater consensus in population ethics.26

Appendix: A Smoothness Axiom and a New Argument for Total Utilitarianism One response to the argument in Section 4 of the chapter would be to agree that the modified axioms in their bounded versions capture some of our important intuitions, but not all of them, because there is a specific intuition that is omit­ ted: that axiology is infinitely continuous. Consider the case in which a family of axiology is chosen, based on axioms some bounded and some unbounded, such that a social welfare function of form W is chosen, with the additional properties that: • Bounded separability is assumed in social evaluation, so that the social ˆ = g (n )h(x ), ­welfare function can be written as a function of two variables: W where n is the expected size of the population and x is the expectation of f (x ) . Then, g and the other functions are functions of all real numbers (not just counting numbers). • f and h are both identity functions, as in Total or Average Utilitarianism or ˆ = g (n ) x , where x is average Theory X’, so the expression simplifies to: W utility. 26  Thanks to Andrea Asker, Drew Burd, Krister Bykvist, Tim Campbell, Diane Coffey, Iwao Hirose, Gerald Lang, Melissa LoPalo, Kevin Kuruc, Tristram McPherson, Josh Petersen, Shlomi Segall, Sangita Vyas, and the audiences at Paris School of Economics, the Australian National University, and the Institute for Futures Studies.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Responses to the Challenges of Population Ethics  133 • g is the identity function on the bounded set, as in Total Utilitarianism, and is any non-­decreasing function outside of the bounded set, so the Repugnant Conclusion is not logically entailed (and therefore may or may not be avoided). This is the sort of family of social welfare functions that Section 4 highlights as possible, but extended for illustration to the case of expectations, in order to cover real numbers (and not only counting numbers of people); this will not appeal to advocates of non-­expected social evaluations. Now consider the intuition that axiology should be infinitely continuous—an intuition that may appear as an experience of unease about the boundedness of axioms. We can formalize this axiom as: Smoothness:  g is C∞, which is mathematical notation for the property of a func­ tion in which each derivative is continuous everywhere. For real-­valued functions, the Smoothness axiom would imply that they are poly­ nomials. Therefore, g must be the identity function everywhere, because it is the identity function in the bounded set. The upshot is that the bounded assumptions ˆ is expected Total Utilitarianism.27 above plus the Smoothness axiom imply that W The Smoothness axiom—and the intuitive response to the boundedness proposal that it captures—is therefore a new, constructive argument for Total Utilitarianism. ˆ implies the Repugnant Conclusion. Therefore, With the smoothness axiom, W the Smoothness axiom introduces a new theoretical cost of avoiding the ˆ If you find Repugnant Conclusion, in the context of the bounded axioms of W. boundedness distasteful because you find infinite continuity to be a plausibly compelling property of axiology, then that intuition—in combination with other axioms—is a new argument counting in favor of Total Utilitarianism and acceptance of the Repugnant Conclusion. Of course, it can also be taken as a new impossibility theorem for those who accept smoothness, the bounded assumptions above, but not the Repugnant Conclusion.

References Anglin, W. (1977). “The Repugnant Conclusion.” Canadian Journal of Philosophy 7(4): 745–754. Arrhenius, G. (2000a). “An Impossibility Theorem for Welfarist Axiologies.” Economics and Philosophy 16(02): 247–266. Arrhenius, G. (2000b). Future Generations: A Challenge for Moral Theory. http://www. diva-portal.org/smash/record.jsf?pid=diva2:170236. 27  Thanks to Kevin Kuruc for suggesting consideration of this argument.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

134  Gustaf Arrhenius, Mark Budolfson, and Dean Spears Arrhenius, G. (2001). “What Österberg’s Population Theory Has in Common With Plato’s,” in Omnium-gatherum. Philosophical Essays Dedicated to Jan Österberg on the Occasion of his Sixtieth Birthday (Vol. 50). Uppsala: Department of Philosophy, Uppsala University: Uppsala Philosophical Studies, 29–44. Arrhenius, G. (2011). “The Impossibility of a Satisfactory Population Ethics,” in Descriptive and Normative Approaches to Human Behavior, ed. E.  Dzhafarov and L. Perry. Singapore: World Scientific Publishing, 1–26. Arrhenius, G. (Forthcoming). Population Ethics: The Challenge of Future Generations. New York: Oxford University Press. Arrhenius, G., J.  Ryberg, and T.  Tännsjö (2014). “The Repugnant Conclusion.” The Stanford Encyclopedia of Philosophy, ed. E.  N.  Zalta. http://plato.stanford.edu/ archives/spr2014/entries/repugnant-conclusion/. Asheim, G.  B. and S.  Zuber (2014). “Escaping the Repugnant Conclusion: Rank Discounted Utilitarianism with Variable Population.” Theoretical Economics 9(3): 629–650. Beckstead, N. (2013). “On the Overwhelming Importance of Shaping the Far Future.” PhD dissertation, Rutgers University, New Brunswick, NJ. Blackorby, C., W.  Bossert, and D.  Donaldson (1995). “Intertemporal Population Ethics: Critical-Level Utilitarian Principles.” Econometrica 63(6): 1303–1320. Blackorby, C., W. Bossert, and D. Donaldson (2005). Population Issues in Social Choice Theory, Welfare Economics, and Ethics. New York: Cambridge University Press. Blackorby, C. and D.  Donaldson (1984). “Social Criteria for Evaluating Population Change.” Journal of Public Economics 25(1–2): 13–33. Bostrom, N. (2013). “Existential Risk Prevention as Global Priority.” Global Policy 4(1): 15–31. Broad, C. D. (1979). Five Types of Ethical Theory. London: Routledge. Broome, J. (1992). Counting the Cost of Global Warming. Cambridge: The White Horse Press. Broome, J. (2004). Weighing Lives. Oxford: Oxford University Press. Broome, J. (2010). “The Most Important Thing about Climate Change,” in Public Policy: Why Ethics Matters, ed. J. Boston, A. Bradstock, and D. Eng. Canberra: ANU E Press, 101–116. Broome, J. (2012). Climate Matters: Ethics in a Warming World. New York: W. W. Norton. Budolfson, M. and D.  Spears (2018a). An Impossibility Result for Decision-Making under Normative Uncertainty. Mimeo. Budolfson, M. and D. Spears (2018b). Methods for Quantifying Animal Wellbeing and Estimating Optimal Tradeoffs against Human Wellbeing—and Lessons for Axiology, Including New Arguments for Separability. Mimeo. Budolfson, M. and D. Spears (2018c). Why the Repugnant Conclusion is Inescapable. Mimeo. Online at http://www.budolfson.com/papers/RC.pdf

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

Responses to the Challenges of Population Ethics  135 Bykvist, K. (2017). “Moral Uncertainty.” Philosophy Compass 12(3): 1–8. Bykvist, K., W.  MacAskill, and T.  Ord (2019). Moral Uncertainty. Oxford: Oxford University Press. Greaves, H. (2017). “Population Axiology.” Philosophy Compass 12(11): 1–15. Greaves, H. and T.  Ord (2017). “Moral Uncertainty about Population Axiology.” Journal of Ethics and Social Philosophy 12(2). Gustafsson, J. E. (Forthcoming). Our Intuitive Grasp of the Repugnant Conclusion. In The Oxford Handbook of Population Ethics, ed. G. Arrhenius, K. Bykvist, T. Campbell, and E. Finneron-Burns. Oxford: Oxford University Press. Hare, R. M. (1988). “Possible People.” Bioethics 2(4): 279–293. Hedden, B. (2016). “Does MITE Make Right? On Decision-Making under Normative Uncertainty,” in Oxford Studies in Metaethics (Vol. 11), ed. R.  Shafer-Landau. Oxford: Oxford University Press, 102–135. Hsiung, W. and C. R. Sunstein (2006). “Climate Change and Animals.” University of Pennsylvania Law Review 155(6): 1695–1740. Huemer, M. (2008). “In Defence of Repugnance.” Mind 117(468): 899–933. Mackie, J. L. (1985). “Parfit’s Population Paradox,” in Persons and Values, ed. J. Mackie and P. Mackie. Oxford: Oxford University Press, 242–248. McMahan, J. (1981). “Review: Problems of Population Theory.” Ethics 92(1): 96–127. McTaggart, J.  M.  E. (1927). The Nature of Existence. Cambridge: Cambridge University Press. Narveson, J. (1967). “Utilitarianism and New Generations.” Mind 76(301): 62–72. Ng, Y.-K. (1989). “What Should We Do about Future Generations?” Economics and Philosophy 5(02): 235–253. Pachauri, R.  K., L.  Mayer, and Intergovernmental Panel on Climate Change (eds.) (2015). Climate Change 2014: Synthesis Report. Geneva: IPCC. Parfit, D. (1982). “Future Generations: Further Problems.” Philosophy & Public Affairs 11(02): 113–172. Parfit, D. (1984). Reasons and Persons. Oxford: Clarendon Press. Parfit, D. (1986). “Overpopulation and the Quality of Life,” in Applied Ethics, ed. P. Singer. Oxford and New York: Oxford University Press, 145–164. Parfit, D. (2016). “Can We Avoid the Repugnant Conclusion?” Theoria 82: 110–127. Scovronick, N., M. B. Budolfson, F. Dennig, M. Fleurbaey, A. Siebert, R. H. Socolow, D.  Spears, and F.  Wagner (2017). “Impact of Population Growth and Population Ethics on Climate Change Mitigation Policy.” PNAS 114(46): 12338–12343. Shiell, L. (2008). “The Repugnant Conclusion and Utilitarianism under Domain Restriction.” Journal of Public Economic Theory 10(6): 1011–1031. Sider, T. R. (1991). “Might Theory X Be a Theory of Diminishing Marginal Value?” Analysis 51(4): 265–271.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 11/02/21, SPi

136  Gustaf Arrhenius, Mark Budolfson, and Dean Spears Sidgwick, H. (1907). The Methods of Ethics. London: Macmillan. Spears, D. (2019). “The Asymmetry of Population Ethics: Experimental Social Choice and Dual-Process Moral Reasoning.” Economics and Philosophy. Forthcoming. Tännsjö, T. (2002). “Why We Ought to Accept the Repugnant Conclusion.” Utilitas 14(03): 339–359.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

SECTION II

C O G N IT ION , E MOT IONS , A ND C L IMAT E C HA NG E

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

6

Way to Go, Me Chrisoula Andreou

1. Introduction The title of this chapter is based on a blog post by the same name that I wrote quite some time ago (Andreou 2010). The post was inspired by the kudos conveyed on a disposable Starbucks coffee cup, which proved significantly more quote-­worthy than your standard disposable cup. I was struck by the slogan on the container (“Way to go, you”) and the accompanying ‘explanation’ because their appearance on a disposable cup suggested an intriguing asymmetry in human psychology, one that seems relevant to understanding and impacting behavior in the face of climate change. Though tongue-­in-­cheek, the short post captures the ideas I want to build on in this chapter, and so I begin by re­pro­du­ cing the body of the post here: In a recent encounter with a $4 mocha latte in a disposable cup, I was quite pleased to find that I received high praise (“Way to go, you”) for “stop[ping] by for a coffee.” “Just by doing that,” I contributed to a cause that’s “bigger than coffee,” namely “responsibly grown, ethically traded coffee.” Though my purchase might on the surface seem questionable, or at best negligible, looked at from a certain perspective (and with a bit of squinting) it becomes clear that I, in partnership with others, am making it possible for a certain coffee company we all know and love to get big enough to use its “size for good” by “working with farmers” to get [a higher percentage of] responsibly grown, ethically traded coffee (and so, I guess, not quite as much irresponsibly grown or unethically traded coffee).  In addition to the satisfaction I got from this pat on the back, it was also philosophically satisfying, given my research interests, to see the accumulation of negligible effects getting some attention. Just think of all the disposable cups circulating, spreading the message of how a little from each of us can really add up to a great deal. So inspiring, especially because it’s all about ME (and teamwork of course, but mostly ME).  Now if only I could put aside the logic of negligible effects in dealing with my disposable cup, the praise for my stewardship of our “shared planet” could go down without too much salt. As things stand, my pioneering stewardship leaves Chrisoula Andreou, Way to Go, Me In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Chrisoula Andreou. DOI: 10.1093/oso/9780198796282.003.0007

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

140  Chrisoula Andreou me so concerned about the disposal of the cup, I can’t quite seem to get rid of it. Maybe I’ll keep it as a Warholesque objet d’art.1 

Of course, I still have the cup and still wonder whether the folks who crafted the slogan and the accompanying ‘explanation’ noticed the seemingly tricky nature of the situation: the consumer, who is credited with caring about our “shared planet” and being concerned with her contribution to “bigger” things, is to pat herself on the back for buying the coffee without feeling bad about using a disposable cup. Perhaps extensive market research suggested they could pull this off; or perhaps the trickiness of the situation went altogether unnoticed, which itself bodes well for the feel-­good-­about-­buying-­our-­coffee campaign working. Either way, the situation is interesting and suggestive. My aim in this chapter is to explore a mindset that I suspect the campaign taps into and to consider the implications of this mindset in relation to motivating environmentally friendly behavior in the face of “creeping environmental problems.”2 In the end, whether or not the campaign taps into the mindset that I will sketch out, there is good reason to believe that the mindset exists, and that it is relevant to creeping environmental problems. Relatedly, although the springboard for my reasoning is the campaign, I will not embark on a close examination of the campaign, or of mostly-­responsibly-­grown coffee, or of disposable cups; my real interest is in using the provocative example of the campaign to illuminate more general issues associated with the perceived significance of individually trivial contributions to momentous outcomes.

2.  Climate Change as a Creeping Environmental Problem I take it that climate change can be helpfully modeled as a creeping environmental problem. By a creeping environmental problem, I mean an environmental problem wherein a good portion of the serious damage at issue is realized via a series of individually trivial contributions. The problem of (anthropogenic) climate change involves a huge number of people contributing via a huge number of actions (or omissions). A single action, or even the set of actions of a single individual, is generally too small to have a significant impact on the problem. This structural feature can prompt problematic preferences, wherein there is a widespread general preference that we show sufficient restraint for major damage to be 1  For a much less feel-­good slogan (and story) relating to our “shared planet” that figures, in a way that I hope will become apparent, as the flip-­side of Starbucks’ “Way to go, you” slogan, consider John D. Sutter’s “You’re making this island disappear,” which is the headline for his piece on climate change and, more specifically, on the “forecast” that “the Marshall Islands likely won’t exist if we warm the planet 2 degrees” (Sutter  2015). My thanks to Ewan Kingston for bringing the headline to my attention. 2  I borrow this term from Glantz (1999). I say a little more about problems of the relevant sort in the next section.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Way to Go, Me  141 avoided, but there is no local preference for, say, me to show restraint now because no individual instance of showing restraint will make a significant difference.3 In maintaining that climate change can be helpfully modeled as a creeping environmental problem, I am committed to the view that it is fruitful to think about the problem of climate change in terms of the structural features and preference combinations I have been describing; indeed, for the purposes of this chapter I will be focusing almost exclusively on these elements. But I do not mean to suggest that other ways of thinking about the problem should, quite generally, be put aside. To the contrary, I assume that it is also helpful to model climate change as, for example, a cost distribution problem, an intergenerational problem, and a Prisoner’s Dilemma situation, where each of these models emphasizes some conflict of interest that can interfere with environmentally friendly behavior.4 I see these models as isolating and focusing on different strands of the problem, and so as complementary rather than competing. In assuming that climate change is a creeping environmental problem, I will put aside skepticism about the proposed dire effects of climate change and accept the currently prevailing view that climate change is very much worth worrying about. I will also assume that the large-­scale production and use of modern conveniences, from little things like disposable cups to big things like fuel-­powered private vehicles, are contributing to the problem, and that people recognize this. (It might be suggested that using disposable cups is not too wasteful if they are recycled. But I have yet to encounter someone who stocks their kitchen with disposable cups with the idea that recycling is so energy-­efficient, it’s the use of nondisposable cups that’s a problem. I am not denying that some such individuals may exist; but, if they do exist, they seem to be the exception rather than the rule.) As will become apparent, although the position I will develop is all about rationalizing, it does not require individuals to, even temporarily, embrace any ­idiosyncratic views regarding what is wasteful and what is not. In any case, the mindset I will explore—and the substantial variations in the perceived significance of individually trivial contributions that the mindset can prompt—will, I  hope, be recognized as clearly relevant in relation to creeping environmental problems regardless of the room there is for tangential debate about the details of specific illustrations.

3  For extensive discussion regarding the connection between individually trivial effects, intransitive preferences, and creeping environmental problems, see Andreou (2006, 2007). For objections to the idea that, in some cases of collective harm, no individual instance of showing restraint will make a significant difference, see, for example, Kagan (2011); for a forceful reply, see Nefksy (2012) and, relatedly, Budolfson (2019). All the same issues arise in relation to the idea that, in some cases of collective harm, no single individual’s contributions (even over the individual’s whole lifetime) make a significant difference. Some additional relevant works are cited in note 5. 4  For interesting discussion concerning these models and more complicated syntheses and vari­ ations on them, see Gardiner (2002).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

142  Chrisoula Andreou

3.  Different Orientations Now suppose someone faced with the creeping environmental problem of climate change reflects on the abundance of unnecessary car drives or disposable cups in her life. It might seem like one’s best bet, in terms of predicting and understanding how the individual will view her behavior, is to look for clues about whether she is individualistic or group-­oriented: if she is individualistic, in that her focus is on what she effects, her take on her behavior will naturally be something like the following: “As anticipated, nothing I have been doing has made any significant difference, so I’m not at fault”;5 if, on the other hand, she is group-­oriented, and so her focus is on what the group she is part of effects (via the use of conveniences like the ones under consideration), her take will naturally be something like the following: “We’re messing things up (in a way that is causing serious harm) and I’m doing my part, so I’m at fault.”6 Relatedly, it might seem like an agent with an affinity toward thinking something like “As anticipated, nothing I have been doing has made any significant difference, so I’m not at fault” in cases involving a negative outcome will also have an affinity toward thinking something like “As anticipated, nothing I have been doing has made any significant difference, so I don’t deserve any credit” in cases involving a positive outcome; similarly, an agent with an affinity toward thinking something like “We’re messing things up and I’m doing my part, so I’m at fault” in cases involving a negative outcome will also have an affinity toward thinking something like “We’re doing great things and I’m doing my part, so I deserve some credit” in cases involving a positive outcome. Return now to Starbucks’ kudos-­bearing cup. If one’s best bet, in terms of predicting and understanding how an individual will view her behavior in cases involving individually trivial contributions, is to look for clues about whether she is individualistic or group-­oriented, Starbucks’ feel-­good-­about-­buying-­ourcoffee campaign would, it seems, backfire, even among individuals who are concerned about our “shared planet” and the various ways in which it is subject to anthropogenic damage. For, although group-­oriented individuals, who focus on and assess their individually trivial actions in terms of group-­level effects, would 5 Note that, although some philosophers maintain that individual contributions to large-­scale creeping environmental problems do make a significant difference (see, for example, Broome 2012: ch. 5; Hiller 2011; and Nolt 2011), all seem to agree that this position deviates from common thinking, which steers those who think individualistically away from finding fault with their contributions to creeping environmental problems. Notably, if one should adopt a group-­oriented stance and think in terms of what we together are effecting, then refusing to fault oneself can be wrong-­headed even if, as I here grant, common thinking is right in taking at face value the idea that, with respect to many creeping environmental problems, the contributions of a single individual (even over the individual’s whole life) do not make a significant difference. See, relatedly, note 3 above. 6  For interesting discussion regarding these two takes, see Sinnott-­Armstrong (2005).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Way to Go, Me  143 be ready to pat themselves on the back for contributing to the cause of r­ e­spon­sibly grown, ethically traded coffee (at least if they are not thwarted from doing so by the fact that the coffee is only mostly responsibly grown and ethically traded), they would also be ready to feel bad about using the disposable cup the feel-­goodmessage is printed on (since the production and use of such conveniences is implicated in a harm the group effects). And, although individualistic individuals would not be prone to feeling bad about using the disposable cup the feel-­goodmessage is printed on (because using the cup has only trivial effects), they would not be in a position to swallow the feel-­good-­message itself (because it prompts the individual to focus on group-­level effects). This ‘worry’ about the campaign backfiring is, however, predicated on arguably unrealistic assumptions regarding individual consistency. The point is not simply that we are less than completely consistent (at least in part because the attentional requirements are beyond us), but that we are systematically inconsistent in ways that facilitate the campaign’s robustness.

4.  Switching Between Orientations and Mindset M Consider the following possibility. Perhaps one’s best bet, in terms of predicting and understanding an individual’s take on a particular matter (or on a particular feature of his situation) in cases involving individually trivial contributions, is not to look for clues about whether the individual is individualistic or group-­oriented, but instead to figure out whether an individualistic take or a group-­oriented take will facilitate his seeing himself as praiseworthy, or at least not blameworthy: if an individualistic take (on a particular matter) will facilitate self-­praise, or avoid selfblame, then, other things equal, the individual will think individualistically (about that matter); if a group-­oriented take will facilitate self-­praise, or avoid self-­blame, then, other things equal, the individual will think in terms of group accomplishments (which can be positive or negative). This mindset, henceforth mindset M or just M, figures as a perfect target for the feel-­good-­about-­buyingour-­coffee campaign. For it allows the agent to feel good about his individually trivial contributions toward positive outcomes (e.g. an improved coffee industry) while allowing the agent to avoid feeling bad about his individually trivial contributions toward negative outcomes (e.g. the large-­scale production and use of wasteful conveniences). (More on this balancing act below.) Might we actually have mindset M? I turn now to some research results that seem to support the possibility. First, there is a great deal of empirical research suggesting that, far from being super-­principled, an individual’s take on a situation (or on some particular feature of a situation) can be influenced by “the motivation to sustain or enhance

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

144  Chrisoula Andreou [his] sense of self-­worth” (Shepperd et al.  2008).7 Consider, for example, that ­people often make “internal attributions for desired outcomes and external attributions for undesired outcomes,” particularly when the outcomes “have implications for self-­worth” (2008: 895, 897). Given that internal attributions appeal to “abilities, skills, personal traits, or effort” whereas external attributions appeal to “the actions or inactions of others, luck, and circumstances such as the weather or economy,” this attribution bias is clearly self-­serving (2008: 895). Importantly, research suggests that the self-­enhancement motivation is cross-­cultural, with cultural differences affecting the manifestation of the motivation: “people from Western cultures are particularly likely to show self-­enhancement on traits and behaviors that are valued within individualistic cultures,” while “people from Eastern cultures are particularly likely to show self-­enhancement on traits and behaviors that are valued within collectivistic cultures” (2008: 897). Similarly, the manifestation of the self-­enhancement motivation can vary depending on other features of an individual’s mental state: for example, “people with low social anx­ iety have an acquisitive style directed toward garnering approval and enhancing identity. By contrast, people with high social anxiety have a cautious, protective style directed toward avoiding social disapproval and protecting identity” (2008: 898). Whatever the mediating factors, the self-­enhancement motivation can, in the face of multiple possible credible takes on a situation, pull one’s attention and allegiance to the most flattering one. It can thus prompt either an individualistic take or a group-­oriented take, depending on which is more flattering. This is consistent with individuals being responsive to various socializing influences. Notice, relatedly, that, despite the characterization of some cultures as individualistic and others as collectivist, things are fairly complicated. Even Westerners, with their famed proclivity to individualism, are willing to sometimes pat themselves on the back for collective achievements; and, even Easterners, with their famed selfesteem-­implicating commitment to collectivism, are open to skewing the causes of their personal successes and failures, particularly when their attributions are made privately, and so do not run the risk of garnering blame for violating the value of modesty (2008: 897, 898). Second, there is a significant amount of empirical research suggesting that, far from being super-­principled, individuals are prepared to ‘spin’ the way they see or describe things when blame is at stake. For example, rather than having a principled stance regarding whether or not the foreseen side effect of an intended act counts as intentionally realized, individuals are more likely to describe foreseen side effects that an agent is clearly indifferent toward as intentionally realized when

7  The quotes in this paragraph are all from Shepperd et al. (2008). The article helpfully surveys recent research and meta-­research on the self-­serving bias, which is at least in part due to the self-­ enhancement motivation of interest here. Another helpful discussion concerning the main strands of research on the self-­enhancement motivation can be found in Sedikides and Gregg (2007).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Way to Go, Me  145 the effects are negative and the agent is seen as blameworthy.8 The mo­tiv­ation to blame can, it seems, at least temporarily draw one from a narrow conception of intentional action, according to which foreseen side effects do not count as intentionally realized, to a broader one, according to which they do. This result fits neatly with the idea that, when there are multiple possible credible takes on a situation, the motivation to blame can impact which take prevails. This adds to the case for thinking that the same is true of the motivation to avoid blame, including self-­blame. According to a particularly interesting related result, the motivation to ease or offset self-­blame after failing a moral test can impact whether a situation is construed as a moral opportunity, i.e., an opportunity to go beyond what is morally required and earn (redeeming) self-­praise. For example, the motivation of “transgressors” to redeem their temporarily diminished moral self-­evaluations can lead them to imbue requests of dubious merit with redemptive moral potential. One example of this is Brock and Becker’s (1966) finding that Ohio State University (OSU) undergraduates induced to break a piece of experimental equipment were more likely than their “nontransgressing” peers to subsequently sign a petition advocating an increase in OSU tuition. It is difficult to imagine that refusing to support the tuition hike would be seen as a failed moral test to an undergraduate (indeed, no students in the control condition signed the petition). It may also seem surprising that undergraduates seeking to enhance their threatened moral image would see any great redemptive potential in this act, but apparently they were able to convince themselves that it was at least above the bar on this score. (Miller and Monin 2016: 46)

This result suggests that, when self-­blame is at stake, it does not take much for interpretations that evade or repair diminished self-­evaluations to count as cred­ ible enough.

8  This description of the relevant empirical results, which can be found in Knobe (2003a), is based on the common idea (whose defense is beyond the scope of this chapter) that whether or not the foreseen side effects of an intended action count as intentionally realized cannot vary depending on whether the side effect is negative, positive, or neutral, because a proper analysis of the concept “intentional action” precludes such variation. Following Knobe (2003b), I allow, at least for the sake of argument, that there are “two way of thinking about the relation between concepts and people’s use of words in ordinary language” (2003b: 314). According to one way of thinking, “when we offer an account of, e.g., the concept of intentional action, we are offering a standard against which people’s ordinary utterances can be judged. If people do not actually use the words ‘intentional action’ in the way specified by the account, we might conclude that they are speaking incorrectly” (2003b: 314). According to another way of thinking, “the concept just is whatever people have grasped” (2003b: 314). Knobe is concerned with elucidating the concept people have grasped; my interest here is on how that concept deviates from more prescriptive analyses of the concept.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

146  Chrisoula Andreou In light of the empirical research just reviewed, it seems quite plausible to suppose that mindset M exists; and, given mindset M, one can easily occupy the sweet spot Starbucks’ feel-­good-­about-­buying-­our-­coffee campaign invites one to occupy. It might be suggested that this appeal to mindset M requires more ‘mental acrobatics’ than it is realistic to postulate. Can an agent really have and sustain an individualistic take with regard to her individually trivial contributions to negative outcomes (accepting stances like “My using this cup is not making any significant difference, so I’m not at fault”), while having and sustaining a group-­oriented take with regard to her individually trivial contributions to positive outcomes (accepting stances like “We’re supporting responsibly grown coffee and I’m doing my part, so I deserve some credit”)? While humans are capable of quite a bit of mental acrobatics, it might seem like maintaining this combination is too amazing a feat to expect of us given the conflicting positions involved. Notice, however, that maintaining two attitudes that are rooted in conflicting views is not at all hard to do, particularly when the following hold: (1) the views are each fairly credible (as is the case with individualistic and group-­oriented takes regarding individually trivial contributions), and so each view is easy to accept when it is prompted by a self-­enhancement motivation; and (2) the views do not figure as explicitly articulated occurrent thoughts and so are not readily available for consistency comparisons. (Presumably, an agent can accept that X even if she is not occurrently thinking that X, and even if X figures as something less than a firm, easily accessible belief, such as, say, the unarticulated, unprocessed idea or representation behind a certain positive or negative feeling.)

5.  Seeking Self-­Praise versus Avoiding Self-­Blame I have been indulging in some speculation about the Starbucks feel-­good-­aboutbuying-­our-­coffee campaign. My aim, however, is not to arrive at some solid conclusion concerning the campaign (or to knock Starbucks); rather, my aim is simply to use the campaign to flesh out a possibility and to consider the implications of that possibility, particularly in relation to situations in which individuals are not subject to external monitoring and constraints but are instead acting privately and guided by their own assessments of their behavior. The possibility, identified above, is that one’s best bet, in terms of predicting and understanding an individual’s take on a particular matter in cases involving individually trivial contributions, is not to look for clues about whether the individual is individualistic or group-­oriented, but instead to figure out whether an individualistic take or a group-­oriented take will facilitate the individual’s seeing himself as praiseworthy, or at least not blameworthy, with respect to the matter at issue: insofar as one of the takes is better suited to facilitating self-­praise, or avoiding self-­blame, with respect to the behavior at hand, it will, other things equal, be adopted.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Way to Go, Me  147 Interestingly, if this mindset is in play, then activating the strong motivation that individuals have to see themselves as praiseworthy can have very different consequences than activating the strong motivation that individuals have to avoid seeing themselves as blameworthy. Relatedly, activating the former motivation can have better results even if (given our loss aversion) the latter motivation is stronger.9 To see why, notice that, given mindset M and a creeping environmental problem, the presence of a strong motivation to see oneself as praiseworthy will, other things equal, prompt one to buy into the significance of doing one’s part, in accordance with a group-­oriented take, and to act accordingly. By contrast, given mindset M and a creeping environmental problem, the presence of a strong mo­tiv­ation to avoid seeing oneself as blameworthy will, other things equal, prompt one to deny the significance of one’s contributions to the problem, in accordance with an individualistic take, and to act accordingly. Because M can prompt one to focus on what one alone is effecting, it provides a system for avoiding self-­blame that does not require convincing even oneself that one is doing one’s part. So if, for example, one is about to take an unnecessary drive, and one is guided by mindset M, then, other things equal, one will not be hampered from taking the drive by the threat of self-­blame, since M allows one to avoid that threat even if one takes the drive. Still, one may be lured (or perhaps just nudged) away from the drive by the promise of self-­praise for doing one’s part. Of course, if the appeal of the drive is sufficiently great, the promise of self-­praise for doing one’s part may be no match for it, particularly if M can be relied on to keep selfblame off the table. Fortunately, there is lots of room for the promise of self-­praise to make a significant difference when the immediate cost of doing one’s part is small and there is little risk of having to incur a large cost for doing one’s part without enough others doing their part. Consider, for example, a scenario in which the immediate cost of voting for an environmentally friendly measure is low and there are no substantial costs that one will have to incur unless the measure is approved and one can count on enough others doing their part as well. Someone who is at least somewhat concerned about our “shared planet” may find the promise of selfpraise associated with voting for the environmentally friendly measure substantial enough to prompt her to buy into the significance of doing her part and voting (even if each vote is individually trivial); she may also have enough in the way of incentives to act in accordance with the measure if it is approved and enforced; moreover, this can be so even if she would not, in the absence of mutually agreed upon constraint, find the promise of self-­praise associated with taking personally costly actions—particularly ones that are individually trivial in terms of addressing the environmental problem at stake—substantial enough to prompt

9  For seminal results on loss aversion, see Kahneman and Tversky (1984).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

148  Chrisoula Andreou her to constrain her behavior accordingly, and so would, other things equal, be pushed by mindset M to think individualistically and deny the significance of her behavior, conveniently fending off any self-­blame. I have been qualifying many of my claims with “other things equal” clauses because, even if mindset M figures in our psychology, it can, I want to allow, be interfered with or overwhelmed by other forces.10 My goal in isolating M is not to suggest that it is the most crucial determinant of our behavior when it comes to creeping environmental problems, but that it is arguably a force to be reckoned with, and that reckoning with it requires understanding how it interacts with our motivations to seek self-­praise and avoid self-­blame. Such understanding can help us figure out when framing a contribution to an environmentally friendly measure in a way that invites self-­praise for participation might be more effective than framing the contribution in a way that invites self-­blame for non-­participation, where the former might figure as a small but sufficiently enticing carrot and latter might figure as a large stick that fails to motivate because it never gets off the ground.

6. Conclusion It might be noted that I have not raised or addressed the question of whether the individualistic stance or the group-­oriented stance regarding individually trivial contributions is correct. I will not delve into my thoughts on the matter here. I do, however, want to entertain the possibility that each stance contains some grain of truth and that the philosophically correct position is a highly nuanced one that reveals the complicated reality that makes each of these stances credible. If this is right, then it is worth noting that, while inconsistency is obviously less than ideal, wavering between two credible but inconsistent stances in response to different cues and contexts is not necessarily practically or epistemically worse than adopting a consistent but oversimplified stance. Given limited cognitive resources, the best strategy, in terms of doing as much justice as we feasibly can to what we care about and to the complicated reality we face, may not be to attempt to eliminate inconsistency, but to work on prompting different thoughts with different grains of truth in different contexts in a way that allows us to adequately achieve our goals even if we cannot expect to arrive at and act on the basis of the whole truth

10  There is an extensive and interesting literature regarding motives that compete with the motive of self-­enhancement, including, for example, the motive of self-­knowledge, which reinforces the plausi­bil­ity constraints on self-­enhancing construals. More controversial is the purported motive of self-­verification, which involves the desire to confirm one’s already existing views about oneself, even if they are negative, but which might be just another manifestation of the plausibility constraints on self-­enhancement. For a sense of the state of play, see Sedikides and Gregg (2007).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Way to Go, Me  149 about individually trivial contributions.11 The sort of ‘happy compromise’ that I have in mind is perhaps easiest to illustrate via familiar intrapersonal cases involving individually trivial contributions to important goals. Consider, for example, the challenges of saving enough for retirement or maintaining a healthy body weight, both of which are normally achieved via a series of individually trivial goal-­directed actions or omissions. With respect to either of these goals, any manageable personal rule (which avoids both endless defeasibility clauses and extreme vagueness) is not likely to be sufficiently nuanced to provide guidance that avoids over-­rigidity on the one hand and over-­permissiveness on the other (Ainslie  1999); as such, staying more or less on track may involve a habitually-­reinforced and/or cue-­responsive mix of lapse-­permitting and lapseprohibiting thoughts that each have some grain of truth but are not easy to square with one another—thoughts like “This won’t make any significant difference, so there’s no problem with doing what I’m inclined to do” or, alternatively, “There’s no chance of success if restraint is not shown in cases of this sort, so I’d better show restraint.” This suggests that, given limited cognitive resources, some inconsistency can facilitate, rather than interfering with our being sufficiently accountable to the complexities of achieving goals that must be realized over time and/or via the actions of multiple individuals. At least part of the feat, in both intrapersonal and interpersonal cases, is to cultivate mixes of lapse-­permitting and lapseprohibiting thoughts that ‘add up’ in a satisfactory way (given what we care about and the complicated reality we face), and that effectively tap into our motivations to seek self-­praise and avoid self-­blame. This can be especially challenging in cases involving individually trivial effects, particularly if the resources we have to avoid self-­blame, which seem to include mindset M, are underestimated.12

References Ainslie, G. (1999). “The Dangers of Willpower,” in Getting Hooked: Rationality and Addiction, ed. J.  Elster and O.-J.  Skog. Cambridge: Cambridge University Press, 65–92. 11  Note that, as Gilbert Harman emphasizes, even with respect to matters where arriving at the whole truth is feasible, uncovering the truth and/or using it in one’s reasoning may not be “worth the effort” if one is managing well enough even though one’s beliefs are inconsistent. See Harman (2004, 49–50 & 55, note 1). 12  My thanks to Mark Budolfson, Tristram McPherson, Elijah Millgram, David Plunkett, Michael White, Mariam Thalos, and the participants at my work-­in-­progress talks at the Center for Human Values at Princeton University and at the Philosophy Institute of the University of Saarlandes for helpful comments on earlier drafts of this chapter. After completing the final version of the chapter, I received some thought-­provoking comments from a group of students at Dartmouth College. I am grateful for the feedback and expect that some of the deep points raised will impact my future work on these issues. My work benefitted from a University of Utah Faculty Research and Creative Grant. I appreciate the support.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

150  Chrisoula Andreou Andreou, C. (2006). “Environmental Damage and the Puzzle of the Self-Torturer.” Philosophy & Public Affairs 34: 95–108. Andreou, C. (2007). “Environmental Preservation and Second-Order Procrastination.” Philosophy & Public Affairs 35: 233–248. Andreou, C. (2010). “Way to Go, Me.” Choosing Well. Psychology Today, April 9. Copyright [Chrisoula Andreou]. Brock, T. C. and L. A. Becker (1966). “‘Debriefing’ and Susceptibility to Subsequent Experimental Manipulations.” Journal of Experimental Social Psychology 2: 314–323. Broome, J. (2012). Climate Matters: Ethics in a Warming World. New York: W. W. Norton. Budolfson, M. (2019). “The Inefficacy Objection to Consequentialism and the Problem with the Expected Consequences Response.” Philosophical Studies 176: 1711–1724. Gardiner, S. (2002). “The Real Tragedy of the Commons.” Philosophy & Public Affairs 30: 388–416. Glantz, M. (1999). Creeping Environmental Problems and Sustainable Development in the Aral Sea Basin. Cambridge: Cambridge University Press. Harman, G. (2004). “Practical Aspects of Theoretical Reasoning,” in The Oxford Handbook of Rationality, ed. A. R. Mele and P. Rawling. Oxford: Oxford University Press, 45–56. Hiller, A. (2011). “Climate Change and Individual Responsibility.” The Monist 94: 349–368. Kagan, S. (2011). “Do I Make a Difference?” Philosophy & Public Affairs 39: 105–141. Kahneman, D. and A.  Tversky (1984). “Choice, Values, and Frames.” American Psychologist 39: 341–350. Knobe, J. (2003a). “Intentional Action and Side Effects in Ordinary Language.” Analysis 63: 190–194. Knobe, J. (2003b). “Intentional Action in Folk Psychology.” Philosophical Psychology 16: 309–324. Miller, D. T. and B. Monin (2016). “Moral Opportunities versus Moral Tests,” in The Social Psychology of Morality, ed. J.  P.  Forgas, L.  Jussim, and P.  A.  M.  Van Lange. New York: Routledge, 40–55. Nefsky, J. (2012). “Consequentialism and the Problem of Collective Harm: A Reply to Kagan.” Philosophy & Public Affairs 39: 364–395. Nolt, J. (2011). “How Harmful Are the Average American’s Greenhouse Gas Emissions?” Ethics, Policy and Environment 14: 3–10. Sedikides, C. and A. P. Gregg (2007). “Portraits of the Self,” in The SAGE Handbook of Social Psychology, ed. M. A. Hogg and J. Cooper. London: Sage, 93–122. Shepperd, J., W. Malone, and K. Sweeny (2008). “Exploring Causes of the Self-Serving Bias.” Social and Personality Psychology Compass 2: 895–908.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Way to Go, Me  151 Sinnott-Armstrong, W. (2005). “It’s Not My Fault: Global Warming and Individual Moral Obligations,” in Perspectives on Climate Change: Science, Economics, Politics, Ethics, ed. W. Sinnott-Armstrong and R. B. Howarth. Oxford: Elsevier, 285–307. Sutter, J. D. (2015). “You’re Making this Island Disappear.” CNN Two Degrees. http:// www.cnn.com/interactive/2015/06/opinions/sutter-two-degrees-marshall-islands/.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

7

The Wages of Fear? Toward Fearing Well About Climate Change Alison McQueen

Scientists, policymakers, and activists have historically used information to prompt action on climate change. They have tried to change our behavior by giv­ ing us facts about the phenomenon’s causes and consequences. This strategy rests on an “information deficit” model, which assumes that inaction on climate change is caused by a lack of information (Bain et al.  2012,  2016; Moser and Dilling 2011; Stern 2012; Sturgis and Allum 2004). This model has informed most of climate change communication for the past three decades. It is also dra­mat­ic­ al­ly incomplete. As the science on the causes and consequences of climate change has become more settled, as more information has been publicized, and as more resources have flowed to powerful climate action campaigns, the proportions of Americans who think that climate change is happening (73%) and is human-­caused (62%) have varied but are now at their highest levels (Leiserowitz et al. 2020a).1 The proportion of Americans who report being alarmed about climate change has doubled since 2015 and currently stands at 26%. Meanwhile, the proportions of American who report being dismissive, doubtful, disengaged, and cautious have all declined over the past five years (Leiserowitz et al. 2020b). The majority of Americans (63%) now report being at least “somewhat worried” about climate change (Leiserowitz et al. 2020a). Yet, despite these changes in belief in and concern about climate change, nearly three-­quarters of Americans report that they are not actively engaging the issue as individuals, consumers, or political actors (Leiserowitz et al. 2020b).2 These find­ ings suggest that while information and understanding may be necessary for 1  The percentage of Americans who think global warming is happening stands at 73% as of April 2020. This is a modest increase from 2009 (71%), but a marked increase from 2015 (63%). The per­ centage of Americans who think that global warming is mostly human-caused stands at 62% as of April 2020. This is a modest increase from 2009 (57%), but a marked increase from 2015 (51%) (Leiserowitz et al. 2020a: 6, 9). 2  Only the “Alarmed” segment (26%) in Leiserowitz et al.’s (2020b) study are taking action to combat climate change.

Alison McQueen, The Wages of Fear? Toward Fearing Well About Climate Change In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Alison McQueen. DOI: 10.1093/oso/9780198796282.003.0008

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  153 motivating action, they are not sufficient (Moser and Dilling 2011; Nisbet and Scheufele 2009; Sturgis and Allum 2004).3 In response to this problem, some scholars have started to focus on the role of emotions in motivating action on climate change (Leiserowitz 2006; Markowitz and Shariff  2012; Moser  2007; Myers et al.  2012; Norgaard  2011; Smith and Leiserowitz 2014). The assumption of a lot of this work is that affective strategies can complement informational approaches to motivate collective action and exert popular pressure for domestic and intergovernmental policy changes (Moser and Dilling 2011). Fear has received the most attention. Appeals to fear have been central to a lot of recent climate change communication. It is not hard to see why. Fear appeals promise to overcome the motivational challenges of climate change by making the issue more salient. However, they have also been criticized as “fear monger­ ing,” antithetical to democratic and civic values, and counterproductive (Ereaut and Segnit 2006; Feinberg and Willer 2011; Gourevitch 2010; Mann et al. 2017; Moser and Dilling  2004, 2011; Nordhaus and Shellenberger  2007, 2009, 2014; O’Neill and Nicholson-­Cole 2009). In part for these reasons, attention has increasingly turned to hope as an alter­ native to fear (Chadwick 2015; Feldman and Hart 2018; Head 2016; Moser 2007; Ojala 2012; Stern 2012; Upton 2015; Williston 2012; Yeo 2014).4 I question this turn from fear to hope. In the case of climate change, fear appeals have more strengths than their critics seem to think, while hope appeals have dangers that their proponents have not yet confronted. Blending philosophical treatments of fear and hope with social scientific research on climate change communication, my argument proceeds as follows. In Sections 1 and 2, I argue that climate change fear appeals have clear benefits and that these appeals can be protected against the standard criticisms. The most ser­ ious danger of climate change fear appeals is they risk engendering a sense of res­ ignation, which in turn undermines individual and collective action. 3  For various explanations of why information and understanding may not be motivationally suffi­ cient, see: Bain et al. (2012); Broomell et al. (2015); Markowitz and Shariff (2012); McCright and Dunlap (2011); McQueen (2017); Myers et al. (2013); Weber (2006). 4  The background assumption behind the turn to hope as an alternative to fear is that the two are inconsistent emotional strategies. This might seem strange if one holds something like the following view. S fears that p when S wishes that not-­p and is uncertain whether p or not-­p will obtain. Understood in this way, fearing that p and hoping that not-­p are compatible. S hopes that not-­p when S wishes that not-­p and is uncertain whether p or not-­p will obtain (Gordon 1980: 564). One needs to specify further preconditions for fear and hope that show why they are incompatible (for two different attempts, see: Descartes 2015: II.58, 221; Gordon 1980). However, I think the primary reasons why many climate change commentators see hope and fear as opposed are empirical and conventional. Some studies of climate change communication have found an inverse relationship between fear and hope. Communicative strategies that increase fear also tend to decrease hope, and vice versa (e.g. Feldman and Hart 2018). Furthermore, in American public discourse about politics generally and cli­ mate change specifically, fear and hope are often opposed (B.  Clinton  1994; H.  Clinton  2016; Debenedetti 2016; Gore 2014, 2016; Mandel 2016).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

154  Alison M c Queen In Sections 3 and 4, I consider the proposed alternative of hope appeals. While climate change hope appeals promise to avoid the danger of resignation, they risk encouraging complacency, which may be equally pernicious. We should therefore be wary of abandoning fear appeals in favor of a strategy of hope. In Section 5, I turn to Aristotle for guidance on how fear about climate change might be cultivated both more responsibly and more effectively. Aristotle’s insist­ ence that we attend to the ethical and political dimensions of rhetoric is a cor­rect­ ive to contemporary debates about climate change communication, which tend to focus more narrowly on questions of the effectiveness of rhetoric (cf. Lamb 2018; Lamb and Lane 2016). I outline Aristotle’s conception of “civic fear,” which I take to be an empirically-­sensitive model for how political collectives can fear well about a common threat (Aristotle 1996: 1308a, 135; Aristotle 2006: 1382b–1383a, 128–131; Pfau 2007).5 A civic fear appeal is designed to reap the salience benefits of fear and the motivational benefits of hope without succumbing to the dangers of resignation or complacency. This, I suggest, is an attractive possibility for cli­ mate change communication.

1.  The Promise of Fear Throughout the chapter, I will understand “fear” as a placeholder for a more gen­ eral kind of powerful negatively valenced response to a possible undesirable out­ come (or threat). Insofar as there is a motivational component to fear it involves prompting the fearer to attend to and avert the threat. A fear appeal is an argument or persuasive message that attempts “to arouse fear in order to promote precautionary motivation and self-­protective action” in the face of uncertainty (Ruiter et al.  2001: 614; see also Witte  1994: 114; cf. Walton 2000: 20). A rhetorically ideal fear appeal does four things. First, it identi­ fies a threat to which recipients or those to whom they have moral or affective ties are susceptible. Second, the fear appeal offers evidence that this threat is severe.6 Combined, these first two elements generate a sense of threat and connect that threat to the target of the appeal. Third, the fear appeal recommends a course of action (or range of actions) that it presents as effective in responding to the threat. Psychologists and communications scholars sometimes refer to this as “response efficacy.” Finally, the fear appeal presents this action (or range of actions) as easy or, at the very least, as within the power of the recipient to undertake. Psychologists and

5  The term “civic fear” is Michael Pfau’s (2007) and I will use and extend Pfau’s analysis in Section 5. 6  This need not amount to showing that the threat is likely to obtain. It might instead amount to showing that the expected disvalue of the threat is high. So, a severe threat could be: (a) improbable but dire, if it obtains, or (b) more probable but comparatively less dire, if it obtains.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  155 communications scholars sometimes refer to this as “self-­efficacy” (Ruiter et al. 2001).7 Combined, these final two elements generate a sense of agency. Fear appeals are most common in public health campaigns. Consider a classic example from an Australian campaign to increase AIDS awareness. A television advertisement depicts a Grim Reaper bowling a ball, which gathers speed and knocks over ten human “pins.” A voiceover states: At first, only gays and IV drug users were being killed by AIDS. But now we know every one of us could be devastated by it. The fact is [that] over 50,000 men, women, and children now carry the AIDS virus. That in three years, nearly 2000 of us will be dead. That if not stopped, it could kill more Australians than World War II. But AIDS can be stopped and you can help stop it. If you have sex, have just one safe partner or always use condoms. Always.

The advertisement ends with the message: “AIDS. Prevention is the only cure we’ve got” (NACAIDS  1987). The advertisement presents HIV infection as a threat to which viewers and those with whom they have close ties are susceptible. It stresses the severity of the threat. It recommends effective response actions (safe monogamous sex and condom use). It suggests that recipients have the power to take these actions. Fear is also the emotional currency of much of the climate change debate. Consider David Wallace-­Wells’s recent book, The Uninhabitable Earth.8 The book, much like the popular 2017 article on which it is based, was widely hailed as ter­ rifying. Wallace-­Wells identifies a severe threat: the global devastation that does and will result from a warmer world. He vividly describes floods, wildfires, dying oceans, declining air quality, food insecurity, plagues, and conflict. For those cur­ rently living in affluent countries that are still protected from some of these hor­ rors, our own judgments about the severity of threat are, he thinks, unreliable. “It is worse, much worse, than you think” (2019: 3). Given that many of the worst effects of climate change will be felt first by the global poor and by future generations, it can be hard to convey a sense of suscep­ tibility to the current inhabitants of affluent countries. Wallace-­Wells attempts to elicit vulnerability by stressing the effects of climate change that already affect us, like deadly heatwaves, increasingly intense hurricanes, and ever-­more devastating 7  I have made two small modifications to Ruiter et al.’s (2001) list of the elements of a fear appeal. First, they stress that an ideal fear appeal presents a threat “to which the recipient is [personally] sus­ ceptible” (614). I see no reason why this need be the case. We can feel fear on behalf or for those with whom we have some kind of moral or affective connection. Second, they stress that the recommended action(s) must be easy to execute. I am not convinced that this need be the case. We might imagine a recommended action that is likely to be effective and, while not easy to execute, is well within our power to execute. 8  Al Gore uses a similar fear strategy in the film An Inconvenient Truth, as well as an article that accompanied the film’s release (Gore 2006).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

156  Alison M c Queen wildfires. These current effects are important to highlight, he argues, because one of the maladaptive ways people “might manage to navigate [the reality of climate change] without crumbling collectively in despair is, perversely, to normalize cli­ mate suffering at the same pace we accelerate it” (Wallace-­Wells 2019: 216). Wallace-­Wells’s fear appeal struggles, as many others on climate change do, with the question of efficacy. He gestures at the familiar policy tools: “a carbon tax and the political apparatus to aggressively phase out dirty energy; a new approach to agricultural practices and a shift away from beef and dairy in the global diet; and public investment in green energy and carbon capture” (Wallace-­Wells 2019: 227). This will require that individuals in affluent and powerful countries make “meaningful political noise” (Wallace-­Wells 2019: 186), by which I take him to mean organizing with others to demand that domestic governments and inter­ nation­ al institutions coordinate to drastically reduce emissions immediately. When it comes to persuading readers that these responses are within their power, Wallace-­Wells uses an interesting strategy. The acknowledgement that climate change is a problem of our own making “should be a comfort, not a cause for despair” (Wallace-­Wells 2019: 30). While we ought to be terrified, “thinking like one ­people, all the relevant inputs are within our control, and there is no mysti­ cism required to interpret or command the fate of the earth. Only an acceptance of responsibility” (Wallace-­Wells 2019: 226).9 The promise of fear appeals is that they can make a threat salient. This is espe­ cially valuable in the case of climate change. Because the causes of climate change are complex and its worst effects probabilistic and spatially and temporally dis­ tant, the issue often lacks salience for those best placed to take action. Even if cli­ mate change were to become more salient, it must compete for attention with other issues. Psychologists and climate change communication scholars have found evidence that people have a “finite pool of worry.” As we worry more about one kind of risk, our concern about other risks goes down (Linville and Fischer 1991; Weber 2006). This may be one explanation for why concern about environmental issues decreased in the United States after the 9/11 terrorist attacks and again in the wake of the 2008 financial crisis (Saad and Jones 2016).10 Fear appeals promise to make climate change salient in the face of these obstacles.

9  I think it would be reasonable to criticize this move. Climate change, like many complex human problems, is the result of numerous uncoordinated actions by many individuals, corporations, and governments. A lot of what makes mitigation and adaption difficult is not that people are blind to the possibility that human action will make a difference; it is the recognition that coordinating human action on this scale is hard. It is far easier to cause a problem through uncoordinated human action than it is to solve or reduce it by coordinated human action. 10  At the time of publication, there is some initial polling data that suggests that the COVID-­19 outbreak did not change how concerned American survey respondents were about climate change (Kennedy 2020; Leiserowitz et al. 2020a). These results, however, are early and tentative. Neither of the cited surveys formally tested the “finite pool of worry” hypothesis.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  157 Successful fear appeals make threats salient by focusing our attention on them. Aristotle understood this. When we face genuine risks and dangers, he argues, focus is a salutary response (Aristotle  2006: 1382b, 130; Leighton  1988). As Howard Curzer puts it, “fear foregrounds certain things, bringing them to our attention, meanwhile backgrounding other things that would just distract us. Fear transforms a charming country landscape, uncluttered with unsightly structures or fences, decorated with placid, color-­coordinated cows and one tall tree, into a nightmarish, flat, unprotected space with only one possible refuge from a furious, charging bull” (Curzer 2012: 59).11 There is substantial empirical evidence that this attentional view is correct. The experience of fear is correlated with an involuntary focus on the perceived threat and the means of escaping it (Faucher and Tappolet 2002: 114–127). More spe­cif­ ic­al­ly, there is evidence that fear appeals increase the salience of climate change. The problem becomes more salient for those presented with images that elicit fear and horror (Metag et al.  2016; O’Neill et al.  2013; O’Neill and NicholsonCole  2009).12 Given the attentional obstacles posed by the problem of climate change, these salience effects are especially valuable.

2.  The Wages of Fear Critics acknowledge that fear appeals may enhance salience but hold that they prevent us from “fearing well.” There are three sorts of worry here. Let us take them in turn and evaluate the extent to which each holds in the case of climate change fear appeals. First, fear appeals are sometimes seen to amount to “fear mongering.” The con­ cern, I take it, is that fear appeals aim to elicit fear in circumstances where rational moral agents should experience none. Consider the discourse on crime in the United States during the 1990s. While violent crime started declining dra­mat­ic­ al­ly at the beginning of the decade, the evening news was punctuated with reports of murders and warnings about the rise of dangerous young “super-­predators.” It is estimated that between 1990 and 1998 there was a 600 percent increase in American network news stories on murder (Glassner 2004: 820). Most Americans 11  In understanding fear in this way, Aristotle anticipates philosophical approaches which concep­ tualize emotions as “species of determinate patterns of salience among objects of attention, lines of inquiry, and inferential strategies” (de Sousa 1987: 196; see also Rorty 1980). According to this view, emotions give certain features of our surrounding context “a weight in our experience” that they would otherwise have lacked (de Sousa 2014). 12  It is worth noting that in a larger-N study of American respondents, Hart and Feldman (2016) did not find any evidence of the influence of imagery or text about climate change’s impact on per­ ceived issue importance. However, Hart and Feldman used a more limited range of imagery in their experiment. The only image that might be classed as one typical of a fear appeal is an image of flooding.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

158  Alison M c Queen are still not aware that violent crime is declining nationally (Koerth and Thomson DeVeaux 2020). Fear mongering induces epistemic irrationality by eliciting predictable human biases.13 A recent and easily recalled news story or emotionally intense invocation of violent crime may cause us to overestimate the frequency of such events. As news coverage of these events increases, our statistical senses become more dis­ torted, our anxiety and fear increase, and our attention to other risks decreases (Kahneman 2013; Kuran and Sunstein 1999). Fear mongering encourages us to overestimate the likelihood of a frightening outcome and to act on this overesti­ mate by, for instance, supporting tougher laws on violent crime. Appeals that “monger” fear prevent an accurate evaluation of our risk. In the case of climate change, it is not clear that most fear appeals are appropri­ ately subject to this criticism. Absent any sort of communicative or deliberative intervention, many individuals are subject to a range of cognitive biases and forms of motivated reasoning that cause them to either ignore or underestimate the risks posed by climate change (see, for instance: Bazerman 2006; Budescu et al.  2009; Markowitz and Shariff  2012; Stoknes  2015). Rather than aiming to induce epistemic irrationality by eliciting biases, many climate change fear appeals are simply trying to correct these problems.14 Second, fear appeals are accused of being inconsistent with the ends of liberal and democratic politics. They close off the deliberation and debate required to secure meaningful consent to political aims. Consider the case of McCarthyism. During the height of the Cold War “Red Scare,” McCarthy and others used fear to 13  In the context of fear and hope appeals, I take the main form of epistemic irrationality to involve beliefs about probabilities that do not track the balance of available evidence. 14  Three important concessions. First, such appeals may nonetheless be manipulative. Whether or not one reaches this conclusion depends on one’s view of manipulation. If one holds a view of ma­nipu­ la­tion that requires that the manipulator deceive the manipulee (e.g. Goodin 1980: 8), a climate change fear appeal need not be manipulative. If the fear appeal were to exaggerate the threat, then it would be manipulative. However, what counts as exaggeration as opposed to, say, emphasizing less probable but more consequential risks is a thorny question (Mann et al. 2017). If one holds a view of manipulation that requires that the manipulator bypass or subvert the manipulee’s rational capacities, a climate change fear appeal might be manipulative (for an evaluation of the range of views here, see Gorin  2014). While the fear appeal might correct for irrational biases and lead the manipulee to a conclusion that is supported by reasons, it does so by appealing to emotions rather than through reason-­giving. But even if we conclude that the fear appeal is manipulative, it does not follow that it is wrong (Baron 2014). It may be that, in believing that climate change is not a serious threat or one that they have a responsibility to mitigate, those targeted by fear appeals may have made themselves mor­ ally liable to manipulation (or potentially even more overt forms of coercion). That is not a view I have the space to defend here. Second, the fact that fear appeals on climate change may not be manipulative in the sense implied by the term “fear mongering” does not mean that they are not perceived as ma­nipu­la­tive by some of their recipients (Moser 2007). On the difficulties here, see: Lamb and Lane (2016). Third, climate change appeals may remain objectionable on paternalistic grounds, even if they are not mongering fear. But they may not be manipulative in the first sense suggested above. It is beyond the scope of this chapter to address the paternalism objection. However, if one is persuaded by arguments for coercive paternalism (Conly  2012) or libertarian paternalism/nudging (Thaler and Sunstein 2009) in areas like public health, it seems reasonable to think that variations of similar argu­ ments could apply to fear appeals in the case of climate change.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  159 make the threat of America’s destruction at the hands of communist sympathizers salient. In so doing, they cast America’s survival as the end to which all other political projects must be subordinated (Morgenthau 1960: 148–152). Ideals like freedom and equality were seen as dangerously indulgent in the face of an over­ whelming threat to collective survival. McCarthy and his allies presented Americans with a coercively dichotomous choice—a choice between collective destruction and conformity to the demands of the security state. This is the fea­ ture of fear appeals that preoccupies many critics of a “politics of fear” (e.g. Furedi 2005; Gourevitch 2010; Robin 2004). On this view, the problem with fear appeals is not only that they monger fear, but that they are coercively dichotomous. They force us to choose between suc­ cumbing to a terrible threat (e.g. destruction at the hands of the Soviet Union) and accepting the proposed means of averting the threat (e.g. hunting down sus­ pected communist agitators, suspending civil liberties, etc.). Is this a conceptually necessary feature of fear appeals? Douglas Walton, who has offered the only sustained philosophical analysis of fear appeals, argues that it is—all fear appeals have a conditional and dichotomous structure. That is, they sharply divide “the respondent’s available options into two mutually exclusive actions (events) where one will (supposedly) occur if and only if the other does not occur. In other words, the upshot of the dichotomized argument is that the respondent has only two options, and that no third option is available” (Walton 2000: 20). Many fear appeals, including some about climate change, have the structure that Walton attributes to them. However, they need not have this structure. For instance, a fear appeal might juxtapose a single threatening outcome, on the one hand, with a range of options that might avoid the outcome, on the other. A politician might attempt to elicit fear by pointing to the threat posed by a hostile neighboring country. She might argue that in order to avert this threat, some sort of action must be taken. She might then  propose several options that could plausibly avert the threat—diplomatic ­negotiations, economic sanctions, a reshuffling of alliances, and military action. Assuming a broad agreement that she has identified a genuine threat, there is still room for elite and/or popular deliberation about which response to choose.15 Third, fear appeals are often criticized as counterproductive. There is substan­ tial empirical evidence suggesting that while terrifying arguments and images may increase the salience of a particular threat, they can also prompt a sense of powerlessness, fatalism, and disengagement on the part of recipients. These 15  One might respond that this is a rather thin notion of deliberation. That is true, especially when compared to an ideal of deliberative democracy. However, when judged against this ideal, virtually all actual deliberation in advanced liberal democracies is rather thin. Because we are concerned about whether fear appeals are appropriate in a world such as ours, the relevant question is how much the deliberation they might encourage diverges from the best forms of deliberation that our world has to offer. I am not convinced they fare too badly on this standard.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

160  Alison M c Queen feelings reduce the motivation for behavioral change and political engagement (Ruiter et al. 2001, 2014; Witte 1994; Witte and Allen 2000). This is because there is no necessary connection between salience and attentional focus, on the one hand, and motivation and behavioral change, on the other. As Aristotle suggests, fear inclines us to “deliberation” about avoiding the threatening outcome (Aristotle 2006: 1383a, 130; Curzer 2012: 59). We are motivated not by fear alone, but by fear in combination with the deliberations that it prompts. If these delib­ erations lead us to think that any actions we can take will be futile, we will not be scared but resigned. The extended parallel processing model (EPPM) of fear appeals builds on this insight. A fear appeal initiates two sorts of appraisals. First, the recipient appraises the threat and makes an assessment about susceptibility and severity. If she deter­ mines that the threat is great, she then begins a second appraisal and makes an efficacy assessment about the proposed response. When the perceived threat is great and her sense of efficacy is strong, the recipient will attempt to control the danger by taking protective action. When the perceived threat is great and the sense of efficacy is weak, the recipient will instead attempt to control her fear. She will typically do this in maladaptive ways. For instance, she may try to avoid thinking about the threat, impugn the messenger (e.g. by accusing him of dishon­ esty or manipulation), or question whether the threat is real (Feinberg and Willer 2011; Witte 1994). This explanation may help to account for a recurrent finding in studies of cli­ mate change communication. While scary images and arguments tend to increase the salience of climate change, they also tend to decrease recipients’ sense of effi­ cacy (Hart and Feldman 2014; Metag et al. 2016; O’Neill et al. 2013; O’Neill and Nicholson-­Cole  2009). Faced with terrifying and often apocalyptic representa­ tions of the effects of climate change, many conclude—not entirely without good reason—that any actions they can take will be futile (McQueen  2017). Climate change fear appeals may face a salience–efficacy tradeoff. The very measures taken to make the issue salient (e.g. apocalyptic imagery) may leave audiences feeling powerless and resigned.16 But must fear appeals face this tradeoff? No. In fact, there is growing evidence that, when paired with efficacy-­enhancing information, salience need not be para­sit­ic on efficacy (de Hoog et al. 2008; Maibach et al. 2008; Nabi et al. 2018; Witte and Allen  2000). Also, appeals that portray a negative prospect as less

16  This issue was precisely what was at stake in the debate over David Wallace-­Wells’ controversial article, “The Uninhabitable Earth.” He describes the worst-­case effects of climate change—from star­ vation and plagues to death smogs and wars—in exceptionally vivid terms to frighten readers out of (what he sees as) a dangerous complacency. Critics of the piece, who included a number of prominent climate scientists, did not primarily take aim at Wallace-­Wells’ presentation of climate science. Rather, they criticized his rhetorical strategy. “Climate doomism,” some claim, breeds resignation. “In many ways,” it is “as pernicious as outright climate change denial” (Mann et al. 2017).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  161 threatening or not threatening at all have even less capacity to motivate action than scary messages that lack clear action recommendations (de Hoog et al. 2007; Witte and Allen 2000).17 In sum, there is nothing inherent about fear appeals that makes them vul­ner­ able to the standard moral criticisms. Climate change fear appeals in particular are relatively well-­insulated from the accusation of fear mongering and ma­nipu­la­ tion. It also seems possible to issue them in a way that invites, rather than fore­ closes, some deliberation. Perhaps the greatest challenge for the use of fear in climate change deliberation concerns the way in which fear appeals may, in prac­ tice, transform fear into resignation. If climate change fear appeals could be designed in ways that preserve their salience effects, do not monger fear, invite deliberation, and avoid engendering resignation, we would be well on our way to fearing well about climate change.

3.  The Possibility of Hope These criticisms of fear appeals have prompted an interest in hope in climate change communication. I understand “hope” as a placeholder for a more general kind of powerful positively valenced response to a possible desirable outcome. Insofar as there is a motivational component to hope it involves prompting the hoper to attend to and work toward that outcome.18 A hope appeal is an argument or persuasive message that encourages the re­cipi­ent to take this action in the face of uncertainty (Chadwick 2015, 600–601). A rhetorically ideal hope appeal does four things. First, it specifies a concrete goal that is personally valuable to recipients. This claim about value is often linked to a claim that the goal will lead to a better future. Second, it gives the recipient ­reasons to think that the achievement of this goal is possible. Combined, these first two elements generate a sense of opportunity and connect that opportunity to the target of the appeal (Chadwick 2015: 600). Third, the hope appeal recommends some course(s) of action that offers a pathway that could be effective in achieving this goal (response efficacy). Finally, it 17  Engendering a sense of efficacy on climate change is challenging. While there are individual behavioral changes that may, in the aggregate, reduce the impact of global climate change, effective action depends much more on some combination of domestic political changes to existing laws and regulations, intergovernmental cooperation, and international agreements and enforcement. Individual citizens have a more indirect role to play in these responses. The astute recipient of a fear appeal on climate change will be making an appraisal of self-­efficacy that includes an assessment of her own capacity for political voice. Her appraisal of response efficacy will include an assessment of whether her political representatives and institutions will respond to her political voice. This sets a high informational and motivational bar for climate change fear appeals (Hart and Feldman 2014: 327). 18 For more specific accounts of hope that are compatible with this placeholder, see: Calhoun (2018: 68–89), Martin (2013), and Pettit (2004). As with fear, the analysis here is restricted to hopes for outcomes that are at least partially subject to agential control or influence.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

162  Alison M c Queen presents this course of action as easy or, at the very least, as within the power of recipients to undertake (self-­efficacy). Combined, these final two elements gener­ ate a sense of efficacy or agency (Chadwick 2015: 600–601; McGeer 2004). For example, consider Barack Obama’s speech to the 2015 United Nations Climate Change Conference in Paris, which in the hindsight afforded by a bleaker present, was a high watermark for hope on climate change. He casts the Paris conference as a turning point in the collective effort to mitigate the worst effects of climate change: “nearly 200 nations have assembled here this week . . . should give us hope that this is a turning point, that this is the moment we finally deter­ mined we should save the planet.” Pointing to American and global successes at reducing (or, at the very least, not increasing) carbon pollution and investing in cleaner power sources whilst sustaining economic growth, Obama suggests that there are effective paths to meeting the challenges of climate change and that states have the power to meet them. He invites state leaders to reach an agreement on increasingly demanding emis­ sions targets and stronger reporting standards. Obama acknowledges that taking this path “will not reward us with moments of victory that are clear or quick. Our progress will be measured differently—in the suffering that is averted, and a planet that is preserved. And that’s what’s always made this so hard.” However, we must take inspiration from the fact that our children and grandchildren will be able to “take pride in our achievement.” The speech concludes with an exhortation: “Let’s get to work” (Obama 2015). While one might take issue with some of the Obama’s claims about the extent of US success on climate change, there seems little doubt that his speech contains all of the elements of a rhetorically ideal hope appeal. The promise of hope appeals is that they increase the motivation to bring about the hoped-­for outcome. This motivational promise is especially valuable in the case of climate change because both the scope of the problem and some of the rhetorical techniques used to deliberate about it have the potential to elicit appraisals of powerlessness, futility, and resignation (Chadwick 2010, 2015). If we assume that hope appeals can successfully elicit hope, there is reason to think that this motivational effect is plausible. Like fear, hope operates by focus­ ing our attention on the hoped-­for outcome, the pathways available toward that outcome, and the strength of our own capacities for effective action. There is extensive empirical evidence that those with high levels of hope are more goaloriented, more capable of “generating workable routes” to their goals and identi­ fying alternative routes in the face of impediments, and more likely to think that achieving their goals is within their capacity (Snyder et al.  2002: 258; see Snyder  1995,  2002 for a summary of these findings). Those with high levels of hope often express a certain determination or resolve. They say things like: “I am not going to be stopped.” “I’ll find a way to get this done” (Snyder et al. 2002: 258). One way to think about such declarations is that they express a kind of cer­ tainty that is inconsistent with the act of hoping. By definition, we hope for

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  163 outcomes that we are uncertain will obtain. If we were certain they would obtain, we would simply await them. However, if we take hope to have a motivational component, it involves thinking, feeling, and planning around the hoped-­for prospect obtaining. This need not involve irrational certainty or unwarranted probability assessments. For instance, an injured hiker in a remote location may correctly believe that she has about a 10 percent chance of surviving. If she is hopeful, survival (as opposed to death) becomes especially salient for her. She thinks, acts, and plans as if she will make it through this ordeal.19 She may think about the vacation she and her family “will” take in six months’ time and consider which colleagues might cover her work duties during her absence. She may fanta­ size or construct narratives about a future in which she survives. These thoughts, feelings, and plans, she thinks, are justified by the possibility of her survival (Martin 2013). She may do all of this whilst correctly recognizing that the prob­ abil­ity of her survival is low (i.e. without falling victim to epistemic irrationality). Expressions of determination or resolve capture this aspect of hope. Ordering one’s thoughts, feelings, and plans in this way has a number of mo­tiv­ ation­al benefits. It protects the agent from paralyzing despair. For outcomes within the agent’s power, this protection can preserve their chances of obtaining. The despairing hiker, who focuses on the probability of survival and envisions a future in which she dies alone in the wilderness, may be discouraged from work­ ing hard to survive. The hopeful hiker, who focuses on the possibility of survival and envisions a future in which she lives and enjoys her family vacation, may be encouraged to work hard to survive—to fight through pain, to investigate new paths to success, etc. (Martin  2013: 85–90). As Philip Pettit notes: “Many of us will treat a 10 percent chance of success in some venture as a depressing, poten­ tially enervating prospect. And if we do, that may well reduce the chance of suc­ cess even further. But if we gain heart by the assertion of will involved in putting hope in success, then we give success the 10 percent chance it really has” (2004: 161; see also Walker 2006: 55–56). In other words, in cases in which an outcome at least partially depend on our own agency, hope may be pragmatically rational— it may encourage us to take the actions necessary to give our desired outcome its best chance of obtaining (Martin  2013: 24; Pettit  2004). In the case of climate change, where pessimistic assessments of the prospects for effective mitigation

19  Philosophers differ in their view of what, exactly, is going on when this happens. For instance, Philip Pettit argues that the hopeful patient sets her probability assessments aside—puts them “offline”—and tries to respond and act as if the treatment will work or has a “good chance” of working (2004: 159, 157, emphasis mine). Adrienne Martin argues that Pettit fails to capture the phenomenon of hope: “The person who takes 1 percent as a reasonable basis for hoping against hope doesn’t thereby think of 1 percent as 25 percent, or anything like that—she simply sees 1 percent as enough to go for­ ward” (2013: 23). Martin argues that the hopeful person orients her thoughts, feelings, and plans around the possibility of the hopeful outcome obtaining. For a similar view (albeit one in which the hoper adopts a different temporal perspective), see Calhoun (2018: 68–89).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

164  Alison M c Queen come on the heels of every international conference (and some American elections), these motivation-­enhancing benefits may be especially valuable.

4.  The Perils of Hope Assuming that hope appeals on climate change can deliver on this motivational promise, should we simply abandon strategies of fear and devote ourselves whole­ heartedly to cultivating hope? I think not. Particularly in the case of climate change, there is reason to worry that hope appeals can lead to wishful thinking and complacency. Wishful thinking is a species of epistemic irrationality that involves an over­ esti­mation of the probability of some good outcome obtaining, despite available evidence (Bovens  1999). How might a hope appeal lead to wishful thinking? Hope appeals stress the personal value of a particular goal, generally by linking it to the achievement of a better future. To the extent that they perform this task successfully, they may encourage a form of motivated reasoning (Kunda 1990). From political to health outcomes, we tend to see the prospects we desire as more likely to obtain (Krizan and Windschitl 2007).20 So, wishful thinking is a risk of any hope appeal that elicits greater desire for an outcome. Might wishful thinking in turn engender complacency? Victoria McGeer sug­ gests that wishful thinking, combined with a failure to formulate paths and plans for achieving the hoped-­for outcome, leads to a kind of complacency that she calls “wishful hoping.” Wishful hopers “generate hopes that are fanciful insofar as they are not grounded in any real understanding of how they will be realized; they are simply the direct output of desires and so undisciplined by knowledge of the world” (McGeer 2004: 113). There is empirical evidence to support this worry. When we fantasize about a (subjectively) likely outcome—when we “mentally enjoy the desired future” now as if it had already obtained—we are less likely to work toward its achievement (Oettingen and Mayer  2002: 1199). Hope appeals that stress the desirability of an outcome without engaging specifically with paths for its achievement may risk engendering complacency. Neither wishful thinking nor complacency are empirically necessary outcomes of hope appeals.21 But the specific features of the problem of climate change might make both more likely. First, much of the messaging around climate change sug­ gests that individuals in advanced industrialized countries are collectively respon­ sible for “environmental damage as an unintended side effect of their behaviour 20  While there is good evidence of a correlation between the desirability of an outcome and esti­ mates of its probability, the evidence that increased (decreased) desirability causes increased (decreased) probability estimates is only weakly positive (Krizan and Windschitl 2007). 21  To my knowledge, there are no empirical findings that point to the frequency with which hope appeals cause wishful thinking and/or complacency.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  165 and lifestyle.” To reduce unpleasant feelings of responsibility, complicity, or guilt, individuals “will often engage in biased cognitive processes to minimize their perceptions of their own complicity” (Markowitz and Shariff 2012: 244). Having minimized their own role in causing the problem of climate change, they will also diminish the role that their motivation and action will have in effecting a solution (Stoll-­Kleemann et al. 2001). Second, empirical research suggests that when faced with any uncertainty about the bad outcomes of climate change, individuals are likely to underestimate the likelihood of these outcomes (Budescu et al. 2009). As Ezra Markowitz and Asim Shariff suggest, “the less definitive and incontrovertible the conclusions, the more room there is for individuals to infer unreasonably optimistic outcomes. As that optimism reduces the gravity of the issue, so too may it reduce the mo­tiv­ ation to act” (Markowitz and Shariff 2012: 244). There are two forms that such wishful thinking might take in the case of climate change. First, it might take the form of a vague techno-­optimism, in which hopeful people depend on technological solutions for generating effective responses to cli­ mate change and therefore fail to take individual or collective action themselves.22 Second, wishful thinking might take the form of dependence on other indi­ viduals or institutions for realizing one’s hopes. Because many climate change hope appeals understandably stress the agency of political and corporate decisionmakers, we might expect individual citizens to be especially prone to becoming wishful hopers in this way. There is a sense in which such wishful hoping is not entirely irrational. Political and corporate decision-­ makers often have more power to take effective action on climate change. Yet to the extent that the beliefs and actions of individual citizens and consumers play some role in influencing and directing political and corporate decisions, this kind of wishful hoping is maladaptive. Without individual motivation, hope in the power of political and corporate decision-­makers is misplaced hope. In sum, while fear appeals seem to be more protected from the standard wor­ ries in the case of climate change, hope appeals appear be less so.

5.  Civic Fear The argument so far has identified the promise and potential dangers of fear appeals in climate change communication. It has also given us reason to be wary of abandoning fear appeals altogether in favor of strategies of hope. Is there a way 22  This is not, of course, to say that all those who think technological solutions have a role to play in addressing climate change are wishful thinkers. However, those who (a) hope that technology will save us without a clear conception of how that might happen and (b) fail to take any individual or collective action because of that hope are both wishful thinkers and wishful hopers in McGeer’s (2004) sense of the term.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

166  Alison M c Queen to envision how we might come to fear well about climate change? In this section, I will argue that Aristotle sketches an attractive model that points us in the right direction. Aristotle’s model of “civic fear” (Pfau  2007) harnesses the salutary focusing effects of fear and the sustaining effects of hope without surrendering to resignation or complacency. In the Politics, Aristotle notes that citizens need little encouragement to protect their polity from threats that are temporally or spatially near. However, it is chal­ lenging for citizens to see these threats to begin with. For those living in advanced industrialized states in the Global North, climate change is just such a threat. In response to threats of this kind, Aristotle contends, it may be entirely appropriate to cultivate fear among citizens. When responding to threats to the polity, he advises rulers to “bring distant dangers near, in order that citizens may be on their guard, and, like sentinels in a night-­watch, never relax their attention” (Aristotle 1996: 1308a, 135). Here, the relevant political agents are democratic rulers, rather than ordinary citizens. The reason that rulers have a special responsibility to cultivate fear about threats to the constitutional order is because “no ordinary man can discern the beginning of evil, but only the true statesman” (Aristotle 1996: 1308a, 135). The underlying point, however, is that the survival of the polity depends on the capacity of citi­ zens to collectively perceive and understand the threats and opportunities that face them. In some cases, leaders may have a special role to play in aiding this collective perception and understanding. However, Aristotle also envisions a broader kind of political deliberation that allows the collective to benefit from the particular strengths, practical wisdom, and judgment of other members of the polity (Aristotle  1996: 1281b–1282a, 76–78). It seems reasonable to expect that some of this strength, wisdom, and judgment would bear on the identification and understanding of potential threats to the polity. Whoever apprehends the threat and attempts to elicit popular fear, their ultimate aim should be to prompt deliberation. As we have seen, for Aristotle, fear makes individuals “inclined to deliberation” (Aristotle 2006: 1383a, 130). It turns our cognitive energies to thinking about how to avert the threat. Assuming that they have identified a threat that is genuine but difficult to per­ ceive and appreciate, how might a leader, expert, or ordinary citizen elicit fear in a way that prompts the members of a polity to fear well? We can read Aristotle’s analysis of fear in the Rhetoric as a set of guidelines on this question (Pfau 2007). Like much of the empirical literature on fear appeals, Aristotle suggests that one should stress that the threat has “great potential for destruction or for causing harms that lead to great pains.” In the case of threats that are spatially and/or temporally distant, Aristotle urges communicators to draw attention to their proximity: “For what is far off is not feared” (2006: 1382a, 129). This need not involve any deception or misrepresentation. When dealing with distant threats, one might emphasize the evidence or signs that these threatening

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  167 outcomes will obtain. For, “even the signs” of harmful and destructive outcomes “are causes of fear” and make the threat seem “near at hand” (Aristotle  2006: 1382a, 129). So far, this Aristotelian conception of civic fear simply affirms what many fear appeals on climate change already do. They emphasize the harmful and destructive impacts of climate change that are likely to affect the audience or those with whom they have close ties, they draw attention to the signs or observ­ able evidence of these outcomes, and they make the danger seem nearer at hand by stressing the local impacts it is already having and the impacts it is likely to have on generations that are currently living. Aristotle’s next moves are what make a civic fear appeal more distinctive. Fear is a painful emotion that is often “accompanied by an expectation of experiencing some destructive misfortune” (Aristotle 2006: 1382b, 130). A communicator should therefore expect her audience to resist a fear appeal. She must guard against two especially dangerous tendencies of resistance. The first comes from those who are “confident.” Like wishful thinkers, the confident “think they experience . . . great good fortune [and] do not think they might suffer; therefore, they are insolent and belittlers and rash” (Aristotle 2006: 1383a, 130). They will be resistant to fear appeals and reluctant to assume the responsibility for deliberation and threat-­avoidance. The confident are in danger of complacency. While Aristotle frames the follow­ ing advice in general terms, it is presumably the confident who should be made to “realize that they are liable to suffering; for [he can say that] others even greater [than they] have suffered, and he should show that there are others like him suf­ fering [now] (or who have suffered) and at the hands of those from whom they did not expect it and suffering things [they did not expect] and at a time when they were not thinking of [the possibility]” (Aristotle 2006: 1383a, 130). It is presumably also to the confident that David Wallace-­Wells directs the story of his own moral transformation. He is not, he explains, an environmental­ ist. He is a meat-­eating city-­dweller who appreciates nature in the abstract, but avoids camping. In short, he is “like every other American who has spent their life fatally complacent, and willfully deluded, about climate change.” He loosened the grip of this complacency not only by familiarizing himself with the science on climate change, but also by “collecting stories of climate change, many of them terrifying, gripping, uncanny narratives” (Wallace-­Wells 2019: 7). His own rhet­ oric­al strategy is aimed at replicating this response by making the current effects of climate change more visible and the prospective consequences apocalyptic. Like Aristotle, Wallace-­Wells recognizes that slow-­moving threats are easy to normalize: “we are always coming to terms with what is just ahead of us, decrying what lies beyond that, and forgetting all that we have ever said about the absolute moral unacceptability of the conditions of the world we are passing through in the present tense, and blithely” (2019: 216). The confident are, on this view, both liable to suffering and susceptible to the sort of blind wishful thinking that pre­ vent them from bearing the responsibilities of action.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

168  Alison M c Queen The second group likely to be resistant to civic fear appeals are the resigned. Aristotle conceives of the resigned as those “who have already suffered all the dreadful things possible and have become coldly indifferent to the future, like those actually being done to death,” or legally executed (Aristotle 2006: 1383a, 130, emphasis mine).23 Those “being done to death” are presumably resigned because the outcome that they feared has now obtained. It is no longer uncertain. Therefore, they do not fear it (Gravlee 2005: 468). However, as we have seen, resignation need not be the result of previous or current suffering. Those who are resigned may simply think that they cannot— either alone or with others—effectively avert the threat. If they are correct in this, they are rightly resigned. If they are not correct—that, is if there is a possibility that they can avert the threat—they are wrongly resigned. For the wrongly resigned, a civic fear appeal must offer hope. “[For fear to continue],” argues Aristotle, “there must be some hope of being saved from the cause of the agony. And there is a sign of this: fear makes people inclined to deliberation, while no one deliberates about hopeless things” (Aristotle 2006: 1383a, 130, emphasis mine) Aristotle’s final observation in this passage is crucial. Without it, one might well assume that just as the message about suffering should be directed specifically toward the confident, the hopeful component of a civic fear appeal need only be narrowly directed at the resigned. If this were the case, a civic fear appeal would— like a good newspaper or biblical prophecy—simply attempt to “comfort the afflicted and afflict the comfortable.”24 Yet Aristotle seems to be suggesting that hope plays a more general role in fear appeals and should not only be encouraged in those who are resigned. To the extent that fear appeals encourage deliberation, they do this by eliciting hope. It is the hope associated with fear, rather than fear itself, that encourages deliberation (Gravlee 2005: 468–472). How does a communicator elicit this hope? At a minimum, she must offer some reason to think that it is possible to be “saved from the cause of the agony.” She must portray the outcome as susceptible and responsive to human action. But even more, she must portray the outcome as within the particular power of the individuals and the political collective she is addressing. For, “we deliberate about what is in our power, that is, what we can do” and each group or polity “deliber­ ates about what they themselves can do” (Aristotle 2000: 1112a, 42). In the case of climate change, there is preliminary empirical evidence that rhetorical strategies that elicit fear, then hope in sequence have a stronger motivational effect than those that elicit either separately (Nabi et al. 2018).

23  A literal translation would be “like those crucified on a plank,” which implies legal execution. Thank you to Josiah Ober for alerting me to this. 24  Finley Peter Dunne originally used this phrase to (mockingly) describe the job of the press. It is repeated more earnestly by E. K. Hornbeck (Gene Kelly) in Inherit the Wind (1960).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  169 How might a communicator elicit a sense of agency on climate change? Empirical studies of fear and hope appeals suggest that communicators invite responses that are well within the power of many individuals. In the case of cli­ mate change, these responses might include home energy conservation measures (e.g. installing solar panels), changes to consumer habits (e.g. not buying dispos­ able plastic water bottles, getting a car with better fuel economy), and measures to “counteract” the environmental effects of one’s actions (e.g. by buying carbon off­ set credits). However, given what we know about the emissions reduction targets that will have to be met in order to avoid the worst effects of climate change, any rational evaluation of the response efficacy of these sorts of individual actions is unlikely to be encouraging. Absent an ambitious and powerful program of domestic, intergovernmental, and international political action, the efficacy of individual responses will be minimal. In order to move her audience to deliberation and action, then, a communica­ tor will not only have to give them reason to think that particular action paths will be effective and that these routes are within their power to take. She will also have to propose pathways for indirect agency through the exercise of political voice (Hart and Feldman 2014). Given the unequal distribution of political voice, particularly in the United States (Schlozman et al. 2012), and the known difficul­ ties of exercising it effectively, the purveyors of civic fear appeals on climate change will be inviting their audiences to take on substantial challenges. They will have to join social movements and press for political change. This will require organization and coordination, to be sure. But it will also require courage. The virtue of courage does not figure in the empirical literature on fear and hope appeals. This literature argues that the way to counteract resignation or complacency is to propose responses that are “easy” for audiences to undertake (Chadwick 2010; Ruiter et al. 2001, 2014; Witte and Allen 2000). Yet facing our collective fears squarely and preparing ourselves for the work of responding to them effectively is not always easy. In order that an audience be prepared to meet their fears without resignation, to hope without complacency, and to face the challenges that may be required by effective responses, they will need to be imbued with courage. For Aristotle, the courageous do not overcome their fear. Rather, they fear what it is rational and appropriate to fear (Aristotle 2000: 1106b, 30; 1115a–b, 48). They face the objects of their fear squarely and with careful attention. They endure the difficulties, pain, and sacrifices that squarely facing fears entails (Aristotle 2000: 1117a, 53; see also Balot 2014). How might a climate change communicator elicit this kind of courage? One promising route would be to invite her audience to consider how they and their actions will be remembered by future generations. Churchill relied on this tech­ nique as he sought to steel his audience for the looming Battle of Britain: “Let us therefore brace ourselves to our duties, and so bear ourselves that, if the British Empire and its Commonwealth last for a thousand years, men will still say, ‘This

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

170  Alison M c Queen was their finest hour’” (Churchill 1940).25 Obama seems to be attempting a similar move when he invites his audience at the 2015 Paris conference to undertake the kind of action that will elicit pride from future generations “when they look back and they see what we did here” (2015). There is some empirical support for the effectiveness of this rhetorical strategy. Preliminary experimental evidence suggests that thinking about one’s legacy increases the motivation to undertake difficult preventive actions on climate change (Zaval et al. 2015). The purveyors of civic fear appeals, then, might not need to hold back from inviting their audiences to respond in ways that may not be easy. However, they must attend to the question of how best to steel their audi­ ences to face these threats and the challenges of meeting them collectively.

6. Conclusion Political fear is not popular. This is especially true now, when the politics of fear has become the currency of populist demagogues and authoritarians (Nussbaum 2018). But fear can be a rational and motivationally powerful response to climate change. I have argued that climate change fear appeals can be protected against the standard criticisms of political fear. Hope appeals, by contrast, seem vulnerable to serious motivational drawbacks in the case of climate change. We should not therefore abandon fear appeals in favor of hope appeals. Instead, we should take our bearings from Aristotle in an effort to cultivate fear more responsibly—in a way that makes room for the best aspects of hope, elicits rather than extinguishes our sense of agency, and invites rather than forecloses deliberation. This chapter raises at least two sets of questions for future research. First, Aristotle writes as if, setting aside tendencies toward confidence or resignation, fear appeals will have a largely uniform effect on all members of the polity. However, empirical findings suggest that the effects of affective climate change appeals vary by political ideology. Liberals, conservatives, and moderates respond differently and with varying strength to invocations of fear and hope (e.g. Feldman and Hart  2018; Hart and Feldman 2016; see also McCright and Dunlap 2011). How might civic fear appeals be tailored for specific audiences for maximal motivational effectiveness? Second, I have stressed throughout the chapter that climate change fear appeals need not elicit epistemic irrationality on the part of their audiences. They need not, for instance, encourage their audiences to overestimate the likelihood of cata­stroph­ic outcomes. Fearing well, I have argued, entails fearing in a way that is both epistemically and pragmatically rational—that is, fearing with an accurate 25  One might imagine a variant of this sort of appeal that dispenses with the disturbing endorse­ ment of empire. My thanks to Danielle Allen for calling this out.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  171 enough threat assessment and in a way that leaves one motivationally prepared to act to avert the threat. Similar commitments seem to lie behind Aristotle’s con­ ception of civic fear. But what if we were faced with empirical evidence that, at least under certain conditions, encouraging false probability assessments or mis­ representing the nature of the threat were more motivationally effective than a strategy of complete accuracy?26 How much (if any) misrepresentation are we willing to accept in exchange for a citizenry more committed to meeting the chal­ lenge of climate change? These are debates that urgently remain to be had.

References Aristotle (1996). The Politics, and the Constitution of Athens. Revised student edition, ed. S. Everson. Cambridge: Cambridge University Press. Aristotle (2000). Nicomachean Ethics. Cambridge: Cambridge University Press. Aristotle (2006). On Rhetoric: A Theory of Civic Discourse, 2nd edition. New York: Oxford University Press. Bain, P.  G. et al. (2016). “Co-Benefits of Addressing Climate Change Can Motivate Action around the World.” Nature Climate Change 6(2): 154–157. Bain, P.  G., M.  J.  Hornsey, R.  Bongiorno, and C.  Jeffries (2012). “Promoting ProEnvironmental Action in Climate Change Deniers.” Nature Climate Change 2(8): 600–603. Balot, R. K. (2014). Courage in the Democratic Polis: Ideology and Critique in Classical Athens. New York: Oxford University Press. Baron, M. (2014). “Manipulation: Theory and Practice,” in The Mens Rea and Moral Status of Manipulation, ed. C.  Coons and M.  Weber. Oxford: Oxford University Press, 98–120. Bazerman, M. H. (2006). “Climate Change as a Predictable Surprise.” Climatic Change 77(1–2): 179–193. Bovens, L. (1999). “The Value of Hope.” Philosophy and Phenomenological Research 59(3): 667–681. Broomell, S.  B., D.  V.  Budescu, and H.-H.  Por (2015). “Personal Experience with Climate Change Predicts Intentions to Act.” Global Environmental Change 32: 67–73. Budescu, D.  V., S.  Broomell, and H.-H.  Por (2009). “Improving Communication of Uncertainty in the Reports of the Intergovernmental Panel on Climate Change.” Psychological Science 20(3): 299–308.

26  For instance, Thomas Schelling has suggested that climate change communicators may have to exaggerate the threat to spur action and has implied that this may be morally permissible (Clarke 2009).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

172  Alison M c Queen Calhoun, C. (2018). Doing Valuable Time: The Present, the Future, and Meaningful Living. Oxford: Oxford University Press. Chadwick, A.  E. (2010). “Persuasive Hope Theory and Hope Appeals in Messages about Climate Change Mitigation and Seasonal Influenza Prevention.” PhD disser­ tation, Pennsylvania State University. Chadwick, A. E. (2015). “Toward a Theory of Persuasive Hope: Effects of Cognitive Appraisals, Hope Appeals, and Hope in the Context of Climate Change.” Health Communication 30(6): 598–611. Churchill, W. (1940). “Their Finest Hour.” Presented at the House of Commons, London. http://www.winstonchurchill.org/resources/speeches/1940-the-finesthour/122-their-finest-hour. Clarke, C. (2009). “An Interview with Thomas Schelling, Part Two.” The Atlantic. https://www.theatlantic.com/politics/archive/2009/07/an-interview-with-thomasschelling-part-two/21273/. Clinton, B. (1994). “Address to the UN General Assembly.” https://2009–2017.state. gov/p/io/potusunga/207377.htm. Clinton, H. (2016). “Transcript: Hillary Clinton’s Speech at the Democratic Convention.” New York Times. https://www.nytimes.com/2016/07/29/us/politics/ hillary-clinton-dnc-transcript.html. Conly, S. (2012). Against Autonomy: Justifying Coercive Paternalism. Cambridge: Cambridge University Press. Curzer, H. J. (2012). Aristotle and the Virtues. Oxford: Oxford University Press. Debenedetti, G. (2016). “Obamas Combine Forces with the Clintons to Fend off Trump.” Politico. https://www.politico.com/story/2016/11/obama-clinton-trumpphiladelphia-230917. Descartes, R. (2015). The Passions of the Soul and Other Late Philosophical Writings. Oxford: Oxford University Press. Ereaut, G. and N. Segnit (2006). Warm Words: How Are We Telling the Climate Story and Can We Tell It Better? London: Institute for Public Policy Research. Faucher, L. and C. Tappolet (2002). “Fear and the Focus of Attention.” Consciousness and Emotions 3(2): 102–144. Feinberg, M. and R. Willer (2011). “Apocalypse Soon? Dire Messages Reduce Belief in Global Warming by Contradicting Just-World Beliefs.” Psychological Science 22(1): 34–38. Feldman, L. and P. S. Hart (2018). “Is There Any Hope? How Climate Change News Imagery and Text Influence Audience Emotions and Support for Climate Mitigation Policies.” Risk Analysis 38(3): 585–602. Furedi, F. (2005). Politics of Fear. London: Bloomsbury. Glassner, B. (2004). “Narrative Techniques of Fear Mongering.” Social Research 71(4): 819–826. Goodin, R. E. (1980). Manipulatory Politics. New Haven, CT: Yale University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  173 Gordon, R. M. (1980). “Fear.” The Philosophical Review 89(4): 560–578. Gore, A. (2006). “The Moment of Truth.” Vanity Fair. http://www.vanityfair.com/ news/2006/05/climate-change-200605. Gore, A. (2014). “The Turning Point: New Hope for the Climate.” Rolling Stone. http:// www.rollingstone.com/politics/news/the-turning-point-new-hope-for-the-climate20140618. Gore, A. (2016). Al Gore: The Case for Optimism on Climate Change. https://www.ted. com/talks/al_gore_the_case_for_optimism_on_climate_change?language=en. Gorin, M. (2014). “Do Manipulators Always Threaten Rationality?” American Philosophical Quarterly 51(1): 51–61. Gourevitch, A. (2010). “Environmentalism: Long Live the Politics of Fear.” Public Culture 22(3): 411–424. Gravlee, G. S. (2005). “Aristotle on Hope.” Journal of the History of Philosophy 38(4): 461–477. Hart, P.  S. and L.  Feldman (2014). “Threat Without Efficacy? Climate Change on U.S. Network News.” Science Communication: 1075547013520239. Hart, P. S. and L. Feldman (2016). “The Impact of Climate Change: Related Imagery and Text on Public Opinion and Behavior Change.” Science Communication 38(4): 415–441. Head, L. (2016). Hope and Grief in the Anthropocene: Re-conceptualising Human– Nature Relations. Abingdon: Routledge. de Hoog, N., W. Stroebe, and J. B. F. de Wit (2007). “The Impact of Vulnerability to and Severity of a Health Risk on Processing and Acceptance of Fear-Arousing Communications: A Meta-Analysis.” Review of General Psychology 11(3): 258–285. de Hoog, N., W. Stroebe, and J. B. F. de Wit (2008). “The Processing of Fear-Arousing Communications: How Biased Processing Leads to Persuasion.” Social Influence 3(2): 84–113. Kahneman, D. (2013). Thinking, Fast and Slow. New York: Farrar, Straus & Giroux. Kennedy, B. (2020). “U.S. concern about climate change is rising, but mainly among Democrats.”PewResearchCenter.https://www.pewresearch.org/fact-tank/2020/04/16/u-sconcern-about-climate-change-is-rising-but-mainly-among-democrats. Koerth, M. and A.  Thomson-DeVeaux (2020). “Many Americans are Convinced Crime is Rising in the U.S.  They’re Wrong.” FivethirtyEight.com. https://fivethir­ tyeight.com/features/many-americans-are-convinced-crime-is-rising-in-the-u-stheyre-wrong/ Krizan, Z. and P.  D.  Windschitl (2007). “The Influence of Outcome Desirability on Optimism.” Psychological Bulletin 133(1): 95–121. Kunda, Z. (1990). “The Case for Motivated Reasoning.” Psychological Bulletin 108(3): 480–498. Kuran, T. and C.  R.  Sunstein (1999). “Availability Cascades and Risk Regulation.” Stanford Law Review 51(4): 683–768.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

174  Alison M c Queen Lamb, M. (2018). “Ethics for Climate Change Communicators.” Oxford Research Encyclopedia of Climate Science. Oxford: Oxford University Press. http://climatesci­ ence.oxfordre.com/view/10.1093/acrefore/9780190228620.001.0001/acrefore9780190228620-e-564. Lamb, M. and M.  Lane (2016). “Aristotle on the Ethics of Communicating Climate Change,” in Climate Justice in a Non-Ideal World, ed. C.  Heyward and D.  Roser. Oxford: Oxford University Press, 229–254. Leighton, S. R. (1988). “Aristotle’s Courageous Passions.” Phronesis 33(1): 76–99. Leiserowitz, A. (2006). “Climate Change Risk Perception and Policy Preferences: The Role of Affect, Imagery, and Values.” Climatic Change 77(1–2): 45–72. Leiserowitz, A., E.  Maibach, S.  Rosenthal, J.  Kotcher, P.  Bergquist, M.  Ballew, M. Goldberg, A. Gustafson, and X. Wang (2020a). Climate Change in the American Mind: April 2020. New Haven, CT: Yale Program on Climate Change Communication. Leiserowtz, A., J. Marlon, X. Wagng, P. Bergquist, M. Goldberg, J. Kotcher, E. Maibach, and S. Rosenthal (2020b). Global Warming’s Six Americas in 2020. New Haven, CT: Yale Program on Climate Change Communication. Linville, P.  W. and G.  W.  Fischer (1991). “Preferences for Separating or Combining Events.” Journal of Personality and Social Psychology 60(1): 5–23. McCright, A. M. and R. E. Dunlap (2011). “The Politicization of Climate Change and Polarization in the American Public’s Views of Global Warming, 2001–2010.” Sociological Quarterly 52(2): 155–194. McGeer, V. (2004). “The Art of Good Hope.” Annals of the American Academy of Political and Social Science 1: 100–127. McQueen, A. (2017). “Salutary Fear? Hans Morgenthau and the Politics of Existential Crisis.” American Political Thought 6(1): 78–105. Maibach, E.  W., C.  Roser-Renouf, and A.  Leiserowitz (2008). “Communication and Marketing as Climate Change-Intervention Assets.” American Journal of Preventive Medicine 35(5): 488–500. Mandel, C. (2016). “Al Gore’s New Message of Hope.” National Observer. http://www. nationalobserver.com/2016/02/24/news/al-gores-new-message-hope. Mann, M. E., S. J. Hassol, and T. Toles (2017). “Opinion: Doomsday Scenarios Are as Harmful as Climate Change Denial.” Washington Post. https://www.washington­ post.com/opinions/doomsday-scenarios-are-as-harmful-as-climate-change-denial /2017/07/12/880ed002-6714-11e7-a1d7-9a32c91c6f40_story.html. Markowitz, E. M. and A. F. Shariff (2012). “Climate Change and Moral Judgement.” Nature Climate Change 2(4): 243–247. Martin, A. (2013). How We Hope: A Moral Psychology. Princeton, NJ: Princeton University Press. Metag, J., M.  S.  Schäfer, T.  Füchslin, T.  Barsuhn, and K.  Kleinen-von Königslöw (2016). “Perceptions of Climate Change Imagery Evoked Salience and Self-Efficacy in Germany, Switzerland, and Austria.” Science Communication 38(2): 197–227.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  175 Morgenthau, H. J. (1960). The Purpose of American Politics. New York: Vintage. Moser, S. C. (2007). “More Bad News: The Risk of Neglecting Emotional Responses to Climate Change Information,” in Creating a Climate for Change: Communicating Climate Change and Facilitating Social Change, ed. S.  C.  Moser and L.  Dilling. Cambridge: Cambridge University Press, 64–80. Moser, S.  C. and L.  Dilling (2004). “Making Climate Hot: Communicating the Urgency and Challenge of Global Climate Change.” Environment: Science and Policy for Sustainable Development 46(10): 32–46. Moser, S.  C. and L.  Dilling (2011). “Communicating Climate Change: Closing the Science–Action Gap,” in The Oxford Handbook of Climate Change and Society, ed. J. S. Dryzek, R. B. Norgaard, and D. Schlosberg. Oxford: Oxford University Press, 161–174. Myers, T. A., E. W. Maibach, C. Roser-Renouf, K. Akerlof, and A. Leiserowitz (2013). “The Relationship between Personal Experience and Belief in the Reality of Global Warming.” Nature Climate Change 3(4): 343–347. Myers, T. A., M. C. Nisbet, E. W. Maibach, and A. Leiserowitz (2012). “A Public Health Frame Arouses Hopeful Emotions About Climate Change.” Climatic Change 113(3–4): 1105–1112. Nabi, R. L., A. Gustafson, and R. Jensen (2018). “Framing Climate Change: Exploring the Role of Emotion in Generating Advocacy Behavior.” Science Communication [online first]: 27. NACAIDS (1987). Grim Reaper [AIDS Advertisement]. https://www.youtube.com/ watch?v=OJ9f378T49E. Nisbet, M. C. and D. A. Scheufele (2009). “What’s Next for Science Communication? Promising Directions and Lingering Distractions.” American Journal of Botany 96(10): 1767–1778. Nordhaus, T. and M.  Shellenberger (2007). Break Through: From the Death of Environmentalism to the Politics of Possibility. Boston: Houghton Mifflin. Nordhaus, T. and M. Shellenberger (2009). “Apocalypse Fatigue: Losing the Public on Climate Change by Ted Nordhaus and Michael Shellenberger.” Yale Environment 360. http://e360.yale.edu/feature/apocalypse_fatigue_losing_the_public_on_climate_ change/2210. Nordhaus, T. and M. Shellenberger (2014). “Global Warming Scare Tactics.” New York Times. http://www.nytimes.com/2014/04/09/opinion/global-warming-scare-tactics. html. Norgaard, K.  M. (2011). Living in Denial: Climate Change, Emotions, and Everyday Life. Cambridge, MA: MIT Press. Nussbaum, M. (2018). The Monarchy of Fear. New York: Simon & Schuster. Obama, B. (2015). “Remarks at the First Session of COP21.” Presented at Le Bourget, Paris. https://www.whitehouse.gov/the-press-office/2015/11/30/remarks-presidentobama-first-session-cop21.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

176  Alison M c Queen Oettingen, G. and D. Mayer (2002). “The Motivating Function of Thinking About the Future: Expectations Versus Fantasies.” Journal of Personality and Social Psychology 83(5): 1198–1212. Ojala, M. (2012). “Hope and Climate Change: The Importance of Hope for Environmental Engagement among Young People.” Environmental Education Research 18(5): 625–642. O’Neill, S. J., M. Boykoff, S. Niemeyer, and S. A. Day (2013). “On the Use of Imagery for Climate Change Engagement.” Global Environmental Change 23(2): 413–421. O’Neill, S. J. and S. Nicholson-Cole (2009). “ ‘Fear Won’t Do It’: Promoting Positive Engagement with Climate Change through Visual and Iconic Representations.” Science Communication 30(3): 355–379. Pettit, P. (2004). “Hope and Its Place in Mind.” The Annals of the American Academy of Political and Social Science 592: 152–165. Pfau, M. (2007). “Who’s Afraid of Fear Appeals? Contingency, Courage and Deliberation in Rhetorical Theory and Practice.” Philosophy and Rhetoric 40(2): 216–237. Robin, C. (2004). Fear: The History of a Political Idea. Oxford: Oxford University Press. Rorty, A. O. (1980). “Explaining Emotions,” in Explaining Emotions, ed. A. O. Rorty. Los Angeles, CA: University of California Press, 103–126. Ruiter, R.  A.  C., C.  Abraham, and G.  Kok (2001). “Scary Warnings and Rational Precautions: A Review of the Psychology of Fear Appeals.” Psychology & Health 16(6): 613–630. Ruiter, R. A. C., L. T. E. Kessels, G.-J. Y. Peters, and G. Kok (2014). “Sixty Years of Fear Appeal Research: Current State of the Evidence.” International Journal of Psychology 49(2): 63–70. Saad, L. and J. M. Jones (2016). “U.S. Concern About Global Warming at Eight-Year High.” Gallup.com. http://www.gallup.com/poll/190010/concern-global-warmingeight-year-high.aspx. Schlozman, K. L., S. Verba, and H. E. Brady (2012). The Unheavenly Chorus: Unequal Political Voice and the Broken Promise of American Democracy. Princeton, NJ: Princeton University Press. Smith, N. and A. Leiserowitz (2014). “The Role of Emotion in Global Warming Policy Support and Opposition.” Risk Analysis 34(5): 937–948. Snyder, C. R. (1995). “Conceptualizing, Measuring, and Nurturing Hope.” Journal of Counseling & Development 73(3): 355–360. Snyder, C.  R. (2002). “Hope Theory: Rainbows in the Mind.” Psychological Inquiry 13(4): 249–275. Snyder, C. R., K. L. Rand, and D. R. Sigmon (2002). “Hope Theory: A Member of the Positive Psychology Family,” in Handbook of Positive Psychology, ed. C. R. Snyder and S. J. Lopez. New York: Oxford University Press, 257–276. de Sousa, R. (1987). The Rationality of Emotion. Cambridge, MA: MIT Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Wages of Fear?  177 de Sousa, R. (2014). “Emotion.” The Stanford Encyclopedia of Philosophy, ed. E. N. Zalta. http://plato.stanford.edu/archives/spr2014/entries/emotion/. Stern, P. C. (2012). “Psychology: Fear and Hope in Climate Messages.” Nature Climate Change 2(8): 572–573. Stoknes, P. E. (2015). What We Think About When We Try Not To Think About Global Warming: Toward a New Psychology of Climate Action. White River Junction, VT: Chelsea Green Publishing. Stoll-Kleemann, S., T. O’Riordan, and C. C. Jaeger (2001). “The Psychology of Denial Concerning Climate Mitigation Measures: Evidence from Swiss Focus Groups.” Global Environmental Change 11(2): 107–117. Sturgis, P. and N. Allum (2004). “Science in Society: Re-evaluating the Deficit Model of Public Attitudes.” Public Understanding of Science 13(1): 55–74. Thaler, R.  H. and C.  R.  Sunstein (2009). Nudge: Improving Decisions About Health, Wealth, and Happiness. Revised & expanded edition. New Haven, CT: Yale University Press. Upton, J. (2015). “Media Contributing to ‘Hope Gap’ on Climate Change.” http:// www.climatecentral.org/news/media-hope-gap-on-climate-change-18822. Walker, M. (2006). Moral Repair: Reconstructing Moral Relations after Wrongdoing. Cambridge: Cambridge University Press. Wallace-Wells, D. (2017). “The Uninhabitable Earth.” New York Magazine. http:// nymag.com/daily/intelligencer/2017/07/climate-change-earth-too-hot-forhumans.html. Wallace-Wells, D. (2019). The Uninhabitable Earth: Life After Warming. New York: Tim Duggan Books. Walton, D. (2000). Scare Tactics: Arguments That Appeal to Fear and Threats. Dordrecht: Kluwer. Weber, E. U. (2006). “Experience-Based and Description-Based Perceptions of LongTerm Risk: Why Global Warming Does Not Scare Us (Yet).” Climatic Change 77(1–2): 103–120. Williston, B. (2012). “Climate Change and Radical Hope.” Ethics & the Environment 17(2): 165–186. Witte, K. (1994). “Fear Control and Danger Control: A Test of the Extended Parallel Process Model (EPPM).” Communication Monographs 61(2): 113–134. Witte, K. and M.  Allen (2000). “A Meta-Analysis of Fear Appeals: Implications for Effective Public Health Campaigns.” Health Education & Behavior 27(5): 591–615. Yeo, S. (2014). “Lord Giddens: Hope Must Trump Fear in Climate Debate.” Climate Home. http://www.climatechangenews.com/2014/10/15/hope-must-trump-fear-inclimate-debate-lord-giddens/. Zaval, L., E. M. Markowitz, and E. U. Weber (2015). “How Will I Be Remembered? Conserving the Environment for the Sake of One’s Legacy.” Psychological Science 26(2): 231–236.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

8

Climate Change and Cultural Cognition Daniel Greco

1. Introduction Much recent epistemology concerns the social dimensions of knowledge. While traditionally epistemology has mainly been concerned with the predicament of the individual knower, considered in abstraction from his or her community, in reality much of the information we have comes not from our own isolated investi­ gations, but instead by trusting the testimony of others.1 An important task for socially situated knowers, then, is to identify knowledgeable individuals whose testimony we should seek out and rely on. Normally this task is straightforward, and doesn’t raise any deep epistemological puzzles. If my toilet isn’t flushing, I’ll ask a plumber to tell me what’s wrong, and how to fix it. And if I don’t already know a reliable plumber, there are obvious, unproblematic strategies I can use to find one. Sometimes, however, this task is less straightforward. Both descriptive ­questions—how do we decide whom to trust?—and normative ones—how should we decide whom to trust?—can be tricky.2 In this chapter, I’ll use the example of climate change to discuss a partial answer to the descriptive question of how we decide whom to trust: the cultural cognition thesis, according to which we treat individuals as knowledgeable experts on factual questions of political import only insofar as they share our cultural worldview.3 I’ll go on to ask some normative questions about cultural cognition: should we think of it as a species of irrational­ ity that must be overcome (or circumvented) if we are to communicate scientific results effectively, or should we instead think of it as an inescapable part of rational belief management? If it is irrational, why is that so? Answering these normative questions will require getting clearer on how and to what extent cul­ tural cognition contrasts with other, putatively distinct strategies for learning 1  See, e.g., Coady (1992); Lackey (2008). 2  See Goldman (2001) for an instructive treatment of some of the difficulties involved in sat­is­fac­ tor­ily answering the latter question. The latter parts of this chapter can be seen as an attempt to apply some of the general lessons Goldman suggests. 3  Kahan and Braman (2006) introduce the concept of cultural cognition, and much of Kahan’s research since has been related to it.

Daniel Greco, Climate Change and Cultural Cognition In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Daniel Greco. DOI: 10.1093/oso/9780198796282.003.0009

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Cultural Cognition  179 from testimony, both rational (e.g., Bayesian conditionalization) and irrational (e.g., confirmation bias). Before introducing the cultural cognition thesis, how­ ever, I want to head off a potential worry. Climate change might seem like a poor example to use to explore normative questions about deference to experts, because the answers can seem boringly, frustratingly easy. For instance, Elizabeth Anderson (2011: 153) writes that “information needed for laypersons to make sound second-­order judgments of the trustworthiness of testifiers about global warming is readily available. The criteria for making such judgments are evident and easy to apply to information on the Web.” While there are certainly residual social and psychological questions about (1) why global warming is nevertheless still controversial, at least in the United States, and (2) how to best change people’s minds about global warming,4 epistemological questions about what we should take the relevant facts to be in the first place can seem philosophically uninterest­ ing—akin to questions about how to find a good plumber. Whether previous worry is apt depends on just what questions about climate change we’re considering. If the question is just whether anthropogenic global warming (AGW) is real—whether temperatures are rising, and will continue to rise, due to past and present human activity—then the worry is on point; not only is this question uncontroversial among those who might plausibly belong to the community of relevant experts, but finding out that it is so is not a difficult matter. On this point, Anderson is persuasive. But questions about the magnitude and costs of AGW, as well as questions about the costs of policy responses to AGW, are more controversial. To take one significant example, some climate scientists have proposed various measures to mitigate the effects of AGW, such as releasing sulfate aerosols into the stratosphere, with the aim of scattering sunlight into space and thereby reducing the amount that reaches and warms earth.5 Such “geoengineering” proposals have been harshly criticized by environmental groups, on the grounds that they are likely to have unintended harmful side effects, as well as to distract political actors from the necessity of reducing carbon emissions.6 4  See, e.g., Oreskes (2011) and Jost (2015) for some discussion of these social and psychological questions. 5  This technique, and more generally any effort aimed at reducing AGW by scattering or reflecting sunlight, is known as Solar Radiation Management (SRM). Along with Carbon Dioxide Removal (CDR), which aims at removing carbon dioxide from the atmosphere, it is one of the two main geoen­ gineering strategies that have been proposed. See Keith (2013) for sympathetic explanation and discussion. 6  For example, the ETC Group’s “Hands Off Mother Earth” campaign, which will be discussed later. I don’t mean to overstate the extent of the controversy—few if any credible voices are calling for geo­ engineering measures as a replacement for reducing emissions, rather than a supplement—they tend to stress that even in the best case scenarios for political action to reduce emissions, temperatures may still rise quite a good deal, making additional responses worth considering. This is the stance taken by Keith (2013).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

180  Daniel Greco Concerning the issue of whether geoengineering7 should be taken seriously as a response to AGW, or dismissed as a dangerous distraction, normative questions about whom to trust are quite tricky. Moreover, as we’ll see, responses to geoengi­ neering proposals provide a nice example of evidence for the cultural cognition thesis. In the next section I’ll introduce that thesis, as well as discuss its particular bearing on responses to geoengineering proposals.

2.  Cultural Cognition The cultural cognition thesis says that individuals’ cultural commitments, includ­ ing values, are explanatorily prior to their beliefs about matters of fact. As Kahan and Braman (2006: 148) put it, “individuals accept or reject empirical claims about the consequences of controversial policies based on their vision of a good society.” In general, individuals will think that the policies that fit naturally into their cultural worldviews have independently good consequences.8 For example, concerning the death penalty, those who favor (oppose) it on moral/cultural grounds will go on to believe that it is effective (ineffective) in deterring crime (Kahan  1999). Moreover, similar dynamics influence perceptions of who is a trustworthy expert. To the extent that one’s cultural commitments fit with a policy of relatively unrestricted gun ownership (gun control), one will regard as more knowledgeable and trustworthy (ignorant and unreliable) putative “experts”— who are reported to hold that laws permitting “concealed carry” decrease crime (Kahan et al. 2011). Of relevance to the present chapter, Kahan et al. report simi­ lar results concerning perceptions of who is an expert about global warming, and nuclear power. While the sizes of cultural cognition effects vary from case to case, they are often substantial.9 To get a clearer picture of just what cultural cognition 7  The National Academies of Sciences, Engineering, and Medicine have recommended that the term “geoengineering” be replaced with talk of “climate intervention” (NRC 2015). While their argu­ ments for this proposal strike me as cogent, the terminological change hasn’t yet been widely adopted, so I use the older term to avoid confusion. 8  I talk here about “cultural worldviews”, but in most of Kahan’s research he talks more specifically about individualist/hierarchical worldviews and communitarian/egalitarian ones. The questions used to measure worldview are available in the appendix of Kahan et al. (2011). 9  For instance, Kahan et al. (2011: 157) report that: a plurality of hierarchical individualists (47%) perceived that most expert scientists agree, a ­plural­ity of egalitarian communitarians (47%) that most expert scientists disagree, that permitting citizens to carry handguns in public reduces crime. Only 10% of hierarchical individualists perceived that expert scientists disagree with this position, and only 12% of egalitarian communitarians that they agree with it. They report similarly large effect sizes concerning perceptions of scientific consensus concerning global warming: Solid majorities of egalitarian communitarians perceived that most expert scientists agree that global warming is occurring (78%) and that it has an anthropogenic source (68%). In contrast, 56% of hierarchical individualists believe that scientists are divided, and another 25% (as opposed to 2% for egalitarian communitarians) that most expert scientists

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Cultural Cognition  181 is supposed to be, and how it is supposed to be distinctive, we can contrast it with the more familiar and general phenomenon of confirmation bias. Confirmation bias is the tendency to interpret new information as confirming the beliefs one already has, and to generally conduct one’s inquiries in ways likely to reinforce, rather than undermine, one’s prior opinions.10 This isn’t entirely adequate as a characterization of a form of bias, where that is construed as a pejorative. Some form of methodological conservatism is widely viewed as good epistemic prac­ tice; extraordinary claims require extraordinary evidence, and sometimes it can be quite appropriate, when faced with phenomena that provide a prima facie chal­ lenge to a firmly entrenched view, to look for ways of interpreting those phenom­ ena so that they fit with the firmly entrenched view.11 But for present purposes we can set aside the problem of distinguishing confirmation bias from sensible con­ servatism; for now, the question is why, or whether, we should think of cultural cognition as anything other than a tendency to stick to one’s preexisting beliefs, even in the face of apparent counterevidence. After all, the subjects in Kahan et al. (2011) likely had prior views about the effectiveness of gun control, the riskiness of climate change, the safety of nuclear waste disposal, etc. If you’re faced with evidence that a putative expert rejects the views you currently hold, but because you’re subject to confirmation bias you’ll try to interpret new evidence in ways unthreatening to the views you already hold, then deciding that this “expert” is in fact no such thing is a natural response.12 Kahan (2016) considers this response, and takes pains to explain how a cultural cognizer will reason differently from a victim of confirmation bias. Roughly, the idea is as follows. If confirmation bias is what explains the results above, then we should expect it will be very hard to change people’s minds about, e.g., the riskiness of global warming. Once someone has an opinion, they’ll inter­ pret all further evidence so as to reinforce, or at least not undercut, that opinion. But if cultural cognition is what’s responsible, then there is a potential path to persuasion. If the “cultural meaning” of an opinion can be manipulated—if sub­ jects can be provided with information that makes the opinion seem more con­ genial to their cultural worldview than it previously did—then it should be possible to change their minds. This is just what Kahan et al. (2015) found evidence for. In their study, subjects read a “Nature Science” article—presented as a stand-­alone publication, but in disagree that global temperatures are increasing. Likewise, a majority of hierarchical indi­ vidualists, 55%, believed that most expert scientists are divided on whether humans are causing global warming, with another 32% perceiving that most expert scientists disagree with this conclusion. 10  See Kahneman (2011) for discussion. 11  This idea is often associated Kuhn (1962). 12  See, e.g., Sunstein (2006) who argues that “cultural cognition” is explicable entirely in terms of the familiar set of phenomena known as “bounded rationality”, of which confirmation bias is a prom­ in­ent example. Kahan and Slovic (2006) reply directly, but the more relevant aspect of the response for present purposes is discussed in Kahan (2016).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

182  Daniel Greco fact composed of passages from genuine papers in Nature and Science—­describing various risks associated with climate change, though not advocating for any par­ ticular policy response to those risks. The general thrust of the piece was bleak; it stated that even with “strenuous efforts to reduce emissions, atmospheric CO2 concentrations will rise beyond 450 ppm, the level commonly thought to be the maximum safe limit.” Before reading the article, however, they read one of three different pieces—two framed as responses to the Nature Science study, and a con­ trol piece on an unrelated topic. One response called for stricter limits on carbon emissions, while the other called for greater investment in geoengineering research as an alternative to stricter emissions limits. Subjects then answered various questions about their opinions of the validity of the Nature Science study, as well as the risks of climate change more generally. Kahan et al.’s hypothesis was that, depending on subjects’ cultural worldviews—measured on two scales, from individualist to communitarian and hierarchical to egalitarian—they’d respond to the experimental manipulation differently. In particular, subjects with hier­arch­ ic­al individualist worldviews tend to have very different attitudes towards com­ merce and industry than subjects with egalitarian communitarian worldviews. Hierarchical individualists tend to value commerce, and to be skeptical of legal restrictions on commercial activities. Egalitarian communitarians, by contrast, “see commerce and industry as sources of unjust disparity and symbols of nox­ ious self-­seeking, and thus readily credit assertions that these activities are haz­ ardous and therefore worthy of regulation” (Kahan et al. 2015: 194). To the extent that the cultural meaning of the Nature Science study is that it favors more limits on commerce and industry—as subjects who read the response piece advocating for stricter emissions caps will feel—hierarchical individualists should tend to be suspicious, while egalitarian communitarians should tend to be sympathetic.13 But if hierarchical individualists interpret the cultural meaning of the Nature Science study differently—which was the purpose of their reading the geoengi­ neering piece—then they might be less resistant to crediting it. As Kahan et al. explain: Researchers who study the impact of narrative on cultural cognition posit that hierarchical individualists assess information against a story template that valor­ izes the use of human ingenuity to overcome seeming natural limits on com­ merce and industry ... By featuring how an innovative technology might itself be used to offset the damaging byproducts of technological innovations on the environment, geoengineering, we posited, would evoke this theme and related affective sensibilities.  (Kahan et al. 2015: 200)

13 Of course, there are also hierarchical communitarians, and egalitarian individualists. Their responses didn’t contrast as sharply with one another as did those of the hierarchical individualists and egalitarian communitarians, though.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Cultural Cognition  183 Sure enough, after reading the geoengineering response piece, hierarchical indi­ vidualists were significantly more likely to credit the Nature Science article about risks of climate change than after reading the emissions limits response piece.14 Perhaps surprisingly, this effect was symmetrical—after reading the geoengineer­ ing response piece, egalitarian communitarians were significantly less likely to credit the Nature Science article than after reading the emissions limits response piece. All this would be difficult to explain, Kahan argues, if it were just con­firm­ ation bias at work (Kahan 2016). Individuals subject to confirmation bias would form firm opinions about climate change and would then resist changes to those opinions, whether prompted by additional evidence of a familiar sort, or ex­pos­ ure to the sorts of response pieces involved in Kahan et al.’s study. The fact that exposure to those responses did prompt people to modify their opinions suggests that confirmation bias isn’t the whole story. Before moving on, let’s take stock. Part of my aim in summarizing the above research was to draw attention to a partial answer to the descriptive question of how we decide whom to trust on controversial matters of political import—we use cultural cognition. But I also hope it makes salient the trickiness of related normative questions. In particular, to the extent that (a) there are controversial factual questions about the efficacy of geoengineering proposals and the danger of their side effects, and (b) our views about whom to trust on these matters are likely shaped by cultural cognition,15 it’s natural to start asking normative ques­ tions about the rationality of cultural cognition. If we have strong initial reactions to geoengineering proposals—as a non-­trivial proportion of the public does when exposed to them (Mercer et al. 2011)—should we distrust those reactions, once we get evidence that they were likely shaped in this way?16 To shed some light on this question, I’ll start by pushing back on Kahan’s argu­ ment that cultural cognition can be cleanly distinguished from confirmation bias. Having done that, I’ll go on to argue that, at least in the relevant cases, it’s also hard to distinguish cultural cognition from simple Bayesian updating, with cer­ tain sorts of priors.

3.  Values or Beliefs? For reasons already explained, Kahan (2016) makes a compelling case that a cer­ tain form of confirmation bias—a bias towards maintaining one’s opinion about 14  Though they still credited the article a lot less than egalitarian communitarians did. 15  To be clear, I don’t know of any studies that directly look for relations between cultural world­ views and opinions on geoengineering—Kahan et al. (2015) never directly ask people their opinions on geoengineering, and while Mercer et al. (2011) provide helpful evidence concerning public opin­ ions on geoengineering, they don’t look at the relationship between those opinions and the cultural worldviews that figure in Kahan’s line of research. 16  Perhaps, e.g., for reasons of the sort discussed by Ballantyne (2012) and Vavova (2018).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

184  Daniel Greco the validity of studies suggesting serious risks from climate change—cannot explain the phenomena described in Kahan et al. (2015). But it would be too quick to conclude from this that such phenomena cannot be explained by any form of confirmation bias. Suppose we think of the subjects as predisposed towards maintaining not their specific beliefs about climate change risks, but their more general beliefs about cultural groups and trust­ worthi­ness. How might this more general disposition explain the behavior of the subjects in the study? Suppose hierarchical individualists tend to distrust en­vir­ on­men­tal­ists—i.e., they think that environmentalists tend to be wrong about the factual issues on which they take distinctive positions. This distrust will lead them, upon reading about how environmentalists take climate change risks ser­ ious­ly, to take those risks less seriously. But when reading that geoengineering advocates take climate change risks seriously, we might naturally expect the opposite reaction. Geoengineering—especially as presented in the study in Kahan et al. (2015), with pictures of giant carbon scrubbers being erected in an other­ wise pristine wilderness, and futuristic aircraft spraying chemicals into clouds— seems like exactly the sort of policy that environmentalists would hate. So when hierarchical individualists learn that among those who take climate change ser­ ious­ly are people who are not culturally associated with environmentalists— indeed, people who propose measures of the sort environmentalists reject—they get reason to take climate change more seriously themselves. And the opposite will hold for egalitarian communitarians. Once they learn that among the people who take climate change seriously are people who would propose such prima facie anti-­environmentalist measures, they take climate change (a bit) less ser­ ious­ly themselves. The crucial point here is that, on the above interpretation, the phenomena are explained as a consequence of well-­entrenched beliefs; values needn’t be playing a fundamental cognitive role. Kahan et al. (2015) isn’t the only study to which Kahan (2016) appeals to distinguish confirmation bias from cultural cognition. But other studies are vulnerable to a similar reinterpretive strategy. For example, in protest, in a between-­subjects design, subjects saw a video that they were either told was of an anti-­abortion protest outside an abortion clinic, or a protest against the “don’t ask, don’t tell” policy, outside a campus recruitment center (Kahan et al. 2012). They were then asked questions about whether the protesters “blocked” and “screamed in the face” of pedestrians, and more generally whether the pro­ testers’ conduct was aimed only at persuasion, or instead intimidation. Subjects’ responses were significantly mediated by their cultural worldviews. For example, hierarchical individualists were significantly more likely to see “blocking” when the protest was supposedly outside a military recruitment center than when it was outside an abortion clinic (83% vs. 62%). Egalitarian communitarians displayed the opposite pattern, seeing more blocking outside the abortion clinic (56% vs 35%) (Kahan 2016: 29).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Cultural Cognition  185 Kahan (2016) argues that these results cannot be explained as the result of overly sticky beliefs (confirmation bias), because the beliefs in question were an artifact of the experiment, and didn’t even exist prior to it—before they saw the video, subjects had no views about the (fictional) protest, or the conduct of the demonstrators therein. But subjects did, presumably, have general beliefs about behavioral tendencies of members of cultural groups, which might have influ­ enced the specific beliefs about the protesters they went on to form. Suppose, plausibly, hierarchical individualists think that anti-­military protesters tend to be unruly and rude, while anti-­abortion protesters tend to be fervent, but respectful. And suppose egalitarian communitarians think that anti-­abortion protesters tend to be heartless and judgmental, while anti-­military protesters tend to be passion­ ate, but peaceful. Then after watching a three and a half minute video, we might expect that these standing views would influence subjects’ recollection and in­ter­ pret­ation of ambiguous visual evidence.17 Just how this influence might work is a complex, and much-­studied topic.18 But what matters is that, on this in­ter­pret­ ation, values enter into the picture only in a derivative way; the phenomena are explained by subjects’ general beliefs about how people with values that differ or align with their own tend to behave. I’ve just argued that the grounds Kahan (2016) offers to think that the phe­ nomena he and his coauthors have identified cannot be explained by con­firm­ ation bias are inadequate. But in fact, the response is more general—nothing in my response depends on the idea that the general beliefs that play a role in explaining such phenomena be biased in some normatively loaded sense. Given an orthodox subjective Bayesian view, hierarchical individualists might be per­ fectly rational in harboring a general distrust of environmentalists, just as cultural egalitarians might be perfectly rational in trusting them.19 And on objectivist ver­ sions of Bayesianism, at least some of these general beliefs about which groups are trustworthy might be rational, though perhaps not all. Once we think of these phenomena as potentially explained by straightforward updating on general beliefs, rather than by some cognitively distinctive process in which values are usurping the roles typically reserved for beliefs, should we continue talking about some distinctive species of “cultural cognition”, and asking questions about its normative status? In the next section I’ll explore some reasons to think the answer might still be “yes”. 17  On the point about ambiguity, there was relatively less polarization on more clear-­cut questions such as whether protesters spat or shoved pedestrians (Kahan 2016: 29). 18 In particular, given a strong “cognitive penetrability” thesis, subjects’ background cognitive states, such as beliefs about who tends to be rude or violent, might directly affect their visual experi­ ences upon seeing a video like the one in the study. While such theses have been widely endorsed, there are some strong grounds for skepticism—see Firestone and Scholl (2015). But even skeptics of strong cognitive penetrability theses grant that background cognitive states often modulate attention, in ways that have significant influences on perceptual judgments and recollections of those judgments. 19  See Howson and Urbach (1996) for a canonical defense of this view.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

186  Daniel Greco

4.  Cultural Cognition and Coincidence The core idea of cultural cognition is that it involves values taking some sort of cognitive priority over beliefs. In the previous section I explored how, in the explanations Kahan offers, general beliefs could take the place of values, appar­ ently depriving values of any cognitively distinctive (and potentially suspicious) role to play. In this section, however, I’ll try to make a case that even if this is right, the sort of general beliefs involved might nevertheless be an appropriate target of normative criticism, due to a kind of dependence on values. To see why, it may help to return to (putative) examples of cultural cognition that directly involve views about public policy, rather than whom to trust about public policy. For example, recall how those whose values oppose (favor) the death penalty tend to believe it lacks (has) a significant deterrent effect, or the analogous relation between holding values congenial or hostile to gun ownership, and beliefs about consequences of concealed carry laws.20 These correlations seem (to me at least, and I hope to the reader) like plausible symptoms of some kind of irrationality; to the extent that I notice such correlations in my own body of belief, I get worried, and suspect that I’ve succumbed to some sort of bias. Why should this be? A natural thought is that, to the extent that this pattern holds quite generally for a subject—that whichever policies she favors on basic values grounds, she also thinks have independently good consequences—this should seem, from the subject’s point of view, like a striking, unexplained coincidence. Suppose the death penalty is both fundamentally unjust, and applied in a racially discriminatory manner.21 Neither of these properties would plausibly figure, however, in any explanation of its deterrent effects; if it lacks a significant deter­ rent effect, the explanation for that absence won’t have to do with its fundamental injustice, or its racially discriminatory application.22 So if one believes both that it is fundamentally unjust and applied in a discriminatory manner, and that it lacks a deterrent effect, one believes a coincidence. In the remainder of this section, I’ll argue as follows. First, I’ll offer a (rough) analysis of the notion of a coincidence. Then, with this analysis in hand, I’ll argue that (1) the general beliefs that would figure in explaining cultural cognition would typically be beliefs in coincidences, and (2) there’s an important sense in which beliefs in coincidences are epistemically suspect. I’ll take the following as a working analysis of the notion of coincidence: P & Q is a coincidence if the best explanation of P & Q is just the best explanation of P,

20  On the former, see Kahan (1999). On the latter, see Kahan et al. (2011). 21  See, e.g., Baldus et al. (1997). 22  This is a bit quick. For example, if the death penalty is never applied to members of some racial group, and this is widely known, then we should expect that it wouldn’t deter crime among members of that group.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Cultural Cognition  187 concatenated with the best explanation of Q.23 It’s not particularly important to me that this capture all cases we’d tend to characterize as coincidences—my aim is just to be able to identify a target notion of coincidence for which both (1) and (2) above hold. But I do want to say a bit about why it might seem like a natural analysis, by taking a paradigm coincidence, and a paradigm non-­coincidence, and showing how the analysis neatly fits them both. First, let’s take a paradigm coincidence. Suppose I run into two whiskey distillers on the same day. Two ver­ sions of the case: 1. They’re in town for independent reasons; one of them is visiting a relative who just had a child, while the other is here to watch a sporting event. 2. They’re both in town for the American Craft Spirits Association (ACSA) convention, and I meet them in the convention hotel.24 Case 1 clearly describes a coincidence. The analysis captures this. We can for­ mulate the target coincidence as a conjunction: I met A, who is a whiskey distiller, and I met B, who is a whiskey distiller. And the explanation of this conjunction will just be the concatenation of the explanation of my meeting A, and the explanation of my meeting B. Case 2 clearly describes a non-­coincidence. The analysis captures the contrast. In case 2, the best explanation of the conjunc­ tion won’t concatenate the best explanations of the conjuncts, since that would involve unnecessary repetition. The best explanation of the first conjunct overlaps substantially with the best explanation of the second conjunct—both ex­plan­ ations crucially involve the ACSA convention.25 Since good explanations aren’t needlessly redundant, the best explanation of the conjunction would only say once, rather than twice, that the convention was in town. Why should we be worried about believing coincidences? In general, we shouldn’t. For example, if I know there is no ACSA convention in town, or any­ thing else that would tend to attract distillers, and yet I still meet two distillers on the same day, there needn’t be anything wrong with believing this to be a coinci­ dence. Nevertheless, there is a special sort of case in which beliefs in coincidences are epistemically suspect. Suppose a subject S believes P & Q, and takes this to be

23  This is very much in the spirit of Owens (1992), who defends the view that for a conjunction to be a coincidence is for the conjuncts to lack a common cause. While my emphasis is different, it’s quite plausible that when events have a common cause, the best explanation of their conjunction will fea­ ture that common cause, and so will be more concise than a concatenation of their best explanations, individually. While there are problems with this account—see Lando (2017) for some ingenious coun­ terexamples, as well as a plausible diagnosis of the problem—it’s close enough for present purposes. 24  A true story of my experience at the 2016 central APA. 25  A more careful analysis would probably focus on degrees of coincidence. The greater the overlap in the best explanations of the conjuncts, the less coincidental the conjunction. Even case 1 would involve some overlap, and so wouldn’t be maximally coincidental. But for present purposes, we can harmlessly treat the notion of coincidence is a binary one, for the sake of simplicity.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

188  Daniel Greco a coincidence. But suppose, in addition, that the conjunction S believes P & S believes Q is itself not a coincidence; that is, the explanation of why S believes both P & Q is better than what you’d get by just tacking an explanation of why S believes P onto an explanation of why S believes Q. To the extent that one sus­ pects one is such a subject, one has reason to suspect one has succumbed to some form of bias or irrationality. The preceding was somewhat abstract, so it may help to consider an example. Suppose Don is prejudiced against Armenians, in the sense that he tends to form negative judgments of individual Armenians he meets. But suppose he doesn’t have any negative general beliefs about Armenians as a group. So Don believes, e.g., that Georgy is dishonest and that Houry is incompetent, but doesn’t think there’s any common explanation of these facts—he takes this conjunction to be a coincidence. Nevertheless, the fact that he believes both of these things is not a coincidence—the best explanation for this conjunction is that Don tends to form negative beliefs about Armenians, and Georgy and Houry are both Armenian.26 The epistemically problematic cases of beliefs in coincidences are cases where there’s no common explanation of the (putative) facts that P and that Q, but there is a common explanation of a subject’s belief in those facts. Why should such cases, in general, be symptomatic of irrationality? In general, we hope that patterns in our beliefs will correspond to patterns in the facts those beliefs are about. To the extent that they don’t, our beliefs are inappropriately unresponsive to reality, and will be inaccurate. When we have evidence that pat­ terns in our beliefs are unrelated to patterns in reality—that the explanations for why we believe what we do are unrelated to the explanations for why the facts are the way they are—we have evidence that our beliefs are inaccurate, which should force some kind of reassessment of those beliefs. It’s hard to extend these sketchy remarks into a fully general rule or test for irrationality. For instance, the prin­ ciple that we should only believe those propositions that play a role in explaining why we have the beliefs we do, is self-­defeating.27 But we needn’t endorse any such fully general thesis to think that, for ordinary, non-­foundational, empirical beliefs, it’s some kind of red flag when the best explanation of the belief is unre­ lated to the best explanation of the (putative) fact it’s a belief about. Let’s return to cultural cognition, as we’re now in a position to state more precisely what sort of epistemological worries it raises. A cultural cognizer believes some normative claim P, and some descriptive claim Q, such that P & Q, if  true, would be a coincidence.28 But there are available various plausible, 26  So by the analysis, an explanation of the conjunction that concatenated the explanations for the conjuncts would be needlessly repetitive, in that it would mention Don’s prejudice twice. 27 See Pust (2001), who attributes similar theses to Gilbert Harman and Alvin Goldman and argues, convincingly to my mind, that the theses are self-­defeating when endorsed in full generality. 28  There are also cases where both claims are descriptive, but the paradigm cases will involve nor­ mative and descriptive claims.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Cultural Cognition  189 non-coincidental explanations for why she believes both P & Q—perhaps involving wishful thinking, or cognitive dissonance reduction, or other related psychological mechanisms. To the extent that the cultural cognizer comes to appreciate her ­situation, some kind of reassessment is called for—she has evidence that there are patterns in her beliefs that are unrelated to patterns in reality. What about when the beliefs at stake aren’t directly about the consequences of public policies, but instead about whom to trust concerning those consequences? Can this sort of cultural cognition also be understood as involving belief in potentially suspicious coincidences? I think it can. The two potentially coinciden­ tal sorts of beliefs will be as follows. First, beliefs to the effect that certain putative experts’ values are the right (wrong) ones concerning the area of policy in ques­ tion. Second, beliefs to the effect that those putative experts are reliable (un­re­li­ able) about descriptive questions concerning the consequences of policies in the relevant area. I don’t mean to imply that such combinations of belief would always be beliefs in coincidences. It’s plausible that, in some cases, there will be common ex­plan­ ations of people’s reliability on evaluative matters, and on factual matters. For example, a capacity for empathy might both enable someone to appreciate certain moral facts (e.g., concerning others’ dignity, entitlement to respect, moral status, etc.), and certain non-­moral facts (e.g., concerning others’ psychological states).29 But such connections shouldn’t be treated as a default; in the absence of some explanation as to why moral and non-­moral reliability in an area should coincide, we should treat their putative coincidence as, well, coincidental.30 And insofar as we find ourselves consistently and across domains thinking that those with the right values are also the most reliable on descriptive questions, we should worry that we’ve succumbed to some form of irrationality—we’ll have noticed a pattern in our beliefs that doesn’t plausibly match a pattern in reality. This diagnosis neatly fits some cases of cultural cognition, but other cases are less obvious, and some don’t fit at all. To see how, it will help to contrast two cases. Imagine two individuals, both of whom have broadly libertarian values—they oppose most economic regulations. They also both believe that economic regula­ tions tend to stifle innovation and to increase poverty—they tend to produce changes that even non-­libertarians would recognize as for the worse. Our first libertarian, however, holds her position for fundamentally deontological reasons—she believes that, irrespective of their downstream consequences, ­

29  Though see Bloom (2016). 30  Street (2006) argues that while there are available good evolutionary explanations of our reliabil­ ity on non-­moral matters, our reliability on moral matters would be inexplicable, in the absence of a constructivist metaethics. And even in the presence of such a metaethics, our reliability on moral matters receives a very different explanation than our reliability on non-­moral matters. I mention her position not to endorse it, but just to make salient that the task of explaining our reliability on non-­ moral matters may look very different from explaining our reliability on moral matters.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

190  Daniel Greco economic regulations infringe on people’s rights, and are intrinsically wrong. Perhaps she was convinced by Nozick (1974). Our second, by contrast, has purely consequentialist, non-­rights-­based grounds for her libertarianism. Now, imagine both of them are confronted with some new proposed regulation. Both of them think it would be wrong to enact, and both of them think it would have bad con­ sequences. If my earlier claims are right, our first libertarian has grounds for sus­ picion about her beliefs. By her lights, the fact that economic regulations are both unjust and have bad consequences is a pure coincidence. The discussion above suggests that her beliefs deserve serious scrutiny, even if they may survive it. But by contrast, our second libertarian needn’t countenance any such suspicious coin­ cidence; while she thinks that economic regulations are unjust, and that they have bad consequences, the former fact is fully grounded in the latter. There’s nothing epistemically suspicious here.31 Both subjects would probably look like cultural cognizers in a social scientific survey; they’d both assent to general statements like “People who are successful in business have a right to enjoy their wealth as they see fit,” and “Government regulations are almost always a waste of everyone’s time and money,”32 and would both also agree to claims about how particular regula­ tions would likely have undesirable consequences. But these similarities conceal crucial epistemological differences. Does this mean that cultural cognition only raises epistemological worries in cases where the “values” in question that correlate with beliefs are genuinely val­ ues, rather than relatively general factual beliefs about what sorts of policies have what sorts of consequences? I don’t think so. Even in the latter case, essentially the same kinds of worries about suspicious coincidences can arise. An example may help make this clear. Suppose a political party releases a platform detailing many positions on a wide range of issues—economic policy, foreign policy, social policy, science policy, etc. Suppose, plausibly, that there is no simple rationale that would collectively justify each of these positions—after all, they’re positions on widely disparate issues. Rather, if they are all good policies, each is good for inde­ pendent reasons. Now imagine someone who’s always voted for this party, call him “Team Player”, who thinks that each of the policies in the platform is a good one, in the sense that each policy would in fact have consequences that are widely recognized as desirable (i.e., even by people of conflicting worldviews). There is at least the threat of suspicious coincidence here. Even though each of the beliefs in this case—beliefs about the likely effects of some or another policy—is strictly factual, their joint truth might look suspiciously coincidental, by Team Player’s lights. Certainly this will be true if Team Player doesn’t think that the process by

31  I’m setting aside, of course, questions about what the evidence actually suggests concerning con­ sequences of economic regulations. 32  These are items from the worldview scales used by Kahan and collaborators. See http://www. sjdm.org/dmidi/Cultural_Cognition_Worldview_Scales.html.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Cultural Cognition  191 which the party platform was generated was particularly reliable—in that case, the situation would be like somebody winning the lottery many times in a row: suspicious. But even if Team Player does think the process by which the party platform was generated was reliable, that belief might itself look like a belief in a coincidence. After all, since the party platform concerned so many diverse areas of policy, the party would have to have reliable methods for discerning facts across unrelated domains, and would have to reliably pick the policies best sup­ ported by those facts across diverse domains. While this certainly isn’t impossible, it’s the sort of thing that should make Team Player wonder—how is it that my party so consistently got things right? And to the extent that Team Player cannot come up with a good answer here, he should worry that this pattern in his beliefs—his consistently thinking that his side got things right—is produced by some kind of bias, rather than an appreciation of a genuine pattern in reality.

5. Geoengineering How should these general lessons inform our thinking about climate change and geoengineering? My aim in this section will be as follows. First, I’ll provide a very brief summary of the range of opinion on the promises and risks of geoengineer­ ing. Then, I’ll contrast how an unreconstructed cultural cognizer would react to these facts, with how we should react to them after learning to be skeptical of beliefs formed via cultural cognition. While there is a great deal of evidence about opinions—both scientific and lay—concerning general questions about the existence of anthropogenic global warming, there is relatively less evidence about the distribution of opinion on geoengineering. Moreover, in the case of lay opinion, responses differ signifi­ cantly depending on how questions are framed—perhaps unsurprisingly, given the relative lack of public familiarity with geoengineering.33 Both the Royal Society (2009) and the National Academies of Sciences, Engineering, and Medicine (NRC 2015) have released reports that provide evi­ dence of some points of scientific consensus. First, geoengineering shouldn’t be seen as an alternative to emissions reductions—at best, it would be a supplement. Second, no geoengineering proposals are well enough understood yet for imple­ mentation on a wide scale. Both reports call for additional research, both 33  Compare the results reported in Mercer et al. (2011), which show a decent amount of support for geoengineering (or at least, for geoengineering research), with those of the National Survey of American Public Opinion on Climate Change—the relevant portions of which are summarized by Borick and Rabe (2012)—which seem to show more skepticism. While the surveys were taken at simi­ lar times by similar populations, the exact questions asked were different. Moreover, Mercer et al. (2011) included information explaining the concept of solar radiation management—a particular ver­ sion of geoengineering—while the National Survey of American Public Opinion on Climate Change did not.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

192  Daniel Greco scientific—e.g., concerning the feasibility and risks of various geoengineering proposals—and political—e.g., concerning how geoengineering proposals might be agreed upon and subsequently regulated on an international scale. But even if there is agreement on those points—and there’s some reason to think that even the call for more research is controversial—it likely conceals a wide range of atti­ tudes, ranging from the very pessimistic to the cautiously optimistic, about the promise of geoengineering.34 While no major scientific body endorses either pole of pessimism or optimism about geoengineering, other groups have taken much stronger positions, both pro and con. The American Enterprise Institute (AEI)—a conservative think tank—launched a geoengineering project. While the AEI doesn’t adopt official positions on policy issues, it’s fair to say that the project was generally progeoengineering. For example, it was co-­directed by Lee Lane, who testified before the House Select Committee for Energy Independence and Global Warming that “research to learn more about geoengineering’s potential and limitations might be one of the very best ways of hedging against the larger risks and uncertainties that surround climate policy” (Lane 2008). On the other side, the Action Group on Erosion, Technology, and Concentration (more commonly known as the “ETC Group”) has launched an anti-­geoengineering campaign called Hands Off Mother Earth (HOME), opposing not only full-­scale deployment, but also preliminary testing of any geoengineering technologies.35 What will a cultural cognizer, upon learning the above, think about geoengi­ neering? One possibility, suggested by the results reported in Kahan et al. (2015), is that geoengineering all on its own—independent of facts about the distribution of opinion about it—would trigger cultural associations, depending on one’s val­ ues either positive or negative. But even in the absence of such associations, knowing that institutions like the AEI and the ETC group stand on opposite sides of the issue would likely be enough to imbue it with a cultural valence, either positive or negative depending on one’s antecedent sympathies and antipathies for conservative pro-­business groups and environmentalist groups. Hierarchical individualists, then, seeing that groups they trust (because of their generally free market stance) support geoengineering, and that groups they distrust (because of support for limits on free enterprise and entrepreneurial culture) oppose it, will 34  I don’t know of any large-­scale survey of scientific opinion on geoengineering, though in 2009 the UK’s Independent conducted a poll of 80 leading scientists, finding that 54% agreed with the need for a geoengineering-­based “plan B” in case efforts to reduce emissions fail, while 35% disagreed, and 11% said they didn’t know whether such a plan was necessary (Green and Connor 2009). See the online debate between Keith and Hulme (2013) for representatives of these two poles. 35  I don’t mean to imply that environmental groups have spoken with one voice. See the website below for a summary of the positions of the Environmental Defense Fund and those of other groups; the ETC group’s opposition is the strongest, with the Environmental Defense Fund and many other groups calling for small-­scale experimentation with geoengineering techniques: https://www.edf.org/ climate/our-­position-­geoengineering.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Cultural Cognition  193 tend to support geoengineering. Faced with disagreement among leading scien­ tists, they’ll likely modify their opinions of which scientists are really experts— taking those optimistic about geoengineering to be more qualified and knowledgeable than those pessimistic. And we would expect a symmetrical and opposite reaction from egalitarian communitarians, resulting in their discrediting those scientists who are optimistic about geoengineering, and trusting those more skeptical.36 But what if we’ve learned to distrust cultural cognition? Whatever our values, we’ll be hesitant to treat those who share (reject) our values as thereby reliable (anti-­reliable) on factual questions. So, whether we share the values of the AEI or not, we won’t think that they are thereby reliable or anti-­reliable on questions about the magnitude of the risks and benefits of geoengineering. Likewise with the ETC group. Moreover, since neither the AEI nor the ETC group have many expert climate scientists as members—where expertise is judged on the largely culturally neutral grounds of the sort discussed by Anderson (2011), involving things like advanced degrees, publication in leading journals, membership in honorary scientific societies, and the like—we won’t take their positions to have a great deal of evidential weight concerning factual questions about the risks and promises of geoengineering proposals. We will take the reports like that of the Royal Society and the National Academy of Sciences very seriously. But as already discussed, both reports are quite cautious, and neither endorses a strong position on geoengineering—either optimistic or pessimistic—that we might adopt as our own. While the reports of those two bodies likely reflect internal disagreement among their members, unlike the unreconstructed cultural cognizer we’ll lack anything we recognize as good grounds for deciding, among those members, which ones are more reliable and trustworthy than others; we won’t be able to identify the more optimistic or pessimistic members as the genuine experts. So we’ll likely end up quite uncertain about the risks and rewards of geoengineering—hoping for more research, but uncertain which (if any) geoengineering proposals will be advisable once such research is conducted. This isn’t to say that we’ll regard the cultural associations of geoengineering as completely irrelevant to questions about its risks. Insofar as a potential partial explanation for the AEI’s support for geoengineering is that they hope it will undermine support for strict emissions limits, and insofar as we agree with the aforementioned reports that strict emissions limits are necessary (regardless of whether geoengineering is or is not pursued), we may acknowledge a political risk of geoengineering to which its cultural associations are relevant. But we shouldn’t be confident about the magnitude of this political risk. Recall the results from

36  This would be the same dynamic as the one reported in Kahan et al. (2011).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

194  Daniel Greco Kahan et al. (2015), in which hierarchical individualists were less skeptical about research warning of harms from AGW after reading about geoengineering. It’s entirely possible that widespread discussion of geoengineering might increase support for strict emissions limits, by undermining the sense among AGW skep­ tics that climate change is a “hoax” cooked up merely to motivate such limits. Absent further similar research, confidence about whether geoengineering research would distract from political action on emissions is unwarranted.

6. Conclusion My aim in this chapter has been to explore the epistemological implications of a psychological thesis about how we decide whom to trust concerning matters of cultural and political significance. While I’ve used the examples of climate change and geoengineering to frame my discussion, the lessons are, I hope, much broader. Much has been written recently on the epistemology of disagreement.37 While a good deal of the discussion has focused on disagreement between “epistemic peers”—roughly, people in an equally good position to get at the truth on some matter—there’s something a bit odd about this.38 After all, in many of the most pressing, real-­life cases, including that of geoengineering, the disagreement that seems most significant is disagreement between our epistemic superiors. And whatever the plausibility of the idea that some kind of agnosticism or conciliatory response is called for when you find yourself disagreeing with your epistemic peer, it’s significantly more plausible that such responses are called for when you find your epistemic superiors disagreeing with each other. But just how common should we think it is that, on descriptive questions about the likely consequences of public policy, there is little consensus among our epi­ stem­ic superiors?39 How we answer this question will likely depend on whether we’re comfortable using cultural cognition to decide who counts as an expert. If we’re unreconstructed cultural cognizers, we’ll find relatively more agreement among those we take to be experts—because we will withhold the status of “expert” from those who don’t share our values—and we’ll therefore have an eas­ ier time identifying and deferring to a robust expert consensus. But if we’re ­skeptical about cultural cognition and manage to stop ourselves from using it, the

37  For a sampling, see the papers collected in Feldman and Warfield (2010) and Christensen and Lackey (2013). 38  This isn’t to say that all of the literature focuses on peer disagreement. See, e.g., Goldman (2001), Frances (2010), Ballantyne (2014) for some examples of papers broadly in the disagreement literature, but which focus on the challenge posed by disagreement from epistemic superiors. 39  See the discussion of “spinelessness” in Elga (2007) to get a sense for some of the complexi­ ties here.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Cultural Cognition  195 pool of people we count as “experts” will be wider, and the opinions in that pool will tend to be more varied. We’ll have a harder time making our way to firm views about the likely consequences of controversial public policies. Much psychological research suggests that refraining from using cultural cog­ nition should be difficult and unpleasant; strong opinions about controversial questions of public policy are markers of group membership and cultural identity, and if we shed them, we may feel as if we are shedding our very selves. I say, so be it! Where is it written that rationality should be easy?40

References Anderson, E. (2011). “Democracy, Public Policy, and Lay Assessments of Scientific Testimony.” Episteme 8: 144–164. Baldus, D., G.  Woodworth, D.  Zuckerman, N.  A.  Weiner, and B.  Broffitt (1997). “Racial Discrimination and the Death Penalty in the Post-Furman Era: An Empirical and Legal Overview with Recent Findings from Philadelphia.” Cornell Law Review 83: 1638–1770. Ballantyne, N. (2012). “The Problem of Historical Variability,” in Disagreement and Skepticism, ed. D. E. Machuca. New York: Routledge, 239–259. Ballantyne, N. (2014). “Counterfactual Philosophers.” Philosophy and Phenomenological Research 88: 368–387. Bloom, P. (2016). Against Empathy: The Case for Rational Compassion. New York: HarperCollins. Borick, C. and G.  Rabe (2012). Americans Cool on Geoengineering Approaches to Addressing Climate Change. Issues in Governance Studies. Washington, DC: Brookings Institution. Christensen, D. and J. Lackey (eds.) (2013). The Epistemology of Disagreement: New Essays. New York: Oxford University Press. Coady, C.  A.  J. (1992). Testimony: A Philosophical Study. New York: Oxford University Press. Elga, A. (2007). “Reflection and Disagreement.” Noûs 41: 478–502. Feldman, R. and T.  Warfield (eds.) (2010). Disagreement. New York: Oxford University Press. Firestone, C. and B. Scholl (2015). “Cognition Does Not Affect Perception: Evaluating the Evidence for ‘Top-Down’ Effects.” Behavioral and Brain Sciences 39. Frances, B. (2010). “The Reflective Epistemic Renegade.” Philosophy and Phenomenological Research 81: 419–463.

40 For helpful comments and discussion, thanks to Nathan Ballantyne, Tristram McPherson, Elizabeth Miller, and David Plunkett.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

196  Daniel Greco Goldman, A.  I. (2001). “Experts: Which Ones Should You Trust?” Philosophy and Phenomenological Research 63: 85–110. Green, C. and S.  Connor (2009). “Climate Scientists: It’s Time for ‘Plan B’: Poll of international experts by The Independent reveals consensus that CO2 cuts have failed—and their growing support for technological intervention.” The Independent, January 2. Howson, C. and P.  Urbach (1996). Scientific Reasoning: The Bayesian Approach. Chicago: Open Court. Jost, J.  T. (2015). “Resistance to Change: A Social Psychological Perspective.” Social Research 82: 607–636. Kahan, D. (1999). “The Secret Ambition of Deterrence.” Harvard Law Review 113: 413–500. Kahan, D. (2016). “The Politically Motivated Reasoning Paradigm,” in Emerging Trends in Social & Behavioral Sciences, ed. S. M. Kosslyn. New York: Wiley Online Library. Kahan, D. and D. Braman (2006). “Cultural Cognition and Public Policy.” Yale Law and Policy Review 24: 147–169. Kahan, D., D. Hoffman, D. Braman, D. Evans, and J. Rachlinski (2012). “‘They Saw a Protest’: Cognitive Illiberalism and the Speech-Conduct Distinction.” Stanford Law Review 64: 851–906. Kahan, D., H. Jenkins-Smith, and D. Braman (2011). “Cultural Cognition of Scientific Consensus.” Journal of Risk Research 14: 147–174. Kahan, D., H.  Jenkins-Smith, T.  Tarantola, C.  Silva, and D.  Braman (2015). “Geoengineering and Climate Change Polarization: Testing a Two-Channel Model of Science Communication.” The Annals of the American Academy of Political and Social Science 658: 192–222. Kahan, D. and P. Slovic (2006). “Cultural Evaluations of Risk: ‘Values’ or ‘Blunders’?” Yale Law School, Public Law Working Papers. Kahneman, D. (2011). Thinking, Fast and Slow. New York: Farrar, Straus & Giroux. Keith, D. (2013). A Case for Climate Engineering. Cambridge, MA: MIT Press. Keith, D. and M.  Hulme (2013). “Climate Science: Can Geoengineering Save the World?” Blog post. https://www.theguardian.com/sustainable-business/blog/climatescience-geoengineering-save-world. Kuhn, T. S. (1962). The Structure of Scientific Revolutions. Chicago, IL: University of Chicago Press. Lackey, J. (2008). Learning from Words: Testimony as a Source of Knowledge. New York: Oxford University Press. Lando, T. (2017). “Coincidence and Common Cause.” Noûs 51: 132–151. Lane, L. (2008). “Geoengineering as an Approach to Climate Change.” Testimony given before House Select Committee on Energy Independence and Global Warming.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Cultural Cognition  197 Mercer, A. M., D. W. Keith, and J. D. Sharp (2011). “Public Understanding of Solar Radiation Management.” Environmental Research Letters 6: 044006. National Research Council (NRC) (2015). Climate Intervention: Carbon Dioxide Removal and Reliable Sequestration and Climate Intervention: Solar Radiation Management. Washington, DC: National Academies Press. Nozick, R. (1974). Anarchy, State and Utopia. New York: Basic Books. Oreskes, N. (2011). Merchants of Doubt: How a Handful of Scientists Obscured the Truth on Issues from Tobacco Smoke to Global Warming. London: Bloomsbury. Owens, D. (1992). Causes and Coincidences. New York: Cambridge University Press. Pust, J. (2001). “Against Explanationist Skepticism Regarding Philosophical Intuitions.” Philosophical Studies 106: 227–258. Royal Society (2009). Geoengineering the Climate: Science, Governance, and Uncertainty. Royal Society Policy Document 10/09. London: The Royal Society. Street, S. (2006). “A Darwinian Dilemma for Realist Theories of Value.” Philosophical Studies 127: 109–166. Sunstein, C. (2006). “Misfearing: A Reply.” Harvard Law Review 119: 1110–1125. Vavova, E. (2018). “Irrelevant Influences.” Philosophy and Phenomenological Research 96: 134–152.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

SECTION III

C L IMAT E C HA NG E A ND IN DIV IDUA L ET H IC S

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

9

Climate Change and Individual Obligations A Dilemma for the Expected Utility Approach, and the Need for an Imperfect View Julia Nefsky

1. Introduction Unless global greenhouse gas emissions are rapidly reduced, the consequences of climate change will be catastrophic.1 What is the significance of this for how or­din­ary people ought to live their lives? Of course, government action is crucial for tackling the climate crisis. But it is also true that if enough individual people reduce their emissions, this would reduce global warming and its harmful consequences. We thus frequently hear calls to change our individual behavior: take a bike or public transit instead of a car, forgo air travel, choose vegan options over meat and other animal products, avoid single-­use plastics, buy energy-­efficient appliances, etc. The suggestion is that we ought to make these sorts of choices—we act wrongly if we do not. Something like this surely must be right. Given the facts about climate change, surely we, as individuals, have moral obligations when it comes to our emissionsproducing lifestyle choices. But there is a problem in understanding how there could be such obligations: while we together could reduce climate change harms if enough of us make low-­emissions lifestyle choices, any one ordinary individual’s actions typically do not seem to make a difference. It seems as though the consequences of climate change are not going to be different, for instance, depending on whether or not I drive to work, or whether or not I fly to Europe this summer. But if climate change harms will be just as bad give or take my individual emissions-­producing activities (driving, flying, etc.), then how can I be said to be acting wrongly, on ­climate change grounds, in doing such things? This is the ‘inefficacy challenge’.2 1  We are already seeing the effects of climate change (extreme weather, droughts, wildfires, etc.). For projections about what is likely to happen in the future see, for instance, the IPCC’s Special Report on Global Warming of 1.5° (2018), https://www.ipcc.ch/sr15/. 2  A different kind of climate-­change-­related obligation that individuals might have is to pressure governments and other organizations to change their policies. It is important to note, though, that the inefficacy challenge can arise with respect to activist activities as well. While very many signatures on

Julia Nefsky, Climate Change and Individual Obligations: A Dilemma for the Expected Utility Approach, and the Need for an Imperfect View In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Julia Nefsky. DOI: 10.1093/oso/9780198796282.003.0010

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

202  Julia Nefsky The inefficacy challenge actually runs deeper than just described. If it really won’t make a difference whether I drive to work or take the subway, not only is it unclear how I could be acting wrongly by driving, it is unclear that there is any point at all (to do with climate change) in refraining. What reason would I have to inconvenience myself in this way, if doing so won’t make a difference for the better? Elsewhere I have focused on this more basic challenge at the level of reasons,3 but my concern in the present chapter is with our moral obligations. The aim of this chapter is not to answer the inefficacy challenge for individual climate-­change-­related obligations, but rather to say something about the sort of answer we should be looking for. And I do this by way of raising a new problem for a popular approach to the inefficacy challenge: the expected utility approach. The first half of the chapter focuses on the expected utility approach. I argue that it faces a problematic dilemma. In the second half of the chapter I diagnose the problem. There we see that the dilemma arises from a very general feature of the view, and thus is shared by various other views as well. I then discuss what a solution to the inefficacy challenge needs to look like if it is to avoid the dilemma. The inefficacy challenge is not specific to climate change. It arises in any ‘col­ lect­ive impact’ context: any context in which people can collectively cause, or fail to prevent, great harm or injustice by acting in a certain way, but in which individual such acts do not seem to make a difference. While the central example of the chapter is climate change, in the final section I discuss how the point extends to other collective impact cases.

2.  A Dilemma for the Expected Utility Approach Importantly, I am approaching the inefficacy challenge without assuming any prior commitment to a general moral theory like consequentialism. That is, I am approaching the problem directly, rather than from the perspective of someone committed to a particular general theory. What I am interested in is how we can give an intuitively plausible answer to the challenge—one that delivers certain intuitively important results. As an illustration, consider Ms. Lavish: Ms. Lavish lives a high-­emissions lifestyle. She has a big home, which she keeps very warm in the winter and very cool in the summer, and which is decked out with lots of fancy energy-­consuming features, like an indoor swimming pool and gas heat lamps for her rooftop deck. She travels frequently for weekend getaways a petition or participants in a protest could make a difference to government decisions, it’s hard to believe that these decisions will go differently give or take any one ordinary individual’s signature or attendance at the protest. The discussion in this chapter applies to that context as well, but for sim­pli­ city I’m focusing my attention on obligations with respect to lifestyle choices. 3  See Nefsky (2017).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Individual Obligations  203 and vacations. She has a gas-­guzzling SUV, which she loves driving. She uses it multiple times a day—to get to the grocery store (even though it is just down the street), to get to work (even though she could easily take public transit), and so on. Indeed she loves driving her SUV so much that on most Sundays she goes for a drive along the nearby scenic highway just for fun.4 I think a skeptical answer to the inefficacy challenge (i.e. one that says that individuals do not have climate-­change-­related lifestyle obligations) is really implausible when one thinks of someone like Ms. Lavish. Surely Lavish should not be living like this, given the facts about climate change. I am interested in how we can give a plausible solution to the inefficacy challenge that delivers this intuitively important result, among others. To be clear: there might be other reasons, unrelated to climate change why Lavish should not be living like this. But for our purposes, we want to leave any such other factors aside. The question is: how can we explain why her behavior is morally problematic on climate-­change-­related grounds? One popular way of answering this is by appealing to expected utility. This expected utility approach (‘EU’) proceeds in two steps:5 first, it says that it is incorrect to say that your individual act won’t make a difference, because there is always at least some chance that it will. In the case of climate change, this step goes something like this: as emissions climb higher, there are going to be various points—or, thresholds—at which higher greenhouse gas concentrations in the atmosphere result in further harmful outcomes (like floods, droughts, hurricanes, and so on, that would not otherwise have occurred, or would not otherwise have been as severe). While it may be very unlikely that your individual emissionsproducing act (e.g. your drive to work) is the one that causes such an emissions threshold to be crossed, there is at least a chance that it does. There is a chance, in other words, that your small individual act pushes things over a tipping point, resulting in further harm that otherwise would not have occurred. So there is at least a small chance that you do make a difference for the worse by driving. The second step is that even if the chance of making a difference is only tiny, this will be a tiny chance at making a large difference for the worse. We are talking about a difference to whether or not severe storms, floods, droughts and so on, occur, or to how severe they are. These things affect large numbers of people and animals. A small chance at making a large difference in harm amounts to a substantial expected harm. As a result, we can say, on expected utility grounds, that it

4  The example of going for a Sunday joy ride in one’s gas-­guzzling SUV comes from Sinnott-­ Armstrong (2005). 5  Many philosophers have adopted EU as an approach to collective impact cases in general. For two general defenses of this approach, see Kagan (2011) and Norcross (2020). For the EU argument applied specifically to the climate change case, see Morgan-­Knapp and Goodman (2015), Lawford-­ Smith (2016), and Broome (2019). Broome (2012, 76–77) also partially appeals to it, and Hiller (2011) does too.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

204  Julia Nefsky is typically impermissible to do a higher emissions act when there are lower emissions options available: it is impermissible unless there is some sufficiently weighty expected benefit to doing it over the alternatives. Applied to Lavish’s case, EU says: take any instance of the activities described. For instance, take one of her Sunday drives along the scenic highway in her SUV. This has a substantial expected harm, because there is a chance (even if only very small) that the emissions produced will push things over some threshold, resulting in harmful outcomes that would not otherwise have occurred, or would not have been as severe. This expected harm is going to clearly outweigh the expected benefit of this drive. After all, the only expected benefit is pleasure for her, and lower-­emissions options could still be quite pleasurable, even if not as much. So, Lavish acts wrongly in driving her SUV along the scenic highway for fun that day. And, analogously, she acts wrongly every other time she does it. The exact same point applies to many of Lavish’s other activities. She is constantly making choices like this: driving when she could easily avoid it, cranking up the air conditioner, heating her indoor swimming pool, flying to yet another holiday destination, and so on, all for the sake of gains in her own pleasure, comfort, or convenience—gains that are minor when compared to the substantial expected disutility of each of these acts. Thus, the view says: Lavish is acting wrongly much of the time, on climate-­change-­related grounds. So far so good. But I am going to argue that EU faces an unhappy dilemma. There are two possibilities for the view: either it cannot really get those nonskeptical verdicts or it yields implausibly strong verdicts. Either way, it does not deliver intuitively important results that we should hope our solution will deliver. So we should be interested in finding a different approach that can do better. Beginning with the first horn of the dilemma: some have argued that the expected utility calculus, done correctly, will not actually give the desired nonskeptical verdicts. For instance, Budolfson points out that even if one’s car trip is in fact the one that causes some emissions threshold to be crossed, you only make a difference for the worse if that threshold would not have been crossed, or would not have been crossed at approximately that time, had you not taken that car trip. But that, he argues, is essentially impossible. This is because there is “an entire planet of other people” constantly engaging in emissions-­producing activities. So, if some emissions threshold is crossed at t1 with your car trip, it would still have been crossed, if not at t1 then at an imperceptible amount of time later without it.6 But this means that there is essentially no chance that any of Lavish’s acts makes a difference to climate change harms. So, since Lavish does at least gain something from these activities, the expected utility of doing them will actually be higher than that of refraining. So, she acts permissibly as far as expected utility is concerned. Kingston and Sinnott-­Armstrong make a similar argument.7 6  Budolfson (n.d.: 36).

7  Kingston and Sinnott-­Armstrong (2018: 180–181).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Individual Obligations  205 If this argument is right, then, the expected utility calculus—done correctly— actually yields the skeptical conclusion in Lavish’s case, and to the question more generally. One reply, which Broome emphasizes, is that it is not true that an im­per­cept­ ible fraction of a second means making no difference in harm. This is because the atmosphere is a chaotic system. Because of this chaos or instability, a minuscule fraction of a second difference in the timing of emissions can make a difference to what happens. For instance, it could make a difference to the location or severity of a storm. This is also known as the butterfly effect: the idea that a small dis­turb­ ance in the atmosphere can have random, unpredictable influences on weather.8 I don’t think this reply works. The chance of making a difference via the butter­ fly effect is no more of a chance of making a difference for the worse than it is a chance of making a difference for the better. So, simply delaying a thresholdcrossing by an imperceptible fraction of a second is no more likely to make the harmful outcomes more severe than it is to make them less severe. In other words, the chances of individually making a difference for the better and for the worse via the butterfly effect should really cancel each other out in an expected utility calculation.9 But a better reply comes from the fact that it is possible that there will be a peak point after which global greenhouse gas concentration levels start declining. Perhaps world leaders will make the necessary changes in policy for this to happen. If the timing of the peak happens to coincide exactly with an emissions threshold being crossed, then each individual drive will count as having made a difference for the worse. This is because it’s not true in this scenario that had your drive not happened, the threshold would still have been reached at basically the same time anyway. This peak-­emissions threshold-­crossing coincidence scenario seems highly unlikely. But, assuming it is indeed possible, this mean that there is a

8  Broome (2019: 112–113). 9  Thank you to Mark Budolfson for a discussion of this by email. Morgan-­Knapp and Goodman consider this sort of reply. They point out that a walk in the park could also affect the weather via the butterfly effect. But they say that in that case the probabilities cancel out because “you have no information to suggest that taking a walk in the park is any more likely to cause a hurricane than it is to prevent one” (2015: 185). But they claim this is not so with a drive: “the carbon dioxide emitted by the car engine does affect the process by which the sun heats the earth, and does change the level of energy in the atmosphere. Here the distribution of effects is not symmetrical. It is slightly skewed towards worse outcomes” (2015: 185). Kian Mintz-­Woo has made a similar point to me. However, I think our discussion above suggests that this way of thinking about the butterfly effect and climate change is not correct. The difference one’s drive could make with respect to the concentration of greenhouse gases in the atmosphere is, if anything at all, purely a difference of a tiny fraction of a second in timing. (This leaves aside one exception, which I discuss next.) I don’t think there is any reason to think that that sort of extremely minuscule difference in timing would be more likely to make a difference for the worse than for the better in the consequences of climate change. In other words, if it makes a difference to harmful outcomes at all, this would be purely because of chaos/the butterfly effect, which is no more skewed one way than another.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

206  Julia Nefsky tiny chance of making a difference for the worse in climate change harms via a single drive.10 Going forward, then, I am going to proceed on the assumption that what we know is that there is a chance of making a difference to climate-­change-­related harms by a single drive, or by any other ordinary emissions-­producing activity, but that this chance is extremely small.11 This vindicates the first step of the EU argument. But it seems questionable now that the second step would go through. It could very well be that, because the chance of making a difference for the worse is so small, the expected harm is not actually going to be substantial enough to outweigh the expected benefits to Lavish of her scenic drive, or her weekend getaway in Napa. If so, then the

10  I have focused on what I think is the best argument found within articles like Broome (2019), Morgan-­Knapp and Goodman (2015), and Lawford-­Smith (2016), for the claim that there is a significant chance of making a difference to climate change harms via a single ordinary activity like driving. But these authors also bring in other points to try to defend the claim. In this note, I briefly address one other such point from Broome (2019), namely that some of the harms of climate change happen through “continuous processes”, as opposed to discrete events. An example he gives of a discrete climate event causing harm is a typhoon. Examples he gives of continuous processes causing harm are: “as water tables sink, people have to do more and more exhausting work to get the water they need; as the sea level rises, arable land is gradually washed away and families grow more hungry” (2019: 112). Broome suggests that in the case of these sorts of continuous processes, it isn’t that each emissions-­ producing act has a chance of crossing a threshold for a big change in harm, but rather that each act does make a difference for the worse but just a tiny, imperceptible one. Because the tiny difference affects a lot of people and animals, though, we should count it as a non-­trivial difference (2019: 122). But I do not think this argument works. Just because the process that causes harm is continuous rather than a discrete event, this does not show that each individual drive makes a difference to how the process unfolds. The continuous process could be causally overdetermined. Also, even if we did concede that each individual act makes some imperceptible difference to the process, I do not think this establishes that each act makes a difference with respect to what matters morally. Maybe, for instance, each additional ordinary act makes the sea a trillionth of a meter higher, but this does not make a difference to the harmful effects of the rising sea level. (See Nefsky 2011: 372–377 for a discussion of why an imperceptible difference to something that causes harm does not imply a morally relevant difference in harm.) In any case, while Broome does spend time on this imperceptible harms argument, he also seems to acknowledge that his stronger argument is the expected utility argument appealing to chances of making larger differences (see Broome 2019: 125). 11  A different way of getting a chance of making a difference might be through the idea of ‘signaling’: by driving less or not at all, by avoiding air travel, by reducing one’s energy use at home, and so on, one sends a political signal—one communicates to the government, and to other people, that one thinks climate action is important, and that one is eager to cooperate in a solution (Cullity 2015: 12). While the relevant people will not necessarily pick up on these signals, they might. This could make a substantial difference for the better. But is this really a significant possibility? The chance of my making a difference to government policy by way of a signal I send when I, e.g., take the subway or forgo my vacation seems extremely small or non-­existent. As Cullity writes, “it is fanciful to think that simply reducing energy consumption itself sends an effective political signal. Joining a political lobbying movement can be a way of participating in sending such signals; simply turning the lights off is not” (2015: 13). On the other hand, it does seem that there is a chance of influencing some other ‘ordinary’ people to change their habits by way of changing our own habits. If a friend comes over, and sees various eco-­friendly products in my house, this might influence her to switch to similar products. But for many of us, the chance of having a large enough influence in this way to actually make a difference to climate change harms seems tiny. (See Nefsky 2019: section 4.3 and also Nefsky 2018: 270–272 for more on this point.) So I think the conclusion remains: we can generally assume that if there is a chance of individually making a difference via an ordinary emissions-­producing activity, it is a very small one.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Individual Obligations  207 expected utility calculus again really gives the skeptical verdicts—it would say that Lavish acts completely fine, as far as climate change is concerned.12 That, then, is the first horn of the dilemma: perhaps EU really just gives us the skeptical verdicts. But now let’s assume that the expected utility calculus does work out as described.13 So, let’s assume that this tiny chance does yield a significant enough expected harm to outweigh the benefits for Lavish of all or most of her luxurious emissions-­producing acts in an expected utility calculation. The problem now is—and this is the second horn of the dilemma—that EU cannot get that result without also getting other implausibly strong verdicts. Consider Ms. Green: Ms. Green has a very low-­emissions lifestyle. She lives in a tiny apartment, with minimal and efficient appliances. She does not use an air conditioner in the summer even though it gets quite hot where she lives, and she delays turning on her heat in the Fall until it is necessary. She gets around almost exclusively by walking, biking, and taking public transit. She does own a car—an old Toyota Camry that her grandma passed down to her. But she barely drives it. She keeps it primarily so that she can drive every once and a while to the woods (about an hour away) to go hiking. She barely travels. Her last trip was two years ago, when she went to Italy with a friend. The problem for EU is that if it is going to come out against Lavish’s driving her SUV for fun along the scenic highway, it also must come out against Green’s occasional drive to the woods to go hiking. And in general, various seemingly permissible actions of Green’s would come out to be wrong on this model. Take one of Green’s rare drives to the woods to go hiking, and compare it to one of Lavish’s many Sunday drives along the scenic highway. Things could easily be such that the numbers that go into the two expected utility calculations are the same, and thus that the two trips have the same expected utility. In particular, we can stipulate that the two trips produce the same amount of total emissions. Even if Lavish’s SUV produces more emissions per mile, we can just stipulate that Green’s round-­trip to and from the woods is longer. Thus, there would not be any difference in the probability of making a difference for the worse, or in the magnitude of that supposed potential difference. So, the two trips would have the same expected harm.

12  Kingston and Sinnott-­Armstrong (2018: 180) and Cullity (2015: 7–8) also make this point. 13  A different issue with the expected utility approach is that we do not and perhaps cannot know which side of these two possibilities it is: we do not, and perhaps cannot, know if the expected utility calculus comes out against or in favour of Lavish’s drive, for instance. Even if we can know that there is a tiny chance at making a difference, and so some expectation of harm, it is arguable we cannot know how to quantify it and weigh it up against the expected benefits. See Cullity (2015, section III) for an argument like this. See also Heal and Millner (2014). They argue that given the depth and nature of our uncertainty about various aspects of the climate change problem, the expected utility approach to decision making under uncertainty is not a good one for use in climate change policy.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

208  Julia Nefsky We can also stipulate that the expected benefit to each of going on their re­spect­ive trips would be the same: the two trips have an equally high probability of giving their respective agents an equally fair amount of pleasure. But if these factors are the same, then the two acts have the same expected utility. Of course the verdicts depend not only on the expected utility of the acts in question but also on the expected utility of the alternatives available to the agents. But there need not be any difference there either. So, the point is: if we say that the calculus comes out against Lavish driving her SUV for fun, we also need to say that it comes out against Green’s driving her car to the woods. So, on EU, if the former is wrong, then so is the latter. The same point can be made with respect to other things that Green does. Take her flying to Italy two years ago. The fact that she only travels once every few years does not enter into the expected utility calculation for that individual trip.14 So, there is no reason that taking this trip couldn’t get the same values in the expected utility calculus as one of Lavish’s many trips. So, if Lavish acts wrongly, on the expected utility view, in traveling yet again, then Green acts wrongly too in trav­el­ ing just that one time. This, intuitively, is not the right conclusion. Given that Green keeps her emissions so very low, and given that any one drive or flight is highly unlikely to make a difference for the worse (via climate change), it seems implausible that Green is violating her climate-­change-­related obligations by taking that one drive or that one rare vacation. Now, if you are starting from a point of having already subscribed, on independent grounds, to an expected utility view about what individuals ought to do in general, this probably is not going to be a convincing argument against your view. You might be happy to just follow the EU approach wherever it leads. But my starting point is different: it is one of looking for an intuitively plausible answer to the inefficacy challenge. And the point is that, if this is what you are interested in, EU does not do well—it is either is too weak, yielding the skeptical conclusion, or it is implausibly strong. But let’s consider some replies. First, advocates of EU might reply: ‘it is true that on our view, Green’s drive comes out wrong, just like Lavish’s. But it is not as though the two people are therefore moral equals. Green does way fewer of these wrongful acts. Whereas, Lavish is acting wrongly on emissions-­related grounds constantly. So, Lavish is deserving of much more criticism than Green.’ I do not think we should be satisfied with this reply. The claim that Green acts wrongly in occasionally driving to the woods for a hike, or in flying to Italy that one summer, is still implausible. This is not to say that there is no 14  Other than if the rarity itself has a relevant effect on how much pleasure she gets from the trip, which it need not. More on this point shortly.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Individual Obligations  209 climate-­change-­related reason for Green not to do such things.15 But to say that these things are impermissible for her is intuitively false. She doesn’t seem morally criticizable, or to be going at all morally astray, in making those individual choices.16 But perhaps the reply is that this intuition is mistaken, and we should not regard the conclusion that Green acts wrongly on those occasions as implausibly strong. One might say: ‘yes, it is demanding, but morality should be very demanding when it comes to stopping severe climate change harms.’ But this reply won’t do. The claim is not that morality should not demand a lot of us when it comes to stopping climate change harms. If it turned out that each time one drives, or flies, one has a fairly high chance at making a substantial difference in harm, via climate change, then it would not be implausible—at least not in the sense that I’m talking about—for morality to demand that one not drive a car just for the sake of the pleasure of going hiking, or for it to demand that one never travel for a vacation. My point is, rather, this: given the assumption that her drive, or her taking that flight to Italy, has no more than a tiny chance at making a difference when it comes to climate-­change-­related harms, and given that Green keeps her emissions so low overall, it seems unfittingly demanding for morality to forbid Green from ever doing those things.17 Perhaps a better way to reply, then, is to dispute my claims about the expected utilities. One might argue that the expected benefit for Green of any one of her trips to the woods must be much higher than the expected benefit to Lavish of 15  What reason there is depends on one’s answer to the inefficacy challenge at the level of reasons. See Nefsky (2017) for my answer (which can be combined with other answers as well). 16  To this they might reply that we need to separate whether Green is criticizable from whether she is acting wrongly. It could be that because she is so good overall, and far better than the average person, it would be inappropriate to criticize her, but that she still acts wrongly. I agree that there are situations in which the fact that someone is generally very good, or much better than average, can make criticizing a relatively small moral failing of theirs inappropriate, even though it is still a failing. For instance, suppose Nessa is normally an exemplary mother, who is extremely patient, kind, and calm, but on one occasion she snaps at her kids—getting overly angry in response to something minor. One might feel as though one cannot criticize her for snapping that one time; it would not be reasonable to criticize her, given how calm and patient she generally is and given that this is a relatively small infraction. And yet, even though criticizing her may be inappropriate, it was still wrong for her to snap on that occasion. But while I recognize that there are cases like that, Green’s is not one of them: it is not that while Green is acting wrongly in driving to the woods to go hiking we should not criticize her for it; it is that she isn’t acting wrongly in doing so. This rare drive of hers to the woods does not seem like an example of impermissible conduct. 17  A related point is that one cannot get out of the problem by opting for a less demanding version of EU. The dilemma arises just as much for a satisficing version, or a hybrid version that incorporates agent-­centered prerogatives. On a satisficing version, the problem will be that wherever you set the bar for impermissibility, if some of Lavish’s individual car trips (or travels) are going to fall below the bar, then so will Green’s. So, for a satisficing view to say that Lavish acts wrongly sometimes (which it must if it is to give a non-­skeptical answer to the inefficacy problem), it must also say that Green acts wrongly with her occasional drive or rare flight. Similarly, for a hybrid view, which allows an agent to give a certain amount of additional weight to their own projects and commitments: if Lavish’s individual car trips are going to fall below the bar, such that they are impermissible despite the extra weight the agent-­centered prerogative assigns to the benefits she gets from them, then so will Green’s.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

210  Julia Nefsky any one of her scenic SUV drives, because Lavish is doing things like this all the time. Similarly, because it is rare for Green to travel, her Italy trip must have been so much more meaningful and pleasurable to her than any one of Lavish’s trips would have been to her. So my stipulation that they get the same expected benefit from their individual acts should be rejected. And if the expected benefit for Green is high enough, this could outweigh the expected harm, making these activities permissible for Green. But the first issue with this reply is that it assumes that scenic drives, hikes in the country, and world travels always have a decreasing marginal utility, and that is not true. For these specific types of activities, whether or not they have a decreasing marginal utility within the particular sorts of choice situations under discussion will depend on the details. For Lavish, it need not be that each of her weekly drives is less pleasurable than it would be if she did it less frequently. It could easily be the opposite: because she does it weekly, the route becomes very familiar and thus the drive becomes more stress-­free and purely pleasurable. We can also imagine that the ‘ritual’ element of it adds a significant amount of pleasure. And for Green, just because her car trips to the woods are rare, we cannot assume that they have some particularly elevated importance for her happiness. Her hikes can just be one among many things she enjoys doing. So, the stipulation that Lavish and Green in fact get the same amount of expected benefit from their respective car trips on a given occasion is entirely fine. The second, more important problem with this type of reply is that the thought that it is only permissible for Green to drive to the woods on a rare occasion if she will get much more pleasure or fulfillment from doing so than Lavish gets from any one of her many trips in her SUV, and that otherwise it is impermissible for her to do so, is itself highly unintuitive. It seems as though even if nothing major is at stake for Green in whether or not she drives to the woods on a given weekend, it should be permissible for her to do so, given that she lives such a lowemissions lifestyle and given that any one such trip is extremely unlikely to make a difference for the worse. We can fill out the scenario to make the stakes for Green obviously low: imagine that there are good hikes close to home that Green can access without driving but that she likes to drive out to this further location occasionally just for a change in scenery. Or even, imagine that she could get to this particular hiking trail without using her car—she could do so by taking a two-­hour trip on a public bus. So, it is just a matter of convenience that she drives there. Even so, it seems unfitting for morality to forbid Green from enjoying this convenience on rare occasions, given that her overall emissions are so low and that her doing so is extremely unlikely to make a difference for the worse.18 18  One might propose that EU can avoid these troubles by focusing on overall patterns of activity or lifestyle choices rather than on individual actions. If you look at the expected utility of Lavish’s and Green’s lifestyles as a whole, instead of focusing on each act separately, we can get that Lavish acts impermissibly while Green acts permissibly. This is because—the argument would go—while each act

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Individual Obligations  211

3. Diagnosis EU, thus, faces a dilemma: either it is stuck with the implausible skeptical verdict that individuals do not have climate-­change-­related obligations with respect to their lifestyle choices, or it must commit to the implausibly strong verdict that even someone leading a very low-­emissions lifestyle cannot permissibly do things like drive or fly on rare occasions, if it is only for pleasure or convenience. What is it about EU that yields this result? It is, in fact, something very general about the approach—something it has in common with many other views. And this is that it approaches the question of our individual obligations in collective impact cases by trying to identify a consideration that, in effect, makes acting in the relevant ways pro tanto wrong. By ‘the relevant ways’ I mean the ways that when enough of us so act, serious harm or injustice results; that is, the ways of acting in a collective impact case that generate the inefficacy challenge. Let’s call these the ‘collectively harmful’ ways of acting. By ‘pro tanto wrong’, I mean wrong on any given occasion unless there is a sufficiently weighty competing consideration speaking in favor of so acting on that occasion. Advocates of EU might not be inclined to use the term ‘pro tanto wrong’ in describing their view, but it is a good way of capturing the very general feature of the view that I have in mind. In the case of climate change: EU claims that any ordinary emissions-­producing activity has significant expected harm, and it says that therefore any such activity will be wrong, unless there is a sufficiently weighty expected benefit to doing it. This has the effect of saying that an emissions-­producing activity is pro tanto wrong. Many responses to the inefficacy challenge have this same general nature, and as a result they face the exact same problem: if they can get non-­skeptical verdicts at all, it is at the cost of having to say, implausibly, that even someone leading a very low-­emissions lifestyle cannot occasionally drive a car for the sake of pleasure or convenience, or cannot even take the rare vacation. of Lavish’s has only a tiny chance at making any difference for the worse, her overall pattern of activity has a substantial chance because the risk is repeated so many times. Green, on the other hand, only takes the tiny risk infrequently, and so her overall pattern does not have a high expected disutility. Repeatedly taking a tiny risk of harm adds up to a substantial risk, whereas taking it only a few times does not. (See Spiekermann 2014: 76–77 for a similar idea.) However, while there might be something to this idea (see note 29 for more on this proposal), this is not the EU proposal under discussion. The authors I have in mind apply EU to individual acts, and not just to overall lifestyle choices or sets of acts. They are, as far as I can tell, working with a standard conception of EU on which it can be applied to any choice situation an agent faces, from individual acts to lifestyle choices. On this conception, there needs to be consistency between what EU says about an act viewed as part of a set and what it says about that same act viewed individually. (For a formal statement of this condition, see “Separability”, McClennan 1990: 122.) So, the EU proposal in question cannot condemn Lavish’s pattern of driving every Sunday, without saying of at least some of those individual drives that they come out as wrong on the expected utility calculus. Since any one of Lavish’s individual drives can have the same utility/disutility specifications as one of Green’s, it cannot condemn Lavish’s lifestyle or pattern over time without also condemning Green’s individual drives.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

212  Julia Nefsky Take, for example, responses that appeal to notions of complicity or participation. Such views approach collective impact cases by saying that even if it makes no difference for the worse, your act can be wrong because it amounts to participation in a system, or way of life, or set of acts—the details vary—that causes great harm or injustice. By doing a collectively harmful act (e.g. driving, or boarding a plane), you are taking part in a system that causes these bad consequences. You are thus complicit, even if you do not yourself make a difference.19 Presumably the claim is that complicity, in the relevant sense, is pro tanto wrong, rather than always all-­things-­considered wrong. Otherwise, these views wouldn’t be able to say that it is permissible to drive someone to the hospital to save them from a life-­threatening emergency. So, these views have that same general nature that I attributed to EU: they say that an emissions-­producing act is wrong on any given occasion, unless there is some sufficiently weighty consideration in favor of doing it on that occasion. They, thus, face the same problem: by stipulation, there are no weighty considerations speaking in favor of Green’s driving to the woods on the given occasion, or at least, nothing that is not also there for Lavish’s driving her SUV along the scenic highway and so nothing that this view should count as ‘sufficiently weighty’ if it wants non-­skeptical verdicts. So, if it is going to get non-­skeptical verdicts, this view must say that Green acts wrongly. This is the exact same result as with EU.

4.  The Imperfect Approach What is the alternative to having a view of that nature? It might seem as though any non-­skeptical answer would have to be like this: if it is going to get obligations not to act in the collectively harmful ways, these would have to take the form of obligations not to do so unless there is a sufficiently strong countervailing consideration in favor of doing so. What else would a non-­skeptical answer to the question look like? But there is an alternative, and one that I think is much more plausible, at least in the case of climate change. The alternative is to argue for what I am going to call an ‘imperfect obligation’: an obligation to avoid acting in the collectively harmful ways—the ways that when enough of us so act, great harm or injustice results—enough of the time. So, rather than arguing that there is a consideration

19  See, for example, Kutz (2000: ch. 6), Parfit (1984: ch. 3), and Driver (2015). The point I am about to make also goes for a view that says that emissions-­producing activities involve complicity in the sense of cooperating with someone else’s wrongful plan—for instance, cooperating with oil companies or the airline industry. For this understanding of complicity, see McPherson (2016: section 3).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Individual Obligations  213 that makes acting in such ways pro tanto wrong (that is, wrong in any instance unless there is, in that instance, sufficiently strong countervailing considerations), this alternative type of approach is to argue that there is a consideration that makes acting in that way too much of the time wrong. On such an approach it can be permissible to act in a collectively harmful way even when there is no weighty moral consideration in favor of doing so. But one must not do so too much. What counts as ‘too much’ would depend on the particular account; I am just describing a general nature that many different accounts could have.20 I am using the term ‘imperfect obligation’, but nothing hangs on this choice of terminology, and I am not claiming that my usage matches how others use the term. There is a relationship between my sense of imperfect obligations for col­ lect­ive impact contexts and what most others have in mind when they talk about imperfect duties. In particular, if we have an obligation in a collective impact case that is imperfect in the Kantian sense, this will also be imperfect in my sense. It will have the feature that I am talking about.21 But there can be views that are imperfect in my sense, without being imperfect in the Kantian sense. In any case, though, this is not important for my present purposes. When it comes to climate change, an imperfect view—in my sense of the term—would say that what one is obligated to do is keep one’s emissions low enough. It is an obligation, in other words, to reduce your emissions (if they aren’t already low enough). This sort of obligation concerns your lifestyle choices taken cross-­temporally. You must make these choices so that you do not emit too much over time. On such a view, the wrongness of Lavish’s activities is found in what she does over time, and not in what she does on any one occasion taken in isolation. So, her driving her gas-­guzzling SUV along the scenic highway a given Sunday need not be wrong. What is wrong is that she does this sort of thing regularly.22 And on 20  Whether emitting ‘too much of the time’ should be understood as a pro tanto wrong or an all-­ things-­considered wrong will depend on the details of the account. There are ways of understanding ‘too much’ such that overriding considerations are built in, and ways of understanding it such that they aren’t. Especially on the former type of account, the content of what counts as ‘too much’ will need to vary depending on the individual. e.g. it might be different for someone poor than for someone wealthy, or for someone who is an ambulance driver than for someone who is a teacher at a school reachable by public transit. 21  At least, on most interpretations of a Kantian ‘imperfect’ duty. The same goes for various other accounts of imperfect duties that are not specifically Kantian. For instance, see Hanser (2014). On Hanser’s account, “an imperfect duty is a second order reason, with requiring force, to act in accordance with first-­order justifying reasons of a certain kind often enough over time” (2014: 311). A duty that qualifies as imperfect on this account will also be imperfect in my simpler, less robust sense. For another example, see Schroeder (2014). 22  Her individual drive might be wrong, but one would not see this if one looked at it in isolation. If it is wrong, on an imperfect view, this would be derivative on its place in the overall pattern (e.g. it is past the point when one has already emitted too much overall). Alternatively, though, there can be imperfect views on which no individual act is wrong, even though the overall pattern over time is wrong. This would be the moral equivalent of Tenenbaum and Raffman’s claim that there can be ‘top-­ down irrationality’: certain patterns of activity can be “irrational without any of the momentary

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

214  Julia Nefsky such a view, Green can permissibly drive to the woods on a given occasion, even if there is no weighty reason to do so, because she can do this without violating her obligation to keep her emissions low enough. Of course, what counts as ‘low enough’ will have to be filled out. But again, this is the job of the particular account of our obligations. I’m just describing a general nature that many different accounts could have.23 The possibility of an imperfect view has often been overlooked in discussions of collective impact cases. Authors often see the task as one of looking for a consideration that makes an instance of the relevant act type wrong (in the absence of strong competing considerations), and resultingly they think that if they can’t find such a consideration then we are left with the skeptical conclusion.24 But there is at least one important view in the literature that does yield imperfect obligations, in my sense of the term. This is the view that our individual obligations in these cases take the form of obligations to do one’s ‘fair share’ of a collective obligation. Whether or not one does one’s fair share of our collective obligation to reduce harmful climate change is a matter of whether one keeps one’s overall emissions low enough, rather than being about what one does on any given occasion taken in isolation.25 One can drive for fun along a scenic highway, or to the actions that compose those patterns being irrational” (2012: 22) One could have an imperfect view in which there is top-­down wrongness: a certain pattern of activity (e.g. Lavish’s) can be wrong without any of the momentary actions that compose it being wrong. I give an example of such a view in note 29 below. 23  One might wonder whether the notion of an imperfect obligation that I am invoking really marks off something distinctive. Is a so-­called ‘imperfect obligation’ not to act in a collectively harmful way really just a perfect obligation not to act in that way too much? I am not going to delve into the details of whether or not this recasting works, because even if it does work, this is fine for me. It is not important for the point I am making that there is a deep, irreducible distinction between these two types of obligations. What is important is rather that there are two different ways a non-­skeptical answer to the inefficacy challenge could go: (1) it could say that acting in collectively harmful ways is pro tanto wrong, or (2) it could say that while there are restrictions for individuals on acting in such ways, such acts are not pro tanto wrong—it can be permissible to do such things, even when there is no weighty moral consideration in favor of doing them. These are two different ways an answer to the problem could go, even if there is no deep category difference in the kind of obligation invoked. 24  This is implicit in the possibilities that many authors consider, and the way they talk about them. But for a more explicit example, see Sinnott-­Armstrong (2005). There Sinnott-­Armstrong says that to investigate whether we have global-­warming-­related obligations as individuals with respect to our lifestyle choices we can focus on a test case: is it wrong to drive a gas-­guzzling SUV just for fun on a particular sunny Sunday afternoon? He looks at many different attempts to explain why this drive would be wrong, and argues against each one, concluding that it is not wrong. He draws from this the skeptical conclusion: individuals do not have obligations with regard to their lifestyle choices and emissions. It is governments, not individuals, who have an obligation to do something about climate change (2005: 311–312). But once one recognizes the possibility of an imperfect view, one sees that this skeptical conclusion about individual obligations does not follow from the arguments about the test case. Even if it is true that an individual joy ride in a SUV is not wrong, this does not mean that individuals are not obligated to reduce their emissions. An imperfect obligation could even require drastic reduction in emissions, without making any one such drive taken in isolation wrong. 25  One author who explicitly points out this difference between EU and the fair share approach is Hiller. Hiller is an advocate of EU, and notes explicitly that on his view a Sunday drive is pro tanto wrong, whereas on a ‘equal shares’ approach it is not (Hiller 2011: 360). Hiller does not see this as a disadvantage of his view, but I am arguing that it is a feature that makes it implausible.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Individual Obligations  215 woods to go hiking, without violating this obligation, and this is so even if there is no weighty consideration in favor of such a drive.26 I give the fair share view, though, just as an example of an existing view that has the general nature that I am talking about.27 I am not specifically advocating for it, and indeed, I myself am inclined to take a different approach.28 But since my goal in this chapter is only to argue that we should aim for a view that yields an imperfect obligation, and not to argue for any one particular such view, I will not get into this here.29

5.  Extension to Other Cases My focus has been on the nature of our obligations as individuals with respect to our lifestyle choices and climate change. I have argued that we should aim for a view in which such obligations are imperfect. I think the same conclusion is correct in various other collective impact contexts as well. Suppose, for instance—as 26  Various authors take a fair share approach to obligations in collective impact cases. Baatz (2014) is one such author who explicitly talks about how the fair share duty he is advocating for is imperfect in the Kantian sense. Baatz argues that there is a collective duty to limit global emissions, and that as a result each person has a duty to not exceed their fair share of the global emissions budget. But how much is one’s fair share depends on the details of one’s circumstances, and all we can really say in general is that each person is morally required to reduce their emissions “as far as can reasonably be demanded of them” (Baatz 2014: 10). This is an imperfect duty in the sense that it is not clear what exactly it requires, and there is some latitude for the individual to decide. As I mentioned, an imperfect duty in this Kantian sense is also an imperfect duty in my sense: on such a view an individual emissions-­producing activity need not be wrong, even if there is nothing substantial speaking in favor of it. Baatz notes this feature of his account. However, he also argues that it is very unlikely that, in the actual world we live in, a drive in a gas-­guzzling SUV would not involve exceeding one’s fair share (see Baatz 2014: 12). 27  A related but different example can be found in Schroeder (2014: 582–583). 28  Very briefly: I’ve argued elsewhere that an act can make significant progress toward changing an outcome even if it cannot (or is extremely unlikely to) individually make a difference. I argue, for instance, that by refraining from driving I make small but real progress toward less harmful climate change outcomes, regardless of whether the harms of climate change will be different depending on my having done so. (See Nefsky 2017.) If I am right about this, then I think this can ground an obligation. I ought to refrain from driving because doing so will make progress toward preventing severe climate change harms. So, in my view, being highly unlikely to make a difference does not undercut having an obligation grounded in what you individually can do toward changing the outcome—in other words, in the instrumental progress you can make. But what it does do is affect what it makes sense for this obligation to be—namely, it makes it an imperfect obligation. Because things do not, or are extremely unlikely to, depend on any one individual such act of yours, the obligation is imperfect: it is an obligation to make such choices ‘enough’ of the time, rather than to do so each time. On my view, fairness to other agents who are doing something toward preventing climate change harms might be a relevant factor in setting how much counts as enough, but—contrary to the fair share approach—fairness to such agents is not the primary grounds for the obligation to reduce one’s emissions. 29  A different possibility for how one might get an imperfect view is that one could argue that, in collective impact cases, we should apply expected utility to patterns of activity over time, or overall lifestyles, but reject the idea that the expected utility analysis works when applied to individual acts. On such an approach, one could say that Lavish’s pattern over time is wrong because she is taking a minuscule but real risk of harm frequently, whereas Green’s is not because she is only taking the risk very rarely. But it would say that if you take any individual act of Lavish’s or Green’s it is not wrong. This would be a major departure from the typical EU proposal, as explained in note 18.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

216  Julia Nefsky might be the case—that a large-­scale change in demand from conventional to fair trade coffee would greatly reduce the exploitation of coffee farmers in developing countries, but that any one purchase won’t, or at least is extremely unlikely to make a difference. The relevant decisions in the supply chain (about labor practices, who is employed on various farms, and so on) will almost certainly not go differently depending on any one ordinary individual’s coffee purchases. Supposing these claims are true, this is a collective impact case. A solution to the inefficacy challenge that yields an imperfect obligation here would say: each of us ought generally to opt for fair trade coffee, but the occasional purchase of nonfair-­trade coffee can be permissible (even if there is no significant moral consideration in favor of it). Occasionally having a cup of coffee at a place that just sells ‘coffee’ (no further information available) is okay. But making a daily habit of it is not. It seems to me that this is the right sort of response to aim for. A response that condemns even the occasional purchase of conventional coffee (in the absence of some unusual weighty reason for the purchase) seems unreasonable, given that such a purchase almost certainly won’t make a difference for the worse to anyone. But a skeptical view which says consumers are under no moral constraints when it comes to such purchases seems like it cannot be right either, if it is really true that collective consumer choices in this domain have a major impact on the lives of thousands of workers. An imperfect account would be ideal. Consider next the idea that most middle-­class individuals in wealthy countries have an obligation to donate to efforts to aid the global poor. While this is often overlooked, the inefficacy challenge comes up here. It is often assumed that if one took the money that one would otherwise choose to spend on, say, a new pair of pants or movie tickets, and donated it instead to a famine relief fund, one would thereby be “preventing another person from starving.”30 But this is doubtful. As Cullity says, faced with a famine, a relief agency would likely: make a large-­scale calculation of the size of the overall need, the amount of money they are likely to be able to raise from various sources to pay for it, and the extent to which it makes sense to draw on contingency funds in the light of other likely calls on them. Guided by these large-­scale considerations, they then arrange the shipment of a large quantity of food.31

With this sort of decision-­making process in mind, it seems highly unlikely that $30 more or less in the agency’s large pool of funds will make a difference to what aid is provided to people. Suppose that is correct. If so, then a view which says that, even if someone is donating a large portion of their own income, they act wrongly if they spend some money on movie tickets or a nice dinner out (unless 30  As Singer (1972: 235) puts it.

31  Cullity (2004: 58–59).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Individual Obligations  217 there is a weighty reason for them to do so, e.g. their own life projects depend on it) is highly implausible. And as with climate change, it is doubtful we can avoid that implausible result without falling into a skeptical conclusion, unless we adopt an imperfect view. Many philosophers have argued for views of beneficence that count as imperfect in the relevant sense: these include both Kantian-­style imperfect duties of beneficence (in which there is some indeterminacy involved), and also ‘cap’ or ‘fair share’ views of beneficence.32 But advocates of these views often assume that when you spend small amounts of money rather than donating it, you are passing up an opportunity to prevent some distant person from suffering or dying. They argue that, nonetheless, because of—for example—considerations of integrity or fairness or self-­ownership, you do not need to donate every time that you could without significant cost to yourself. You just need to do so ‘enough’ overall. Wherever you stand on those arguments, my point is simpler and weaker. It is not about whether it is okay to pass up a clear opportunity to prevent death or suffering. Instead, it is restricted to what we should say about our duty to make mon­et­ ary contributions that are such that, while very many would make a large difference, any one is highly unlikely to make a difference itself. The point is that at least if that is the sort of context we are in, an imperfect obligation (in the broad sense specified) has great intuitive plausibility. But while an imperfect view is highly intuitive in various collective impact cases, our intuitions might not go the same way in certain other cases. One case in which one might not find the imperfect view attractive or intuitive is that of meat consumption and factory farming. This is a collective impact case because, while a large shift in demand away from factory-­farmed meat would result in many fewer animals being tortured on factory farms, it seems that any one purchase is not going to, or is very unlikely to, make a difference to the suffering of animals. The factory farm, which operates at a massive scale, is almost certainly not going to be

32  For an example of the second sort of account, see Murphy (2000). For an argument for the first sort of account, see Noggle (2009). See also Portmore (2012, especially 35–36). Noggle raises a similar ‘either too weak or too strong’ dilemma with various moderate principles of beneficence to support the need for the sort of Kantian imperfect view that he describes. But I am making quite a different case here. Crucially, I am appealing to the claim that an individual contribution has a low probability of making a difference. This provides quite a different (or, if you want, an additional) basis that the ‘too strong’ side is indeed too strong. It no longer is (just) a point about how much sacrifice morality can demand of you. It is instead about what a fitting format for demands look like when the chances of individually making a difference by contributing to a collective pool are very low. Note that Noggle objects to the ‘fair share’ or ‘cap’ views on the basis that they allow someone to not save a person drowning in a pond in front of them as long as they do their share of beneficent contributions overall. (The idea is that an ‘ends-­based’ imperfect conception of the sort that Noggle endorses avoids that result.) However, this argument against fair share or cap views does not have the same force insofar as we think that a cap or fair share duty is restricted to contexts in which one’s individual contributions are highly unlikely to make a difference themselves (but in which very many such contributions by very many people will). This is why I am open to its being either the cap/fair share-­style view or the Kantian sort of imperfect duty.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

218  Julia Nefsky making different production decisions in virtue of a single purchase.33 But even so, for many of us, it might seem as though the facts about the suffering of animals on factory farms should make it impermissible to ever purchase factory-­farmed meat—at least unless there is some unusual, strong reason to buy it (e.g. your life depends on it). An imperfect view might, thus, seem too weak. There are two possibilities: this is right and we need a view that yields a perfect obligation (saying purchasing factory-­farmed meat is pro tanto wrong), or this isn’t right and it should be an imperfect obligation there too. Begin with the first possibility. So, suppose we should get a perfect obligation: buying factory-­farmed meat is always wrong, in the absence of strong moral considerations in favor of doing so. Even if that is right, this is not an argument in favor of EU, or any other view (e.g. complicity) that is only capable of yielding a perfect obligation or none. After all, in the climate change case and others, a perfect obligation remains implausible. What we should aim for instead is a view that can differentiate the two kinds of cases: one that can yield an imperfect duty in the climate change case and a perfect duty in the meat case. EU cannot do that. There are two ways one could go about differentiating the two kinds of cases. One is to come up with a single source of our obligations that yields an imperfect obligation in some cases, like the climate change case, but a perfect obligation in others, like the meat case.34 The second is to argue that there are different moral factors at play in the two cases, and thus different sources for the obligations, such that in the one case there is a perfect obligation and the other an imperfect one. Neither of these paths is available within EU. But now consider the other possibility: could it be that the imperfect view is right in the factory-­farming case too? Keep in mind that just because we have an imperfect view, it need not be very permissive. How much counts as refraining ‘enough’ will depend on the details of the account. It could be quite restrictive, such that one must not eat factory-­farmed meat almost all the time. What an imperfect view denies is that one can never permissibly purchase factory-­farmed meat unless there is a strong morally relevant consideration in favor of doing so. 33  Or at least, so it seems. I will not argue for this here. See Budolfson (2019) for an argument that this is indeed correct. Whether or not one agrees with the full argument in his paper, I think it at least shows that the chance is very small. However, see Hedden (2020) for a reply to Budolfson. 34  I’m not sure whether or not the fair share approach can get a perfect obligation in the meat case, but it does provide an example of a single account of our obligations that can sometimes yield perfect obligations and sometimes imperfect. The fair share view can, for example, explain why the obligation is perfect in ‘one-­off ’ cases: cases in which each person only has one chance to contribute or not in the relevant way, rather than repeated chances over time. Voting in an election is like this. The fair share view can say that because we have an obligation to elect the better/less harmful candidate, each of us has a duty to do our share of that. And because each of us has only one chance to either vote or not in a given election, doing one’s share requires voting (for the better/less harmful candidate), even though doing so almost certainly won’t make a difference to whether or not that candidate wins. Thus, you are obligated to vote for the better candidate (or, perhaps, to abstain if you cannot come to a sufficiently informed decision), provided there is no weighty competing consideration. This counts as a perfect obligation in my sense. See Brennan (2009) for the fair share reply to the inefficacy objection to voting, including the idea that it might make abstaining obligatory for some people.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Individual Obligations  219 A view could deny that, and thus count as imperfect in the relevant sense, while still saying that it is wrong to buy factory-­farmed meat any more than very rarely. Of course, it could be more permissive. But the point is that exactly how re­strict­ ive or permissive it is depends on the particular account, and is not settled by the claim that the duty is imperfect. With this in mind, I don’t think we should find it necessarily implausible that our obligations in the factory-­farming case are imperfect too. Or at least, I don’t think we should find it implausible if it is indeed true that any one purchase is extremely unlikely to make a difference to the suffering of animals (and to other bad consequences, like climate change). Many people who are vegetarian, and vegetarian on broadly consequentialist grounds having to do with animal suffering and wanting to prevent it, regard it as forbidden to ever purchase meat, and yet are totally comfortable taking at least occasional car trips just for convenience, or taking occasional airplane flights just for the sake of pleasure or career advancement. They are comfortable doing so, despite knowledge of the looming catastrophic threats to humans and animals of climate change. Why is that? Why do they have this combination of attitudes? I am doubtful that it is because they know that the expected harm of a single meat purchase is much higher than the expected harm of an individual car trip or airplane flight. Nor, do I think it is because they are confident that the expected benefit of their driving or taking an airplane in each instance is substantial enough to outweigh the expected climate change harm. The difference in attitude is likely instead because of a phenomenological difference. The idea that it is very unlikely that one will individually make a difference is phenomenologically salient in the case of climate change, and is not so in the meat-­purchasing case. Buying some chicken from the butcher does not feel like the kind of act that is very unlikely to make any difference at all to chickens. If, however, one internalizes the idea that a single purchase of factory-­farmed chicken is extremely unlikely to make a difference for the worse to any chickens, but also that many such purchases by very many people make a huge difference for the worse, then I think an imperfect obligation35 becomes intuitive here too.36 35 Note that if the obligation is imperfect, abstaining completely is (most likely) morally even ­ etter—it is just not required. Also, for many individuals it might be psychologically easier to abstain b entirely than to moderate their consumption. For individuals who are like that, the most promising strategy for meeting an imperfect obligation not to eat factory-­farmed meat would be to treat it as if it were a perfect obligation. 36 Many thanks to Zach Barnett, Mark Budolfson, Matthew Hanser, Kian Mintz-­Woo, Theron Pummer, Alex Rennet, Hamish Russell, Andrew Schroeder, Sergio Tenenbaum, and Jordan Thomson. I am also very grateful to audiences at The Centre for Ethics, Philosophy and Public Affairs at the University of St. Andrews; the Popper Seminar at the London School of Economics; the 2017 Meeting of the Pacific APA; the 2017 Meeting of the Canadian Philosophical Association; and, the 2018 Halbert Network Workshop on Collective and Temporarily Extended Rights and Wrongs. Thanks also to the participants in the 2016 Bled Conference on Ethics, where I gave a talk that was very different from this paper but a precursor to it. My work on this paper was supported by an Insight Grant from the Social Sciences and Humanities Research Council of Canada; I am very grateful for this support.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

220  Julia Nefsky

References Baatz, C. (2014). “Climate Change and Individual Duties to Reduce GHG Emissions.” Ethics, Policy & Environment 17(1): 1–19. Brennan, J. (2009). “Polluting the Polls: When Citizens Should Not Vote.” Australasian Journal of Philosophy 87(4): 535–549. Broome, J. (2012). Climate Matters: Ethics in a Warming World. New York: W. W. Norton. Broome, J. (2019). “Against Denialism.” The Monist 102: 110–129. Budolfson, M.  B. (n.d.). “Collective Action, Climate Change, and the Ethical Significance of Futility.” Manuscript, version 5.2. Budolfson, M.  B. (2019). “The Inefficacy Objection to Consequentialism and the Problem with the Expected Consequence Response.” Philosophical Studies 176: 1711–1724. Cullity, G. (2004). The Moral Demands of Affluence. New York: Oxford University Press. Cullity, G. (2015). “Acts, Omissions, Emissions,” in Climate Change and Justice, ed. J. Moss. Cambridge: Cambridge University Press, 148–164. Driver, J. (2015). “Individual Consumption and Moral Complicity,” in The Moral Complexities of Eating Meat, ed. B.  Bramble and B.  Fischer. New York: Oxford University Press, 67–79. Hanser, M. (2014). “Imperfect Aiding,” in The Cambridge Companion to Life and Death, ed. S. Luper. Cambridge: Cambridge University Press, 300–315. Heal, G. and A. Millner (2014). “Uncertainty and Decision Making in Climate Change Economics.” Review of Environmental Economics and Policy 8(1): 120–137. Hedden, B. (2020). “Consequentialism and Collective Action.” Ethics 130: 530–554. Hiller, A. (2011). “Climate Change and Individual Responsibility.” The Monist 94(3): 349–368. Intergovernmental Panel on Climate Change (IPCC) (2018). Special Report: Global Warming of 1.5°C. https://www.ipcc.ch/sr15/. Kagan, S. (2011). “Do I Make a Difference?” Philosophy & Public Affairs 39(2): 105–141. Kingston, E. and W.  Sinnott-Armstrong (2018). “What’s Wrong with Joyguzzling?” Ethical Theory and Moral Practice 21: 169–186. Kutz, C. (2000). Complicity. Cambridge: Cambridge University Press. Lawford-Smith, H. (2016). “Difference-Making and Individuals’ Climate-Related Obligations,” in Climate Justice in a Non-Ideal World, ed. C. Hayward and D. Rosers. Oxford: Oxford University Press, 64–82. McClennen, E. F. (1990). Rationality and Dynamic Choice: Foundational Explorations. Cambridge: Cambridge University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Climate Change and Individual Obligations  221 McPherson, T. (2016). “Why I am a Vegan (and You Should be One Too),” in Philosophy Comes to Dinner, ed. A.  Chignell, T.  Cuneo, and M.  Halteman. New York: Routledge, 73–91. Morgan-Knapp, C. and C. Goodman (2015). “Consequentialism, Climate Harm and Individual Obligations.” Ethical Theory and Moral Practice 18(1): 177–190. Murphy, L. (2000). Moral Demands in Nonideal Theory. Oxford: Oxford University Press. Nefsky, J. (2011). “Consequentialism and the Problem of Collective Harm: A Reply to Kagan.” Philosophy & Public Affairs 39: 364–395. Nefsky, J. (2017). “How You Can Help, Without Making a Difference.” Philosophical Studies 174: 2743–2767. Nefsky, J. (2018). “Consumer Choice and Collective Impact,” in The Oxford Handbook of Food Ethics, ed. A.  Barnhill, M.  Budolfson, and T.  Doggett. Oxford: Oxford University Press, 267–286. Nefsky, J. (2019). “Collective Harm and the Inefficacy Problem.” Philosophy Compass 14: e12587. Noggle, R. (2009). “Give Till It Hurts? Beneficence, Imperfect Duties, and a Moderate Response to the Aid Question.” Journal of Social Philosophy 40(1): 1–16. Norcross, A. (2020). “The Impotence of the Causal Impotence Objection.” Southwest Philosophy Review 36(1): 161–168. Parfit, D. (1984). Reasons and Persons. Oxford: Clarendon Press. Portmore, D. (2012). “Imperfect Reasons and Rational Options.” Noûs 46(1): 24–60. Schroeder, A. (2014). “Imperfect Duties, Group Obligations, and Beneficence.” Journal of Moral Philosophy 11(5): 557–584. Singer, P. (1972). “Famine, Affluence, and Morality.” Philosophy and Public Affairs 1(1): 229–243. Sinnott-Armstrong, W. (2005). “It’s Not My Fault: Global Warming and Individual Moral Obligations,” in Perspectives on Climate Change: Science, Economics, Politics, Ethics, ed. W.  Sinnott-Armstrong and R.  B.  Howarth. Amsterdam: Elsevier, 285–307. Spiekermann, K. (2014). “Small Impacts and Imperceptible Effects: Causing Harm with Others.” Midwest Studies in Philosophy XXXVIII: 75–90. Tenenbaum, S. and Raffman, D. (2012). “Vague Projects and the Puzzle of the SelfTorturer.” Ethics 123(1): 86–112.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

10

The Puzzle of Inefficacy Tristram McPherson

1. Introduction Consider two of the great tragedies of our age: Animals  The production of meat and other animal products inflicts massive quantities of horrifying suffering on non-­human animals Climate  Absent improbably dramatic changes in our behavior, human-­caused climate change will (on balance) have massive bad effects on both humans and non-­human animals Animals and Climate are great tragedies in large part because of the extremely bad outcomes they involve. Many of us take it to be obvious that these bad outcomes give each of us reasons to perform certain actions.1 For example, perhaps the bad effects mentioned in Animals give me reasons to eat less meat, or to support political organizations that work to protect animals, or to educate people about facts like Animals and their ethical implications. (Parallel examples suggest themselves in light of Climate.) Suppose that each of us has such reasons. We can ask: what explains why these bad outcomes give us such reasons? One possibility is that by performing these actions I can mitigate the bad effects that make Animals and Climate tragic. For example, perhaps by eating less meat I can reduce the suffering of non-­human animals. This would suggest a straightforward explanation of my reasons in these cases: I have reasons here because I have reasons quite generally to reduce bad effects in the world through my actions. It is, however, very controversial whether I can expect to mitigate the bad effects mentioned in Animals and Climate through my actions. And these bad effects appear to give me the sorts of reasons mentioned above even if I cannot

1  Distinguish this from another plausible idea: that such tragedies have ethical implications for us considered collectively. For relevant discussion of the latter issue, see e.g. Dietz (2016).

Tristram McPherson, The Puzzle of Inefficacy In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Tristram McPherson. DOI: 10.1093/oso/9780198796282.003.0011

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  223 mitigate them.2 This generates a striking puzzle. Crudely: how could I have ­reasons grounded in the (dis-) value of outcomes, if I cannot make a difference to those outcomes? This sort of question has received considerable (and valuable) philosophical attention in the context of food ethics and the ethics of climate change.3 However, in my view, the puzzle has not yet been addressed in its most forceful general form. To see the generality of this puzzle, note that indirect consequentialist views in normative ethics, such as rule-­utilitarianism, face a version of the same general challenge. This is because, on such views, all of our reasons are grounded in facts about the value of outcomes, and many options for which we possess such reasons would fail to promote the relevant valuable outcomes. I dub the general form of this puzzle the puzzle of inefficacy. This chapter aims to set out this puzzle, and to offer the general form of its most principled solution, for the case of reasons generated by Animals and Climate. This solution takes the form of a vindicating explanation of the intuitive idea that the bad effects mentioned in Animals and Climate give individuals reasons to perform actions, even if those actions do not mitigate those bad effects. I do not proceed in this way because such intuitive ethical ideas are sacrosanct. Rather, I take identifying the most promising vindicating explanation of our intuitive ethical ideas to be a crucial step in the larger project of assessing whether they are true.4 In light of its generality, my solution is not straightforwardly a competitor to important recent approaches to the puzzle in practical ethics. It is better understood as providing an attractive explanatory framework within which we could potentially evaluate such recent approaches. I set out the puzzle in§2. I argue that the puzzle has two dimensions. The first is explanatory: it can seem difficult to offer a satisfactory explanation of the reasons just mentioned. The second dimension concerns the weight of those reasons. Simple arguments seem to entail that however such reasons are explained they will be of negligible weight, in such a way that makes them irrelevant to practical ethical deliberation. The remainder of the chapter develops the general form of a solution to this puzzle that is fully responsive to the features that make the puzzle compelling. Shelly Kagan has introduced the useful idea of an evaluative focal point in order to clarify the theoretical structure of our ethical views. However, no one has 2  This idea is often treated as obvious in discussions in practical ethics. For one canonical example, see Rachels (1997: 106–107). 3  Several philosophers have followed Singer (1980) in appealing to expected effects to explain our reasons in cases like Animals and Climate. Attempts to vindicate the idea that I can have reasons in cases like Animals and Climate, even when I cannot mitigate the relevant bad effects include Cullity (2000), Kutz (2000), Sandler (2010), Lepora and Goodin (2013), Driver (2016), McPherson (2016: §3), Nefsky (2017), Lawford-­Smith (2018), and Budolfson (Ms-­b, n.d.). 4  Compare Foot’s comment in “Euthanasia” that “We are not trying to make new judgments pos­ sible, but rather to find the principle of the ones we do make” (1977: 92). I understand Foot as identifying an important phase of ethical inquiry, as opposed to its ultimate goal.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

224  Tristram M c Pherson attempted to explain what would make a candidate evaluative focal point correct. I offer a novel sufficient condition for identifying a genuine evaluative focal point, and show how this can allow us to solve the explanatory side of the puzzle (§3). Because this solution is very abstract, I illustrate it by sketching the principles of what I call contribution ethics, which identify reasons to contribute to—or to resist—important social patterns (§4). With this illustration in hand, I show how the negligibility dimension of the puzzle can potentially be solved (§5).

2.  Introducing the Puzzle of Inefficacy In this section, I state a precise version of the intuitive ethical claim at the heart of the puzzle of inefficacy. I then spell out the two dimensions of the puzzle, and explain why it is unappealing to evade the puzzle by rejecting the intuitive claim. I begin with a precise statement of the claim that generates the puzzle: Intuitive  The bad effects mentioned in Animals and Climate give one (­non-­negligible) reasons to perform certain actions, even if those actions will not mitigate those bad effects Intuitive should be read in light of the following three clarifications. First, in the Introduction I gave examples of the sorts of actions that the reasons mentioned in Intuitive might support. However, my aim in this chapter is not to address the important question of precisely which actions these reasons support. It is to understand how such reasons are possible. Second, Intuitive does not purport to identify the only reasons that we can have in light of facts like Climate and Animals. For example, we may quite generally have reasons to promote the satisfaction of our desires. And some of us have desires to perform the sorts of actions referenced by Intuitive. I will ignore such reasons in what follows. Third, some of us are inclined to think that facts like Animals and Climate often give us (perhaps defeasible) obligations as opposed to simply reasons. Here I focus on explaining reasons, and the possibility of their non-­negligible weight. I take this to be sufficiently challenging, without broaching the further difficult question of when reasons entail the existence of an obligation.5 Before proceeding, it will be useful to explicitly distinguish two very different sorts of explanatory relations that will be invoked both in the exposition of the puzzle, and in my solution. We can introduce these relations by considering a simple example. Suppose that you tell me a joke, and I then experience a burst of

5  On this focus, compare Nefsky (2017: 2744–2745).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  225 amusement. Now consider the question: what explains my amusement? We can hear this question in two ways. On one way of hearing this question, your telling the joke explains the amusement. This is especially clear if I would not have been amused if you had not told the joke. Call this sort of explanation difference-­making explanation.6 It is difference-­making, because the presence or absence of the joke makes a difference to its coming about that I am amused. Such explanation is often diachronic: what does the explaining (the joke) often occurs prior to what it explains (the amusement). On a different way of hearing the initial explanatory question, you could explain my amusement by appealing to the facts or states which ground, constitute, or realize my amusement. Call this sort of explanation grounding ex­plan­ ation. Grounding explanations are often synchronic: for example, on a simple physicalist hypothesis, my amusement at a certain time is grounded in certain of my brain states at that time. The aim of this section is to explain why the truth of Intuitive can seem dis­ tinct­ive­ly puzzling. The core puzzle arises from two implications of Intuitive: Outcomes Ground  The reasons mentioned in Intuitive are partly grounded by facts about the badness of outcomes Inefficacy  Some of the actions for which you have such reasons would make no difference to the relevant outcomes These implications put to work the two kinds of explanation just distinguished: Outcomes Ground offers a partial grounding explanation, while Inefficacy denies a difference-­making explanation. Neither Outcomes Ground nor Inefficacy are puzzling by themselves. Consider Outcomes Ground: the idea that we can have reasons grounded in the value of outcomes is extremely natural.7 That I will enjoy eating the cake is a reason to eat it. That it will make me feel bloated and uncomfortable shortly afterwards is a reason not to. That an encouraging word will lift your spirits is a reason for me to offer it. That tripping you as you walk by risks inflicting pain and injury is a reason not to do it, etc. 6  Many difference-­making explanations are also causal explanations. But for my purposes, it is important to distinguish these notions: in cases of causal over-­determination, C may be a cause of E but not a difference-­making explanation for E. And difference-­making is often especially straightforwardly ethically relevant, in ways that over-­determining causes are not. 7  Those favoring buck-­passing accounts of value (see Jacobson 2011 for an introduction) can redescribe the grounding claims just made in the text: on such views it is not, strictly speaking, the goodness or badness of outcomes that grounds the reasons; rather it is certain features of outcomes (which we would correctly deem good or bad) that ground the reasons. For simplicity, I will ignore this in the text below; buck passers can, however, read this redescription into relevant parts of the discussion throughout.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

226  Tristram M c Pherson Nor is the puzzle that we are too physically or temporally distant from the good or bad effects mentioned in Animals or Climate, or that we cannot hope to completely “solve” the problems described in those cases. For example, a $1000 donation to the Against Malaria Foundation will not “solve” malaria, and its central good effects will occur far away from me. Nonetheless, if we suppose that such a donation will save many people from the misery and medical complications of untreated malaria, it is unmysterious why this gives me a reason to make that donation. (At least, to the extent that any reason is unmysterious.) The puzzle also does not arise simply from Inefficacy: that one can have a reason to do something that does not improve a relevant outcome. For example, it is plausible that we have reasons to refrain from certain paternalistic acts, even when we know that performing them would have relevant good effects. In such cases, our reasons to refrain appear to be explained in terms independent of the outcomes of our acts. The explanatory dimension of the puzzle facing Intuitive is that we appear to have reasons that are grounded in facts about the value of outcomes (like the donation example, and unlike the paternalism example), to do things that do nothing to improve those outcomes.8 Relative to the bad outcomes mentioned in Animals and Climate, our individual actions appear futile. It might seem that the goodness or badness of an outcome could not give us reasons to do something that is futile with respect to that outcome.9 After all, how could the value of that outcome explain such reasons? There is another side to the puzzle facing Intuitive, which can be illustrated by a stylized case: Oracle  Consider the things that Animals might give you reason to do, given Intuitive: eating less or no meat, supporting certain political organizations, etc. The infallible oracle tells you that, surprisingly, your doing any of these things will lead to increased suffering in non-­human animals. In this case the sorts of reasons proposed by Intuitive weigh against relevant ­reasons which are explained by your ability to make a difference to the relevant bad outcomes. And it seems very plausible that the latter reasons are sys­tem­at­ic­ al­ly stronger: insofar as you care about animal suffering, you seem to have net reasons not to eat less meat, etc. Further, this seems true even if your doing these things will only increase animal suffering by a small amount. This might seem to

8  Sandler (2010) calls something very close to this side of the puzzle the “problem of inconsequentialism.” Shafer-­Landau (1994) calls a closely-­related challenge the “inefficacy challenge.” 9 For a compelling prosecution of related claims about the ethical significance of futility, see Budolfson (Ms-­a, n.d.).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  227 suggest that even if the sorts of reasons mentioned by Intuitive exist, they must be negligible in weight.10 That our reasons grounded in facts like Animals and Climate are vanishingly weak is unintuitive. (I have stated Intuitive to be inconsistent with negligibility for just this reason.) It also makes it difficult to see how such reasons would be of any interest to ethicists interested in informing our practical choices: even if such ­reasons exist, it seems that we could legitimately ignore them for practical purposes. But how—in light of cases like Oracle—could these reasons fail to be neg­li­ gible in weight? I call this the negligibility dimension of the puzzle. The two dimensions of the puzzle interact. An initially attractive explanation of why the difference-­making reasons are always weightier in Oracle-­style cases is that the appropriate relation of actions to the value of outcomes is an act-­promotion relation. But given our apparent inefficacy, it is impossible to explain Intuitive via this sort of act-­promotion relation. In this chapter, I do not aim to defend Intuitive, but rather to provisionally assume it, and explore how it might be vindicated in light of the puzzle just sketched. However, some readers may take the puzzle just sketched as a reason to simply reject one of the two apparent facts—Outcomes Explain and Inefficacy— that generate the puzzle. In the remainder of this section, I explain why we should at least hesitate to avoid the puzzle in these ways. First consider Outcomes Ground. As I suggested in the Introduction, it is very plausible that we have reasons in light of facts like Climate and Animals. But surely it is the badness of the outcomes mentioned by these facts that explain these reasons: if these facts mentioned only good outcomes, we would, plausibly, have very different reasons. Suppose that we set this direct point aside, and consider the attempt to argue that the reasons we have in light of Climate and Animals are explained by Kantian respect, or by related contractualist ideas.11 It is famously challenging to implement this strategy satisfactorily with respect to Animals. This is because many of the touchstones of broadly Kantian ethics are poor candidates to explain our reasons with respect to non-­human animals. Next consider seeking to avoid the puzzle by rejecting Inefficacy. Several philo­ sophers have followed Peter Singer’s lead in arguing that (for example) there are ethically significant expected effects to ordering meat at a restaurant.12 Simplified 10  The negligibility side of the puzzle is inspired in part by reflection on Parfit-­style mineshaft cases (Parfit 1984: §§25–26). Parfit’s cases compare reasons grounded in what difference an individual can make with that individual’s (alleged) reasons grounded in one’s contribution to a group action that makes a difference, seemingly suggesting that the former systematically outweigh the latter (or perhaps that the latter do not exist). 11  Relevant literature includes Baxter (1974), Carruthers (1992), Wood (1998), Rowlands (2002), Korsgaard (2004), Talbert (2006), Baatz (2014), Suikkanen (2014), and Calhoun (2015). 12  Singer (1980). This reasoning has been endorsed or developed by Matheny (2002), Norcross (2004), Kagan (2011), and Huemer (2018). Sometimes the significance of Singer’s reasoning is framed

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

228  Tristram M c Pherson somewhat, Singer’s argument is that at a sufficiently coarse grain, the production of meat must be sensitive to demand. If it is not typically sensitive to individual acts of consumption, it must therefore be sensitive to aggregate consumption crossing some threshold. Our particular choices are highly unlikely to trigger such a threshold. But if a choice does trigger such a threshold, the effects of doing so will be very large. Multiplying this large effect by the small chance of triggering it results in our individual choices having substantial expected effects, according to Singer. Singer’s reasoning, and the various attempts to develop it, are controversial.13 I take controversy over this question to be largely grounded in the expectations we should have in light of difficult-­to-­resolve empirical questions.14 However, I take Intuitive to express the plausible idea that we have reasons of the relevant kind to perform certain actions, even if those actions make no difference to the relevant bad effects. To make this thought vivid, consider the following broadly analogous case: Restaurant  There is a new restaurant in town: the food is sensational, and the prices are very low. How do they do it? The owner kidnaps world-­class chefs, and enslaves them at the restaurant. The owner is connected, and going to the police would just get you killed. Your patronizing the restaurant will not lead to further enslavement, or any additional mistreatment of the enslaved chefs.15 It is plausible that you have reasons not to patronize the restaurant, and that these reasons are explained by the badness of the treatment of the enslaved chefs. But by hypothesis one’s patronage makes no difference to the badness of such treatment. To be clear, I have already granted that we have many reasons that appear to be well-­explained by appeal to difference-­making (as in the donation example) and to something broadly like respect (as in the paternalism example). And we may have some such reasons in light of facts connected to Animals and Climate. My point here is that even granting these points, it appears strikingly implausible to reject Intuitive (hence the name!). That warrants the project of this chapter: to identify the general form of a solution to the puzzle sketched in this section which assumes the truth of Intuitive.

within an act-­consequentialist framework, but its significance is potentially much broader, given that it is very plausible that expected effects have (at least derivative) ethical significance, even given the falsity of act-­consequentialism. 13 Direct or indirect challenges to Singer-­style reasoning includes Frey (1983,  2004), Sinnott-­ Armstrong (2005), Chartier (2006), Sandberg (2011), and Budolfson (2019). 14  See McPherson (2018b: 222–224) for a brief introduction to the key dimensions of this debate. 15  This example is adapted from McPherson (2017: 824–825).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  229

3.  Ethical Structure and Social Structure This section sketches a principled solution to the explanatory aspect of the puzzle of inefficacy. To briefly restate the puzzle: Intuitive suggests that we have reasons that are grounded in the badness of the outcomes mentioned in Animals and Climate, to perform actions that do nothing to improve those outcomes. But it is puzzling how such reasons could make sense. My solution appeals to a central dimension of ethical structure which Shelly Kagan has helpfully dubbed evaluative focal point. I explain what evaluative focal points are, and show that solving the explanatory dimension of the puzzle of inefficacy requires that we provide a principled explanation for a relevant focal point: specifically, a focal point other than a particular action. I then develop such an explanation, which supports a sufficient condition for counting as an evaluative focal point. In a slogan, my account suggests that, in this case, ethical structure is grounded in explanatory structure. I then show that social structures can satisfy the sufficient condition I have defended. I conclude by showing how treating social structures as evaluative focal points permits an elegant explanation of the reasons mentioned in Intuitive, and hence a solution to the explanatory dimension of the puzzle. I begin by further clarifying the explanatory task. As I understand it, the task is not simply to propose an ethical principle that entails the existence of the relevant reasons. To see this,consider an ethical principle according to which each of us has reason to “act against” tragedies such as those named by Animals and Climate. This principle has some claim to encapsulate the intuitive basis of the idea that we have reasons here. But to my eye, it fails wholly as a solution to the explanatory puzzle. This principle is close to simply renaming the explanatory problem I seek to address: given that “acting against” the tragedies will not mitigate them, why does the badness of the tragedies give us reasons to act against them? What I seek to provide in this section is an explanation that can do better than this: a prin­ cipled ethical grounding explanation that illuminates why we have such reasons. My explanandum is the thesis I dubbed Intuitive: that the bad effects mentioned in Animals and Climate give one reasons to perform certain actions even if those actions do not mitigate those bad effects. Notice that this explanandum is inconsistent with some influential views about ethical structure. For example, given plausible assumptions about value, Intuitive is inconsistent with a simple act-­consequentialist theory of reasons for action. On this view, the strength of ­reasons in favor of an option is proportional to the net good effects promoted by selecting that option (and symmetrically, net bad effects will entail proportionally strong reasons against). The proponent of Intuitive agrees with the act-­ consequentialist that our ­reasons in light of Animals and Climate are grounded by facts about the value of outcomes. They must thus reject the act-­consequentialist’s account of structure of

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

230  Tristram M c Pherson the grounding explanation of such reasons. There are two natural general ­strat­egies for rejecting this account. The first rejects the act-­consequentialist’s idea that promotion is the relevant explanatory relation to value. The second rejects the idea that we should evaluate acts directly in terms of their outcomes. I will not pursue the first strategy because I do not see how to make it promising in this context. Let me briefly sketch one example of why. Suppose that one rejected the promotion relation to value by appealing instead to the idea that certain actions are directly fitting responses to the value of outcomes. The problem with this idea is that the idea of a fittingness relation between actions and the value of outcomes is opaque. Most of the literature on fittingness focuses on the idea that certain of our emotions and attitudes can be fitting responses to certain properties. For example, amusement seems like a fitting response to certain jokes, and on some accounts, this fact will ground the fact that those jokes are funny.16 It is not clear how to understand the fittingness of actions on this model.17 In light of this, I pursue the second strategy. This strategy involves addressing a question about ethical structure that is most familiar from debates between act- and rule-­consequentialists. To illustrate this question, consider a pair of such consequentialists who agree completely concerning what is valuable, and also agree that facts about reasons for action are grounded in facts about ­value-­promotion. The central disagreement between our imagined act- and rule-­consequentialists is this: in determining an agent’s reasons, what is it that gets  directly assessed in terms of value-­promotion? This is an instance of what  Shelly Kagan (1998: Ch. 6) calls an evaluative focal point question. The act-­consequentialist answers that the agent’s options constitute the evaluative focal point: reasons to select options are a function of how well those options promote value. The rule consequentialist answers instead that social rules with certain properties are the evaluative focal point: reasons to select options are a matter of whether those options conform to social rules, the adoption of which (e.g.) would promote value. Consider a (simplified) example of how this contrast in focal point might matter. Here is a social rule the acceptance of which would plausibly promote value: a rule against assaulting random strangers. Given this, a simple rule consequentialism implies that one always has a reason to refrain from such assault. Now consider an unusual case where such assault will in fact promote value: perhaps such an assault is the only way to save someone’s life. Act-­consequentialism implies that in this case one has a reason to perform the assault. 16  Nye (2018) is an excellent introduction to fitting-­attitude analyses of normative structure. 17  The most plausible way of explaining reasons in terms of the value of outcomes via fittingness involves also deploying the second strategy of evaluating acts indirectly. For example, consider a view on which certain moral emotions are fitting responses to the value of certain outcomes, and we have moral reasons to perform actions that would express those moral emotions, in circumstances where those emotions are fitting. I set aside such hybrid accounts here.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  231 While the debate between act- and rule-­consequentialists is the most familiar context where the question of evaluative focal point arises, the question is not limited to this debate. Let me illustrate the generality of the issue in successively broader terms. First, consequentialists are not limited to taking options or rules as their focal points: for example, there are versions of consequentialism that take the focal point to be psychological features like motives or character traits.18 Second, the question is not limited to consequentialist theorizing.19 To see this, consider a familiar contractualist thesis, according to which facts about rightness are grounded in facts about reasons to reject certain hypothetical agreements. Such a grounding thesis leaves open the evaluative focal point question: what do we directly test by considering the reasons that persons have for rejection? The most familiar form of contractualism tests certain general principles via reasons for rejection. On this view, an act is right insofar as it is compatible with a prin­ ciple that cannot be reasonably rejected. But one could instead be an “­act-­contractualist,” applying the contractualist test directly to an agent’s options in fully-­specified contexts. And of course, as with consequentialism, these are only two salient examples of the evaluative focal points that a contractualist might appeal to (compare Kagan 1998: 242). Third, I have so far offered as examples theories that are usually treated as candidate fundamental and general ethical or moral theories. However, the question of evaluative focal point can arise wherever there are ethical grounding ex­plan­ ations. And such grounding explanations need not mention the most fundamental features: rather, grounding explanations connect the less to the more fundamental. For example, some philosophers have suggested that certain contractualist principles might entail act-­consequentialism as a derivative ethical theory. Such a derivative act-­consequentialism would still provide a grounding explanation that conflicts with Intuitive. Nor must ethical grounding ex­plan­ations be as general as the act­consequentialist explanation, in the sense of purporting to explain all of our reasons. For example, one sort of ethical pluralist will think that there are different sorts of reasons that have different grounding explanations. These observations about evaluative focal points are crucial background to my strategy in what follows. I propose to solve the explanatory dimension of  the puzzle of inefficacy by developing a principled explanation of the

18  The locus classicus for a focus on character traits as focal point is Adams (1976); Sandler (2010) offers a virtue consequentialist solution to what I am calling the explanatory side of the puzzle of inefficacy. There are less familiar options as well: for example, a consequentialist could adapt central ideas from Darwall’s “second personal” approach to ethics (Darwall  2006) by taking the evaluative focal point to be practices of interpersonal accountability. 19  Although I have adopted Kagan’s useful terminology here, it is worth noting that his word ‘evalu­ ative’ might mislead by seeming to suggest that the relevant sort of structure only arises in consequentialist theories. For discussion of the related question of whether contractualism is best understood as a theory of an action’s being wrong, or of its wronging a particular person, see Katz (2017: §2).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

232  Tristram M c Pherson evaluative focal point structure that partly grounds the reasons that arise from Animals and Climate. If we could simply posit evaluative focal points willy-­nilly, the explanatory puzzle would be easy to solve. For example, one can easily imagine simply stating a rule-­consequentialist explanation of why Intuitive is true. In my view, however, this sort of “explanation” would, by itself, merely relocate the explanatory ­puzzlement I aim to address. For we can ask: why are rules (with such and such properties) the correct evaluative focal points in our theory of reasons for action? More generally, the crucial question before us is this: what explains why some feature is an evaluative focal point? This question has been surprisingly neglected in the literature. To illustrate, consider the most familiar case: rule-­consequentialism. A natural motivation for endorsing rule-­consequentialism is that one is compelled by the consequentialist idea of explaining reasons and rightness in terms of value, but one rejects act-­consequentialism as extensionally inadequate. For example, consider Judith Thomson’s famous “transplant” case. The rule consequentialist might hope to vindicate the idea that it is wrong to carve up the one person to save five in such a case.20 This sort of consideration may provide some reason to believe that rule-­consequentialism is more likely to be true than an otherwise similar ­act-­consequentialism. But it does nothing to explain why rules are the evaluative focal point. Some rule-­consequentialists have claimed that a rule-­based structure is essential to morality, because it is allegedly essential to moral norms that they satisfy some sort of publicity constraint.21 This at least appears to provide an explanatory consideration. But (to be undeservedly brief) I find it implausible for two reasons. First, I see no special reason to think that a publicity constraint holds for morality. Consider a possible world in which an evil demon caused anyone who believed (what would otherwise be) a moral truth to become homicidal. This is surely a world in which the moral truths should remain unknown, not a world in which there are no moral truths. Second, publicity constraints would implausibly threaten the ethical significance of morality. To see this, suppose that there are weighty ethical considerations that fail to satisfy the publicity constraint. (Again to keep it simple: if most people accepted certain ethical considerations, the evil 20  Thomson (1985: 1396). One leading illustration: Hooker (2000) structures the evaluation of his rule-­consequentialist moral theory around five plausible desiderata. But arguably it is only with respect to extensional adequacy that the choice of rules as a focal point distinguishes itself from an  otherwise similar act-­ consequentialism (e.g. Ch. 7). Acute readers will notice that an act­consequentialist could vindicate our judgment about the transplant case by adopting a suitable theory of value. 21  Compare Brandt (1992: 136); Hooker (2000: 85), characterizes failure of publicity as morally objectionable. Copp (1995) argues that a rule-­based structure is essential to morality for very different reasons: on his view, such structure is a consequence of what he takes to be the most promising account of how moral norms can be non-­circularly justified.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  233 demon would kill everyone.) Consider any circumstance where one ought to act on such a consideration. If morality is subject to a publicity constraint, this could entail that in such circumstances one ethically ought to act immorally. Conversely, act-­consequentialists have famously taken there to be something objectionably indirect about other candidate focal points. Recall here famous charges that rule-­consequentialists are guilty of “rule-­worship.”22 Notice that these charges are most forceful if extensional adequacy is one’s only motive for rejecting the idea that the agent’s options are the evaluative focal point. If we possessed a satisfying explanation for why rules were the uniquely correct evaluative focal point, then the charge of rule worship would be hard to make persuasive. This suggests two quite general reasons for ethicists to seek a compelling ex­plan­ation of why a certain feature counts as an evaluative focal point. First, identifying such an explanation seems essential to understanding the rele­vant ethical structure. Second, it should put us in a better position to evaluate arguments for or against certain focal points. It might be doubted that we can provide such an explanation. Arguably, ex­plan­ ations in ethics must come to an end somewhere. And it might seem hard to see what an adequate explanation of evaluative focal point would look like. So perhaps the literature has neglected this question because a candidate focal point can be more or less plausible, but cannot be explained. I argue against this explanatory pessimism by developing an instance of the relevant sort of explanation. I do not aim here to give a completely general explanation of evaluative focal point. Instead, my proposal is tailored to the project of this chapter. I aim to offer a sufficient ethical explanation of evaluative focal points in explaining facts about reasons in terms of facts about the value of outcomes. I thus do not claim that the explanation I offer is either ethically fundamental or fully general (in the sense of illuminating the structure of the explanation of all of our reasons). One reason for this agnosticism is epistemic: it is typically harder to credibly defend more general or fundamental ethical theses than it is to defend less fundamental or general theses.23 Another reason is that I take what I say below to be compatible with a range of important and quite different views about fundamental normative ­ethic­al theory. I thus want to present my account in a way that highlights its interest to a wide range of philosophers. In a slogan, I claim that (at least in this case) ethical structure is grounded in explanatory structure. Let me now spell out this slogan. Call a feature evaluatively relevant if it is the sort of feature that makes a state of affairs good or bad. For example, containing a lot of suffering makes a state of affairs bad (even if that state is overall good), and hence is an evaluatively relevant feature. There are facts, both specific and general, about the distribution of evaluatively relevant features. And 22  The locus classicus is Smart (1956). 23  See McPherson (2018a: §6) for some reasons why this might be.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

234  Tristram M c Pherson these are (at least almost) never brute facts: for any such fact, there will be a difference-­making explanation of why that fact holds. Within such explanations, there will be features which, to put it intuitively, “do the explanatory work” in the explanation. I call these explanatory focal points. These ideas allow me to make the idea suggested by my slogan more precise: Explanatory Link  It is sufficient for something to be an evaluative focal point that it is an explanatory focal point in a difference-­making explanation of the instantiation of evaluatively-­relevant properties Consider a simple example of Explanatory Link in action. Suppose that a certain kitten is in agony. This agony is an evaluatively relevant property. So Explanatory Link instructs us to look for the difference-­making explanation of this agony. Suppose that the summary explanation is: that I stepped on the kitten’s tail. The complete explanation of the kitten’s agony would, of course, involve many facts: facts about my mass and the velocity of my foot; facts about the way that the kitten’s nervous system responds to the tissue damage in the kitten’s tail, facts about the unbroken nerve pathway between the site of damage and the brain, and possibly facts about the lack of anesthetic in the kitten’s nervous system. But in many versions of the case, the summary fact mentioned is a good explanatory summary, precisely because the feature mentioned—my action—appears central to the difference-­making explanation in a way that these other features do not. In light of this centrality, I call my action in this case the explanatory focal point in the explanation of the kitten’s agony. One heuristic for verifying this is modal: for example, it is tempting to think that in the nearest worlds where the kitten is not in agony, I did not step on its tail. It is worth emphasizing that this modal heuristic is not infallible: suppose that I only stepped on the kitten’s tail because you pushed me, and that caused me to lose my balance. In this case, the modal condition is satisfied, but the summary explanation I initially offered is plausibly misleading. An adequate summary explanation should mention your pushing me, which is a better candidate to be the explanatory focal point in this version of the case. Explanatory Link is extremely plausible in these cases. Consider the first variant of the kitten’s tail case: here, the role of my stepping on the kitten’s tail as an explanatory focal point in a difference-­making explanation of the kitten’s agony plausibly explains why facts about the kitten’s agony directly ground a reason for me to refrain from performing that action. Generalizing, Explanatory Link straightforwardly explains why evaluative features can give us reasons to perform (or refrain from performing) actions that are difference-­makers for those evalu­ ative features. Indeed, Explanatory Link is plausible in part because it is a gen­er­al­ iza­tion of a link between difference-­making and reasons for action that we find explanatory in a range of uncontroversial cases. (Consider, for example, the

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  235 unpuzzling cases of outcome-­based reasoning mentioned in §2, such as reasons to donate to an ethically effective charity.) It may seem less clear how Explanatory Link could permit us to explain the reasons mentioned in Intuitive. The key idea needed to complete my explanation is that a social structure or pattern can function as a relevant explanatory focal point. This idea can be motivated by appealing to a familiar idea in metaphysics and the philosophy of science: that there are legitimate difference-­making ex­plan­ ations that appeal to “higher-­level” facts. Consider, for example, the prevalence of a given mutation in a given population of a species in an environment at a time. This fact could be explained by an extremely complicated set of particular facts about the population’s environment, and the particular actions of members of previous generations of the species in that environment. However, in at least a range of cases, we can give a more compelling higher-­level explanation: certain general facts about the environment and the generally prevailing genetic facts in the species at a prior time combine to explain the prevalence of the mutation, by explaining why the mutation was selected for by the environment. The sort of higher-­level explanation just sketched has more going for it than its intuitive plausibility and elegance. It can also be considerably more modally robust than an explanation that appeals to the relevant host of particular facts. For ex­ample, the explanation of the prevalence of the mutation in terms of particular facts will have to appeal to facts such as that Al mated with Bo at time t1, and that Cass starved to death at time t2. But it could well be that, given the general facts in the background, the mutation would have had approximately the same prevalence even were a host of such particular facts different.24 When we test counterfactually for difference-­making causes of certain outcomes, we will often find that higher-­level facts (e.g. about the general character of the environment and the genetic character of the species) will count as difference-­making causes. Because difference-­making is a modal notion, this modal robustness means that higher-­level explanatory features can count as explanatory focal points in difference-­making explanations. And this in turn means that Explanatory Link entails that they can count as evaluative focal points. This point is ethically relevant because of an idea that is familiar from (e.g.) feminist theory, anti-­racist theory, and both Marxist and neoclassical economics: that there are important structural explanations of the instantiation of evalu­ atively relevant properties.25 Consider two standard examples. 24 For one way of understanding higher-­level explanations congenial to this sketch, see Pettit (2007: 217–228). Notice that, because agential properties are plausibly fully grounded in lower-­level properties, the proponent of explanations in terms of acts will need something like the idea of genuine higher-­level explanations as well. 25  Compare the classical Marxist theory of exploitation, and the broader Marxist thesis that economic structure determines the ‘superstructural’ features of social life, in Marx (1992 [1867]). Compare also (e.g.) Frye (1983) and Young (1990) on gender oppression and Fredrickson (2002) and Mills (2015: 29) on racism.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

236  Tristram M c Pherson First, many theorists insist that inequalities among genders or races are not best explained simply by the attitudes and choices of individuals. Rather, such inequalities are better explained by the existence and nature of patriarchal and racist social structures that incite and reproduce the relevant attitudes and choices. Generalizing on these sorts of insights, Iris Marion Young (2011) has argued powerfully for the idea that social structures are the central subject of ­justice. Young argues for this thesis in large part by emphasizing the importance of social structures to the explanation of ethically significant facts about lives. Second, consider a standard bit of explanation from neoclassical economics. Consider a limited fish stock, access to which is uncontrolled by direct regulation, informal coercion, or effectively enforced property rights. Assuming a certain range of facts about supply and demand, economists will predict that such a stock will eventually collapse. This is explained by a structural fact about such a case: that it is an instance of the tragedy of the commons pattern (Hardin  1968), in which individual actors each have strong incentives to act in ways that, col­lect­ ive­ly, will lead to worse outcomes for each. As with the gene mutation example, structural explanations like these will typ­ ic­al­ly be both more elegant and more modally robust than explanations that appeal to collections of individual actions or events. Tragically, no one woman’s resistance will end patriarchy, and in the circumstances the economist imagines, no one angler’s restraint will prevent the collapse of the fish stocks, etc. So each individual action that makes up such a collection will be dispensable to the difference-­making explanation. I am now in a position to provide the schematic form of my solution to the explanatory puzzle of inefficacy. The solution begins by noting that Animals and Climate mention patterns of instantiation of certain evaluatively relevant properties. It is the instantiation of these properties that makes Animals and Climate tragic. Next, it is very plausible that facts about social structures or patterns are focal points in the explanation of these tragic facts, just as they are for a range of facts about racism, sexism, and resource depletion. The plausibility of individual inefficacy with respect to the bad effects mentioned in Animals and Climate entails that individual actions are poor candidates to serve as focal points in difference-­making explanations of these facts. By contrast, it is extremely plaus­ ible that a complex set of social, economic, and legal institutions and practices structure our collective emissions of greenhouse gases (e.g.) in a difference­making way. Explanatory Link proposes that being such an explanatory focal point is sufficient to make a feature an evaluative focal point. And this in turn explains how we can have reasons in light of Animals and Climate, even when we cannot make a difference to the bad effects they mention. We can have such ­reasons because our options can bear the right relation to the social structures or patterns that explain those bad effects. I will have more to say about what the “right relation” might be, in the next section. The general point being made here

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  237 is illustrated by the (different) case where certain rules are explanatory focal points. In such a case, the right relation is plausibly conformity to those rules, even where such conformity will fail to make a difference to the outcomes which ground the rules. To fully address the explanatory side of the puzzle of inefficacy, we need to: 1. vindicate the idea that we have the sorts of reasons mentioned in Intuitive, and 2. show how facts about the badness of outcomes play a central grounding role in this explanation, while 3. explaining how such outcomes can ground reason to perform options that do not mitigate those outcomes The solution just sketched satisfies each of these three conditions. It satisfies the first condition by offering an explanation of how we can have the relevant reasons. It satisfies the second condition because it entails that my reasons are explained by the relations of my options to certain social structures and patterns. And those being the relevant social structures and patterns in this explanation is itself explained precisely by the fact that those patterns produce the relevant bad outcomes. And it satisfies the third condition by offering a principled explanation for why social structures are explanatory focal points, which is the crucial assumption required to meet the first two conditions. Consider two further attractive features of the solution just proposed. First, as I mentioned in §2, the two sides of the puzzle of inefficacy are linked. One danger for any solution to the explanatory dimension of the puzzle is that it will fail to explain how we can have reasons to avoid acts because of the bad effects of those acts. My solution does not have this difficulty. This is because Explanatory Link implies, plausibly, that whenever an act is the best explanation for an evaluatively relevant effect (as in the example of stepping on the kitten’s tail), one’s act is itself a direct evaluative focal point. Thus, in the surprising condition mentioned in Oracle, my account implies that, because (e.g.) the choice to eat less meat would be an explanatory focal point in a difference-­making explanation of animal suffering, I have reasons to refrain from that choices. Third, the solution just proposed is abstract and schematic. This is appealing because there are a range of familiar proposals to explain our reasons in light of cases like Animals or Climate, in terms of ideas like complicity or benefitting from injustice.26 The abstract form of my solution allows us to ask of each of these proposals: can this proposal satisfy Explanatory Link? If not, can this proposal be 26  For complicity, see e.g. Kutz (2000), Schwartz (2010: ch. 4), Lepora and Goodin (2013), Driver (2016), and McPherson (2016: §3); on benefitting from injustice, see e.g. Butt (2007), Goodin and Barry (2014), and Pasternak (2014). For a distinct explanatory competitor see Budolfson (Ms-­b, n.d.).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

238  Tristram M c Pherson amended in a way that allows it to satisfy Explanatory Link, while retaining its intuitive appeal? The schematic solution thus provides a framework within which we can potentially make progress on developing these proposals into their most compelling form, and adjudicating among them. The schematic solution may also strike some readers as elusive, however. And it may be instructive to see a clear example of how it could be implemented. In the next section, I briefly sketch what I take to be the most natural way of implementing the schematic solution.

4.  Contribution Ethics: A Sketch In this section I sketch what I take to be a natural way of implementing the schematic solution to the explanatory puzzle offered in the previous section. I propose a clear way of thinking about the relevant social structures that the solution appeals to. Then I will propose and illustrate principles of what I dub contribution ethics, which link individual behavior to these social patterns. A complete implementation of the solution I have proposed would include considerable social philosophy. For example, it would spell out a theory of social structures, and their precise explanatory powers.27 Here I aim to sidestep these important debates, and focus on one clear and important type of social structure, which has the following features: A. There is a pattern of human behavior B. In order for the pattern to persist, individuals are required to act in a predictable, systematic set of ways: the ways of acting that sustain the pattern C. The pattern itself includes elements that give participants incentives to contribute to sustaining the pattern D. The pattern is resilient over small-­scale resistance Call social structures characterized by these four features self-­sustaining social patterns. (Such patterns need not be indefinitely self-­sustaining. For example, some patterns can be self-­sustaining for a long time, but collapse quite predictably because they rely on consuming some finite resource.) I take there to be a wide range of self-­sustaining social patterns, including (for example) implicit conventions, such as linguistic conventions, or conventions about how close to stand to someone you are talking to; a range of patterns in economic markets; and many laws and other informally enforced social norms.

27  One helpful discussion of related issues is Haslanger (2016).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  239 Consider two examples of self-­sustaining social patterns in slightly more detail. First, consider an implicit convention that arises for driving on one specific side of the road, which is not legally enforced. In such a case, there can be a robust pattern of driving on the right (feature A). This pattern requires individual acts to sustain it (feature B): if everyone got up tomorrow morning and started driving on the left, the pattern would disappear. Of course, with the pattern stably in place, anyone who wishes to drive on the road has enormous incentives to drive on the right, anyone who wishes someone well has strong reasons to tell them to drive on the right, etc. (feature C). And the pattern is resilient over small-­scale resistance (feature D): if a small minority start driving on the left, the right-­side driving pattern is likely to persist, provided that it continues to be safer to drive on the right. Unfortunately, not all social patterns are benign in the way this convention is. For example, a similar story can be told about (part of) patriarchy. There are persistent patterns of behavior and expectations that systematically disadvantage and harm women (feature A). These patterns require individual behavior to sustain them: it is metaphysically possible that we wake up tomorrow, and through our individual actions collectively transform the world in short order into a feminist utopia (feature B). Individuals have incentives to sustain the patriarchal patterns (feature C) that range from the subtle (for example, acting in sustaining ways can feel “natural”) to the blunt (such as patterns of threat and violence). And patri­ archy has proven distressingly resilient in the face of considerable resistance (feature D). It is plausible that self-­sustaining social patterns provide difference-­making explanations for much of what is good and bad in our contemporary world. According to Explanatory Link, this entails that such patterns count as evaluative focal points: we can explain the existence of a range of reasons individuals have for acting in terms of the relationship of those actions to the relevant social patterns. One crucial way in which the solution to the explanatory puzzle offered in the preceding section was schematic was that it spoke of options bearing “the right relation” to the focal social structures and patterns. This raises a crucial question: what is the right relation to an evaluative focal point? We could offer an intuitive answer here, or step back and ask: what explains what relation an option must bear to an evaluative focal point in order for one to have a reason to adopt that option? Like the explanation of evaluative focal point, this question is rarely addressed. Perhaps this is because in a given case, some relation typically seems obviously appropriate. For example, if the relevant evaluative focal point is a rule it may seem obvious that one has a reason to conform to the rule. Explaining what relation an option must bear to an explanatory focal point would be an important part of providing a complete implementation of my proposed solution to the puzzle of inefficacy. However, my aims in this section are

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

240  Tristram M c Pherson instead broadly illustrative. In light of this (and the fact that I am currently unsure of what such a general explanation would look like) I will follow common practice and offer an intuitively plausible account of this relation. This account begins by noting that certain actions are of a type which, col­lect­ive­ly, (i) ground the existence of the relevant pattern, and (ii) function dia­chron­ic­al­ly to sustain that pattern. I will say that such actions contribute to the pattern. For example, driving on the right, telling others about the driving convention, and censuring those who don’t abide by the convention, are each types of behavior that contribute to the convention of driving on the right. Next consider what it would be for this convention to fail: that would be constituted by a critical mass of people abandoning the convention in favor of driving on the left, or (cata­stroph­ ic­al­ly) any which way. There are two ways for an agent to contribute to such abandonment: by withholding sustaining behaviors, and by actively resisting the pattern. For example, someone who simply withdraws from driving on roads, and from advertising or advocating the right-­side driving norm has withheld sustaining behaviors. On the other hand, someone who drives on the left, advocates doing so, and censures those who drive on the right, is actively resisting the right-­driving social pattern. The ethical evaluation of a given social pattern is clearly relevant to the ethical evaluation of contribution to that pattern. Some social patterns are overall virtuous (their good effects are striking and/or uniform enough to warrant uncomplicated positive appraisal), others are overall vicious (their bad effects are striking and/or uniform enough to warrant uncomplicated negative appraisal). Besides these clear cases, there will be a range of more difficult cases. For example, it has been strikingly common for social patterns to have deeply mixed effects. Two examples: some influential anti-­racist movements have been systematically sexist, and despite being awash in injustice, economic globalization has improved the ability of billions of people to meet many of their basic needs. Other patterns will tend to preserve a good but strikingly suboptimal equilibrium. It is a difficult question what reasons we have in relation to such mixed or suboptimal patterns. Here, I focus on the clearly virtuous and vicious patterns, leaving mixed and suboptimal patterns as important hard cases to theorize later.28 With these assumptions in place, there is a natural way to implement the idea that contribution is the ethically relevant relation to self-­sustaining social patterns qua evaluative focal point. We can sum up the natural idea in an ethical principle:

28  One might think that Climate is a mixed case: in our current circumstances, any policy sufficient to prevent catastrophic climate change would negatively impact billions of vulnerable actually existing people. While this is tragic, I deny that it makes Climate a mixed case. These impacts are best understood as (alarming) transition costs for moving away from a vicious pattern. That a pattern possesses such costs in no way renders it less vicious. For example, attempts to reduce sexism and racism predictably generate backlash phenomena that harm many vulnerable people. This does not by itself make the prior racist and sexist patterns any less uniformly vicious.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  241 Contribution  Each agent has prima facie reason to perform an action in virtue of that action: •  contributing to a virtuous social pattern, or •  withholding contribution from a vicious social pattern, or •  resisting a vicious social pattern As I explained above, a token behavior counts as contributing, withholding, or resisting in virtue of instantiating a behavioral type, patterns of which do or could sustain or undermine the relevant social pattern. Consider two examples to illustrate Contribution. First, consider a standard toy case of contribution to a virtuous pattern: a local bus system which is funded by riders paying a fare, which is paid on the honor system. Such a bus system can be a virtuous social structure: as a result of its persistence, many people (e.g. persons with certain disabilities, those who lack their own vehicles) have greater mobility, and traffic congestion and air pollution from cars is reduced, benefitting almost everyone in town. Contribution implies that each has a prima facie reason to contribute to sustaining this bus system. Paying one’s fare when riding is the most salient form of contribution, but there are presumably other ways one could contribute. For example, an expert mechanic could volunteer to maintain some of the buses; a citizen could campaign politically on behalf of the bus system, etc. Next consider a case involving a vicious pattern: Sewage  There are two communities along a river, which is their only source of water. Members of Upstream typically dispose of their sewage in the river. As a result, members of Downstream are constantly painfully and dangerously ill from drinking the polluted water. However, the river is so uniformly polluted by this sewage that the sewage dumped by any individual in Upstream makes no difference to the number or severity of the illnesses in Downstream. Contribution implies that each denizen of upstream has reason to withhold contribution to the vicious social pattern. This means that each denizen of Upstream has reason to stop dumping their sewage in the river, perhaps entailing that they have reasons to build septic fields instead. It also implies that everyone has reason to resist the vicious social pattern. One salient way for a member of Upstream to do so would be to campaign politically for an effective sewage treatment plant in Upstream.29 29 Examples like the bus system case are often used to motivate fair play principles (e.g. Klosko  2004). It may thus be useful to briefly contrast Contribution with fair play principles. One contrast with Contribution is that is that the fair play principle allegedly explains the existence of prima facie duties while Contribution only entails the existence of reasons. A second is that

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

242  Tristram M c Pherson Finally, consider how Contribution explains Intuitive. The exceptionally bad outcomes mentioned in Animals and Climate are plausibly largely explained by the existence of self-­sustaining social patterns. In each case, these patterns are enormously complicated, including, for example, all of: certain market patterns, economic incentives, regulatory (in-)action and other structuring government interventions, and prevailing ethical, social, and cultural assumptions. A complete story about the implications of Contribution for these cases would require an extremely difficult tracing of all of these patterns. However, it is not hard to adduce some plausible implications. For example, it is plausible that by buying or eating meat, one contributes to the persistence of the vicious pattern of treatment of non-­human animals in contemporary animal agriculture: a pattern of such meat-­eating is required to sustain demand for such agriculture. Contribution entails that one has a reason to withhold contribution to this vicious social pattern, for example by refraining from eating meat. Contribution also entails that we have reasons to resist this social pattern, for example by engaging in certain forms of political activity. (Similar points apply to Climate.) In this section I have illustrated the general solution to the explanatory dimension of the puzzle of inefficacy developed in the previous section I have done so by proposing an ethical principle (Contribution). I have introduced plausible assumptions about the explanation of the bad effects mentioned in Animals and Climate. And I have sought to explain how, with those assumptions in hand, a principle like Contribution implements the explanatory strategy abstractly characterized in the previous section.

5.  Negligibility and Interaction The previous section concludes the main task of this chapter: to offer and illustrate an attractive solution to the explanatory side of the puzzle of inefficacy. However, as I emphasized in §2, there is a second dimension to the puzzle: intuitive considerations can suggest that even if there are reasons of Contribution, they must be vanishingly weak, and hence irrelevant for practical purposes. Further, the explanatory and negligibility sides of this puzzle interact. To put the point sharply: if the negligibility side of the puzzle cannot be answered, this suggests that any solution to the total puzzle will have to be significantly

Contribution can provide a reason to contribute to a virtuous social pattern that does not benefit oneself, while fair play obligations are allegedly triggered by benefitting from cooperative patterns. A third contrast is suggested by Sewage: here there is no existing cooperative enterprise, so there are no fair play reasons. (The point of this note is to clarify Contribution, not to criticize fair play, which may also be a legitimate ethical principle.)

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  243 revisionary. And that makes it more attractive to take seriously alternatives to my approach which reject either Inefficacy or Outcomes Explain. In light of this, I briefly address the negligibility side of the puzzle in this section. I do not aim to develop a satisfactory explanatory account of the relevant reasons here. Instead, I aim to show how the account I have sketched thus far is compatible with a promising intuitive reply to the negligibility side of the puzzle. I begin by more clearly motivating the negligibility idea by focusing on a variant of the Sewage example introduced in the previous section. Suppose that, to cease contributing to the vicious pattern that is sickening the denizens of Downstream, I would need to spend $10,000 to build a septic system. Suppose that I can do this. Suppose that I can also invest time and money campaigning politically to change the overall pattern in Upstream. Arguably, by doing these things, I would have done what Contribution entails that I have reason to do: withhold and resist. Suppose further that I realize that, rather than spending $10,000 to build a septic system for my sewage, I could instead donate this money to a charity which distributes effective water filters in Downstream. These water filters function to mitigate the harms of the sewage from Upstream. I understand this, and am also certain that my building the septic field will make no difference to the amount of suffering etc. in Downstream. Suppose that I nonetheless choose to use my $10,000 to build the septic system, rather than donating to the charity. In this case, I have plausibly acted badly.30 Contribution is a thesis about reasons, not a thesis about acting well. Plausibly, whatever reasons Contribution gave me to build the septic system in this case were outweighed by reasons I had to donate filters. It is, of course, wholly compatible with Contribution that the reasons it explains are sometimes outweighed. In this context, the negligibility challenge is that the reasons provided by Contribution are always outweighed by reasons grounded in individual difference-­making. The case being discussed makes the negligibility challenge seem plausible. To see this, consider altering the case just described by varying the efficacy of the water filters (or the charity that provides them). As their efficacy declines, donating $10,000 to the charity will do less to mitigate the relevant harms. This has implications for our reasons to donate to this charity. It is plausible that such ­reasons become weaker as the filters become less effective. However, it is not clear that at any positive level of filter efficacy, I have stronger reasons to build the septic field (which, remember, will not avert any harm) than I do to donate to the charity. And this might seem to suggest that the weight of the reasons that are explained by Contribution is negligibly small. 30  The badness is in some ways structurally similar to the badness of Parfit’s imagined agent who stays to help a group of people save one hundred miners (who would all have been saved without the agent’s help), instead of leaving to save ten miners who will die without the agent’s help (1984: 67–68).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

244  Tristram M c Pherson Once the discussion turns to the weight of reasons, the theoretical structure I have developed in this chapter is not obviously illuminating. (That structure sought to explain the existence of certain reasons.) Nor do I know how to construct a systematic and principled account of how various classes of reasons interact. So what follows will be largely a matter of advertising the plausibility of a collection of linked intuitive ideas that are compatible with Contribution. These ideas can be illuminated by considering pairwise comparisons among various options. The idea is that these comparisons help to clarify how we should think about the strength of the reasons favoring each option. Consider four possible options we can imagine being available to a denizen of Upstream confronting a version of Sewage, with $10,000 to spend: (a) Donating this $10,000 to the water filter charity (b) Using this $10,000 to build the septic field (c) Using this $10,000 to procure an absurdly expensive, yet only somewhat tasty and nutritious, sandwich (d) Destroying this $10,000 My procedure is to imagine a series of cases in which one has a forced choice between a pair of these possible options. (If it helps, you can imagine an evil demon will bring about an apocalyptically bad outcome, if one fails to select one of the pair of options.) The negligibility side of the puzzle was motivated by a choice between options (a) and (b), where it seems clear we have more reason to select (a). Now consider (c): presumably, the fact that the sandwich would be (somewhat) tasty is a reason to select this option. Think of this option as producing a small gustatory improvement over a baseline; selecting an option other than (c) isn’t going to make you starve. You might be tempted to deny that you had any reason to select (c), arguing that that’s no reason to spend a lot of money on a mediocre sandwich! But the comparison with (d) is telling: surely the tastiness and nutrition provided by the mediocre sandwich give you more reason to buy it than you have to destroy the money. Next, consider a forced choice between (b) and (c). Here, it seems plausible that we have more reason to build the septic field. Recall that a central feature of the objection we are considering is that the reasons favoring (a) remained weightier than the reasons favoring (b) even when we altered relevant facts about (a) to make those reasons less weighty. Similarly, it is plausible that we can substantially increase the tastiness of the sandwich, and the number of enjoyable sandwiches that $10,000 will buy you, and still have more reason to select (b) than to select (c). I take these pairwise comparisons to reveal something about the weights of the relevant reasons independently of the pairwise comparison.31 Thus it seems 31  Some philosophers are contrastivists about the weight of reasons (e.g. Snedegar  2017). I am inclined to prefer a non-­contrastive theory of weight. However, what is important for my purposes is

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  245 plaus­ible to me that, if we were presented with all of the options at once, the ­reasons favoring (a) are systematically weightier than the reasons favoring any of the other options. It is also plausible that those favoring (b) are systematically weightier than those favoring (c) or (d). If correct, this conclusion matters for real life ethical deliberation. Consider Animals for a moment. In many realistic cases, it will be far from clear whether we face choices between (i) doing things that will make a positive difference to the well-­being of factory-­farmed animals, and (ii) withholding contribution to the relevant vicious social structure, for example by being vegan. This is for two reasons. First, we may sometimes be ill-­placed to mitigate the bad effects of many of the vicious social structures we contribute to. (Relative to our toy example, that is like lacking option (a).) Second, where we can mitigate those bad effects, it will rarely be that we need to (e.g.) eat animal products in order to do so. Because of this, in either of these sorts of cases, reasons of contribution will not weigh against difference-­making reasons grounded in the same values. And the picture just sketched suggests that they could be quite weighty in such cases. Of course, there are less ordinary cases (like the Oracle case mentioned in §2), where difference­making reasons weigh against reasons of Contribution. But the existence of such cases in no way undermines the practical significance of reasons of Contribution. An act-­consequentialist will not be impressed by this reasoning. Suppose that, in Sewage, there is no water filter charity, but one could donate one’s $10,000 to the Against Malaria Foundation (mentioned in §2), thereby saving many people from the awful effects of contracting malaria for which they cannot afford ef­fect­ ive medication. Surely, the act-­consequentialist will say, one has more reason to donate than to build the septic field. And this case generalizes: in our connected world, those of us in positions of privilege virtually always have options that en­able us to make a difference to valuable outcomes. And it might seem that we always have more reasons to perform such actions, than we have to perform actions for which we have reasons of Contribution. This response is powerful. However, it is also controversial in ways familiar from the literature on our reasons to make the world better. It is plausible that my reasons to mitigate the bad effects of the social pattern to which I contribute (e.g. by donating to the water filter charity) systematically outweigh my reasons to withhold contribution to that vicious social pattern. It is not nearly as clear that my reasons to make a difference to bad outcomes per se always outweigh my ­reasons to withhold contribution to a vicious pattern.32 that the reasons favoring the options I have been discussing appear to behave in a way that can be naturally modeled non-­contrastively. 32  Pogge (2007: 16–17) famously argues that each affluent person is very likely to be causally implicated in the suffering and death of some vulnerable person (which the affluent person has no way of identifying). If correct, this might suggest an interesting middle case between cases of known contribution and cases with a known lack of contribution.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

246  Tristram M c Pherson This is a very rough picture of how different sorts of considerations might interact, on which reasons explained by Contribution have a non-­trivial role. Of course, this picture is sketchy and speculative. I have not offered any principled explanation for why the reasons of these various kinds would interact in the ways I have suggested. But that is not my aim here. My aim is rather defensive: to undercut the appearance that—in light of the sorts of considerations suggested by cases like Oracle and the modified version of Sewage—that Contribution could only explain the presence of reasons so weak as to be irrelevant to our de­lib­er­ ation. The discussion of cases (a)–(d) suggests to me that this is in fact not prima facie plausible. And this in turn means that controversial theoretical premises will be needed to transform the intuitive negligibility side of the puzzle into a serious objection to Intuitive.

6. Conclusions This chapter has argued for an explanation of how certain bad outcomes—such as the horrifying effects of factory farming and climate change—can give us reasons to perform certain actions, even if those actions cannot mitigate those outcomes. This intuitive idea faces a deep puzzle. The first dimension of the puzzle is ex­plana­tory. Roughly: how could facts about the goodness or badness of outcomes give us reasons to do things that do not affect those outcomes? The second dimension of the puzzle is that reflection on certain cases can make it seem that even if we had such reasons, they would be negligible: so weak, we could sys­tem­at­ ic­al­ly ignore them in our deliberation. The heart of this chapter focused on addressing the first challenge, taking ser­ ious­ly (i) that it is an explanatory challenge, and (ii) the level of generality characteristic of that challenge. A satisfactory explanation at this level of generality requires that we can give a principled explanation of why certain things are evalu­ ative focal points: the core links between the evaluation of states of affairs and (in the relevant case) the explanation of why there are reasons for certain of agents’ options. I have offered what I take to be an appealing proposal at this level, which I called: Explanatory Link  It is sufficient for something to be an evaluative focal point that it be an explanatory focal point in a difference-­making explanation of the instantiation of evaluatively-­relevant properties I argued that Explanatory Link, together with some plausible claims about the  explanatory potency of social structures or patterns, enables us to see how the explanatory dimension of the puzzle can be solved. I illustrated this

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  247 schematic solution by introducing and motivating the ethical principles of Contribution Ethics: Contribution  Each agent has prima facie reason to perform an action in virtue of that action: •  sustaining a virtuous social pattern, or •  withholding contribution from a vicious social pattern, or •  resisting a vicious social pattern After explaining the distinctive implications of Contribution, I offered a speculative discussion of the interacting weights of reasons. The aim of this discussion was to show that the threat of negligibility can be answered by locating the initially worrying cases within a broader set of cases that at least initially suggest a more nuanced account of how different sorts of ethical consideration interact. Once we do this, it is much less plausible that cases like Oracle are best explained by the fact that our reasons grounded in facts about the value of outcomes are always related to those outcomes via an act-­promotion structure. The puzzle addressed by this chapter is generated by assuming Intuitive: Intuitive  The bad effects mentioned in Animals and Climate give one (­non-­negligible) reasons to perform certain actions, even if those actions will not mitigate those bad effects Although I have sought to motivate this claim, I have certainly not established it. As I explained in the Introduction, my aims are more modest. The puzzle of inefficacy provides a substantial reason to doubt whether we should continue to accept Intuitive. In this chapter, I have sought to clarify the general form of an explanatory vindication of Intuitive in the face of this puzzle. And this, I take it, is a crucial step in the larger project of assessing whether Intuitive is true. This chapter leaves much to be done before we have a fully developed account of this fragment of ethics. One especially pressing task, mentioned in §4, is to develop a principled explanation of why reasons are grounded in one rather than another relationship between an option and an evaluative focal point. Achieving that would put us in a position to systematically evaluate Contribution against competing attempts to vindicate Intuitive, such as complicity principles, or prin­ ciples which instruct us to avoid benefitting from injustice. Even if the Contribution structure advertised in §4 does prove promising, much needs to be done to develop it. For example: (i) considering social structures that are apt to enter into difference-­making explanations of evaluatively rele­vant facts, but which are not self-­sustaining social patterns, in the sense

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

248  Tristram M c Pherson discussed; (ii) identifying a principled account of implications for individual ­contribution of social patterns that are neither unproblematically virtuous nor uniformly vicious; and (iii) theorizing the ethical significance of a range of ways in which individual psychologies and behavior can link to such social patterns, beyond those mentioned in Contribution. Finally, this chapter advertises the significance of systematic attention to explanatory questions concerning the conditions under which a feature can be an evaluative focal point. This question is important for a very wide range of projects in normative ethics, and yet it has been strikingly neglected. While I find Explanatory Link plausible, we would be in a much better position to evaluate it if the question it purports to answer received more systematic philosophical attention from diverse perspectives.33

References Adams, R. (1976). “Motive Utilitarianism.” Journal of Philosophy 73(14): 467–481. Baatz, C. (2014). “Climate Change and Individual Duties to Reduce GHG Emissions.” Ethics, Policy & Environment 17(1): 1–19. Baxter, W. F. (1974). People or Penguins: The Case for Optimal Pollution. New York: Columbia University Press. Brandt, R. (1992). Morality, Utilitarianism, and Rights. Cambridge: Cambridge University Press. Budolfson, M. (Ms-a, n.d.). “Collective Action, Climate Change, and the Ethical Significance of Futility.” Manuscript. Budolfson, M. (Ms-b, n.d.). “The Inefficacy Objection to Deontology: What It Is, Why It Is Important, and How to Respond to It.” Manuscript. Budolfson, M. (2019). “The Inefficacy Objection to Consequentialism and the Problem with the Expected Consequences Response.” Philosophical Studies 176: 1711–1724. Butt, D. (2007). “On Benefiting from Injustice.” Canadian Journal of Philosophy 37(1): 129–152. Calhoun, C. (2015). “But What About the Animals?” in Reason, Value and Respect: Kantian Themes from the Philosophy of Thomas  E.  Hill, Jr., ed. M.  Timmons and R. M. Johnson. New York: Oxford University Press, 194–214. Carruthers, P. (1992). The Animals Issue. Cambridge: Cambridge University Press.

33  I am grateful for illuminating discussion to audiences at the Vermont Food Ethics Workshop and at my keynote address to the Bowling Green Graduate Conference in Applied Philosophy. I am also indebted to Mark Budolfson, Justin D’Arms, Tyler Doggett, David Faraci, Coby Gibson, Eden Lin, Michael Morck, Josh Peterson, David Plunkett, Abe Roth, and Evan Woods for helpful comments.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  249 Chartier, G. (2006). “On the Threshold Argument against Consumer Meat Purchases.” Journal of Social Philosophy 37(2): 233–249. Copp, D. (1995). Morality, Normativity, and Society. Oxford: Oxford University Press. Cullity, G. (2000). “Pooled Beneficence,” in Imperceptible Harms and Benefits, ed. M. Almeida. Dordrecht: Kluwer, 1–23. Darwall, S. (2006). The Second Person Standpoint. Cambridge, MA: Harvard University Press. Dietz, A. (2016). “What We Together Ought to Do.” Ethics 126(4): 955–982. Driver, J. (2016). “Individual Consumption and Moral Complicity,” in The Moral Complexities of Eating Meat, ed. B.  Bramble and B.  Fisher. New York: Oxford University Press, 67–79. Foot, P. (1977). “Euthanasia.” Philosophy & Public Affairs 6(2): 85–112. Fredrickson, G.  M. (2002). Racism: A Short History. Princeton, NJ: Princeton University Press. Frey, R.  G. (1983). Rights, Killing, and Suffering: Moral Vegetarianism and Applied Ethics. Oxford: Blackwell. Frey, R.  G. (2004). “Utilitarianism and Moral Vegetarianism Again: Protest or Effectiveness?” in Food for Thought: The Debate over Eating Meat, ed. S. F. Sapontzis. Amherst, NY: Prometheus Books, 118–123. Frye, M. (1983). The Politics of Reality. Freedom, CA: Crossings Press. Goodin, R.  E. and C.  Barry (2014). “Benefiting from the Wrongdoing of Others.” Journal of Applied Philosophy 31(4): 363–376. Hardin, G. (1968). “The Tragedy of the Commons.” Science 162: 1243–1248. Haslanger, S. (2016). “What Is a (Social) Structural Explanation?” Philosophical Studies 173(1): 113–130. Hooker, B. (2000). Ideal Code, Real World. Oxford: Oxford University Press. Huemer, M. (2018). “Dialogues on Ethical Vegetarianism.” Between the Species 22(1): article 2. Jacobson, D. (2011). “Fitting Attitudes Theories of Value.” The Stanford Encyclopedia of Philosophy, ed. E.  N.  Zalta. https://plato.stanford.edu/archives/spr2011/entries/ fitting-attitude-theories/. Kagan, S. (1998). Normative Ethics. Boulder, CO: Westview Press. Kagan, S. (2011). “Do I Make a Difference?” Philosophy & Public Affairs 39(2): 105–141. Katz, C. (2017). “Contractualism, Person-Affecting Wrongness and the Non-Identity Problem.” Ethical Theory and Moral Practice 21(1): 103–119. Klosko, G. (2004). The Principle of Fairness and Political Obligation. Lanham, MD: Rowman & Littlefield. Korsgaard, C. (2004). “Fellow Creatures: Kantian Ethics and Our Duties to Animals,” in Tanner Lectures on Human Values, vol. 24, ed. G.  B.  Peterson. Salt Lake City: University of Utah Press, 77–110.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

250  Tristram M c Pherson Kutz, C. (2000). Complicity. Cambridge: Cambridge University Press. Lawford-Smith, H. (2018). “Does Purchasing Make Consumers Complicit in Global Labour Injustice?” Res Publica 24(3): 319–38. Lepora, C. and R. E. Goodin (2013). On Complicity and Compromise. Oxford: Oxford University Press. McPherson, T. (2016). “Why I am a Vegan (and You Should be One Too),” in Philosophy Comes to Dinner, ed. A.  Chignell, T.  Cuneo, and M.  Halteman. New York: Routledge, 73–91. McPherson, T. (2017). “How to Argue for (and against) Ethical Veganism,” in Food, Ethics, and Society, ed. A.  Barnhill, M.  Budolfson, and T.  Doggett. New York: Oxford University Press, 822–845. McPherson, T. (2018a). “Naturalistic Moral Realism, Moral Rationalism, and NonFundamental Epistemology,” in The Many Moral Rationalisms, ed. K.  Jones and F. Schroeter. New York: Oxford University Press, 187–209. McPherson, T. (2018b). “The Ethical Basis for Veganism,” in The Oxford Handbook of Food Ethics, ed. A.  Barnhill, M.  Budolfson, and T.  Doggett. Oxford: Oxford University Press, 209–240. Marx, K. (1992 [1867]). Capital Vol. 1, trans. B. Fowkes. Harmondsworth: Penguin. Matheny, G. (2002). “Expected Utility, Contributory Causation, and Vegetarianism.” Journal of Applied Philosophy 19(3): 293–297. Mills, C. (2015). “Racial Equality,” in The Equal Society, ed. G.  Hull. Lanham, MD: Lexington Books, 43–72. Nefsky, J. (2017). “How You Can Help, Without Making a Difference.” Philosophical Studies 174: 2743–2767. Norcross, A. (2004). “Puppies, Pigs, and People: Eating Meat and Marginal Cases.” Philosophical Perspectives 18(1): 229–245. Nye, H. (2018). “The Wrong Kind of Reasons,” in Routledge Handbook of Metaethics, ed. T. McPherson and D. Plunkett. New York: Routledge, 340–354. Parfit, D. (1984). Reasons and Persons. Oxford: Clarendon Press. Pasternak, A. (2014). “Voluntary Benefits from Wrongdoing.” Journal of Applied Philosophy 31(4): 377–391. Pettit, P. (2007). “Joining the Dots,” in Common Minds: Themes from the Philosophy of Philip Pettit, ed. G.  Brennan, R.  Goodin, F.  Jackson, and M.  Smith. Oxford: Clarendon Press, 215–344. Pogge, T. W. (2007). “Severe Poverty as a Human Rights Violation,” in Freedom from Poverty as a Human Right: Who Owes What to the Very Poor?, ed. T. Pogge. New York: Oxford University Press, 11–53. Rachels, J. (1997). “The Moral Argument for Vegetarianism,” in Can Ethics Provide Answers? And Other Essays in Moral Philosophy. Lanham, MD: Rowman & Littlefield, 99–107.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

The Puzzle of Inefficacy  251 Rowlands, M. (2002). Animals Like Us. New York: Verso. Sandberg, J. (2011). “‘My Emissions Make No Difference’: Climate Change and the Argument from Inconsequentialism.” Environmental Ethics 33(3): 229–248. Sandler, R. (2010). “Ethical Theory and the Problem of Inconsequentialism.” Journal of Agricultural and Environmental Ethics 23: 167–183. Schwartz, D. T. (2010). Consuming Choices: Ethics in a Global Consumer Age. Lanham, MD: Rowman & Littlefield. Shafer-Landau, R. (1994). “Vegetarianism, Causation, and Ethical Theory.” Public Affairs Quarterly 8(1): 85–100. Singer, P. (1980). “Utilitarianism and Vegetarianism.” Philosophy & Public Affairs 9(4): 325–337. Sinnott-Armstrong, W. (2005). “It’s Not My Fault: Global Warming and Individual Moral Obligations,” in Perspectives on Climate Change: Science, Economics, Politics, Ethics, ed. W.  Sinnott-Armstrong and R.  B.  Howarth. Amsterdam: Elsevier, 285–307. Smart, J. J. C. (1956). “Extreme and Restricted Utilitarianism.” Philosophical Quarterly 6(25): 344–354. Snedegar, J. (2017). Contrastive Reasons. Oxford: Oxford University Press. Suikkanen, J. (2014). “Contractualism and Climate Change,” in Canned Heat: Ethics and Politics of Climate Change, ed. M.  Di Paola and G.  Pellegrino. New York: Routledge, 115–128. Talbert, M. (2006). “Contractualism and Our Duties to Nonhuman Animals.” Environmental Ethics 28: 202–215. Thomson, J. J. (1985). “The Trolley Problem.” The Yale Law Journal 94(6): 1395–1415. Wood, A. (1998). “Kant on Duties Regarding Non-Rational Nature.” Proceedings of the Aristotelian Society, Supplementary Volume 72: 189–210. Young, I.  M. (1990). Justice and the Politics of Difference. Princeton, NJ: Princeton University Press. Young, I. M. (2011). Responsibility for Justice. New York: Oxford University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

11

On Individual and Shared Obligations In Defense of the Activist’s Perspective Gunnar Björnsson

1.  The Activist’s Perspective Issues like climate change, massive inequality, or threatening fascism pose ­col­lect­ive practical problems: problems that will only be resolved if a great many people contribute. People who make substantial efforts to help resolve such prob­ lems often take what I will call “the activist’s perspective,” thinking that: collective capability:  Together, we can significantly resolve the issue (at an acceptable cost). collective obligation:  In virtue of collective capability, we have a moral obligation to significantly resolve the issue. resulting individual obligations:  The considerations that ground this collective obligation also ground individual prima facie moral obligations to help resolve the issue as best they can, at least when there is a high enough chance that it will be significantly resolved. The first two elements of the activist’s perspective attribute capability and obliga­ tions to some relevant group of agents. In the case of climate change, for example, the group might include all affluent contemporaries; in the case of racism, mem­ bers of the racist society in question. Importantly, these attributions seem to be non-­distributive: they do not imply that each individual member can significantly resolve the issue or has an obligation to do so. By contrast, the third element does attribute ­obligations to the individuals in the group. collective capability is exactly the sort of fact that motivates activism (see e.g. Van Zomeren et al. 2008), likely in part because it is seen as giving rise to collective and individual obligations. Admittedly, explicit ideas about “obliga­ tions” might play little role in some activists’ thinking about collective practical problems. But consider failures to resolve such problems: to significantly decrease the risk or extent of climate catastrophe, or prevent a fascist takeover of a consti­ tutional democracy, say. My impression is that the typical activist thinks that Gunnar Björnsson, On Individual and Shared Obligations: In Defense of the Activist’s Perspective In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Gunnar Björnsson. DOI: 10.1093/oso/9780198796282.003.0012

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  253 many such failures would be failures to live up to what is morally required of us: we have to resolve these issues and would be to blame if we failed. Similarly, the perspective involves the sense that one would be at least somewhat to blame if one did nothing to help in spite of being aware of the issue, particularly well placed to contribute to its resolution at a low cost, and not already occupied with similarly important tasks.

2.  Two Problems of Individual Incapability and Group Agency The activist’s perspective seems both prevalent and important in motivating col­ lect­ive action. But it can also seem theoretically deeply problematic. The first problem has to do with the sheer scale of the collective issues in question: individual incapability:  With few exceptions, no action by an indi­ vidual human agent can significantly affect whether the issue is resolved. For example, however successful we will be in preventing or reducing the threat of global climate catastrophe, the extent of that success is unlikely to have been much affected by my actions. The same seems true for most moral agents, includ­ ing most that the activist’s perspective is concerned with. In light of individual incapability, it might also seem that no individual obligation:  Because of individual incapability, the importance of resolving the issue at hand and our collective capacity to do so cannot ground individual obligations to act in ways aimed at achieving ­resolution (see e.g. Sinnott-­Armstrong 2005; Nefsky 2011; but cf. Nefsky 2017). This directly contradicts resulting individual obligations. Of course, individuals might have other reasons to contribute: perhaps contributing might show solidarity with others who contribute, or be a way to take on a fair share of the burdens involved (for discussion, see e.g. Brennan and Sayre-­McCord 2015). But this does not seem to fully capture the obvious teleological element of activ­ ism: the activist’s efforts are meant to effectively contribute to the collective end. The second problem with the activist’s perspective concerns collective obligation. As already noted, it should not be understood as implying that any individual in the group has an obligation to resolve the issue, as such reso­ lution is beyond the capacity of any or most individual agents. But the claim also seems deeply problematic if it attributes obligations to the group. So understood, it seems to conflict with the appealing thought that agency requirement:  Only moral agents are subject to moral obligations.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

254  Gunnar Björnsson The agency requirement does not rule out all attributions of obligations to groups. Many have argued that groups can constitute bona fide moral agents: cor­ rectly organized, they can have their own beliefs, goals, rational point of view, and capacity for self-­regulation, free choice, moral thinking, and moral motivation (see e.g. French 1984; List and Pettit 2011; Hess 2014; Björnsson and Hess 2017; for an overview, see Tollefsen 2015). It is also clear that massive problems have been resolved when people have organized in such complex ways, not least by forming states. However, it is not clear how a group can have obligations when, like the groups of collective obligation, it has not yet organized in this way (see e.g. Collins  2013; Lawford-­Smith  2015; for discussion of the problem and proposed solutions, see Björnsson 2020). Moreover, even if the groups in question can be understood as agents or potential agents, this looks unlikely to help make sense of the activist’s perspective. Activists are concerned with how to reach out and engage other individuals in these groups, have in mind how diffi­ cult it is for individuals to contribute, and think about what can be achieved if individuals can be motivated by the moral urgency of the issue at hand. It is not central to the perspective that all relevant agents organize around some one plan or accept the same decision-­making procedure. Climate catastrophe prevention, for instance, will realistically happen through a wide variety of actions motivated by the threats of climate change, actions which, although often locally co­ord­in­ ated, will not necessarily be guided by one global master plan or procedure to which all or even most agents sign on. It is just not part of this picture that the group is or will become a moral agent in its own right, much less that it already satisfies standard requirements of moral agency.1 The task of this chapter is to explain how, in spite of problems of individual incapability and group agency, collective obligation and resulting individual obligation are coherent and sensible, in particular with respect to climate obligations. In other work, I have suggested that we should understand the relevant kind of collective obligation as shared obligation (Björnsson 2014a, 2020; cf. Aas  2015; Schwenkenbecher  2014). Though the groups in question are not themselves moral agents, they can straightforwardly have obligations because their members are such agents—or so I have argued. In Sections 3 and 4, I begin by briefly recounting and motivating a general understanding of ­individual obligations before explaining how it naturally yields an account of shared obligations. In Sections 5 and 6, I then explain how this understanding of obligations lets us make sense of the activist’s perspective, in spite of individ­ ual incapability.

1  For one recent book-­length treatment of climate obligations that attributes important such obli­ gations to non-­agential groups, see Cripps (2013).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  255 My ambition is to show that the activist’s perspective on our climate-­related ­ bligations is coherent and sensible in spite of the apparent problems, not necessarily o that it is correct. For this reason, I take a number of things for granted, most obvi­ ously the reality of anthropogenic climate change and the moral importance of the threat it poses. In Sections 5 and 6, I rely on further normative assumptions. I pro­ vide no defense for these assumptions apart from appealing to some initial intuitive support or suggestive examples, as one can agree that the perspective is coherent and sensible without ultimately accepting its normative content.

3. Obligations Can collectives that do not themselves constitute moral agents have moral ­obligations? To answer this question, we would be helped by a clear general understanding of what moral obligations are. Unfortunately for our purposes, there are several notions of moral obligation, and philosophers variously appeal to ob­ject­ive, prospective, and subjective obligations, and to all-­things-­considered, prima facie, and pro tanto obligations. On top of that, talk of what one must do is famously context dependent.2 In what follows, I will therefore propose an explica­ tion of one sort of phenomenon that we have in mind when talking of obligations in moral contexts. The analysis of collective obligations that I propose based on this will then have to be expanded to cover other species of moral obligation. Abstractly, we can say that our obligations are what can be properly demanded of us, with moral and legal obligations being what can be properly morally and legally demanded of us, respectively. What can be morally demanded is a matter of what can be demanded on moral grounds, but demands come in different forms, where failures to satisfy the demand have different consequences. Trivially and generically, one must satisfy moral ideals on pain of being morally non-­ideal. Certain more specific moral demands must be satisfied on pain of making one a fitting target of moral blame in the form of guilt (in the first-­personal case) or indignation. (To the extent that such demands permit actions that are morally ­non-­ideal, they allow for supererogatory actions: actions that are morally better than what is demanded.) Other demands must be satisfied on pain of making one a fitting target of sanctions or attempts to force one to comply that are more hands­on than basic expressions of negative reactive attitudes. (To the extent that such demands allow for acts that merit blame, they allow for suberogatory actions:

2 The locus classicus is Kratzer (1977); for my own understanding of the distinction between must and ought, see Björnsson and Shanklin (2014).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

256  Gunnar Björnsson actions that are permitted but still blameworthy.3) Partly for convenience, and partly because I think these to be an important part of common sense as well as the activist’s perspective, my focus here will be on demands the violations of which normally ground blame. In this section, focus is on individual obligations: individual obligation & blame:  X has a moral obligation to φ if and only if X would be morally to blame for not φ‑ing.4 To better understand this notion of obligation, we will be helped by an account of when one is to blame for not φ-ing. Ideally, such an account should help us explain a number of general facts about moral obligations. One is that someone is required to do something only if doing it is in some relevant sense under her control (“ought implies can”). Another is that obligations take a variety of objects, at least as far as ordinary moral thinking goes: ordinary actions and omissions as well as various cognitive and emotional reactions or lack thereof. Such general facts, I have suggested, can be accounted for if we follow an ­influential tradition of taking blameworthiness to be grounded in the agent’s “quality of will”: quality of will:  One is blameworthy insofar as the object of blame— what one is blamed for—is an “expression” of a bad quality of will (e.g. Strawson 1962; Arpaly 2006; McKenna 2012). As this tradition has made clear, paradigmatic cases of blaming involve thinking that someone did something bad because she didn’t care to the right extent about morally relevant considerations. Conversely, a bad action is typically excused when the agent did it because of external forces or because she was ignorant of crucial information, rather than because of a lack of concern with or respect for the values involved. Though the general idea is powerful in itself, some precisifications and qualifi­ cations are in place.

3  For the category of the suberogatory, see Driver (1992). There are more possibilities: demands that must be satisfied on pain of (i) being a fitting target of moral protest, where this need not involve moral indignation (Smith 2013), (ii) not living up to a standard that should be actively encouraged, where encouragement is different from threats of blame or protest should one fall short, or (iii) not doing what an ideal moral advisor, or any morally reasonable person, would want one to do in the circumstances. 4  This notion of obligation leaves no room for failures to discharge an obligation that are excused due to stress, duress, or disease. If such conditions leave X blameless for not φ‑ing, then they also render it false that X had to φ in the relevant sense. For notions that allow for excuses, see the end of this section and note 20.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  257 First, there are various ways in which one might understand the will that ­ etermines one’s quality of will. I take it that blameworthiness is grounded in d failures to care about certain things in the right way, where caring:  To care about some object in some regard is to be disposed to notice factors relevant to how well it goes with the object in that regard and to invest resources (cognition, action, material and social means) to promote it in light of these factors. On this dispositional account of caring, whether and in what way a certain degree of caring has cognitive or behavioral upshots depends on information and resources available to the agent as well as competing concerns. It should also be noted that this understanding of the will in terms of caring can be combined with different ideas about what objects one needs to care about for one’s will to be good, including both impersonal values (total happiness, justice, beauty) and agent-­relative matters (not lying, not stealing, not killing, being a good human being, citizen, friend, parent, spouse, colleague). Second, not any non-­ideal caring is bad or lacking in a way that grounds blame. We are not blameworthy for every action falling short of ideal virtue. Rather, the caring must fall below some relevant standard, with normative force to back up demands that agents satisfy it and blame when their behavior expresses substand­ ard caring. Moreover, standards do not apply uniformly to all agents or at all times. Most obviously, indignation seems misplaced when directed at an agent who lacks the capacities required for caring about certain things—certain core elements of social and moral cognition, say. More subtly, such capacities come in degrees, and the degree of caring that is demanded might vary accordingly. Furthermore, that degree will depend on the importance of the object of care, and on one’s relation to that object (for an overview of relevant relationships, see Björnsson and Brülde 2017). Quality of will accounts can accept all this but differ with respect to whether they take standards for caring to be grounded in some further considerations, such as rational agency, hypothetical contracts, or costs and consequences of imposing such standards. Third, not any event expressive of substandard caring is a relevant object of blame. The object of blame must itself be morally bad, i.e. of a kind that we have moral reason to prevent. For example, an agent might not be to blame for an expression of substandard caring about his father during free association therapy. However, just as we can be required to care about a variety of objects, a quality of will account can accept a variety of objects of blame, including not only clear cases of harm, rights violations, or suboptimal outcomes, but also attempts to cause harm and failures to notice or be motivated by some morally relevant fact.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

258  Gunnar Björnsson Finally, the object of blame needs to be explained in a normal way by the s­ ubstandard caring, not through some random chain of events (Björnsson 2014a: 114–115; Björnsson and Persson 2012: 330–331). The normal way in which sub­ standard caring explains a bad event (action or outcome) is by explaining agents’ failures to recognize or be motivated by the fact that their behavior would either contribute to this bad event or allow for other, preventable, events that do so. In the simplest case, the bad event is explained by a single instance of substandard caring. But many bad events, such as the deterioration of trust among people or failures to complete a task at a promised time, are upshots of several such instances, sometimes on part of several people.5 The resulting version of the quality of will account is that individual blameworthiness:  X is morally to blame for Y if and only if Y is morally bad and explained in a normal way by X’s substandard caring (Björnsson and Persson 2012; Björnsson 2017a, 2017b). As I have argued elsewhere, this proposal (together with a companion char­ac­ter­ iza­tion of degrees of blameworthiness) accounts for everyday attributions of blameworthiness as well as the role of ordinary excuses and justifications in miti­ gating blameworthiness to various degrees (Björnsson 2017a, 2017b). When we conjoin blameworthiness with obligation & blame, it fol­ lows that X has a moral obligation to φ if and only if X’s not φ‑ing would be bad and explained, in a normal way, by X’s substandard caring. When is X’s not φ‑ing guaranteed to be so explained? Exactly when the satisfaction of these standards would ensure, in a normal way, that X φ.6 If we say that X’s φ‑ing is “morally important” exactly when X’s not φ‑ing would be bad, we get: individual obligation:  X has a moral obligation to φ if and only if X’s φ‑ing would be (i) morally important and (ii) ensured, in a normal way, by X’s car­ing as can be morally demanded (Björnsson 2014a; Björnsson and Brülde 2017). 5  Caring about a value to a certain extent can be understood as a form of skill in promoting that value to the right extent, a skill upheld in part by noticing how well one’s caring does promote the value in question. Following Ruth Millikan (1984), we can then take as normal explanations those that are of the kinds that typically explain the reproduction of such skills. (Given that skills are honed in response to concrete lessons provided by good and bad concrete upshots of levels of skill, we can also say that one is blameworthy for something bad when it is due to one’s substandard caring in such a way that it provides a concrete lesson against that substandard caring; see Björnsson, forthcoming.) 6  That X’s caring appropriately would ensure X’s φ‑ing does not mean that if X cared appropriately, X would φ because of this caring. (The presence of a lifeguard might ensure that no one drowns with­ out actually preventing anyone from drowning; correspondingly, individual obligation allows that one does what one morally must do for non-­moral reasons.) Rather, it means that (i) X φs in all relevant possibilities where X cares appropriately and (ii) in at least some such possibility, X φ-ing is explained by X’s caring appropriately. That it is ensured in a normal way further requires that this explanation is relevantly normal.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  259 Apart from capturing the connection between obligation (of the relevant sort) and blame, individual obligation gives content to the agency require­ ment: for an individual to be subject to moral obligations, it must be subject to moral standards of caring, which presumably requires the capacity to care about morally important matters to the relevant degrees. Relatedly, it straightforwardly accounts for the fact that one has an obligation to φ only if φ‑ing is under one’s control, or something one can do, in the sense that caring appropriately would ensure, in a normal way, that one φs. In addition, individual obligation leaves room for the full range of objects of obligations. That one cares appropriately can ensure that one acts or refrains from certain actions, but also that one notices or ignores certain things, or has (or lacks) certain cognitive or emotional reactions. Moreover, what is ensured might be a series of coordinated actions guided by an explicit intention, such as the act of walking over to the stereo and turning down the volume, as well as a series of separate and individually motivated actions and omissions that has an overall effect, such as the act of letting the neighbor sleep through the night by avoiding activities that make much noise throughout the night.7 This concludes my introduction of the notion of obligation that will be central to our discussion. I stress again that this notion is one among many. Specifically, what individual obligation aims to capture is a notion of all-­things-­considered obligations that depend on available information and current levels of non-­culpable confusion, stress, or duress, rather than prima facie, pro tanto, or objective obligations. To further bring out what sort of notion I have and do not have in mind, I offer the following rough analyses of some neighboring notions for contrast: prima facie:  X has a prima facie obligation to φ when considerations obtain that would characteristically ground a regular (all-­things-­considered) obligation in the absence of contrary considerations. pro tanto:  X has a pro tanto obligation to φ if and only if X has a prima facie obligation to φ based on which it can be properly morally demanded that X cares about φ‑ing.8 objective excuse-­friendly:  X has an objective excuse-­friendly moral ­ bligation to φ if and only if X’s φ‑ing would be (i) morally important and (ii) o ensured, in a normal way if X (a) cared as can be morally demanded, (b) were 7 Notably, individual obligation is structurally similar to virtue ethical identifications of right actions with actions performed by the virtuous (see e.g. Hursthouse 1999: ch. 2), though it is concern with obligations rather than rightness, and appeals to appropriate caring rather than (pre­ sumably more demanding) agential virtue. For a brief discussion of how individual obligation might handle some problems raised for such accounts, see Björnsson (2014a: 116–117, n13). 8  Pro tanto obligations are normative responsibilities to (not) perform certain actions, in the ter­ min­ology of Björnsson and Brülde (2017).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

260  Gunnar Björnsson governed in a normal way by her caring, and (c)  had all morally relevant information.9 non-­objective excuse-­friendly:  X has a non-­objective excuse-­friendly moral obligation to φ if and only if X’s not φ‑ing would be (i) morally important and (ii) ensured, in a normal way if X (a) cared as can be morally demanded and (b) were governed in a normal way by her caring. If something that I say about obligations seems counterintuitive, I urge the reader to make sure that the intuitions concern the intended notion of obligation.

4.  Irreducibly Shared Obligations Return now to the difficulty of understanding obligations of groups that are not themselves moral agents. What I suggest is that such obligations can be adequately understood in terms of the following simple generalization of individual obligation, treating the latter as an account of the special case involving a sin­ gle individual: obligation:  One or more individuals have a moral obligation to φ if and only if their φ‑ing would be (i) morally important and (ii) ensured, in a normal way, by the appropriate caring of them all (Björnsson 2014b, 2020).10 No further adjustments seem to be required to capture obligations of non-­agential groups, as we see when considering the various parts of the account. First, the behavior of a plurality might obviously be morally bad, i.e. the sort of thing that there are moral reasons to prevent. For example, there might be moral reason to prevent a group of teenage yahoos from waking up your sweet elderly neighbor, or from abandoning a cat helplessly stuck in a tree after they chased it up there, just as there might be reason to prevent a single individual from doing those things. Second, obligation allows for the attribution of obligations to groups that are not themselves moral agents in any qualified sense: what is required for group 9  Unlike violations of the individual obligation variety, an agent need not be blameworthy for violating her excuse-­friendly obligations: she might fail to do so because she was not governed in normal ways by her caring or lacked relevant information. These notions of excuse-­friendly obligations rely on a notion of normal governance that might best be understood in terms of Ruth Millikan’s (1984) notion of a Normal explanation. Also see note 5. 10  obligation attributes obligations non-­distributively: obligations of groups are held together, not individually. For related views, see Schwenkenbecher (2019); Aas (2015); Wringe (2016); for dis­ cussion of these, see Björnsson (2020). Corresponding to this account of obligations for both individ­ ual agents and groups is the following account of blameworthiness: blameworthiness: One or more individuals are morally to blame for Y if and only if Y is morally bad and explained in a normal way by substandard caring among those individuals (Björnsson 2011, 2020).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  261 obligations is merely that individual members of the group are subject to moral standards of caring. (Though leaving room for obligations of non-­agential groups, obligation neatly explains why the agency requirement holds for the  cases where it seems most clearly correct, namely for individual obligation ­bearers, if we equate moral agents with agents that are subject to moral standards of caring.) Third, appropriate caring by all members of a group of agents can ensure, in normal ways, a certain behavior from the group. If all the teenagers care appro­ priately about the neighbor, this might straightforwardly ensure that each keeps reasonably quiet in the proximity of her bedroom and that they thereby keep the noise they collectively make below a certain level. If they all care appropriately about the cat, this might similarly ensure that they organize to help it down. Notice that whereas coordination and mutual awareness might be required in the second case, the first case is different. Here, as in countless other cases, appropri­ ate caring avoids cumulative problems by separate efforts of numerous individuals, efforts that need not be made in awareness of the other efforts. This is directly parallel to a typical individual case. It is not the teenager’s caring about the neigh­ bor at one specific time that ensures that the neighbor gets to sleep; it is her caring throughout the night and making efforts to keep the noise down at various times, without necessarily being aware at any of those times of her other efforts.11 If individual obligation is correct and generalizes in this way, the teen­ agers’ moral obligation to help the cat down, or not keep the neighbor awake seems perfectly intelligible, without assuming that the teenagers together consti­ tute a fully-­fledged moral agent. Although the standards that ultimately ground obligations—standards for caring—always apply to moral agents at times, col­lect­ ive obligations of this sort are functions of such demands in the same way as indi­ vidual obligations are. Moreover, these collective obligations imply capability in just the same way as individual obligations: if a group has an obligation to φ, the group’s φ-ing would be ensured, in normal ways, by one or more instances of appropriate caring.12

11  One might worry in general whether the possibility of bringing about some important effect through a series of uncoordinated actions really provides the sort of control needed to ground obliga­ tion (Lawford-­Smith 2015: 233–235; cf. Pinkert 2014; Schwenkenbecher 2014: 68–69). But notice that obligation is friendly to stronger control requirements. In note 6, I said that for X’s caring appro­ priately to ensure, in a normal way, that X φ, there has to be some relevant possibility in which such caring explains, in a normal way, that X φ. It might perhaps be argued that in order for a group’s φ-ing to be explained in a normal way of the appropriate caring of its members, there must be some co­ord­ in­ation between them, though perhaps not of a centralized or very sophisticated sort. 12  Though unstructured groups can be subjects of the obligation variety of moral obligations, only groups of which it can be properly demanded that they care about doing certain things are sub­ ject to the pro tanto variety, as defined above. Some resistance to the idea that unstructured groups can have obligations (e.g. that of Collins 2013) might be best understood as concerned with some­ thing like the latter rather than with obligation obligations. I thank Stephanie Collins for discus­ sion of this issue.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

262  Gunnar Björnsson Because obligation takes group obligations to consist in a number of agents together standing in a certain relation to φ‑ing, I will describe these as “shared” obligations. Notice that such obligations are irreducibly shared. To see this, com­ pare two cases where an individual teenager cannot help the cat down the tree because at least two people are required to erect the ladder and because there is no help available. In the first case there is no help available because the teenager is there alone; in the second case because everyone else present is unwilling to help. In neither case does she have a regular individual obligation to help the cat—she cannot—but in the second, she might share an obligation with the others to do so. The latter possibility also illustrates how shared obligations do not consist in regu­lar individual obligations to achieve some end together with others: given the unwillingness of others, she has no individual obligation to help (Björnsson 2014b, 2020). Thus far I have argued that given an independently supported understanding of obligations, collective obligations of a certain kind—shared obligations—are entirely non-­mysterious. But three questions remain, neither of which has an obvious answer: Shared obligations to resolve collective problems depend on what appropriate caring among individual agents would ensure. But given that an appropri­ ately caring individual is ­unable to affect whether such obligations are discharged—given individual incapability—why would she be motivated to contribute? In answering this question, I will argue that such an individual is largely motivated by her own ability to improve the expected value of the outcome. But given this, why are activists motivated by the group’s ability to resolve the issue? Given that an individual is incapable of affecting whether a collective problem is resolved, how can we account for the activist’s motivation to help resolve it, and to do so in effective ways? I address these questions in the following two sections, focusing on the issue of threatening climate catastrophe.

5.  Collective Obligations in Spite of Individual Incapability Our account of shared obligations might seem to make straightforward sense of collective obligation for the case of threatening climate catastrophe. First, according to obligation, we have a moral obligation to significantly reduce that threat insofar as our caring appropriately would ensure, in a normal

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  263 way, that we do. Second, it seems plausible that we can be morally required to care about things threatened by climate change, in particular the wellbeing of people in areas most likely to be hit the hardest by flooding, drought, disease, social unrest, and war. Of course, our production and use of energy will have to change radically in a generation, imposing significant costs on many people. But key technologies have developed swiftly even without costly measures, and more forceful policy changes, should such receive widespread support, would speed up the process even more. Given the magnitude of what is at stake, surely the required sacrifices can be demanded.13 On reflection, though, individual incapability might seem to under­ mine this conclusion. Given how small an impact we can expect of an individual agent’s actions on overall preventive success—given individual incapability— it is unclear why climate threat should prompt even moderately costly action among appropriately caring individuals. That issue is the topic of this section. An obvious first answer is that although we are extremely unlikely to make any major difference with respect to the grand outcome that is the object of our shared obligation, the difference we can make is not thereby morally irrelevant. We might vote for politicians that support important climate legislation, favor climate friendly technologies when choosing energy company and retirement plan, decrease air travel and car use, support emission reduction schemes to offset our remaining carbon footprint, and help spread awareness of the risks and what needs to be done. Taking these steps, one might think, would likely make some contribution in positive direction. After all, whatever threat reduction is eventu­ ally achieved will in significant part be achieved by a multitude of such steps. Given appropriate caring, this could motivate enough action to ensure significant overall reduction. I will argue that this is indeed a sensible perspective, but it too might seem to exaggerate the role we can play as individuals. The first problem is that many of the systems that we operate in have what we might call “buffer zones,” i.e. ranges within which differences in input values make no difference to the overall out­ come. Voting is the most familiar example: unless an election is as close as can be, no one individual vote will make a difference as to who wins. The likelihood that my vote will decide an election with millions of voters is thus typically vanish­ ingly small, especially where reliable advance polling indicates a clear winner. But numerous natural and human systems display similar buffering, with correspond­ ing effects on likelihoods: it might seem extremely unlikely that the vanishingly 13  Exact estimates of required sacrifices vary considerably depending on what assumptions are made. But estimates for quick and decisive global action suggest that individual burdens would fall far short of costs regularly paid by countless people when changes in labor market conditions force re­ loca­ tion or periods of joblessness. For one recent argument that costs are reasonable, see Stern (2015).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

264  Gunnar Björnsson small increase in temperature or sea level rise that might be caused by my ­emissions of greenhouse gases will make a difference to whether an area is hit by drought or forest fires, whether someone dies from dehydration or overheating, or whether flood barriers resist the forces of waves. Moreover, it might seem that unless one’s actions make some concrete identifiable difference with respect to some value—unless it harms an identifiable human being, say—that value ­cannot make performing or abstaining from these actions morally obligatory (Sinnott-­Armstrong 2005: 289–294). All this might suggest that focus on the effects or difference-­making of indi­ vidual actions is a mistake, as well as a bad fit with the activist’s perspective. Perhaps we should also care about not adding to sets of actions that together cause significant identifiable harm, and care about adding to sets of actions that together delivers significant identifiable goods. To capture this wider sense of causal relevance, let us say that an individual action “contributes” to some out­ come if it is part of why the outcome came about. (Understand “action” so as to include omissions.) When a set of actions overdetermine an outcome, individual actions can thus contribute to the outcome without making a difference to it. With this terminology, the suggestion might be put as follows: mere contribution:  If one cares appropriately about some important outcome and thinks that the collective behavior of a group of people can sig­ nificantly promote (obstruct) that outcome, one cares about (not) contributing to that behavior, even if one will not thereby significantly affect the likelihood that the outcome is realized.14 Given mere contribution, it is clear why individuals would care about not adding to the excess of greenhouse gases and about supporting pro-­environmental policies. Depending on the strength of the required caring, this could conceivably ensure a significant reduction of climate threat. For cases involving collective benefits rather than harms, Julia Nefsky (2017; cf. Parfit 1984: 83) has defended a principle related to mere contribution. But she notes that without restriction such principles imply that a valuable outcome can provide an agent with reasons to contribute even when it is known that enough has already been done to secure that outcome. Finding this consequence implausible, Nefsky suggests that the principle applies only to contributions that satisfy non-­superfluity:  An action is non-­superfluous with respect to an ­outcome if and only if it is possible at the time of action that the action will 14 For considerations that might seem to motivate mere contribution, see Brennan and Sayre-­McCord (2015: 53–54); Petersson (2004); Parfit (1984: ch. 3).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  265 contribute to the outcome and possible that the outcome fails to materialize for lack of such ­contributions (Nefsky 2017: §4). Perhaps a similar restriction makes sense for cases of foreseen collective harm. If the actions that others have already taken have ensured the harm, it is unclear why concern with that harm should motivate an agent not to perform another action of the relevant kind. (The agent might of course have reason to avoid such actions under certain circumstances—perhaps doing so will have power­ ful symbolic effects, say—but our concern here is with the activist’s teleological reasons.) For our purposes, adding a requirement of non-­superfluity makes no differ­ ence to the overall relevance of mere contribution, as climate related out­ comes are still very much dependent on future actions. Why, though, should we think that mere contributions matter? The most straightforward answer is that this explains the rightness and wrongness of actions in a variety of cases. The most obvious cases might be ones where the contributor intends the collective harm, such as: Car Push:  A member of the opposition has been trapped in a car. Four ­people are pushing the car over the edge of a cliff, more than enough to get the work done. You cannot prevent the outcome or increase its likelihood, but you nevertheless join in, intending to contribute to resulting harm.15 As Chad Vance (2017) makes plausible, however, contributing to a harm that is merely foreseen can be as serious as contributing to one that is intended: consider a version of Car Push where you join in to get some exercise, not caring about the harm that results from their action. In light of this, it might seem that we should accept mere contribution for the case of (non-­superfluous) collective harms generally. If we care appropriately about climate harms, we will be disinclined to significantly contribute to those harms just to reap minor personal benefits, and inclined to make some personal sacrifice to contribute to climate solutions. While cases like Car Push suggest that mere contributions matter, other cases suggest that mere contributions count for nil, even when they are non-­superfluous. Consider: Unhelpful Contribution:  Approaching the rocky shore, I see two people trying to save a drowning man from the violent waves, with uncertain success. A bystander who I recognize as an off-­duty lifeguard is clearly conflicted about get­ ting involved, screaming that the man in the water bankrupted his family. If the 15  See Sinnott-­Armstrong (2005: 289–290), who advocates a principle that arguably implies mere contribution for intended harms specifically.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

266  Gunnar Björnsson bystander were to join the effort it would almost certainly be successful. I could try to convince him, but this would likely increase his reluctance. I could also contribute to the rescue effort myself, but given my physical limitations and lack of experience, that wouldn’t affect the chance of success. Instead, I’m considering making things easier for another man, about to make it ashore on his own. In this case, my contribution to the rescue effort could very well be non-­superfluous: if I get involved and the person is saved, it could be because of our efforts, and if he isn’t saved, it could be because too few people got involved. Still, it seems that however marginally I could improve the final stretch of the other man’s struggle to get out of the water, it would be enough to trump any reason I would have to contribute to the ongoing saving effort. And the most straightforward explanation for this seems to be that such contributions add no value of their own (cf. Parfit 1984: 83; Dietz 2016: 974–975). The problem is also illustrated by cases involving mere contributions to harm: Unhelpful Abstention:  In large amounts, Medicine minimizes otherwise deadly seizures in certain patients. But some of my colleagues at the town clinic where I work as a nurse have used small amounts of Medicine to effectively help patients with stubborn headaches. As supplies are running low, a group of doctors at the clinic will together decide whether to continue this practice. If they decide to, the one seizure patient in town will be on her own until stocks are replenished; otherwise, supplies will be sufficient for her by some margin. I am aware of all this as I’m considering whether to use a small dose to help that very patient with a headache caused by her recent Medicine-­minimized seizure. Suppose that I use Medicine to relieve the patient’s headache. Then I might ­non-­superfluously have contributed to depriving her of protection against deadly seiz­ures: if she loses the protection, it will be because we have handed out Medicine to relieve headaches, but enough of us might still refrain from doing so, in which case she keeps it. Still, if my action doesn’t increase the risk that she loses her protection, it seems that any degree of headache relief I can offer her out­ weighs whatever reasons I have not to non-­superfluously contribute to depriving her of that protection. Again, this suggests that caring appropriately about moral values does not imply caring about whether one non-­superfluously contributes to good or bad outcomes. Intuitively, cases like Car Push suggest that intended or foreseen non­superfluous mere contributions matter, while cases like Unhelpful Contribution and Unhelpful Abstention suggest the opposite. But this apparent conflict can be straightforwardly explained, I will now suggest, with reference to the following assumption:

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  267 instinctual attraction/aversion:  If one cares about some value, one will be instinctually attracted to (averse to) saliently contributing to what promotes (impedes) that value, i.e. attracted (averse) in ways not relying on a re­flect­ive assessment of how so contributing promotes (impedes) that value compared to alternative actions.16 To say that one is instinctually attracted (or averse) to salient contributions to what promotes (or impedes) one’s values is not to say that one will always try (not) to so contribute. But if one rejects the option one is attracted by or takes the option that one is averse to because one thereby hopes to promote some of one’s other values, one will understand this as in tension with one’s commitment to the first value. Importantly, instinctual attraction/aversion makes good sense given some basic constraints on practical reasoning. At almost any moment, there are potentially infinite different things that we could do if we decided to, but our capacity to consider and carefully compare these alternatives is limited. Because of this, we need pre-­reflective grounds for considering some alternatives rather than others. Part of the solution to this problem is to focus on salient kinds of options that tend to contribute to what we care about and discarding salient kinds that have the opposite tendency. Notably, options that saliently contribute to what is likely to help and harm, respectively, form exactly such kinds: what saliently contributes to what has a good chance of helping or harming itself has a pretty good chance of helping or harming, respectively. It would thus make sense to have mechanisms that pre-­reflectively push us towards considering the former kind of options and away from the latter. And given such mechanisms, we would be relevantly attracted to the former and averse to the latter. instinctive attraction/aversion says that we can be attracted or averse to mere contributions. But it would also explain why a simple principle like mere contribution is too strong. The reason is that the instinctual attraction or aversion will weaken when the action is no longer saliently seen as contribut­ ing to promoting or impeding the object of caring. Most obviously, this will hap­ pen when the harm is already seen as a done deal, in line with Nefsky’s observation. But it will also happen when the very value that grounds the attrac­ tion or aversion prompts consideration of options that might better promote that same value: such options will not be seen as in tension with one’s commitment to the value in question. This, I suggest, is why the possibility of contributing to the saving effort in Unhelpful Contribution seems to carry no weight against the 16  On this view, what one is averse to is being implicated in the collective harm in a particular way. Elsewhere (Björnsson forthcoming), I argue that being implicated in such a way grounds reactive atti­ tudes like indignation and guilt over that harm.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

268  Gunnar Björnsson option of helping the man who is saving himself: the latter option is superior in relation to the very same concern that would make contributing attractive, namely the concern for the people in the water. It is also why the possibility of contributing to the shortage of Medicine in Unhelpful Abstention seems to carry no weight against the option of using Medicine to relieve the seizure patient’s headache: here the latter option is superior in relation to the same concern that would make abstaining from using Medicine attractive, namely concern for the seizure patient. Return now to our initial conundrum. For obligation to yield the result that we have an obligation to significantly reduce the threat of climate catastro­ phe, it needs to be the case that if we each individually cared appropriately about the relevant values, this would ensure such a reduction. The problem was to explain why appropriately caring individuals would take action jointly sufficient for significant threat reduction given that no such individual can affect whether that reduction takes place or not. Though mere contribution seemed implausible, instinctual attraction/aversion has a better chance of filling that gap. Moreover, it identifies the relevant motivation as based on instru­ mental concern with the outcomes at stake, thus capturing an important teleological element of the activist’s perspective. On reflection, though, it is not clear that this source of motivation is sufficient. One immediate problem is that the strength of the relevant attraction and aver­ sion relies on the salience of the contribution. Because individual contributions to climate-­related harms are indirect and diffuse, the obligations will likely be weak, leaving us shy of a full explanation. Another problem arises for those who are interested in promoting the good overall and are seeing climate-­related benefits as elements of that good. Given that perspective, if a mere contribution to climate harm ever so slightly improves the situation with respect to the overall good, it will not be seen as in tension with commitment to that good: the situation will be analogous to that of Unhelpful Abstention. Because of these limitations of our attraction and aversion to certain mere contributions, more needs to be said to ground collective obligation in relation to climate catastrophe prevention. The literature contains the basic elements of what I think is the right comple­ mentary response. Reflecting on worries about buffer zones, many have pointed out that even highly unlikely outcomes matter if they are significant enough (see e.g. Parfit 1984: 73–75; Hiller  2011; Kagan  2011; Singer  1980; Lawford-­Smith 2016). To balance the significance of possible outcomes with their likelihood, our decisions should arguably be based on something like the relative expected values of available alternatives (i.e. by the sums of the products of the value of each of the possible outcomes and their probability for each alternative). Assuming that this is correct, suppose that the minute difference in greenhouse gases that I might affect has some non-­zero likelihood of affecting the occurrence or extent of forest fires, droughts, or whether someone dies of dehydration or overheating, and so

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  269 forth. Even if each such effect is extremely unlikely, there are plausibly a staggering number of possibilities for such effects, each individually of considerable importance. Moreover, unlike in cases of elections with reliable forecasts, it seems that individuals have no way of knowing whether they find themselves in buffer zones with respect to the vast majority of these possible effects. Given this, the difference in expected value that the individual can make by affecting his own or others’ greenhouse gas emissions might still be significant. Of course, in light of the vast complexities, any calculation of expected value will have to rely on reasonable heuristics. A first approximation might be to look at something like the projected average marginal contribution of a given amount of pollution in the range of ­sufficiently likely outcomes. We might then find that the expected negative climate-­related value of an afternoon worth of gas guzzling joy riding is roughly equivalent to that of ruining someone else’s afternoon (cf. Hiller  2011: 355–361). That should be reason enough to consider carbon offsetting or alternative sources of enjoyment.17 Still, one might think that because of the vast complexity and magnitude of the systems at work, no concrete harmful or beneficial effects can be traced back to an individual’s contribution such that this particular contribution can be said to be causally responsible for those effects (Sinnott-­Armstrong 2005: 289–291). Perhaps causation of climate-­related harm and benefit is an emergent phe­nom­ enon that cannot be pinned on the quickly dispersed contributions of any single individual (Kingston and Sinnott-­Armstrong  2018: 175–176). And perhaps the climate-­related consequences of our actions are metaphysically indeterminate rather than merely in principle individually unknowable. If so, then even if the expected degree of climate-­related harm is a positive function of the totality of individual contributions, perhaps there is not only low but zero chance that one individual’s contributions will actually have any definite harmful or beneficial climate-­related consequences. If only the comparative likelihood of concrete, identifiable harmful or benefi­ cial consequences of action were relevant for decision-­making, this could under­ mine duties to reduce one’s own carbon footprint or to otherwise contribute to prevention. On reflection, though, such consequences are just not required for decision-­making relevance, as illustrated by: Carcinogen:  It is in your power to determine whether the population in a town of 35,000 people will be exposed to a carcinogen. Among the people who have the gene in virtue of which they risk developing and dying from a certain form of

17  Some take these and similar assumptions about expected value to be insufficiently motivated (see e.g. Maltais 2013: 593–594; Kingston and Sinnott-­Armstrong 2018: 178–181). However, since my ambition is to make the activist perspective intelligible rather than showing that it is correct, I will not further defend them here.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

270  Gunnar Björnsson lung cancer, exposure increases the average risk of contraction from 50% to 52%, depending in each individual case on numerous factors and indeterministic pro­ cesses. As the expected number of townspeople with the gene is 200, exposure increases the expected number of lung cancer deaths in the town by 4. In Carcinogen, it is not clear that exposing people to the carcinogen would in con­ crete ways harm anyone identifiable. For if you did, we cannot say of any of the people who died from lung cancer that they would have been better off if you had decided against exposure, only that they might have been. Similarly, if you didn’t, we could not say of anyone not developing lung cancer that they would have been worse off if you had allowed exposure, only that they might have been. Nevertheless, the difference in expected value between the alternatives seems to provide an obligation to protect people from exposure absent very strong con­ trary reasons. In terms of obligation: Acting in ways that increase the number of expected deaths is morally bad, and if you cared appropriately about the values and risks at hand, this would ensure, in normal ways, that you decide against exposure. Consequently, you have an obligation not to expose people to the car­ cinogen. Moreover, although this obligation is grounded in the general harmful­ ness of the carcinogen, it does not depend on any requirement to avoid identifiable harm—unless increasing the risk of harm itself constitutes a harm.18 One might agree with me about a case like Carcinogen, but object that in the case of climate change, the effect that a normal individual’s actions has on the expected value of any one individual will be much lower. If the negative expected value of an afternoon’s gas guzzling is the equivalent of a ruined afternoon, for example, the decrease in expected value per individual will not even be one bil­ lionth of that. This, one might think, makes these effects on individuals not only very small, but normatively completely insignificant.19 Something close to this thought is plausibly correct: we cannot be morally required to take one of these effects, on its own, into account in our deliberation. On reflection, though, even in­di­vidu­al­ly morally insignificant differences add up. To see this, consider 18  Whether an increased risk of harm itself constitutes harm, or perhaps more weakly injustice, might matter for the strength of the resulting individual obligations (see e.g. Broome 2012: chs. 4–5). Perhaps, if the effects of my carbon footprint on expected value constitute neither harm nor injustice and avoiding carbon emissions is merely a matter of improving the state of the world, one might think that there are more effective allocations of the resources one would spend: malaria prevention, say. Now, because marginal effects on expectations of large numbers of people add up (as I briefly argue below), and because the poor will bear most of the negative effects of climate change, I do think that sustaining an uncompensated-­for carbon footprint typical of citizens of rich countries typically does constitute injustice. But for the issue that we are currently considering—what individuals would do on the assumption that all cared appropriately—this might not matter much: the marginal expected value of support for what currently yields most good per dollar would likely quickly go down with enough contributions, leaving enough resources to significantly cut net carbon emissions. 19  Cf. Nefsky’s (2011) argument that minor differences in the distribution of goods might be taken not to make a difference as to whether the distribution is fair or not.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  271 Wheel of Fortune: Consider two possible states of a world of 10 billion ­inhabitants. For each, we can represent the relative expected value of an inhab­ itant’s remaining life—their “expectations”—as a position on a circle. At 12 o’clock we find the individual whose expectations are the highest, in the next position, counting clockwise, the individual whose expectations are second highest, and so forth, with the 10 billionth individual, whose expectations are the lowest, just next to the first, closing the circle. (For simplicity, suppose that there are no ties.) The difference in expectations between the first and the last individual is huge, but the difference between any other two adjacent individuals is plausibly morally negligible when taken on its own. The two states of this world, A and B, differ in this regard: In B, the ordering is turned one step compared to A such that each individual has the expectations that the individual in the next position had in A. Imagine that we have to decide whether A or B is realized and have no further information about the two states. Then both the following seem undeniable: no reason:  Since A and B are qualitatively identical with respect to all known morally relevant properties, we do not, on balance, have reason to choose one over the other. better off:  Since the person worst off in A is so much better off in B, we do have a reason to choose the latter over the former. Why is better off not a conclusive reason to choose B? The most straightfor­ ward answer is that the individually insignificant differences add up, without loss. Whatever claim the person who is worst off in A would have on you to choose B, it can be thus outweighed by the fact the everyone else has their expectations ­lowered ever so slightly. (For the argument on which Wheel of Fortune is modeled and further discussion, see Zach Barnett 2018.) Suppose, as Carcinogen suggests, that we can be required to care about effects on expected value even when concrete effects cannot be pinned on us. Suppose further, as Wheel of Fortune suggests, that the effects of our actions on individual expectations add up. And suppose, as previously suggested, that individual con­ tributions to climate catastrophe prevention can significantly affect expected climate-­related harm. It follows that we can have individual obligations to con­ tribute even given individual incapability. It also plausibly follows that we—affluent people living today, say—share an obligation to significantly reduce the threat of climate catastrophe. If all of us care as can be reasonably required of us, we will care about our marginal contribution to such threat reduction, and such caring would likely ensure significant threat reduction, especially in con­ junction with instinctual attraction/aversion.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

272  Gunnar Björnsson Of course, I have not yet said anything about how strongly individuals should care about such effects, or how much action could be expected. The caring required might depend on individuals’ capacity to contribute, their degree of (retro­spect­ive) responsibility for the situation, and the extent to which they have disproportionally benefitted from the system that has given rise to the threat (for a catalogue of possible factors, see Björnsson and Brülde 2017). How much action will be ensured given a certain level of caring will in turn depend on the costs involved and competing considerations. Though it is not part of the current argu­ ment, it is entirely possible that for those of us who belong in the global 1%, car­ ing appropriately will dispose us to considerably lower our material standards of living. But setting this possibility aside, it is easy to see how uniform moderate caring could ensure significant effects. Little or no sacrifice of personal wellbeing is involved in staying somewhat informed or helping raise awareness and spread relevant information, on one’s own or together with others. Similarly, little sacri­ fice is involved in ranking climate policy as particularly important when casting votes in close elections, offsetting one’s carbon emissions, or choosing the climate friendly alternative when options are otherwise roughly equivalent. For most affluent people, performing such actions out of caring about threat reduction would impose limited material hardships and might well involve positive net effects on happiness characteristic of involvement in altruistic causes. Given this, it is not wildly implausible that if everyone cared appropriately, this would ensure that a vast number of individual actions and widespread support from pro-­ environmental policies would greatly speed up the world’s transition towards carbon neutrality, thus significantly reducing climate threat. And from this, together with obligation, it would follow that we have a shared obligation to significantly reduce climate threat.20

20  I’ve suggested that appropriately caring individuals would take steps jointly sufficient for signifi­ cant threat reduction. But might not individuals who care about climate-­related harms fail to draw the conclusion that they have considerable reason to contribute? After all, a number of philosophers reject that very conclusion. And if sufficiently many draw the same conclusion, might appropriate caring fail to ensure significantly prevent catastrophic climate change, thus undermining the idea that preven­ tion is a shared obligation? One response to this worry is to say that the relevant standards for caring require caring about the expected value of one’s actions, independent of identifiable concrete harm or benefits resulting from these actions. On this view, those failing to contribute because they take their inaction to cause no identifiable harm and not caring about expected value would be morally to blame. Another response accepts that the worry might indeed undermine shared obligations of the obligation variety and instead say that it is moral demands on caring together with teleological rationality that require giving weight to expected value. Based on this, one could then say we have a shared obligation under the following, obligation-related, excuse-­friendly, notion of obligation: obligation*: One or more individuals have a moral obligation to φ if and only if their not φ‑ing would be morally bad and their φ‑ing would be ensured, in a normal way, by the appropri­ ate caring and teleologically rationally of them all.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  273

6.  Remaining Problems of Collective Capability and Individual Incapability Thus far we have seen how the activist’s attributions of collective obligations can be understood as concerned with shared obligations. We have also seen reasons to think that we have such a shared obligation in the climate case: if we all cared appropriately, this could well ensure, in normal ways, that the climate threat is significantly resolved. But questions remain. First, to make full sense of the activist’s perspective, we need to explain how our collective ability to resolve the issue can (i) ground our shared obligation (as postulated by collective obligation), and thus (ii) indirectly ground our indi­ vidual obligations to help (as postulated by RESULTING INDIVIDUAL OBLIGATION). Importantly, obligation itself guarantees that we can discharge any shared obli­ gation, in the sense that we would discharge it if we cared as can be properly demanded of us. But as I understand the activist’s perspective, it involves some­ thing more: our collective capacity to resolve an issue is part of what should motivate us to contribute. By contrast, the considerations of capability mentioned in the previous section were concerned primarily with the individual’s capacity to marginally affect expected climate value, contribute to climate related benefits, or avoid contributing to climate related harms. More thus needs to be said. Second, according to the activist’s perspective (as stipulated by resulting individual obligation), our shared obligation to resolve the issue grounds individual (prima facie) obligations to help resolve it as best one can. The prob­ lem here is that activists don’t generally think that a given individual’s actions can affect whether the issue is resolved, or think that this is what grounds the indi­ vidual’s obligations: such a resolution is too vague and too grand a matter, espe­ cially if an individual’s actions might have no determinate climate-­ related upshots.21 A first answer, in line with instinctual attraction/aversion, would be that to help is to contribute to a resolution. But this fails to account for the activist’s sense that individual contributions can be more or less effective in helping to resolve the grand issue. Again, more needs to be said. In this final section, I explain how concerns about important large-­scale out­ comes and considerations of collective capability with respect to such outcomes can intelligibly enter into ordinary individual practical reasoning and make intel­ ligible individual efforts to efficiently help realize these outcomes. Such concerns and considerations, I suggest, play natural roles in what might be the most 21  Given that agents can marginally affect climate-­related expected value, it might be hard to deny that they can marginally affect the likelihood of resolution given various complete precisifications of what counts as a “resolution.” But I do not take that thought to be a natural part of the activist’s perspective.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

274  Gunnar Björnsson common mode of practical reasoning: what I will call issue-­oriented reasoning. Simplifying somewhat, such reasoning paradigmatically proceeds through the following steps: issue-­oriented reasoning Identification of an issue, i.e. some good or bad possibility whose realization might be helped or hindered. Identification of one or more possible means by which that realization might be helped or hindered. A means may be non-­agential: the threatening forest fire might be put out by rain; the tyrant’s reign ended by cardiac arrest. But it will typically involve an agent or group of agents, which may or may not be identified and may or may not involve the agent engaged in reasoning.22 Means might also be more or less abstractly understood, ranging from the broadest outline of what needs to happen to help or hinder the possibility in question to plans painstakingly detailing what would be done by everyone and everything involved. Evaluation of means with respect to effectiveness, likelihood of realization (perhaps conditional upon the active involvement of the reasoner) and various cost and benefits. Costs saliently include resources used, and both costs and benefits can include moral considerations, such as fairness in the distribution of costs or in how people are used, and whether the actions involved go beyond or is in accordance with what can be reasonably demanded of people. In the case of climate threat reduction, the fairness or justice of means might depend on assessments of the individual agents’ capacity to help effectively as well her prior reliance on more than her fair share of Earth’s limited capacity to harmlessly absorb greenhouse gas emissions (see e.g. Baatz 2014; cf. the exten­ sive discussion of “polluter pays” and “beneficiary pays” principles in environ­ mental economics, law, and ethics). Evaluations might lead to rejection or refinement of means, and the identification of the best possible or only accept­ able means. If moral and other costs of available means do not stand in pro­ portion to the target outcome—to keep the risk of the direst catastrophe scenarios under 1%, say—or if the likelihood of success is insignificant, the target might be raised or lowered and the costs recalculated. Identification of ways that the individual can further a certain means, by mak­ ing it more likely to be implemented, raising its likelihood of success, or other­ wise improving on it. Such ways can be more or less abstractly understood. Evaluation of ways of promoting means, with respect to the same range of ­considerations as evaluation of means. Can lead to rejection or refinements of 22 As should be clear from this outline, the sort of reasoning I have in mind here is neither “­we-­mode” nor “I-­mode” reasoning (in the terms of Tuomela  2007), though it might conceivably involve either.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  275 certain ways. Can also lead to reevaluation of means, as prior evaluations might have relied on now rejected assumption about reasoner’s input. Individual decision about what to do about the issue, if anything. We can easily see how both considerations of collective capability and judgments of shared obligations have a natural home in such reasoning. When appropriately caring individuals become aware of the sort of issue that the activist’s perspective concerns, they will tend to engage in issue-­oriented reasoning. Because of the scale of the issue, potentially effective means characteristically involve great many people, making considerations of collective capability highly relevant. At this stage, it might also matter whether reasonable demands on the agents involved would ensure realization of the means—whether the realization is a shared obli­ gation—as this might affect whether the means can likely be realized in a morally acceptable fashion. To see how individuals who have taken this perspective can also rightly see their own actions as more or less effective and helpful contributions to such col­ lect­ive means, we need two more general ideas. The first, which has already been alluded to, is that: helpful contribution:  A contribution to an important outcome can be helpful both by increasing the likelihood of the outcome and by improving the means by which the outcome is reached. Correspondingly, appropriately caring agents will be motivated to contribute not only in ways that make a valuable outcome more likely, but also in ways that make reaching it easier, faster, cheaper, safer, involving a fairer distribution of efforts, and so forth. In line with this, we have already noted that an individual agent can make a difference to how fairly the threat of climate catastrophe is resolved, by not using more than her fair share of a vital common resource and by not unjusti­ fiably reducing the expected wellbeing of others. How, though, can the individual contribution be rationally directed at and ­tailored to efficiently help achieve a certain end when the contribution cannot significantly affect the likelihood that it is achieved? The answer is provided by something like the following plausible principle: effective contribution:  An action or omission is an effective contri­ bution to an outcome to the degree that it increases the extent to which the outcome is realized. Some outcomes are perhaps naturally understood as being either fully realized or not at all: preventing a fascist politician from getting a cabinet position, say, or becoming pregnant. For such outcomes, a contribution is either fully effective,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

276  Gunnar Björnsson or not at all. Other outcomes, however, are naturally understood as involving the realization of enough of some underlying dimension that can be realized to degrees. Some such outcomes are probabilistic: your goal to decrease the likeli­ hood that a fascist gets a cabinet position or that you or your partner becomes pregnant. If so, a contribution is effective to the extent that it affects the likelihood in question. Others are non-­probabilistic. You might strive to be well-­educated, fair-­minded, or wealthy, and contributions to those ends are more effective the more they add to your degree of education, fair-­mindedness, or wealth. Or you might be working for a decisive election win, or a cancer treatment that prevents or reverses tumor growth, in which case a contribution might be more effective the more votes it adds to the victory or the more it prevents or reverses growth, respectively. Given this understanding of effectiveness, the efficiency of a contri­ bution can simply be understood in terms of the costs involved in achieving a certain degree of effectiveness. Importantly, this understanding of degrees of effectiveness and efficiency applies even when, because of the large-­scale vagueness of the outcome, no indi­ vidual action will decide whether the outcome is realized: whether you will now be well educated, or whether the victory was now decisive. For this reason, effective contribution also makes clear how contributions towards sig­ nificantly reducing climate threat can be more or less effective and efficient depending on how they affect expected climate-­related value. We have now seen how considerations of collective capability can ground both a shared obligation to significantly reduce climate threat and individual obliga­ tions to contribute to such reduction, in spite of individual incapability. Appropriately caring individuals who become aware of an issue like this will con­ sider means of resolution that involve groups of people who can resolve the issue. Such individuals will also be motivated to perform actions that are helpful—effi­ cient and fair—contributions to threat resolution even if they do not see their own contribution as increasing its likelihood. Let me end this section by considering an objection to these last claims. It seems painfully obvious that many people care very little and are unwilling to contribute, and one might worry that this undermines what the appropriately caring individual can be expected to do. The worry concerns both efficiency and fairness. If others are not doing their part, it is both unclear whether doing my part will be a comparatively efficient way of contributing to a resolution and unclear how things are made fairer if I incur the cost of doing my part.23 On reflection, though, matters are not quite that bleak. Not only is there still a good chance that enough individuals will each act in ways that marginally

23  For general concerns about obligations under conditions of futility, see Budolfson (n.d.); cf. Sinnott-­Armstrong (2005). The argument of Section 5 rejected the idea that individual contributions to climate threat resolution are futile given that others care appropriately.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  277 increase climate-­related expected value and together significantly reduce climate threat, judging by progress over the last few years. Much has also been done to provide cognitive and material infrastructures that make helpful contributions convenient and inexpensive, at least for the affluent among us. There are increas­ ingly efficient renewable energy sources, improved low-­emission tech­nolo­gies, and detailed sophisticated scientific assessments of both risks and the effects of various policies. There are also policies and services in place that make it easy for us to choose climate friendlier options, support climate-­friendly politicians and policies, or inexpensively offset our carbon footprint. Efficiency should thus not in general be a worry. Considerations relevant to fairness also do not seem signifi­ cantly affected by current reluctance to contribute. That others are using or bene­ fitting from more than their fair share of a common resource without trying to compensate others does not obviously mean that I should be willing to use more than my fair share. It thus makes good sense to assume that, in spite of fairly widespread tendencies not to take the threat of climate catastrophe seriously, and in spite of our individual inabilities to affect whether the threat is averted or not, we do have individual prima facie obligations to help achieve resolution, obliga­ tions grounded in the same facts as our shared obligation. This concludes my attempt to make sense of central aspects of the activist’s per­ spective, with a specific focus on our climate obligations. In Sections 3 and 4, I explained why there need not be anything inherently mysterious about shared obligations. Such obligations have the same basic structure as individual obliga­ tions and do not require that the group in question is itself a bona fide moral agent. In Section 5, I argued that we likely have a shared obligation to significantly resolve the threat of climate catastrophe, as many of us can significantly affect climate-­related expected value. In this section, finally, I have explained how this shared obligation would be based in part on the recognition that we are col­lect­ ive­ly able to significantly resolve the issue, and explained how individuals can have obligations to help achieve climate threat resolution even though they can­ not affect its likelihood. Both considerations of collective capability and attempts to help achieve a collective solution to a large-­scale problem have natural homes in ordinary issue-­oriented practical reasoning. If the arguments are correct, the activist’s perspective thus remains reasonable in spite of the size and complex nature of the threat at hand and the largely disorganized group of agents to which it attributes both shared and individual obligations.24 24 This chapter has benefitted greatly from comments from audiences at seminars at Umeå University and the University of Gothenburg, the Metaphysics and Collectivity workshop at Lund University, the Trondheim ISP-­FIDE PhD forum, the Philosophy and Climate Change workshop at Princeton University, Collective Intentionality X, and MANCEPT 2016. I’m particularly grateful for written comments from Ragnar Francén, Brian Epstein, Anne Schwenkenbecher, Stephanie Collins, Olof Leffler, Olle Blomberg, Stefano Cossara, Daniel Telech, Holly Lawford-­Smith, Erik Brandstedt, Mattias Gunnemyr, Olle Risberg, Richard Endörfer, Mark Budolfson, Tristram McPherson, Michael

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

278  Gunnar Björnsson

References Aas, S. (2015). “Distributing Collective Obligation.” Journal of Ethics and Social Philosophy 9(3): 1–23. Arpaly, N. (2006). Merit, Meaning, and Human Bondage: An Essay on Free Will. Princeton, NJ: Princeton University Press. Baatz, C. (2014). “Climate Change and Individual Duties to Reduce GHG Emissions.” Ethics, Policy & Environment 17(1): 1–19. Barnett, Z. (2018). “No Free Lunch: The Significance of Tiny Contributions.” Analysis 78(1): 3–13. Björnsson, G. (2011). “Joint Responsibility Without Individual Control: Applying the Explanation Hypothesis,” in Moral Responsibility: Beyond Free Will and Determinism, ed. N.  Vincent, I.  van de Poel, and J.  van den Hoven. Dordrecht: Springer, 181–199. Björnsson, G. (2014a). “Essentially Shared Obligations.” Midwest Studies in Philosophy 38(1): 103–120. Björnsson, G. (2014b). “Incompatibilism and ‘Bypassed’ Agency,” in Surrounding Free Will, ed. A. Mele. New York: Oxford University Press, 95–122. Björnsson, G. (2017a). “Explaining (Away) the Epistemic Condition on Moral Responsibility,” in Responsibility: The Epistemic Condition, ed. P.  Robichaud and J. W. Wieland. New York: Oxford University Press, 146–162. Björnsson, G. (2017b). “Explaining Away Epistemic Skepticism about Culpability,” in Oxford Studies in Agency and Responsibility, ed. D.  Shoemaker. Oxford: Oxford University Press, 141–164. Björnsson, G. (2020). “Collective Responsibility and Collective Obligations Without Collective Moral Agents,” in The Routledge Handbook of Collective Responsibility, ed. S. Bazargan-Forward and D. Tollefsen. New York: Routledge, 127–141. Björnsson, G. (forthcoming). “Being Implicated: On the Fittingness of Guilt and Indignation over Outcomes,” Philosophical Studies. Björnsson, G. and B.  Brülde (2017). “Normative Responsibilities: Structure and Sources,” in Parental Responsibility in the Context of Neuroscience and Genetics, ed. K.  Hens, D.  Cutas, and D.  Horstkötter. Cham: Springer International Publishing, 13–33. Björnsson, G. and K.  Hess (2017). “Corporate Crocodile Tears? On the Reactive Attitudes of Corporate Agents.” Philosophy and Phenomenological Research 94(2): 273–298. Björnsson, G. and K.  Persson (2012). “The Explanatory Component of Moral Responsibility.” Noûs 46(2): 326–354.

Morck, Adrian Russian, Coby Gibson, Maxwell Frye, and Ira Richardson. Work on this paper has been funded by the Swedish Research Council.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

On Individual and Shared Obligations  279 Björnsson, G. and R. Shanklin (2014). “‘Must’, ‘Ought’ and the Structure of Standards,” in Deontic Logic and Normative Systems, ed. F. Cariani, D. Grossi, J. Meheus, and X. Parent. Cham: Springer International Publishing, 33–48. Brennan, G. and G. Sayre-McCord (2015). “Voting and Causal Responsibility.” Oxford Studies in Political Philosophy 1: 36–59. Broome, J. (2012). Climate Matters: Ethics in a Warming World. New York: W. W. Norton. Budolfson, M. (n.d.). “Collective Action, Climate Change, and the Ethical Significance of Futility.” Manuscript. Collins, S. (2013). “Collectives’ Duties and Collectivization Duties.” Australasian Journal of Philosophy 91(2): 231–248. Cripps, E. (2013). Climate Change and the Moral Agent: Individual Duties in an Interdependent World. Oxford: Oxford University Press. Dietz, A. (2016). “What We Together Ought to Do.” Ethics 126(4): 955–982. Driver, J. (1992). “The Suberogatory.” Australasian Journal of Philosophy 70(3): 286–295. French, P.  A. (1984). Collective and Corporate Responsibility. New York: Columbia University Press. Hess, K.  M. (2014). “The Free Will of Corporations (and Other Collectives).” Philosophical Studies 168(1): 241–260. Hiller, A. (2011). “Climate Change and Individual Responsibility.” The Monist 94(3): 349–368. Hursthouse, R. (1999). On Virtue Ethics. Vol. 2. Oxford: Oxford University Press. Kagan, S. (2011). “Do I Make a Difference?” Philosophy & Public Affairs 39(2): 105–141. Kingston, E. and W.  Sinnott-Armstrong (2018). “What’s Wrong with Joyguzzling?” Ethical Theory and Moral Practice 21(1): 169–186. Kratzer, A. (1977). “What ‘Must’ and ‘Can’ Must and Can Mean.” Linguistics and Philosophy 1(3): 337–355. Lawford-Smith, H. (2015). “What ‘We’?” Journal of Social Ontology 1(2): 225–249. Lawford-Smith, H. (2016). “Difference-Making and Individuals’ Climate-Related Obligations,” in Climate Justice in a Non-Ideal World, ed. C. Hayward and D. Rosers. Oxford: Oxford University Press, 64–82. List, C. and P.  Pettit (2011). Group Agency: The Possibility, Design, and Status of Corporate Agents. Oxford: Oxford University Press. McKenna, M. (2012). Conversation and Responsibility. New York: Oxford University Press. Maltais, A. (2013). “Radically Non-Ideal Climate Politics and the Obligation to at Least Vote Green.” Environmental Values 22(5): 589–608. Millikan, R.  G. (1984). Language, Thought, and Other Biological Categories: New Foundations for Realism. Cambridge, MA: MIT Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

280  Gunnar Björnsson Nefsky, J. (2011). “Consequentialism and the Problem of Collective Harm: A Reply to Kagan.” Philosophy & Public Affairs 39(4): 364–395. Nefsky, J. (2017). “How You Can Help, Without Making a Difference.” Philosophical Studies 174: 2743–2767. Parfit, D. (1984). Reasons and Persons: Oxford: Clarendon Press. Petersson, B. (2004). “The Second Mistake in Moral Mathematics is not about the Worth of Mere Participation.” Utilitas 16(3): 288–315. Pinkert, F. (2014). “What We Together Can (Be Required to) Do.” Midwest Studies in Philosophy 38: 187–202. Schwenkenbecher, A. (2014). “Joint Moral Duties.” Midwest Studies in Philosophy 38: 58–74. Schwenkenbecher, A. (2019). “Collective Moral Obligations: ‘We-Reasoning’ and the Perspective of the Deliberating Agent.” The Monist 102(2): 151–171. Singer, P. (1980). “Utilitarianism and Vegetarianism.” Philosophy & Public Affairs 9(4): 325–337. Sinnott-Armstrong, W. (2005). “It’s Not My Fault: Global Warming and Individual Moral Obligations,” in Perspectives on Climate Change: Science, Economics, Politics, Ethics, ed. W.  Sinnott-Armstrong and R.  B.  Howarth. Amsterdam: Elsevier, 285–307. Smith, A. (2013). “Moral Blame and Moral Protest,” in Blame: Its Nature and Norms, ed. D. J. Coates and N. A. Tognazzini. Oxford: Oxford University Press, 27–48. Stern, N. (2015). Why Are We Waiting? The Logic, Urgency, and Promise of Tackling Climate Change. Cambridge, MA: MIT Press. Strawson, P. F. (1962). “Freedom and Resentment.” Proceedings of the British Academy 48: 187–211. Tollefsen, D. P. (2015). Groups as Agents. Cambridge: Polity Press. Tuomela, R. (2007). The Philosophy of Sociality: The Shared Point of View. Oxford: Oxford University Press. Van Zomeren, M., T.  Postmes, and R.  Spears (2008). “Toward an Integrative Social Identity Model of Collective Action: A Quantitative Research Synthesis of Three Socio-Psychological Perspectives.” Psychological Bulletin 134(4): 504–535. Vance, C. (2017). “Climate Change, Individual Emissions, and Foreseeing Harm.” Journal of Moral Philosophy 14(5): 562–584. Wringe, B. (2016). “Collective Obligations: Their Existence, Their Explanatory Power, and Their Supervenience on the Obligations of Individuals.” European Journal of Philosophy 24(2): 472–497.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

12

How Much Harm Does Each of Us Do? John Broome

1.  Sorts of Harm and Their Quantity Several moral philosophers have argued that the greenhouse gas emissions of a single individual do no harm. I think they are mistaken, and I have opposed their arguments in a paper I called ‘Against Denialism’.1 Now I shall give some positive account of the quantity of harm that each of us does. Many accounts already exist. First, there is a very large literature in economics on the ‘social cost of carbon’ (SCC), which is supposed to measure in terms of money the harm done by a tonne of carbon dioxide. Multiply the SCC by the total number of tonnes emitted by a person during her lifetime, and we get a money value for the total harm she does. Estimates of the SCC vary greatly, but the average of all the estimates surveyed in a comprehensive meta‑analysis is $55.2 This implies that a person who emits 1200 tonnes over her lifetime, which is typical for an American, causes $66,000 of harm. But a lot is left out of calculations of the SCC.3 Anything whose value can‑ not be made commensurate with money is inevitably omitted or poorly taken into account. This includes the wellbeing of animals and whatever intrinsic value natural objects—such as natural species and individual trees—may have. It may also include human cultural goods, such as the culture of Arctic peoples and monuments that may be lost to the sea. Climate change kills very many people. It kills them directly in droughts, floods, and storms. It kills them less directly through increasing the range of tropical diseases and by impoverishing the people who struggle to live in less hos‑ pitable parts of the world. Estimates of the SCC in principle take into account the harm of killing, but in practice they generally do so badly. The value they assign to the loss of a life is generally based on people’s willingness to pay for extending their lives, and it does not properly recognize the different value of money to dif‑ ferent people. Poor people are willing to pay less than rich people to reduce their risk, but this is not because their lives are less valuable. It is because they have 1  Broome (2019). 2  Wang Pei et al. (2019). 3  See the full discussion in Fleurbaey et al. (2019). John Broome, How Much Harm Does Each of Us Do? In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © John Broome. DOI: 10.1093/oso/9780198796282.003.0013

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

282  John Broome other, more pressing uses for their money. Economists often ignore this simple point.4 Also, they generally discount for time, which means giving less value to later lives than earlier ones. Some evidence will emerge in Section 2 that killing is a major part of the harm that climate change will do. Since existing estimates of the SCC take it into account badly, it is important to pay special, separate attention to this harm. The SCC conveys some information about the harm done by climate change, but we separately need information about killing. This chapter will focus on estimating the quantity of killing we do through climate change. There are precedents for this estimation too. One figure for the amount of this harm is frequently quoted by moral philosophers. It originates from calculations of John Nolt’s.5 Nolt started by working out carefully that an average American is responsible for about the fraction 5 × 10–10 of the climate change caused by green‑ house gas emissions up to 2040. Next he calculated the number of people who will live during the next millennium as 100 billion. Then he says: If over the next millennium as few as four billion people (about 4% [of the num‑ ber who will live during that period]) are harmed (that is, suffer and/or die) as a result of current and near‑term global emissions, then the average American causes through his/her greenhouse gas emissions the serious suffering and/or deaths of two future people.6

Nolt did not try to justify the figure of 4% for the proportion of people who will be harmed. He was not aiming to estimate the amount of harm so much as to illustrate what it might be. Nevertheless ‘the serious suffering and/or deaths of two future people’ is frequently quoted as his estimate—recently in the New York Times.7 We shall see that, as an estimate, it is far too big. At about the same time, I published an estimate based on figures from the World Health Organization.8 My estimate was that a typical Westerner takes away more than six months of human life altogether.9

2.  New Data and Estimates Those figures are now very much out of date. Much better ones are becoming available. A major report, ‘Valuing the Global Mortality Consequences of Climate Change’ (VGMC),10 derives conclusions on the basis of extremely extensive and

4  The IPCC asks them not to ignore it. See IPCC (2014), Summary for Policymakers, p. 5. 5  Nolt (2011). 6  Nolt (2011: 9). 7  Newman (2019). Thanks to Douglas MacLean for this reference. 8  WHO (2009). 9  Broome (2012: 74). 10  Carleton et al. (2019).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

HOW MUCH Harm Does Each of Us Do?  283 detailed data about the effect of weather on death rates at a very local level. The authors divided the land surface of the Earth into 24,378 areas and assembled data on a 38% sample of them. By means of sophisticated statistical analysis, they have derived from their data authoritative estimates of the global effects of cli‑ mate change on mortality. I shall base my conclusions on these estimates. To put it very crudely, VGMC regresses death rates on temperature. This means it takes account of all causes of death—all the various means by which the warm‑ ing of the planet kills people. It includes deaths in heat waves, deaths resulting from the spread of tropical diseases, and so on. It is not limited to particular causes of death, as are earlier studies from the World Health Organization.11 VGMC also takes account of the ages of people who die, so it can calculate the number of life-­years lost as well as crude death rates. These are much more informative. Many people die in heat waves, and this is one of the significant ways in which climate change kills people. But many of those people are elderly12 and many are already suffering from chronic diseases. When a heat wave kills them, it may be shortening their lives by only a few years, months, or days. Climate change also increases the prevalence of diarrheal diseases; this is another significant means by which it kills people. It is mainly children who die from these diseases,13 and they lose many years of life. It would be misleading to count a child’s death and an elderly person’s death the same, and VGMC does not do so. All in all, VGMC data is very valuable. Still, estimates of harm from climate change can never be certain. The science of climate change is very uncertain, and the spread of possibilities is very wide. For example, it is possible that climate change will lead to a catastrophe for humanity, and even to our extinction. It may even be that in responding to cli‑ mate change we should care more about this unlikely possibility of catastrophe than about what is likely to happen. Quite generally, an unlikely but very bad event may be more important for our planning that what is likely to happen. That is why ships ought to carry lifeboats. A ship is unlikely to sink, so its lifeboats are unlikely to be used. But if it does sink the consequences of having no lifeboats will be so dire that they make the expense of carrying lifeboats worthwhile. The econo­mist Martin Weitzman argues that our response to climate change should be like our response to the possibility of a ship’s sinking: directed towards the unlikely but very bad consequences of catastrophe.14 I have no way to estimate the harm that a person’s emissions will do if the results of climate change are catastrophic. I am therefore forced to limit my

11  For example, WHO (2009, 2014). 12  WHO (2014: 17). 13  WHO (2014: 37). 14  Weitzman (2009). See also Wagner and Weitzman (2015). My paper ‘The Most Important Thing about Climate Change’ (Broome 2010) explains that Weitzman’s argument is insufficient for his conclusion.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

284  John Broome estimates to the amount of harm that is likely, rather than take account of all the possibilities. Even the likely harm done by a person’s emissions is very contingent; it depends on the emissions of other people. This is because the relation between tempera‑ ture and mortality is very non-­linear. Its graph is U-­shaped. Both low tem­per­at­ ures and high temperatures increase the number of deaths. As the temperature increases starting from a low level, the death rate decreases until it reaches a minimum at around 20°C.  Then it starts to increase at an accelerating rate. Consequently, an increase in temperature when the temperature is very high causes much more harm than the same increase would do were the initial tem‑ perature lower. An emission of greenhouse gas therefore causes much more harm if other emissions are high than if they are low. This contingency is handled in climate-­change science by means of ‘scenarios’. Each scenario describes a particular possible future development of emissions together with the growth of the world’s population and economy. So when I refer to the harm that is likely to result from a person’s emissions, I mean the harm that is likely given a particular scenario. The VGMC study works with two scenarios known as ‘RCP 4.5’ and ‘RCP 8.5’.15 Perforce, I copy it in this respect. RCP 4.5 is a moderate scenario in which emissions of greenhouse gases begin to decline around the middle of this century. Nevertheless, the temperature under RCP 4.5 is likely to reach 2.4°C above pre-­industrial levels, which is well above the target set in the Paris Agreement negotiated in 2015 by the United Nations Framework on Climate Change. So this is by no means an optimistic scenario. RCP 8.5 is intended to be a baseline that might be considered ‘business as usual’. It should be treated as a basis for comparison rather than a prediction of what will happen. It assumes high growth of population with slow economic growth, and limited technical progress. In RCP 8.5, emissions increase through the century, and the temperature is expected to reach almost 5°C above pre-­industrial levels. This might fairly be counted as catastrophic. RCP 8.5 is a very pessimistic scenario. VGMC calculates what it calls the ‘mortality-­related’ harm that will result from emitting one tonne of carbon dioxide in 2020. By means I shall explain, it expresses the result in terms of dollars. To cut a long story short, its conclusion is that the dollar value of the harm is $18.9 under RCP 4.5 and $98.9 under RCP 8.5 (VGMC, p. 46). These are the figures I shall work with. They assume a 2.5% dis‑ count rate on commodities, and a ‘globally uniform valuation of mortality risk’. I shall explain these two assumptions in Sections 4 and 3 respectively. Compare these figures with $50, which is a typical estimate for the SCC as a whole. The comparison supports the assertion I made in Section 1, that mortalityrelated harm is at least a major part of the harm that climate change will do.

15  See Wayne (2013).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

HOW MUCH Harm Does Each of Us Do?  285

3.  Lives for Money VGMC presents its conclusions in terms of money values. But many philosophers including me are dubious about translating the value of lives into money. We would prefer to see the result in terms of quantities of life itself. This raw informa‑ tion is embedded in the VGMC calculations, but not in a transparent way. The authors of the report are in a position to extract and present it, but only by means of a substantial amount of computation. I believe they will do so in due course. In the meantime this volume goes to press. In order to give readers some rough idea of the quantity of life we take away through our emissions, I have extracted estimates of this quantity from the figures presented in the existing report, using the best means I have available. These means are frankly very crude. The outcome will be very approximate, but it is the best I can do. The authors of the report bear no responsibility for my figures, and mine will be totally superseded by theirs when they are published. I first adjust the figures by subtracting adaptation costs from them. As tem­per­ at­ures increase, people adapt to them. Their bodies acclimatize and they take steps to avoid the heat. VGMC uses sophisticated methods to account for adapta‑ tion in its estimates of the number of people killed by climate change. It also rec‑ ognizes that adaptation often costs money, and it includes this cost in its figures for mortality-­related harm. I want to estimate the actual amount of killing that climate change does, so I need to subtract the adaptation costs. VGMC states that on average 14% of mortality-­related costs are adaptation costs (VGMC, p. 5). I have to use this average figure because I cannot find figures in VGMC related to the particular costs I am working with. I therefore reduce those costs by 14% and get $16.3 for RCP 4.5 and $83.1 for RCP 8.5. These are now the dollar values of life actually lost. Next I work back from these dollar values to calculate the actual quantities of life that they represent. The dollar values are based on the monetary value of life that is standardly used in cost–benefit analysis in the US. This is $10.95 million for a life (VGMC, p. 121). VGMC converts it to a value for a year of life by using the life expectancy of a median-­aged American (VGMC, p. 120). The text does not state what this life expectancy is, so I have to recover it. The median age of Americans in 2018 was 38.2.16 Life tables for 2016 show life expectancy at 38.2 as 40.23 for men and 44.20 for women.17 I shall assume an average life expectancy of 42.2. The result is that a life-­year is valued at $259,000. In principle, VGMC values lives or life-­years on the basis of what people are willing to pay for them, or more exactly what they are willing to pay to improve their chances of living longer. It assumes that what people are willing to pay is 16 https://www.statista.com/statistics/241494/median‑age‑of‑the‑us‑population/. 17 https://www.ssa.gov/oact/STATS/table4c6.html.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

286  John Broome proportional to their income (VGMC, p. 39). In practice, willingness to pay is ­averaged across the population. The value of $259,000 is the average across the US population. The VGMC figures for ‘globally uniform valuation of mortality risk’ are based on average willingness-­to-­pay across the whole world population, under the assumption that willingness to pay is proportional to income. This is what ‘globally uniform’ means. The consequence is that the figure of $259,000 needs to be reduced by the ratio of global average income to American average income. From World Bank data18 in 2018 I find this ratio to be .287. This makes the value of a life-­year $74,300. This value allows us to take the above-­quoted dollar values of killing caused by a tonne of carbon dioxide and convert them back into numbers of life-­years. The result is .000219 life-­years in RCP 4.5 and .00112 in RCP 8.5. Those are rates per tonne. Next we have to multiply these quantities by the number of tonnes of carbon dioxide emitted by a person during her lifetime. There is a further complication here. These quantities I have derived from VGMC measure the harm done by a tonne of carbon dioxide that is emitted in 2020. Later emissions are done at a time when the global temperature is higher. Because of the non-­linear relationship between temperature and deaths, later emissions therefore do more harm. Earlier emissions do less harm. People who are young now will do more harm by their emissions than older people who emit the same in total. Because I do not aim at precision, I shall ignore this complication. I shall focus on a person who emits 1200 tonnes during her lifetime, which is about average for an American. Multiplying the rates of death per tonne by this amount, we might conclude that this person’s lifetime emissions cause the loss of .263 lifeyears in RCP 4.5 and 1.34 life-­years in RCP 8.5.

4.  The Consequences of Discounting Sadly, this is still not correct. These figures are a serious underestimate because they incorporate some temporal discounting of life-­years. So I turn to the difficult issue of discounting. The VGMC figures assume a discount rate of 2.5% on commodities. Strictly, this is the discount rate on money values after canceling out inflation. This means it is the discount rate on the bundle of commodities that are used as the basis for measuring inflation. It is correct to discount future commodities—which means giving less value to future commodities than to present ones—because both scen­ arios RCP 4.5 and RCP 8.5 assume that economic growth will continue. That is to say, they assume people will become progressively richer. Therefore, the value of commodities to them at the margin (the value of adding to their stock of 18 https://data.worldbank.org/indicator/NY.GDP.PCAP.PP.CD.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

HOW MUCH Harm Does Each of Us Do?  287 commodities) will progressively diminish. This is the consequence of the dimin‑ ishing marginal value of income, which has been recognized in economics at least since the time of Alfred Marshall. The more commodities you are already con‑ suming, the less you will benefit from consuming more commodities. VGMC does not discount life-­years at 2.5%, because it gives a progressively increasing value to future life-­years. I have not mentioned this before. The reason it does so is that it takes the value of life-­years to be proportional to income, and income increases with economic growth. The average rate of economic growth through this century is 2% per year in RCP 4.5 and 1.35% in RCP 8.5.19 The value of life-­years is increased at these rates. So in effect VGMC discounts life-­years at a rate equal to the difference between these rates and 2.5%. The upshot is that the figures for life-­years I have been working with incorporate discount rates of 0.5% for RCP 4.5 and 1.15% for RCP 8.5. The reason I gave for discounting commodities does not apply to life-­years. It is most implausible that life-­years lived later in history are really less valuable than ones that are lived earlier. Discounting life-­years is an instance of what is called ‘pure discounting’. Pure discounting has received some support from eco‑ nomic theorists—notably Kenneth Arrow—but not much.20 It does not sit well with the globally uniform valuation embedded in the VGMC figures I have been using. If, at each date, everyone’s life-­years are given the same value, why should life-­years at one date be given a different value from those at another date? Moreover, VGMC’s way of discounting has a peculiar consequence. The loss of a person’s life-­year is discounted according to the date of the person’s death rather than the date when the life-­year would have been lived. Suppose a 20-­year-­old dies now and loses 60 years of life. The loss of the years she would have lived between the ages of 60 and 80 gets full value in the calculation. But if someone else born at the same time as this 20-­year-­old lives to 60 and then dies, losing 20 years of life, the loss of those 20 years is given less value than the loss of the 20-­year-­old’s, even though they would have been lived at exactly the same time. A life-­year plausibly has a constant value, whoever lives it and whenever it is lived. Because of this, it provides a plausible basis for interpersonal comparisons of the value of commodities, as the Intergovernmental Panel on Climate Change explains.21 It also provides a plausible basis for the intertemporal comparisons of the value of commodities that appear in the discount rate. Plausibly, commodities should be discounted at whatever rate implies a constant value for a year of life. If we maintain VGMC’s assumption that the value of a life-­year is proportional to income, this means discounting commodities at the rate of growth of income. That is: 2% in RCP 4.5 and 1.35% in RCP 8.5. 19  I derive these figures from Figure 12 of Wayne (2013). 20  Arrow (1999). See the discussion in my ‘The Wellbeing of Future Generations’ (Broome 2016). 21  IPCC (2014: 226 (box 3)).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

288  John Broome I urge the authors to provide figures for years of life lost corresponding to these discount rates. At least, they should not treat the discount rate in the way they do, as exogenously given independently of the growth rate. The correct discount rate is a function of the growth rate, as the famous Ramsey equation shows.22 The estimates obtained at the end of Section 3 incorporate discount rates on life-­years of 0.5% in RCP 4.5 and 1.15% in RCP 8.5. In the absence of undis‑ counted figures in VGMC, I need to cancel out the discounting as best I can. How badly do the discounted estimates underestimate the true quantity? This depends on how the killing caused by an emission of carbon dioxide is distributed over time. This information is implicit in the work of VGMC, but I cannot extract it from the paper. I can make only guesses. When a tonne of carbon dioxide is emitted, it causes the atmosphere’s tempera‑ ture to rise soon afterwards, and that raises the death rate. The tonne begins immediately to be absorbed by the land and oceans, so its effect on the death rate will begin to fall too. About half the tonne will fall out of the atmosphere within 50 years. However, perhaps 20% of it will persist for hundreds and even thou‑ sands of years.23 Furthermore, its effect on temperature will lag behind the quan‑ tity of carbon dioxide itself. I am not in a position to judge the extent of the lag; doing so would require running a model of the atmosphere.24 But it is plain that, were the killing to decline only with temperature, it would continue for a very long time and its total would be very large. But actually the killing will be progressively reduced by people’s adaptation to the heat. The VGMC data contains some information about adaptation, but I have not been able to use it at this point in the calculation. In any case, it could tell us very little about the development of humanity several centuries from now. Human life will be so different in three hundred years that it is hard to know even how adaptation could be identified by then. I have to fall back on little more than guesswork. Bearing in mind that the quantity of an emitted tonne of carbon diox‑ ide will be reduced by about a half in half a century, I shall assume that its effect on the death rate will be reduced to half within a century. I assume that the effect on the death rate will be very small after three centuries. (Three centuries is the horizon set on the calculations in VGMC.) For RCP 4.5, the estimate I obtained at the end of Section 3 for the life-­years taken away by a person who emits 1200 tonnes is .263. The discount rate is 0.5%. This amounts to a discount of 40% over 100 years, 63% over 200 years, and 78% over 300 years. This suggests to me that discounting at this rate is not likely to underestimate the total of harm by more than 50%. I guess therefore that in RCP 4.5 the amount of killing done by a person who emits 1200 tonnes is in the region of half a life-­year. 22  IPCC (2014: 229). 23  See the graphs in box 6.1 on p. 473 of IPCC (2013). 24  IPCC (2013: 1102–1105).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

HOW MUCH Harm Does Each of Us Do?  289 For RCP 8.5, the estimate I obtained at the end of Section 3 is 1.34 life-­years. The discount rate is 1.15%, which amounts to 68% over 100 years, 89% over 200 years, and 96% over 300 years. This suggests an underestimate of perhaps 75% or 80% in the amount of killing done. I guess that in RCP 8.5 the amount of killing done by a person who emits 1200 tonnes is in the region of 6 or 7 life-­years.

5.  Conclusion and Why It Matters My attempt in Section 4 to cancel out discounting from the figures is the most speculative part of my calculation. I was forced to speculate about the adaptive success of human beings centuries in the future. Since the VGMC estimates for RCP 4.5 imply only the small discount rate on life of 0.5%, they are less vulnerable to a mistake about this. The higher discount rate of 1.15% implied in the estimates for RCP 8.5 makes a much greater difference. Remember that in any case RCP 8.5 does not represent a prediction so much as a worst-­case baseline. RCP 4.5 is more like a prediction, and I put much more trust in the RCP 4.5 figures. There is anyway a great deal of uncertainty in any quantitative predictions involving climate change. I have tried to work out only the harm that is likely to arise from emissions; much greater harm is possible. Furthermore, remember I am only trying to produce interim results, while I wait in hope that the authors of VGMC will produce much more accurate ones in due course. With all these cav­ eats, my best estimate of the amount of life you are likely to take away by emitting 1200 tonnes of carbon dioxide is half a year. Why does it matter? It helps to position the harm we do through climate change on the scale of all the good and bad things we do. It is important to recog‑ nize that the harm an individual does by her emissions—and correspondingly the good she can do by reducing her emissions—though definitely significant, is not large in comparison to other means of doing good. Some ways of reducing emissions, such as eating less meat and turning down the air conditioning, are easy and cheap. Others, such as insulating your house, are expensive. One of the cheaper ways is to offset your emissions. You can offset by planting trees or by paying for projects that reduce emissions elsewhere. The cost of offsetting is in the region of $10 per tonne. According to my figures, if you were to spend $12,000 on offsetting your lifetime emissions of 1,200 tonnes, you would save perhaps half a life-­year in RCP 4.5 and 6 or 7 life-­years in RCP 8.5. By contrast, the organization GiveWell lists on its website charities that, on its calcu‑ lations, can save a person’s whole life for a donation of $2,000 or $3,000.25 Among

25 https://www.givewell.org/how‑we‑work/our‑criteria/cost‑effectiveness/cost‑effectiveness‑ models.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

290  John Broome them are charities that fight malaria. These are plainly more effective ways of using money to do good. Why, then, should you reduce your emissions? Mainly because justice requires it. You emit greenhouse gas to benefit yourself, but in doing so you harm other people. It is an elementary principle of common-­sense justice that, with certain exceptions such as self-­defense, you should not harm other people for your own benefit. On this point I agree with Nolt.26 It is also true that climate change is in aggregate doing immense harm in the world. Although reducing emissions is not the most effective way of doing good, it is well worth the cost. For you as an individual, this is not unqualifiedly so because you have better ways of using your money. If you use your money in the best ways, starting with the best means of doing good and working down to less good means, you will run out of money long before you get to reducing your emissions much. But governments are different because they control vastly greater resources. They have coercive power over their people’s behavior, by means of taxes and regulations. It is true for a government as it is for an individ‑ ual, that it should first direct resources in more effective ways such as fighting malaria. But when all that is done, a government should still direct a vast amount of further resources towards reducing emissions of greenhouse gas. An appropriate means of doing good for an individual is therefore political action aimed at getting governments to reduce emissions. This is a further reason for reducing your own emissions. Doing so is a sort of political action. It shows that you care. It may induce others to follow you and to vote for reducing emissions.27

References Arrow, K. (1999). “Discounting, Morality, and Gaming,” in Discounting and Intergenerational Equity, ed. P. R. Portney and J. P. Weyant. New York and London: Resources for the Future, 13–21. Broome, J. (2010). “The Most Important Thing about Climate Change,” in Public Policy: Why Ethics Matters, ed. J. Boston, A. Bradstock, and D. Eng. Canberra: ANU E Press, 101–116. Broome, J. (2012). Climate Matters: Ethics in a Warming World. New York: W. W. Norton.

26  (Nolt 2011: 7). See my Climate Matters (Broome 2012), chapter 4. 27  My thanks for very useful comments, information and advice from Mark Budolfson, Tamma Carleton, Maxwell Frye, Coby Gibson, Daniel Gun Lim, Douglas MacLean, Michael Morck, Tristram McPherson, John Nolt and Adrian Russian. Research for this paper was supported by ARC Discovery Grants DP140102468 and DP180100355.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

HOW MUCH Harm Does Each of Us Do?  291 Broome, J. (2016). “The Well-Being of Future Generations,” in The Oxford Handbook of Well-Being and Public Policy, ed. M.  Adler and M.  Fleurbaey. Oxford: Oxford University Press, 901–928. Broome, J. (2019). “Against Denialism.” The Monist 102: 110–129. Carleton, T., M. Delgado, M. Greenstone, T. Houser, S. Hsiang, A. Hultgren, A. Jina, R. Kopp, K. McCusker, I. Nath, J. Rising, A. Rode, H. K. Seo, J. Simcock, A. Viaene, J.  Yuan, and A.  Zhang (2019). “Valuing the Global Mortality Consequences of Climate Change Accounting for Adaptation Costs and Benefits.” BFI Working Paper, August 12. Chicago, IL: Becker Friedman Institute for Economics at the University of Chicago. Fleurbaey, M., M.  Ferranna, M.  Budolfson, F.  Dennig, K.  Mintz-Woo, R.  Socolow, D.  Spears, and S.  Zuber (2019). “The Social Cost of Carbon: Valuing Inequality, Risk and Population for Climate Policy.” The Monist 102: 84–109. Intergovernmental Panel on Climate Change (IPCC) (2013). Climate Change 2013: The Physical Science Basis. Cambridge: Cambridge University Press. Intergovernmental Panel on Climate Change (IPCC) (2014). Climate Change 2014: Mitigation of Climate Change. Cambridge: Cambridge University Press. Newman, A. (2019). “If Seeing the World Helps Ruin It, Should We Stay Home?” New York Times, June 3. Nolt, J. (2011). “How Harmful Are the Average American’s Greenhouse Gas Emissions?” Ethics, Policy & Environment 14: 3–10. Wagner, G. and M. Weitzman (2015). Climate Shock: The Economic Consequences of a Hotter Planet. Princeton, NJ: Princeton University Press. Wang, P., X. Deng, H. Zhou, and S. Yu (2019). “Estimates of the Social Cost of Carbon: A Review Based on Meta-Analysis.” Journal of Cleaner Production 209: 1494–1507. Wayne, G.  P. (2013). “The Beginner’s Guide to Representative Concentration Pathways.” https://skepticalscience.com/docs/RCP_Guide.pdf. Weitzman, M.  L. (2009). “On Modeling and Interpreting the Economics of Catastrophic Climate Change.” Review of Economics and Statistics 91: 1–19. World Health Organization (2009). Global Health Risks: Mortality and Burden of Disease Attributable to Selected Major Risks. Geneva: WHO. World Health Organization (2014). Quantitative Risk Assessment of the Effects of Climate Change on Selective Causes of Death, 2030s and 2050s. Geneva: WHO.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

SECTION IV

CLIM AT E C HA NGE A ND P OLI T ICS

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

13

How Quickly Should the World Reduce its Greenhouse Gas Emissions? Climate Change and the Structure of Intergenerational Justice Lucas Stanczyk

1. Introduction How quickly should the world reduce its greenhouse gas emissions? On the one hand, unless dramatic reductions in greenhouse gas emissions are put in place quickly, then millions of people in the future are likely to suffer climate-change induced catastrophes, including deaths from extreme weather, drought, famine, disease, and migration-induced widespread human conflict.1 On the other hand, if the present generation effects drastic emissions reductions quickly, then the rate of global economic growth is almost certain to fall significantly in the short-run.2 As a result, there will be millions of additional premature deaths from prolonged hunger and disease in the developing world, where billions continue to lack ad­equate nutrition, sanitation, medical care, and even electricity. So, how quickly should the world reduce its greenhouse gas emissions? And what is the right approach to thinking about this question? 1  At today’s still accelerating rate of emissions, the trillionth ton of carbon will be emitted in no more than seventeen years and possibly in as few as nine years from today. On moderate projections, the resulting increase in global average surface temperature will before long breach two degrees Celsius above the preindustrial level, widely regarded as a threshold for devastating climate impacts as well as a potential trigger for catastrophic positive feedbacks such as the rapid melting of the permafrost in the Arctic. Even ignoring such feedbacks, the world can expect hundreds of millions of add­ ition­al climate refugees in the second half of the twenty-first century. For a running estimate of the carbon already emitted, see Trillionthton: http://www.trillionthton.org. On the relationship between increases in global temperature and human population displacement, see Hsiang and Sobel (2016). 2  It is sometimes suggested that there need not be any significant costs from rapidly slashing global greenhouse gas emissions, even in the short-run. This suggestion is unpersuasive. There is presently no way for the global economy to stop burning enormous quantities of fossil fuels, short of grounding virtually all planes, ships, trucks, trains, heavy machinery, power plants, and electrical grids. Alternatives for many essential combustion engines have yet to be invented, and the research that is necessary to mass-produce them is likely to take decades. A large and rapid cut in global greenhouse gas emissions is therefore almost certain to be accompanied by a dramatic decline in the global rate of economic growth. Lucas Stanczyk, How Quickly Should the World Reduce its Greenhouse Gas Emissions? Climate Change and the Structure of Intergenerational Justice In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Lucas Stanczyk. DOI: 10.1093/oso/9780198796282.003.0014

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

296  Lucas STANCZYK In this chapter, I would like to explain why I think the dominant approach to this question will have to be rejected, and to outline the structure of what I have come to think is the correct one. Among policymakers, the dominant approach to thinking about the optimal emissions pathway continues to be normative welfare economics. However, as currently conceived and practiced, welfare economics is not suitable for settling how quickly to reduce our greenhouse gas emissions. My discussion will highlight two key issues. First, because of the non-identity problem, it is not true that our runaway emissions are inefficient. Therefore, the goal of securing greater efficiency cannot actually recommend significantly cutting our runaway emissions. Second, the approach that economists are wont to and do fall back on—subjective preferencesatisfaction sum-total utilitarianism—shipwrecks on variants of the Repugnant Conclusion. There is nonetheless a solution to both of these problems. However, it requires parting ways with the dominant approach. More precisely, we should reject the idea that what we owe to (present or future) people is to maximize the sum of welfare. Instead, we should think of our duties of intergenerational justice as requiring us to preserve the social and environmental basis of just or rightsrespecting institutions. Finally, I will argue that thinking about intergenerational justice in this way allows us to side-step the non-identity problem. Even in the many cases in which our runaway emissions will not actually harm any unborn future people, we can still say that at all times humanity is subject to a suitably high environmental conservation standard. The chapter is organized as follows. In the next section, I first review some of the basic building blocks of welfare economics. In Section 2.2, I explain why the global status quo in greenhouse gas emissions is not actually Pareto-inefficient. Section 2.3 explains why the looser Kaldor-Hicks criterion of efficiency likewise cannot recommend significantly cutting our emissions. The rest of the section is dedicated to showing that welfarist analysis of climate policy is subject to variants of the Repugnant Conclusion. In Section 3, I lay out the elements of what I believe will be a more adequate approach. On this approach, requirements of intergenerational justice are understood, first and foremost, not as the means to but as the most important ethical constraints on maximizing intertemperal welfare. I explain why the main content of these constraints can be given by the theories of social and international just­ ice. Finally, I explain why the non-identity problem does not undermine this way of thinking about intergenerational justice. Even if the business-as-usual baseline in greenhouse gas emissions will never harm any unborn future people, we can still say that humanity is forever subject to a suitably high environmental conservation standard.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  297

2.  Reflections on the Dominant Approach 2.1  Basic Notions of Welfare Economics Recall some of the basic notions of welfare economics.3 Consumer surplus is the difference between what a buyer of a good is willing to pay for a good and the price that she actually pays. Producer surplus is the difference between what the seller would be willing to accept and the payment that he actually gets. Both of these quantities are said to be measures of “welfare.” But what is welfare? Officially, welfare economists do not take a stand on this question—except to say that individuals are normally best equipped to judge the answer for themselves. Thus, consumer surplus is said to be a measure of the welfare benefit that a buyer gets from an exchange, by her own lights. She is willing to pay $10 for a good, given all the other things she might do with this money. So, when the market price is $6, she has money in her pocket, so to speak. By her own lights, she gets a surplus benefit equivalent to $4 from buying the good at the market price. Now, competitive markets are said to allocate the supply of goods to those who value them the most, as measured by their willingness to pay. And, competitive markets are said to allocate the demand for goods to those who are able to prod­ uce them at the lowest cost. Therefore, competitive markets are said to maximize “total welfare,” or the sum of consumer and producer surplus. However, officially, welfare economics is not concerned with maximizing anything but with avoiding waste or inefficiency. An allocation is Pareto optimal, or efficient, if and only if it is impossible to make at least one person better off (by her own lights) without making another person worse off (by her own lights). The absence of competitive markets is said normally to be inefficient in this sense, because there are bound to be as-yet unexploited welfare gains from trade. For example, as the President of the World Bank once put it, people in the industrialized countries would be willing to pay quite a lot for unpolluted “pretty air.”4 By contrast, peasants in African countries would rather have better-paying but heavily polluting factory jobs. Unfortunately, there is no market for pretty air, for the simple reason that Africans cannot bottle their skies and sell them to wealthy people in Los Angeles or Detroit, in exchange for well-paying but dirty factory jobs. Nonetheless, there is a way to make the allocation of pretty air and factory jobs more efficient, since it is possible to make everyone better off as judged by her own lights. Namely, the World Bank can use its 3  For elaborations, see any standard microeconomics textbook. 4  See Hausman et al. (2017: 19–20).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

298  Lucas STANCZYK considerable influence to encourage the migration of dirty industries to the developing world!

2.2  Our Runaway Emissions Are Not in Fact Pareto-Inefficient From the perspective of welfare economics, the most important fact about climate change is that our present level of greenhouse gas emissions is seriously inefficient.5 This is the starting assumption of all economic assessments of the existing level of greenhouse gas emissions. Moreover, from the very definition of what constitutes a Pareto inefficiency, we apparently already know something truly remarkable.6 Namely, as various economists have written, it must in prin­ciple be possible to put the world on an efficient pathway “without anyone’s having to make a sacrifice.”7 After all, things are Pareto-inefficient just in case it is possible to make some people better off without making anyone worse off—or in other words, just in case there is truly unnecessary social waste. So, if our present level of greenhouse gas emissions is indeed inefficient, it must be possible to put us on an efficient pathway without anyone’s being made worse off (even) by her own lights. How could this be possible? The answer, widely taking for granted in environmental economics, is that future generations—who will have to live with the devastation that our runaway emissions will eventually wreak—all have some positive willingness to pay to get us to change our polluting ways. At the same time, there must be some amount of money that we in the present generation would happily accept, and which we would consider adequate compensation for curbing our climate-endangering activities to some significant extent. Thus, the only practical question that remains is what people from the distant future might be able to do to pay us off—or equivalently, by what steps we might take resources from the future to compensate ourselves for the costs of our steep emissions cuts. Perhaps we could try depleting the world’s remaining fishstocks? Or if that is not enough, perhaps we could cut emissions drastically but leave future generations with mountains of public debt? Unfortunately, this standard economic reasoning does not work. This is because future generations do not have positive willingness to pay to get us to  change our ways. The so-called non-identity problem ensures that this is 5  Equivalently in the economic jargon, the present level of emissions represents a deeply inefficient externality. 6  My elaboration of this remark in the next two paragraphs draws on the illuminating discussion by Kelleher (2015). 7  Broome (2010: 102). See also Stern (2014: 425, 427) and IPCC (2014: 227): “GHG is an externality: a person whose activities emit GHG does not bear the full cost of their activities; some of the costs are borne by those who are harmed by the emissions. Consequently, climate change causes Pareto inefficiency, which means that a Pareto improvement would in principle be possible. Indeed it would be possible to remove the inefficiency in a way that requires no sacrifice by anyone in any generation, compared to business-as-usual.”

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  299 not the case.8 To illustrate, if we continue with business as usual, there will sooner or later be climate-related disasters, forcing millions of people to leave their homes and migrate elsewhere. Some of these displaced persons will meet for the first time in environmental refugee camps, and will have children together that they otherwise would not have had. And yet no matter how unfortunate their living conditions, these children will not have positive willingness to pay to get us to stop what we are presently doing. For if we were to put an end to our disastrous polluting, the parents of these children would never end up meeting, and consequently these children would end up never existing. So if anything, these children will have negative willingness to pay to get us to curb our runaway emissions! After all, no one who enjoys living at least a little is willing to pay any positive amount to put an end to her own existence before it even begins. Instead, she will be eager to spend whatever she can for the opposite end. The upshot of these reflections is that the existing level of greenhouse gas emissions is already Pareto-efficient. There is no way of reining in our emissions without making members of future generations “worse off,” in exactly the way that welfare economists understand those words: i.e. the directional sign of each person’s presumed willingness to pay for the contemplated change to the status quo. As a result, being already satisfied with the status quo, the Pareto standard can hardly settle how quickly the world should leave behind its present course.9

2.3  Kaldor-Hicks Efficiency Likewise Cannot Recommend Significantly Cutting Our Emissions Of course, it has long been recognized that the Pareto standard of efficiency is too restrictive, because virtually all policies can be expected to have winners and

8  A version of this argument is elaborated in Kelleher (2015). For more extended discussion, see Broome (2018). 9  If the present generation continues with the current status quo, then, at some point in the future, children will be born to persons who met only because some catastrophe forced them from their sep­ ar­ate homes. Yet if, as a consequence of changes to the current status quo, these particular children would end up never being born at all, should we really say that these children would thereby be made “worse off ” in any way that should give us moral pause? This is a perfectly reasonable question, which I take up later in this chapter, but it is not a question that is relevant to the critical analysis in this section. The reason is that, when economists describe the status quo in greenhouse gas emissions as being strongly Pareto-inefficient, they typically do not acknowledge that meaningful departures from the status quo will change who will ultimately be born. Instead, presuming that who will be born is fixed, they ask how much the very people who will suffer from environmental damage in the future are willing to pay us to prevent the damage from happening in the first place, (incorrectly) assume that this figure must be positive for each such future person, and regard any departure from the status quo that a person disprefers as making her worse off by definition. It is this standard practice that is subject to the criticism in the text, not every philosophical use to which any idea structurally similar to the economist’s concept of a Pareto improvement might be put. For one such further philosophical use, which “allows the lessons [of economics] to survive, but at the cost of reducing their practical significance,” see Broome (2018).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

300  Lucas STANCZYK l­osers. This recognition has led some welfare economists to adopt the ­Kaldor-Hicks criterion, which looks for “potential” Pareto improvements when evaluating policy proposals. According to this criterion, “a policy should be adopted if and only if those who will gain could fully compensate those who will lose and still be better off.”10 Hence the question is always whether the pro­spect­ive gains to the winners from a policy are large enough so that there would be no ­losers, if compensation out of these gains were to be paid after the fact. From this canonical formulation, it should be clear that the Kaldor-Hicks cri­ ter­ion aspires to be a principle of mere efficiency or non-waste. For even though the criterion does not itself require that compensation actually be paid, it nonetheless insists that fully adequate compensation of the losers—from their own point of view—must always be possible before a change to the status quo will be approved. This shows that the criterion is, at the very least, uneasy with incompensable interpersonal tradeoffs and the simple utilitarian aggregation of claims. Nonetheless, in environmental economics, utilitarian aggregation is exactly where the Kaldor-Hicks criterion has led. The reason for this is a certain sleight of hand in measuring the costs and benefits of policies that will predictably lead to loss of human life. To make a long story short,11 by putting a dollar value on a lost statistical life, using a Kaldor-Hicks-inspired compensation test,12 economists say they are merely measuring what would compensate a person for a “small” increase in her risk of death.13 However, any serious program of emissions cuts will lead to more than merely “small” increases in mortality risk. In addition, because the rate of global economic growth is bound to fall significantly in the short-run,14 any serious program of emissions cuts will lead to substantial increases in mortality risk, as well as some number of actual premature deaths from prolonged hunger and disease in the developing world. And yet, it is not as if any amount of money will be able to compensate those who end up dying for their own deaths, after the

10  Boardman et al. (2005: 31). 11  For a longer treatment, see Ackerman and Heinzerling (2004: 61ff.). 12  In order to evaluate policies that will predictably lead to deaths, economists need a way to put a dollar value on the loss of a statistical life. The method they use for this purpose is a mortality-risk subjective compensation test. Developed specifically for Kaldor-Hicks efficiency analysis, this method involves two main steps. First, the method asks what the average person would be willing to accept as fully adequate compensation for a small increase in her risk of death. Next, the method extrapolates an average subjective dollar value for a lost statistical life in the following way. In the United States, the average person is prepared to switch to a job with a one-in-one-thousand higher probability of death as long as he is paid roughly six thousand additional dollars per annum. Welfare economists infer that the average value of a statistical life lost or saved in the United States is on the order of six million dollars. In modern cost–benefit analysis, this is, accordingly, the amount that is used to compute whether a life-endangering alternative to the life-endangering status quo is or is not recommended by the Kaldor-Hicks criterion. 13 See, for example, US Environmental Protection Agency (2010), Appendix B, Mortality Risk Valuation Methods. 14  See the text of footnote 1. See also section 3.2 below.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  301 fact.15 On the contrary, all of the money in the world will be no good to them when they are dead. Therefore, the Kaldor-Hicks criterion, which permits departures from the status quo only if complete ex post compensation of all of the eventual losers is possible in principle, cannot recommend any serious program of emissions cuts—unless the actual loss of human life is totally ignored in determining an “optimal” emissions path.16

2.4  The Collapse of Climate Economics into a Narrow Form of Total Utilitarianism In practice, however, economists do make forceful recommendations in favor of substantial emissions reductions. Moreover, they do not deny that the recommended cuts will lead to numerous premature deaths. Economists thereby ef­fect­ ive­ly subsume the costs of all mortality risks, and all premature deaths, under the conventional measure of the value of a statistical life: a little over six million dollars in the United States, and varying amounts in the rest of the world.17 If a given global emissions cut is predicted to maximize future economic gains notwithstanding these amounts (summing across all foreseeably affected lives), then the cut is recommended as “optimal” even though it will inevitably lead to thousands of premature deaths down the line. These intellectual operations have logical consequences. Most importantly they show that, as currently conceived and practiced, economic analysis of climate policy is ultimately a form of preference-satisfaction sum-total utilitarianism. In other words, when used to discern the “optimal” greenhouse gas emissions pathway, economic analysis is no longer concerned merely with avoiding waste or inefficiency, on any conception of efficiency. Instead, having turned into a narrow 15  “No finite amount of money could compensate a person for the loss of his life, simply because money is no good to him when he is dead. There is nothing esoteric about this; it is an application, if an unexpected one, of the very orthodox notion of value which I have outlined” (Broome 1978: 92). 16  It is true that we cannot know the identity of these victims at the outset or, likely, ever. However, in contrast to other ethical approaches, such as the ex ante variant of contractualism defended in Frick (2015), our ignorance of the identity of future victims is simply not relevant to determining whether significant emissions cuts are recommended by the Kaldor-Hicks criterion. What is relevant is merely whether the projected costs to the losers are outweighed by the projected benefits to the winners, counting all foreseeable costs irrespective of the identity of the losers. Yet, to repeat, the foreseeable costs of a program of steep and rapid global greenhouse gas emissions cuts will not be limited to “small” increases in mortality risk. These costs will include substantial increases in mortality risk as well as, eventually, actual premature deaths from prolonged hunger and disease. Clearly, then, these very important costs must also be accounted for and weighed, if Kaldor-Hicks efficiency analysis is to yield even remotely plausible recommendations. Yet, for this purpose, Kaldor-Hicks analysis offers only the subjective compensation test described in footnote 12. 17  The sentence is the text is not quite right because some economic analyses deploy measures that are distinct from the value of a statistical life, such as the value of a statistical life-year. This qualification will not materially affect the present argument, however.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

302  Lucas STANCZYK form of total utilitarianism, it tells us to maximize an additive intergenerational social welfare function that takes as inputs all present and future preferences over consumption. Indeed, this aim is now perfectly explicit in the maximands of the leading integrated assessment models.18

2.5  Economic Analysis of Climate Policy Is Subject to the Repugnant Conclusion How, then, does contemporary economic analysis of optimal climate policy escape a version of the Repugnant Conclusion? “If we can take steps to create a global population of many tens or even hundreds of billions of people, all of whom consequently have lives that are barely worth living, then we ought, nonetheless, to do it, just as long as moving the world in this grisly direction would add to the total amount of welfare summed across persons.”19 The answer is that economic analysis of climate policy does not escape this conclusion. Instead, it treats future population size as exogenously given to the analyst. But in the present context, this is illicit, for three reasons. First, global population size is a principal determinant of greenhouse gas emissions and, therefore, of the likelihood of future catastrophic climate changes. Second, all serious efforts at curbing greenhouse gas emissions are bound to affect global population size, directly or indirectly. Finally, emissions-reducing methods of managing population size must be evaluated, both because managing population size will be essential for preventing catastrophic climate change, and because many feasible methods of population control are potentially deeply wrong. For these reasons, if economic analysis hopes to yield an even close to allthings-considered judgment regarding what should be done about greenhouse gas emissions and how quickly, it cannot treat population size as exogenous to the analysis. And yet if it does not treat population size as exogenous, then, as a form of total utilitarianism, it is subject to a version of the Repugnant Conclusion: in setting climate policy, it will tell us to aim for a world that has a higher total amount of welfare simply in virtue of having many more billions of people, even if everyone who is born into this world will consequently have a life that is barely worth living. Obviously, then, for the economic approach to provide anything close to reasonable all-things-considered guidance, something sensible will have to be said to address this problem.

18  See, for example, Nordhaus (2017), which describes the influential Dynamic Integrated model of Climate and the Economy (DICE model). 19  For Parfit’s original axiological statement of the Repugnant Conclusion, see Parfit (2004).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  303

2.6  The Need for Discounting Consumption Possibilities in the Future Now, mainstream economic analysis of climate change does have a way of dealing with a closely related problem. Recall that for any total utilitarianism, the size of a population is merely one type of real-life “utility monster.”20 Another embarrassing utility monster always exists in the form of the countless number of future generations—whatever their size—each of whose aggregate welfare we, the living, can also clearly affect and even wipe out. Whereas economists have tended to shy away from addressing the problem of population size, they have long understood and been willing to address the problem for optimal policy created by the prospect of an indefinite series of future generations. In a nutshell, the problem is this. For every dollar that we can consume now or save and invest for the future, the objective of maximizing an additive intergen­er­ ational social welfare function will demand that we deny ourselves the use of this dollar and invest it for the future, just as long as this would produce positive net returns to aggregate welfare summed across an indefinite series of future gen­er­ ations. Thus, the recommended consumption and emissions pathway may look, not merely somewhat, but drastically different than under a policy of business as usual, even if we abstract from questions of the optimal size of future populations. Indeed, the people living today may be required to cut their emissions all the way down to subsistence level, if only this step, too, would provide positive net bene­ fits to a boundless series of future generations. Figure 13.1a–c illustrates. If the world continues with business as usual (as in scenario A), runaway climate change may eventually lead to a permanent collapse in human welfare through migration-fueled global conflicts that, for all we know, may turn out to be even more destructive than the wars that nearly destroyed modern civilization in the twentieth century. Presumably, there are emissions cuts that the present generation could undertake that, while slowing the global rate of economic growth for a time, would prevent truly catastrophic warming and put the human species on a more sustainable consumption path (scenario B). However, the present generation could in theory also cut its greenhouse gas

20  Explanation: because population size is a variable open to manipulation, it, too, must be evaluated, according to a total utilitarian perspective. But, if producing an even more massive population would contribute positively to the sum of welfare in history, then total utilitarianism will not care if, in the process, the welfare of already existing individuals is sure to decline precipitously, even to the point that their lives will be barely worth living: for the sake of higher total welfare, it will recommend taking steps that will produce a truly massive population nonetheless. Moreover, this implication holds, not merely with respect to science fiction examples of individuals who are implausibly im­agined to be utility monsters, but also with respect to a natural phenomenon—population size—that undeniably always exists, and which people are undeniably always able to manipulate. In this sense, population size is a “real-life” utility monster that is hard for total utilitarians to escape. For the original notion of a utility monster, see Nozick (1974: 41). For further discussion, see Parfit (2004).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

304  Lucas STANCZYK a.

Figure 13.1a  Total welfare in time in time under business as usual b.

Figure 13.1b  Total welfare in time under serious cuts c.

Figure 13.1c  Total welfare in time under enormous cuts

emissions drastically, to subsistence level or even lower for many millions of ­people, thereby affording the biosphere time to recover from such ongoing processes as the sixth extinction, and thereby making possible even higher consumptions paths in the more distant future (scenario C). As long as such cuts are technically feasible and would maximize total welfare summing across many generations, an intergenerational utilitarian social objective will require the present generation to undertake the severest possible cuts irrespective of the degree of implied sacrifice to people living in the present. Recognizing this problem, most economists have concluded that the only reasonable way to proceed is to systematically discount the vast amount of future

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  305 consumption (i.e. the much larger area under the curve in scenario C when the curve is extended indefinitely) that could be enabled by imposing extreme sacrifices on the present generation. To be sure, a variety of justifications have been proposed for deploying a positive social discount rate, with no clear consensus emerging about the weight that each deserves.21 Nonetheless, there is no dispute concerning what would be the practical consequence of failing to find any the­or­ et­ic­al­ly acceptable justification. Namely, as Eric Posner and David Weisbach have recently put it, paraphrasing the Nobel prize winner Tjalling Koopmans, “a failure to discount effectively means that the current generation must starve itself to bene­fit the future.”22 Since this implication is widely regarded as indefensible,23 economists have long supposed that some justification must be available for discounting future consumption.24 Otherwise, it is feared, optimal climate policy will require that the present generation should literally starve itself to benefit the future.

2.7  Welfarism Is At Best a Class of Pro Tanto Reasons This reasoning rests on a mistake, however. The conclusion that climate economists seek to avoid follows only if we, the people living today, have an overriding duty to maximize the sum of welfare across all of time. Yet this assumption misdescribes our duties of intergenerational justice in two important ways. First, we do not owe it to future people to help them get more of whatever they happen to want. Welfarism, at best, expresses a class of pro tanto reasons that we 21  The standard justifications for deploying a positive social discount rate include (a) the claim that wealth has declining marginal utility, and that future generations will be richer, (b) the need to take into account the possibility of humanity’s going extinct in the future, in which case present sacrifices would turn out to have been pointless, (c) the alleged importance of policymakers deferring to their current constituents’ pure social rate of time preference, and (d) the importance of discounting the welfare gains expected from climate change mitigation projects at the projected long-run market rate of return as a means of taking opportunity costs into account (assuming that growing conventional forms of capital at the market rate of return will benefit the future, too). For a general overview and a defense of the fourth approach, see Posner and Weisbach (2010: ch. 6). 22  Posner and Weisbach (2010: 149). Or as Kenneth Arrow once put it, “strictly speaking, we cannot say that the first generation should sacrifice everything, if marginal utility approaches infinity as consumption approaches zero. But we can say that given any investment, short of the entire income, a still greater investment would be preferred” (Arrow 1999: 14). 23  Though not unanimously so: for further discussion, and a defense of the revisionist view that “the axiomatic requirement to avoid the repugnant conclusion should be dropped from population ethics,” see Budolfson and Spears (Unpublished). See also Arrhenius, Budolfson, and Spears (2021) in this volume. 24  And indeed, it is no wonder. The IPCC estimates that, with the exception of parameters that represent the climate’s sensitivity to increased concentrations of greenhouse gases, the value assigned to the social discount rate is by far the most important factor determining what the leading integrated assessment models tell us to do about climate change. See the IPCC’s Fourth Assessment Report (IPCC 2007), cited in Posner and Weisbach (2010: 149). This is the straightforward result of so many attempts to escape the variant of the Repugnant Conclusion just noted.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

306  Lucas STANCZYK have. Moreover, welfarist reasons as a class are much weaker than other pertinent concerns. The desires of future people for robotic maids and elaborate computer games do not have the same urgency as the needs of poor people today for clean water, electricity, sanitation, medicine, the rule of law and other basic goods. Conversely, irreversible losses in the future of arable land and other life-sustaining resources cannot be offset by the ability of today’s wealthy consumers to continue their frequent air travel to distant locales. By treating present and future “preferences” for these diverse goods as if they were all on a moral par, and asking us to maximize the extent of preference satisfaction across all of time, the welfarist approach to climate policy as currently deployed misrepresents what we owe to our contemporaries as well as to others more distant in time. In thinking about the morally required emissions pathway, then, the first step must be to redefine the “currency” of intergenerational justice. Duties of intergen­ er­ational justice are not best understood as aiming directly at the satisfaction of an undifferentiated heap of preferences backed by purchasing power. Instead, distinctions must be made regarding the relative importance of the diverse goods that people want, as well as the relative importance of the various misfortunes they wish to avoid. For example, ensuring that people are able to avoid severe drought and famine is more important than ensuring that they or others are able to take more frequent trans-continental vacations. This is true, moreover, no matter how intensely anyone feels either of these preferences, and no matter how much money anyone is able and willing to pay to satisfy his particular preference. In the present context, these elementary observations have far-reaching consequences. They indicate that when thinking about the optimal emissions pathway, welfare economics can ever only have a highly subordinate role to play. This is, of course, not to say that economic analysis has no use whatsoever in this context. On the contrary, modern welfare economics is best understood as attempting to order the reasons we have to help present and future people get more of whatever they happen to want. Understood in this way, welfarist analysis of climate policy is a legitimate exercise and may be seen as having considerable normative force. At the same time, however, the reasons we have to help people get more of whatever they happen to want are far from the only reasons to care about how laws and policies are arranged. Much weightier reasons derive from the fundamental rights and entitlements that everyone should be understood to have, irrespective of what people (now or in the future) happen to want. The principles that summarize these reasons must be understood as ethical constraints on the objective of giving (more) people more of whatever they happen to want. These constraints must be applied to filter out unfair, unjust, undemocratic, and illiberal policies before welfarist analysis of climate policy has any place. Once this is done, welfarist analysis in the climate change context becomes a problem of maximizing preference satisfaction subject to prior ethical constraints, and the problems that plague the unconstrained variety of this approach will become that much less pressing.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  307 Indeed, in some ways, the picture I have outlined describes what is already the standard practice of economists. Thus, for example, economists are loathe to assign different values to white and African American statistical lives, even though there is abundant evidence that the average black person in the United States is willing to accept the same dangerous job for a fraction of the average white person’s reservation wage.25 This reluctance constitutes a clear practical recognition, not only that black people in the United States have unmet claims to equal opportunity and non-discrimination, but also that these claims are morally prior to the economist’s aim of maximizing the sum of preference satisfaction. Similarly, climate economists almost invariably treat population size as exogenous in their models, even though a policy of directing people to have more and more children could grow the total sum of welfare exponentially. (Conversely, in the absence of a similar external constraint, a focus on maximizing the average welfare would recommend discouraging further births in regions, such as sub-Saharan Africa, that are clearly driving down the global average.26) To avoid any hint of such policies, climate economists standardly treat population size as a variable that is to be determined “exogenously,” meaning that it is not to be manipulated in the name of optimal utilitarian policy. This, too, is an implicit recognition of a claim that has moral priority. People have a right to decide how many children they will have, and this right is—correctly—taken to be a constraint on the economist’s aim of maximizing aggregate preference satisfaction. To be sure, rights to procreative freedoms must be delimited and specified, particularly when population growth puts severe pressure on the carrying capacity of the earth.27 However, rights as basic as those protecting our freedoms to make decisions about childrearing and procreation are not best articulated by appealing solely to how a given specification would affect the aggregate sum of preference satisfaction. On the contrary, prior and more fundamental values are at stake, and the practice of economists treating population size as “exogenously” fixed is arguably a tacit recognition of this important moral truth. In another way, however, the practice of economists is inconsistent with the theoretical picture I have outlined. After all, economists continue to grasp for ways to justify a positive social discount rate, supposing that unless we can find a justification we will be required to endorse policies that effectively starve everyone alive today. What they fail to see is that principles of justice that are prior to efficiency already also rule out this possibility. And the reason why some

25  See for example Viscusi (2003). 26  For the recommendation that wealthy countries such as the United States should stop paying for food aid, permanently close their borders, and let the people born in the poorest and most populous countries kill each other or starve to death, all in the name of preserving a high average standard of living, see Hardin (1974) and Hardin (1968). The latter is remarkably one of the most widely cited articles in the social sciences ever. 27  I describe a framework that can be used for this purpose in section 3 below.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

308  Lucas STANCZYK economists fail to see this is that they think that sensible climate policy must ul­tim­ate­ly aim to maximize (the sum of) something. On this assumption, unless we can find a justification for discounting the interests of future people, our approach will shipwreck on some version of the Repugnant Conclusion no matter what currency of intergenerational justice is finally adopted. For example, even if we say that the goal of climate policy should be to maximize the total number of valuable opportunities in existence, we will be led to recommend a world in which there are many more billions of people each limited to a set of op­por­tun­ ities barely worth having, just as long as the total number of valuable op­por­tun­ ities summed across persons is larger. Moreover, even if we hold fixed the optimal size of future populations, the sheer number of future generations will continue to represent a similar “utility monster” for any version of total utilitarianism. It is no wonder, then, that so many economists insist that, in matters of intergen­er­ ational ethics, “the discount rate we choose is all important.”28

2.8  We Do Not Owe It to Future People to Bring Them into Existence This inference, however, is mistaken for a second reason. Namely, we do not owe it to future people to bring them into existence as a means of maximizing something. Indeed, a moment’s reflection shows that we do not owe it to future people to bring them into existence at all. This statement is true both at the level of individual ethics and as a matter of political morality or social justice. First, no fertile individual alive today morally owes it to any of his or her numerically distinct, merely possible future offspring to take steps that would bring any of them into existence. To be sure, some particular persons alive today may have promised their intimate partners that they will one day be willing to have children, and the reliance generated by this promise may, in some cases, even be enough to put these individuals under a moral obligation, owed to their intimate partners, to have children with them some day. Similarly, we may well owe it to each of our contemporaries, as a matter of basic justice, to take steps to ensure that none of them dies a horrible death once they have grown infirm in old age. Since to avoid this outcome, there will need to be younger people around to care for those who have grown infirm with age, there may well be a duty of justice, binding on everyone alive today, to create and support institutions that will ensure that a sufficient number of fertile persons alive today go on to have enough children to make it possible for each of our contemporaries to avoid horrible deaths once they have grown old. However, if there is indeed such a duty of justice to take steps to provide for the orderly exit of living 28  Weitzman (2007: 704).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  309 persons out of existence over time, as I am convinced there is, then clearly it is a duty of justice owed by each of us to our contemporaries, understood as the persons who are alive with us at any given time. It is not a duty that any of us presently owes to any merely possible future persons, that is to say, persons who do not presently exist and some of whom we merely might bring into existence by deciding to have additional children. The reason is that we do not owe it to any of these countless, merely possible future persons to bring any of them into existence in the first place. Instead, what is true is that we will owe it to these people to provide for their orderly exit out of existence, if they ever come to exist. Indeed, if and when add­ ition­al people come to exist, we will owe them many other things—ranging from police protection, the rule of law, and a system of functional educational institutions, to all of the other rights and entitlements that living persons have as a matter of domestic and international justice. Once additional people exist, their rights will then set constraints on what institutions may ever do to them in the pursuit of the maximum aggregate welfare, just as every living person’s basic rights and entitlements set strict limits on the pursuit of efficiency at this very moment. However, the moral force of the constraints on action represented by people’s basic rights and entitlements is always conditional on their coming into existence.29 After all, we do not owe it to any of our countless, merely possible future offspring to take their “rights” and “entitlements” into account when deciding how to live our lives. Indeed, we do not owe our merely possible offspring this kind of consideration even when deciding whether we will have any children at all.

3.  Outline of a More Adequate Approach 3.1  The Real Puzzle of Intergenerational Justice The real puzzle of intergenerational justice, then, is not how we can avoid starving the living in the name of creating a maximum amount of welfare in the future. The real puzzle is why we have a duty to preserve the environmental and social basis of just institutions, so that whoever comes to exist in the future will then have her rights respected and otherwise be treated fairly. More precisely, the puzzle is how we ought to understand this duty so that it will be appropriately demanding. It is again the non-identity problem that generates this puzzle. For, even if our failure to rein in greenhouse emissions will have truly dire consequences in the

29  For a defense of the more general claim that all of our practical reasons to care about how people fare are conditional on their ever existing, see Frick (2020).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

310  Lucas STANCZYK future, it is not clear that we will be wronging anyone by continuing with business as usual. This is so for two reasons, not always clearly appreciated. (a) On the one hand, it does not appear to be the case that our failure to restrict our emissions will undermine the ability of future generations to live together under just institutions. Even when the environment has been despoiled in (let us suppose) a couple of hundred years, to the point that not everyone alive can be adequately fed, it will continue to be possible for institutions to fairly ration the remaining arable land and resources, and otherwise treat all living persons in what will then be the morally required way. It is merely that, because the environment will have been ruined, maintaining just institutions in the future will at best afford everyone a much lower quality of life than some of us are able to enjoy today. (b) On the other hand, however, it does not appear to be the case that we will be wronging anyone by lowering the standard of living attainable in the future to a very low level. For, the present generation is in fact benefitted by not having to make enormous sacrifices for the sake of the distant future. And, even if we were to make radical changes to our energy infrastructure, this would not actually help the people who will see the world’s largest cities consumed by oceans several centuries from now. On the contrary, these particular people would end up never being born at all. After all, who meets whom and decides to have children together strongly depends on the cost of air travel, road transport, and countless other elements of our energy infrastructure. Accordingly, if governments were to make the use of fossil fuels much more expensive in the near term, before abundant substitutes can be mass-produced cheaply, people would stay closer to home when they went on vacation or when they traveled at other times and places. In turn, these changes in individual behavior would inevitably affect the identity of the people who meet, and therefore also the identity of the individuals who fall in love and decide to have kids. But just as you would never have been born, had your mother never met your father and instead had children with some other individual, so, too, the persons who will live if we continue with business as usual will not be one and the same as the persons who will be born if the world makes rapid and extensive changes to its energy infrastructure. Like ripples in a pool, the behavioral changes that will then inevitably follow will alter not only who meets whom and decides to have children together in the proximate future, but also, consequently, whose children will exist and go on to have children of their own in the next generation. Thus, if governments put in place heroic cuts to greenhouse gas emissions today, this step will not actually help the people who, several centuries from now, will experience the loss of the world’s major cities from dozens of feet of sea level rise. On the contrary, the behavioral changes that governments will occasion by making the required cuts today will, in due course, ensure that the people who would otherwise live through such disasters will end up never being born at all.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  311 So, from the point of view of intergenerational justice, it is not clear why we ought to make deeply painful cuts to greenhouse gas emissions today. It seems that the people who will live through climate catastrophes will not be benefitted if we do, whereas the duty not to undermine the possibility of just institutions in the future does not appear to rule out our reckless behavior either. Why, then, should we be morally required to depart dramatically from business as usual?

3.2  Understanding the Full Measure of the Problem In fact, stating the problem in this way considerably understates the difficulty. For it is not at all clear that a refusal by governments to rapidly cut emissions to net zero would indeed be morally reckless. To think this—to think that rapid reductions in greenhouse gas emissions are clearly morally required, and to dismiss the philosophical problem on this basis—is to misunderstand the severity of the global energy challenge and the extremely difficult dilemma that faces us. After all, as I noted at the outset of this chapter, there are billions of people around the world who lack access to goods as basic as clean water, adequate nutrition, sanitation, medicine, shelter, and even electricity. And yet, there is no way to alleviate these deadly hardships in the short-run without substantially growing the global level of greenhouse gas emissions. Indeed, there is no way to continue operating the industrial, shipping, construction, and other sectors of the global economy even at current levels of output without using vast quantities of fossil fuels for many more decades. The reason is as tragic as it is simple: alternatives to many essential combustion engines, petrochemicals, steel, concrete and other carbon-intensive construction materials have yet to be invented, and the research that is necessary to mass-produce them will, in many cases, take decades even in the most favorable of circumstances. This, then, is the full practical force of the philosophical problem. Without substantially adding to the stock of greenhouse gases in the short-run, it will be impossible to grow economic output enough to provide every living Indian and African with such basic goods as indoor plumbing, reliable electricity, and access to sterile hospital operating rooms. Yet it is not clear what can be said to justify to these persons, so desperate for such energy-intensive basic goods, a policy of large and rapid global greenhouse gas emissions cuts. For even if rapidly industrializing Africa and India today is likely to produce catastrophic climate impacts in the future, it appears that we cannot tell our fellow Indians and Africans that the present generation will be wronging the future by failing to make deeply painful cuts to emissions today. To be sure, unless global emissions are quickly reduced to net zero, some of the world’s most populous cities may well be swallowed up by the oceans eventually. However, it seems that the people who will live through

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

312  Lucas STANCZYK such catastrophes will not be benefitted if we slash our emissions today, since the behavioral changes that will then inevitably follow will mean that these particular people will end up never being born at all. And, it seems that by continuing with business as usual, we will also not be undermining the possibility of just institutions in the future. For, even once the biosphere has been despoiled, institutions in the future will still be able to ration the remaining resources and otherwise treat people in whatever will then be the morally required way. Conversely, if providing some good to everyone will no longer be possible, then institutions will no longer required to do so: ought implies can, after all. Either way, carrying on with business as usual does not seem to undermine the very possibility of just institutions in the future. Why, then, should the present generation be required to impose on itself such deeply painful cuts to greenhouse gas emissions today?

3.3  Elements of a Solution: Two Supporting Considerations I believe it is possible to solve this puzzle, and at once to vindicate the idea that humanity is at all times subject to a suitably high environmental conversation standard. In order to do so, we ought to appeal to the following combination of arguments.30 First, we should note that it is possible to wrong persons in the distant future by allowing them to be brought into a terrible existence—if their lives will be so bad as to be not even worth living. Of course, by itself this observation is compatible with a great deal of reckless fossil fuel use in the present. For conditions on

30  To be clear, I understand the following combination of arguments to be complementary to any other promising strategies. At the same time, however, I am not wholly satisfied with the alternatives that have been offered. One such strategy is to argue that failing to cut global greenhouse gas emissions quickly is a form of victimless wrongdoing. I doubt that any such claim could, by itself, be enough to justify imposing deeply painful emissions cuts on the least advantaged living persons. Another strategy, bearing similarities to mine, is Anca Gheaus’ argument that (a) every living person capable of parenting has a right to the conditions necessary for materially adequate parenting, and that (b) this right requires that every person capable of parenting be able to have children in conditions good enough for their future children, too, to enjoy the conditions necessary for materially ad­equate parenting (and so on). As I understand it, this argument fails as a response to the problem that I have explicated. For (a) if, as a consequence of our continuing with business as usual, institutions in the future will no longer be able to secure for everyone the conditions necessary for materially adequate parenting, then institutions in the future will also no longer be required to do so: ought implies can. For this reason, it appears that our continuing with business as usual will not undermine the possibility of just institutions in the future. At the same time, (b) it is not clear why we will be wronging anyone by depriving future people of the material conditions necessary for adequate parenting. After all, the present generation appears to be benefitted by not having to make enormous sacrifices for the sake of the distant future, whereas the people who will eventually be deprived of the conditions for adequate parenting as a result of our present course would end up never being born at all, were the world to change its present course. For Gheaus’ argument, see Gheaus (2016). For another instructive approach that bears structural similarities to mine, but from which my approach departs for similar reasons, see Mazor (2010).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  313 earth will have to get very bad before human life becomes not even worth living. So, we cannot stop there. Second, we should point out that it is false that climate disasters attributable to today’s emissions will befall only people who will be born in the future. On the contrary, some children who are already alive will be gravely harmed because of the emissions that the global economy will put out tomorrow. Therefore, whoever now jointly holds power over the content of public policy and the operation of the global economy has an unconditional duty to take the long-run interests of these children into account when deciding how the global economy should be run. Obviously, this does not mean that fossil fuel use should be forbidden or that net emissions should suddenly be cut to zero. After all, the very survival not only of these children but of billions of other people depends on the continued op­er­ ation of an economy that currently cannot operate without the extensive use of fossil fuels. However, since many children who will outlive us are already alive, it would be clearly wrong for existing power-holders to assign zero weight to the consequences of our emissions in, say, eighty years. Instead, the long-run interests of these children must now be taken into account, in whatever ways are demanded by the best theories of social and international justice. So, for example, if our theories tell us that persons have a life-long right to institutions that protect them from unnecessary famine, and that this right is more basic than anyone’s claim to travel long distances by plane for vacation, then, to that extent, there is already strong moral reason to restrict today’s commercial aviation emissions because of the drought and famine they will help to occasion in eighty years. It makes no difference that most of the people alive today will be dead by then. Some persons who already exist will survive for another eighty years, and so there is now a duty to ensure that institutions which give the appropriate weight to every living person’s diverse claims are erected and con­ tinu­ous­ly maintained for at least that long. In short, failing to moderate our emissions today wrongs children who were born around the world yesterday. For it will create a situation in which, at the end of their lives, not all of these children will be able to escape conflict, famine, and drought, seemingly for no reason more important than the present consumption of luxury goods. Letting this happen now is incompatible with our duty to ensure, as far as we can, that just institutions are in place and at all times give the appropriate weight to every living person’s diverse claims. In making this claim, I am assuming that some of the world’s existing emissions could be reduced at the expense only of wealthy people’s access to unambiguously luxury goods. As will be clear by now, this is not a trivial assumption. It is a regrettable but important fact that, under global capitalism, the wellbeing of the world’s poorest people often depends on more rather than less consumption by already wealthy people. Indeed, in some places, the very survival of poor ­people may well depend on the growth of luxury tourist travel. Nonetheless,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

314  Lucas STANCZYK I  assume that even under the contemporary global economic status quo, requiring for its very maintenance perennial positive economic growth, some of the world’s emissions presently fuel the consumption of only unambiguously luxury goods, and therefore that some of the world’s emissions are properly described as worsen­ing climate change only for this relatively unimportant purpose. To the extent that the empirical element of this claim is true, the moral conclusion that I have drawn will hold. Even if the disastrous consequences will not be felt for another eighty years, letting the world’s emissions run wild now is incompatible with the duty to ensure, as far as we can, that just institutions are in place and at all times give the appropriate weight to every living person’s diverse claims. The reasons for this conclusion, in sum, are twofold. First, there are children already alive who can be expected to live at least that long. Second, the claims of each of these children to be able to go through life without having to become the victims of conflict or drought are more important or basic than the claims of wealthy people to heat and cool oversized houses and travel long distances for vacations.

3.4  Elements of a Solution: Two Ideas Needed to Side-step the Non-Identity Problem What these reflections show is that there is no deep puzzle of intergenerational justice when thinking about how the global economy should be arranged over periods of time approximating a single human life span. The duty to ensure, as far as we can, that just institutions are continuously maintained for whoever exists, if and as long as they exist, already imposes on us the appropriately demanding requirements over this particular time-horizon. The content of these requirements is straightforwardly given by the best theories of social and international justice. However, the central idea at work in the above argument—that there are different types of moral claim picked out by a theory of justice, and that these different types of claim have varying levels of importance—can be deployed in a third distinct argument, generating a requirement on every generation in history to show appropriate concern at every stage in history. To see this, imagine a family with a mother and father and two children, who live on some land that has for many generations been over-farmed. They now live at the edge of their means, in the following sense: if they were to have additional children, then it would no longer be possible for everyone to be adequately fed. The question is why it would be wrong to allow this to happen, assuming that the additional child would have a life that is still worth living and doing otherwise would mean that he or she would end up never existing. The answer is that the parents would thereby wrong one another and their existing children. The ex­plan­ ation for this judgment involves two distinct set of ideas. I will now show that, in

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  315 conjunction with the third set of ideas below, the explanation can be generalized to produce an unending chain of duties of intergenerational justice untouched by the non-identity problem. (a) Here, then, is the first idea: as a matter of justice, persons have claims to the protection of diverse goods and interests, and these claims have different degrees of importance. Therefore, the rules of a society ought to avoid indulging claims to less important interests at the expense of claims to interests that are normally even more important. Yet, this is precisely what the parents would be doing by failing to put in place a moderate restriction on their procreative freedoms (“moderate,” because they have already allowed themselves to have two children). By failing to introduce a moderate restriction on the rights defining their procreative freedoms, the parents would create a situation in which an even more serious restriction—on a right that normally protects an even more urgent interest—would be needed to meet the diverse claims of everyone in existence. For then there would be five mouths to feed, each with identical claims on sustenance, and so access to (even a bare minimum of) food would now have to be strictly rationed. (b) In permitting this situation to arise, the parents would have wronged everyone who was alive before the arrival of their third child. This is because of what I will call the principle of disjunctive wrongs: in the absence of an adequate justification, it wrongs people to act in ways that will necessitate either harming or wronging them in the future. Yet, once again, this is precisely what the parents would be doing by allowing themselves to have additional children. Not only would they be (i) creating a situation in which they could no longer avoid harming or wronging one another and their first two children, but they would also be (ii) avoidably creating this situation without adequate justification. To see why (i) is true, notice the dilemma that would face the parents after the arrival of their third child. To feed all five mouths, the parents would need to begin strictly rationing access to food, with even a fair allotment leaving everyone with a less than adequate amount. By introducing the requisite system of rationing, then, the parents would be seriously (physically) harming one another as well as their first two children. At the same time, however, it would wrong everyone not to introduce the fair system of rationing and instead simply let the third child starve. It is clear why this course of action would wrong the third child after its birth: everyone who comes to exists has, ignoring special needs, identical claims to food. However, by refusing to provide for the third child, the parents would simultaneously wrong one another and their first two children as well. For, it clearly wrongs a person to make her live together with others under unjust rules and institutions. Indeed, the wrong is especially grave when the rules disregard the most urgent claims of others to her considerable material benefit. Accordingly, by failing to restrict everyone’s access to food after the arrival of their third child, the parents would seriously wrong one another as well as all three of their kids.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

316  Lucas STANCZYK In sum, then, after the arrival of their third child, the parents would find themselves in a situation in which it would no longer be possible to avoid either harming or wronging, not only their third child, but also one another and their first two children as well. By refusing to place restrictions on the latter’s (hitherto ad­equate) access to food, the parents would wrongly make everyone live together with others under unjust rules. However, by changing the rules and restricting their own and their older children’s access to even the bare min­ imum of food, the parents would seriously physically harm one another and their first two kids. (ii) Now, to be sure, it can sometimes be permissible to create a situation in which the only way to avoid wronging someone will be to harm people. As we have already seen, however, anyone doing so in circumstances akin to the parents’ situation is bound to lack sufficient justification. Let us state once more the general explanation. As a matter of justice, the rules of a society should normally avoid indulging claims to less important interests at the expense of claims to interests that are in the circumstances even more important. Yet this is precisely what the parents would be doing if, prior to there being a third child, the parents were to fail to put in place a moderate restriction on their procreative freedoms (“moderate,” because they have already allowed themselves to have two children). Namely, the parents would avoidably be creating a situation in which an even more serious restriction, on a right that normally protects an even more urgent interest, would be needed to satisfy the diverse claims of everyone already alive prior to there being any third child. For then there would be five mouths to feed, and so not merely the procreative freedoms but even the access to food of the parents and their first two children would have to restricted, on pain of wronging these very same people (plus one other, new person). Accordingly, in creating the need to harm-or-wrong one another and their first two children in the future, the parents would be acting without adequate justification. By the principle of disjunctive wrongs, then, in failing to restrict their procreative freedoms, the parents would also be wronging one another and their first two children.

3.5  Extending the Reasoning Indefinitely Backwards and Forwards in Time Finally, let me now illustrate how this reasoning can be extended indefinitely backwards and forwards in time. To see this, notice that the parents themselves will have been wronged by their parents, if, through the grandparents’ riotous living, the grandparents avoidably contributed to over-farming and otherwise reduced the earth’s carrying capacity. The explanation for this judgment appeals to the same two ideas as above. First, by failing to put in place restrictions on less

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  317 important, luxury consumption-related freedoms, the grandparents will have avoidably saddled everyone (including one another) with a situation in which an even more serious restriction would be needed—on a right that normally protects even more important, procreative freedom-related interests—in order to meet the diverse claims of everyone already in existence at the time that the still-youthful grandparents were disastrously over-farming. Second, already at this time, the grandparents will have wronged one another as well as the parents through their riotous living. For, without adequate justification (as indicated by the first observation), the grandparents will thereby have created a situation in which the grandparents could no longer avoid harming-orwronging one another as well as the parents in the future. After all, thanks to the grandparents’ riotous living, the grandparents and the parents alike would now be forced to restrict everyone already alive to having at most two children, on pain of acting wrongly for a structurally identical reason: i.e. failing to put in place a restriction on the rights defining everyone’s procreative freedoms; thereby creating a situation in which an even more serious restriction would be needed to meet the diverse claims of everyone already in existence (namely, a painful food rationing system); and thereby producing, without adequate justification, a situ­ation in which it would no longer be possible to avoid either ser­ ious­ly harming already existing people or wronging them for a structurally identical reason. In this way, then, the chain of reasoning that I have illustrated can be extended indefinitely backwards and forwards into the future. The illustration shows that, at every stage in history, the people alive at the time can be said to have strong moral reason to put in place restrictions on rights and freedoms whose exercise affects the earth’s carrying capacity, if failure to put in place these restrictions will produce a situation in which even more serious restrictions will be necessary to meet the diverse claims of everyone in existence (even more serious because touching a right that protects an even more urgent interest). Moreover, my illustration shows that failure to put in place such restrictions at any point in history wrongs people already in existence who will live to experience this necessity, including the members of already existing younger generations. Finally, my illustration shows that the corresponding duties of intergenerational justice are not undermined by the non-identity problem. On the contrary, when the failures of one set of people seriously wrong other already existing persons, it is simply ir­rele­vant to the existence of the wrong that, as a consequence, the identity of asyet unborn future generations will change. So, too, it is simply irrelevant to the existence of duties of intergenerational justice that, were each generation in history to honor these duties and therefore avoid wronging the members of alreadyexisting overlapping generations, the identity of the people who will be born in the future (and how many) would likewise inevitably change.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

318  Lucas STANCZYK

3.6  The Duty to Avoid Allowing Catastrophes to Take Place in the Distant Future This formulation brings me to a final observation. As will be clear by now, the reasoning that I have explicated in this essay is able to generate duties of intergen­ er­ational justice irrespective of the non-identity problem in part for the reason that there are never any gaps in humanity’s existence. It is never the case that humanity ceases to exist temporarily. Instead, as long as humanity exists, it is always made up of an overlapping chain of generations. However, the example above is notable for lacking a second, distinct kind of gap: a gap of time, spanning the existence of more than a few overlapping generations, between when some irresponsible behavior takes place and when its damaging effects will finally occur. What if, then, humanity’s present-day failure to, say, take steps to reduce its rapidly expanding air travel will not give rise to the need to put in place even more serious restrictions on human freedom until the distant future, when every last person who is alive today will long be dead? Using the reasoning that I have explicated in this chapter, can we still say that we will be wronging people now by failing to reduce aviation emissions today? The answer is yes, because the possibility of the relevant gap is ultimately illusory. More precisely, there is no way to create the need to put in place severe restrictions on human freedoms in the distant future without also affecting what the present generation is required to give up today, in light of the catastrophe now looming in the future. On the contrary, creating the need for humanity to make enormous sacrifices in the distant future always has the further effect of bringing into existence a requirement on all those who are already alive (a) to make whatever sacrifices are necessary to plan for, help avert, or mitigate the catastrophe later, (b) just as long as the sacrifices that can now be made are necessary to avoid having to inflict even greater sacrifices later for this purpose. By the principle of disjunctive wrongs, moreover, the failure to make these no-less serious sacrifices always wrongs the members of overlapping generations, since, without adequate justification, it creates a situation in which it will no longer be possible to avoid harming or wronging one’s contemporaries. Finally, this reasoning can once again be extended backwards and forwards indefinitely. To illustrate, consider the extremely dangerous waste produced in nuclear re­actors. At some point in the distant future, the storage containers that have been built to protect human beings from this waste will fail. Suppose, then, that, at this final stage, truly great sacrifices will have to be made in order to mitigate the resulting catastrophic damage, including a great many people having to risk and lose their lives. In that case, for the reasons given above, it will be wrong for the members of the generation who first experience this catastrophe to fail to implement immediately the rules that will then be necessary to avoid even greater sacrifices having to be made later to contain the fallout. For structurally identical reasons, however, it will also have been wrong for the members of any overlapping older

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  319 generations to fail to make the sacrifices necessary to plan for, help avert, or ­mitigate the catastrophe earlier. For even if, at a pre-catastrophe penultimate stage, averting the catastrophe would have required re-directing a truly massive amount of resources to disaster planning and engineering research, the members of the overlapping older generation would nonetheless have been wrong to refuse to do so. After all, once the waste systems finally failed catastrophically, the failure of the older generation to pay for this research will have led to a situ­ation in which they and others would be required to impose even more serious sacrifices on already existing people. Without adequate justification, then, the members of the previous generation will have created a situation in which they could no longer avoid harming or wronging their contemporaries: arranging for the imposition of the required sacrifices in the event of the catastrophe, or wronging everyone by failing to do so. The same chain of reasoning can be extended indefinitely backwards in time. For, if it was wrong for the oldest generation on the scene during the final crisis to fail to make no-more serious sacrifices at an earlier time to try the avert the final catastrophe, then, for the same reason, this generation will itself have been wronged by an even older generation who was still on the scene at that earlier stage, when a massive research mobilization was already required. More precisely, the members of the gen­ er­ation who failed to redirect massive resources to planning and research well before the final crisis happened will have been wronged by an even older generation who was still alive when this particular failure happened, just as long as at any, even earl­ ier point in time during which these two generations overlapped, the latter gen­er­ ation could have planned for—helped to avert or mitigate—the eventual catastrophe without any of their contemporaries later having to mobilize quite as many resources quite so rapidly for this purpose, and therefore without having to sacrifice quite as much as their contemporaries would later then be required to sacrifice. But, of course, the same can be said about this, and indeed every other, “even older” generation. For the reasons given in Section 3.4, they, too, will have been wronged by an even older generation, if, at some point in their shared history, the latter generation failed to redirect extensive social resources to planning and research, just as long as this meant that even greater resources would have to be mobilized even more rapidly in the future. Since this latter generation could very well be us, however, we have now completed our demonstration. For we can now see that even very long gaps between the causes of and the onset of catastrophic climate impacts do not undermine the recommended way of thinking about intergenerational justice.

4. Conclusion As I have argued, then, the chain of reasoning described in the second half of this chapter can be extended indefinitely backwards and forwards into time. It shows that, at every stage in history, the people alive have strong moral reason to put in

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

320  Lucas STANCZYK place restrictions on rights and freedoms whose exercise affects the earth’s ­carrying capacity, whenever failure to put in place these restrictions will produce a situ­ation in which even more serious restrictions will be necessary to meet the diverse claims of everyone in existence—even more serious because touching a right that protects an even more urgent interest. The analysis also shows that failure to do so wrongs people already in existence who will live to experience this necessity, including the members of already existing younger generations. I have argued that, by adopting this structure of reasoning, reflection on intergenerational justice can fully side-step the non-identity problem, in the sense of making it strictly irrelevant to specifying the core content of duties of intergen­er­ ational justice. Indeed, I have suggested that reflection on intergenerational just­ ice can draw its core content directly from the theories of social and international justice. Even then, such reflection can still yield the conclusion that at every stage in history, sacrifices are required to be made by the living that, if made, will prevent even more serious sacrifices having to be made in the future. This, then, is what I have come to think is the correct form of answer to the question of how quickly the world should reduce its greenhouse gas emissions. For the reasons articulated in the first part of this chapter, the correct form of answer to this question is not whatever will produce the greatest intergenerational sum of welfare. Instead, the world should put in place all those restrictions on existing rights and freedoms that, if put in place, will prevent even more serious sacrifices having to be made in order to meet, for everyone who ever lives, the several types of claims picked out by the theories of social and international just­ ice. One important theoretical implication of this answer may well be that a certain form of sustainability is the central imperative of intergenerational justice, notwithstanding the challenges that have been leveled at this notion.31 The most important practical implication may well be a requirement on existing governments and other powerful people to mobilize massive new resources for research. I plan to think more about both of these implications in the future.32

31  For an influential discussion of this notion, see the report of the Brundtland Commission (1987). 32 For helpful feedback on the ideas presented in this chapter, I am indebted to audiences at Princeton’s University Center for Human Values, the Departments of Philosophy and Political Science at Stanford University, the Philosophy Department at UCSD, the Law and Philosophy Workshop at Berkeley, the Law and Philosophy Colloquium at Queen’s University, the Philosophy Colloquium at Northeastern University, the Montreal Social Justice Theory Workshop, the Murphy Institute at Tulane University, and Harvard’s Kennedy School of Government. For generous and incisive comments, I am especially grateful to Johann Frick, David Estlund, Frances Kamm, Tim Scanlon, LouisPhilippe Hodgson, Eric Beerbohm, Charles Beitz, Marc Fleurbaey, Andrew Williams, Paula Casal, Seana Shiffrin, Niko Kolodny, Anca Gheaus, Nicholas Vrousalis, Julie Rose, Sean Ingham, Nathan Lee, Galen Hall, Britta Clark, Bethany Cates, the students in my graduate seminars on non-consequentialist approaches to intergenerational justice, and the editors of this volume.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

How Quickly Should the World Reduce its EMISSIONS?  321

References Ackerman, F. and L. Heinzerling (2004). Priceless: On Knowing the Price of Everything and the Value of Nothing. New York: New Press. Arrhenius, G., Budolfson, M., and D.  Spears (2021). “Does Climate Change Policy Depend Importantly On Population Ethics? Deflationary Responses to the Challenges of Population Ethics for Public Policy”, ed. M. Budolfson, T. McPherson, and D. Plunkett. New York and Oxford: Oxford University Press, pp. xx–xx. Arrow, K. (1999). “Discounting, Morality, and Gaming,” in Discounting and Intergenerational Equity, ed. P. R. Portney and J. P. Weyant. New York and London: Resources for the Future, 13–21. Boardman, A.  E., D.  H.  Greenberg, A.  R.  Vining, and D.  L.  Weimer (2005). CostBenefit Analysis: Concepts and Practice, 3rd edition. Upper Saddle River, NJ: Prentice Hall. Broome, J. (1978). “Trying to Value a Life.” Journal of Public Economics 9: 91–100. Broome, J. (2010). “The Most Important Thing about Climate Change,” in Public Policy: Why Ethics Matters, ed. J. Boston, A. Bradstock, and D. Eng. Canberra: ANU E Press, 101–116. Broome, J. (2018). “Efficiency and Future Generations.” Economics & Philosophy 34: 221–241. Brundtland Commission (1987). Our Common Future: The World Commission on Environment and Development. Oxford: Oxford University Press. Budolfson, M. and D.  Spears (Unpublished). “Why the Repugnant Conclusion is Inescapable.” Mimeo. Frick, J. (2015). “Contractualism and Social Risk.” Philosophy & Public Affairs 43: 175–223. Frick, J. (2020). “Conditional Reasons and the Procreation Asymmetry.” Philosophical Perspectives: Ethics, 34: 53–87. Gheaus, A. (2016). “The Right to Parent and Duties Concerning Future Generations.” Journal of Political Philosophy 24: 487–508. Hardin, G. (1968). “Tragedy of the Commons.” Science 162: 1243–1248. Hardin, G. (1974). “Living on a Lifeboat.” Bioscience 24: 561–568. Hausman, D., M.  McPherson, and D.  Satz (2017). Economic Analysis, Moral Philosophy, and Public Policy, 3rd edition. New York: Cambridge University Press. Hsiang, S. M. and A. H. Sobel (2016). “Potentially Extreme Population Displacement and Concentration in the Tropics Under Non-Extreme Warming.” Scientific Reports 6: 1–7. IPCC (2007). Climate Change 2007: Synthesis Report, Working Group II: Impacts, Adaptation, and Vulnerability. Geneva: IPCC. IPCC (2014). Climate Change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Core Writing Team: R. K. Pachauri and L. A. Meyer. Geneva: IPCC.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

322  Lucas STANCZYK Kelleher, J. P. (2015). “Is There a Sacrifice-Free Solution to Climate Change?” Ethics, Policy & Environment 18: 68–78. Mazor, J. (2010). “Liberal Justice, Future People, and Natural Resource Conservation.” Philosophy & Public Affairs 38: 380–408. Nordhaus, W.  D. (2017). “Revisiting the Social Cost of Carbon.” Proceedings of the National Academy of Sciences 114: 1518–1523. Nozick, R. (1974). Anarchy, State, and Utopia. New York: Basic Books. Parfit, D. (2004). “Overpopulation and the Quality of Life,” in The Repugnant Conclusion: Essays on Population Ethics, ed. J. Ryberg and T. Tännsjö. Dordrecht: Kluwer, 7–22. Posner, E.  A. and D.  Weisbach (2010). Climate Change Justice. Princeton, NJ: Princeton University Press. Stern, N. (2014). “Ethics, Equity, and the Economics of Climate Change 1: Science and Philosophy.” Economics and Philosophy 30: 397–444. US Environmental Protection Agency (2010). Guidelines for Preparing Economic Analyses. Washington, DC: National Center for Environmental Economics. Viscusi, W. K. (2003). “Racial Differences in Labor Market Values of a Statistical Life.” Journal of Risk and Uncertainty 27: 239–256. Weitzman, M. L. (2007). “A Review of The Stern Review of the Economics of Climate Change.” Journal of Economic Literature 45: 703–724.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

14

Political Realism, Feasibility Wedges, and Opportunities for Collective Action on Climate Change Mark Budolfson

1. Overview According to an influential realist argument about what it is politically feasible to do about climate change, regimes are infeasible if they make current citizens of powerful nations worse off, and thus an intuitively unjust global response to the problem of climate change that involves compensating large emitters for reducing emissions is the best we can realistically hope for and is thus the solution that we should actively promote even from an ethical point of view. The realist conclusion of this kind of argument has been endorsed by a wide range of commentators in philosophy, economics, law, and international affairs such as John Broome, Eric Posner, David Weisbach, Cass Sunstein, and others. For example, Broome argues that the costs of mitigation must be shifted to richer future generations to ensure that no one now is made worse off by climate policy, whereas Posner, Weisbach, and Sunstein suggest that poor nations more vul­ner­ able to climate change must make large transfer payments to rich nations who have less to gain to ensure that the rich nations are made no worse off. In this chapter I examine the substance of the realist conclusions that these authors reach, and I identify a number of worries that differ depending on whether the implementation involves the kind of intergenerational transfers that Broome has in mind, or the intratemporal transfers that Posner, Weisbach, and Sunstein have in mind. As a more general point, I also argue that the realist argument is invalid, and that the reason why it fails also points the way toward a more desirable realist response than these commentators endorse. The argument is invalid because it overlooks the fact that what it is in the interest of a nation to do can change depending on the actions of other nations. In particular, even if nations are as invariably self-­interested as the realist premise of the argument assumes, other nations can still change what it is in their interest to do via threats, sanctions, and Mark Budolfson, Political Realism, Feasibility Wedges, and Opportunities for Collective Action on Climate Change In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Mark Budolfson. DOI: 10.1093/oso/9780198796282.003.0015

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

324  Mark Budolfson other measures, and can therefore make it in their interest to comply with ­inter­nation­al standards that it would not have been in their interest to comply with in the absence of those measures. As a concrete illustration of how this ­creates t­ rouble for the argument in the particular case of climate change, I discuss an international climate treaty that requires signatories to impose globally optimal emissions taxes within their territory and empowers compliant nations to impose tariffs and other penalties on non-­compliant nations as retaliation for non-­compliance, somewhat akin to proposals by Joseph Stiglitz and William Nordhaus. I then offer a positive proposal focused on adding a number of feasibility enhancing features and complementary measures that I call ‘feasibility wedges’, involving creative diplomacy and political strategy that may be required in order to achieve anything in the direction of optimal policy outcomes. I identify a strategic dynamic that allows a ‘meta-­architecture for agreement’ including intraand intertemporal transfers to be incorporated into the emerging post-­Paris ‘­bottom-­up regime’ for climate change, with the goal of incentivizing deeper emissions reductions more in the direction of optimal cooperative policy.

2.  The Realist Argument and Efficiency Without Sacrifice According to many, the primary problem standing in the way of an effective global response to climate change is a problem of political feasibility at the level of nations. The basic problem is that it is simply not in the interest of the current citizens of many nations to make substantial emissions reductions, and so, realis­ tically, we should not expect substantial emissions reductions unless those people are ‘bribed’ into making them. This is a large part of a realist explanation why no meaningful action on cli­ mate change has been observed despite its dramatic threat to future generations, and why more substantial action will not happen unless something substantial is done to make it in the interest of current people to make such reductions. Realists often also cite the prisoner’s dilemma-­like nature of the situation, whereby if an individual nation makes large reductions unilaterally without the cooperation of the other nations—or, if even a large but insufficient number of nations join together to make such reductions—they will suffer significant economic losses, and the climate-­related outcome will not be importantly different either for their own citizens or for others around the world who are vulnerable to the effects of climate change. This is partly because of the phenomenon of ‘leakage’, whereby unilateral regulation to reduce emissions in one part of the world could create large incentives for industries there to simply relocate to the unregulated areas of the world and continue emitting in those unregulated areas—thereby increasing the costs to the unilaterally regulating regions, while undermining the realization

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Political Realism, Feasibility, and Collective Action  325 of any global benefits due to the leakage, and thereby reducing the value of the already small sliver of the global benefits for regulating nations.1 From these premises about our circumstances together with the characteristic view of realism, the Realist Argument I will focus on here reaches a conclusion we can call Efficiency Without Sacrifice: Realist Feasibility Constraint: Nations act only in the interests of their current citizens, so a response to climate change is infeasible if it requires a nation to act contrary to the interests of its current citizens. Circumstances: An optimal response to climate change requires substantial emissions reductions from the no policy status quo from many rich nations, and the current citizens of those rich nations would be made much worse off relative to business as usual by such reductions unless they were compensated for ­making them. Therefore, given the Realist Feasibility Constraint and our Circumstances, the best feasible response is Efficiency Without Sacrifice, which involves re­du­cing emissions to a level that is optimal,2 while compensating current people, including citizens of rich nations, for the cost of making such reductions, thereby ensuring that no one including the rich has to make any sacrifices, and that the costs are instead borne by those who are more vulnerable to climate damages, because only in this way can substantial emissions reductions become feasible.

This argument has been endorsed by a wide range of commentators in philoso­ phy, economics, law, and international affairs such as John Broome, Eric Posner, David Weisbach, Cass Sunstein, and many others.3 It depends crucially on the idea that there are transfers (‘side payments’) between people that are both ­feasible and yield a pareto improvement over the no policy ‘business as usual’ status quo. In what follows I first evaluate the different substantial versions of Efficiency Without Sacrifice that have been offered by these commentators. I suggest that the Realist Argument is invalid because it depends on a mischaracterization of

1  Rendall (2015), Budolfson (2012), Nordhaus and Boyer (2000), and Böhringer et al. (2012) on leakage. 2  Following standard usage, the so-­called ‘optimal’ level of emissions is one at which the marginal cost of further reductions equals the marginal benefit from avoiding the damage from further emissions. 3 For example Posner and Weisbach (2010: 6, 86, and 143); Posner and Sunstein (2008: esp. 1569–1570) (but see also Cass Sunstein, “US Should Act Unilaterally on Climate Change,” http://www. bloomberg.com/news/2013-­01-­23/u-­s-­should-­act-­unilaterally-­on-­climate-­change.html); Stewart and Wiener (2003: 102–103); Wiener (2007: 75–76); Broome (2012: 44–47).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

326  Mark Budolfson the feasible options. Ultimately, this suggests complementary additional ­components that should be included in order to design the best feasible response to climate change.

3.  Posner, Weisbach, and Sunstein’s Version of Efficiency Without Sacrifice: Intratemporal Transfers In this section, I raise some concerns about the substance of Posner, Weisbach, and Sunstein’s version of Efficiency Without Sacrifice. On their version of the view, the way to achieve efficiency without sacrifice is for poor nations such as Tuvalu that have relatively more to lose from climate change to make large trans­ fers to rich nations such as the United States so that the current citizens of the rich nations can be made better off vs. the no policy status quo by the com­bin­ ation of those transfers and incurring their share of the costs of a global har­mon­ ized carbon tax. This is a striking view about what is the best feasible option for combatting climate change because it is massively unjust.4 But the idea is that although it is unjust, it is nonetheless better than doing nothing about climate change, and it is as good as we can hope to do, because transfers to rich nations from the poor and vulnerable is the only way of achieving efficient emissions reductions that is a pareto improvement (and thus satisfies the Realist Feasibility Constraint), and it at least makes the poor better off than they would be without efficient reductions. An important problem with their argument is that economic models suggest that it is impossible to have a pareto improvement in this way if the relevant transfers are supposed to involve simply redistributing national outputs in our lifetime, as Posner, Weisbach, and Sunstein seem to assume. For example, William Nordhaus’s multi-­region RICE model implies that under optimal policy with a single global harmonized carbon tax all regions of the world have lower economic output than under the no policy status quo until nearly 2100. This means that even before making the transfers that Posner, Weisbach, and Sunstein have in mind, under optimal policy the current citizens of poor nations (along with

4  These authors do not see this arrangement as involving injustice, but I will set aside that aspect of their argument in what follows. I also set aside for now the important problem of how to realistically ensure universal compliance among nations even given transfer payments, given that many nations would have strategic reason to refuse the offer of a pareto-­improving transfer in order to bargain for a much larger payment—as it would be common knowledge that the success of the entire emissions reductions scheme could in such a way be held hostage by one large nation or at least a small coalition of nations. A similar problem is that unless a self-­enforcing incentive structure is somehow created, it will be common knowledge that a coalition of nations can always scuttle the agreement at a future date by pulling out perhaps in a similar strategic move to bargain for even larger side payments. I return to this kind of problem near the end of the chapter, where I offer some substantive proposals for dealing with it.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Political Realism, Feasibility, and Collective Action  327 every­one else) still have lower GDP than under business as usual.5 If we add large transfer payments from these poor nations on top of that, then they would do even worse, and would certainly not be better off within these models than under the no policy status quo. So, the Posner, Weisbach, and Sunstein view seems based on a basic misunder­ standing of the relevant economic facts, since it seems to assume that there is a possible reshuffling between nations of national economic outputs over our life­ time that in conjunction with optimal emissions reductions yields a pareto improvement for citizens of all nations over the no policy status quo. There is simply no such reshuffling that is even remotely possible in models such as RICE that offer mainstream economic analyses of the sort that these authors otherwise cite with approval, because everyone is a net loser in the coming decades under optimal emissions reductions. Again, the key point is that the most basic tradeoff in the mainstream economic models of climate change is that in all regions ­near-­term output must be sacrificed in our lifetimes to prevent larger sacrifices of output in the further future. Furthermore, these near-­term sacrifices of GDP will be particularly large and painful for developing nations in the near-­term if they are accomplished with a global harmonized carbon tax. So, it is mistake to ­suggest—as many glibly do—that the optimal emissions reductions recom­ mended by economists would amount to nothing more than a small tax on the rich. The reality is that with a harmonized global carbon tax, the tradeoff we face according to mainstream models is to decide how much to sacrifice the wellbeing people in our lifetime—especially how much to sacrifice the desperately poor now who are the poorest people in the whole story—in order to protect more people in the further future.6 If one endorses a global harmonized carbon price—as do Posner, Weisbach, and Sunstein (and Broome, discussed later)—then this serious tradeoff looms.7 However, even with a global harmonized carbon tax there is at least a theoretical way out of the problem that would allow us to protect the current poor while at the same time making optimal emissions reductions. This theoretically possible way out is to make transfers of future economic output from the future to the present (transfers of output across time), rather than merely transfers of eco­ nomic output across nations within our lifetime as considered above. However, Posner, Weisbach, and Sunstein’s view has no compelling rationale if we endorse these kind of intertemporal transfers as realistic. That is because if we assume intertemporal transfers are sufficiently realistic, then all individual nations would be able to secure a pareto improvement over the no policy status

5  Nordhaus (2010); Nordhaus’s RICE 2010 model is available on his website. 6  See Anthoff and Tol (2012); Schelling (1995). 7  For alternatives to a harmonized tax that are superior on both utilitarian and distributive justice grounds, see Chichilnisky and Heal (1994) and Budolfson and Dennig (2020).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

328  Mark Budolfson quo without receiving transfers from other nations by transferring within their own nation from their own future richer citizens to their current comparatively poorer citizens—because even in rich nations the no policy status quo ultimately leads to long-­run GDP losses that are larger in present value terms than the cur­ rent costs of their share of optimal mitigation under a global harmonized carbon tax.8 This means that if intertemporal transfers are realistic, then there is no longer any good argument from the premises of the realist argument above for the dis­tinct­ive conclusion of Posner, Weisbach, and Sunstein that the best feasible global climate regime must involve large transfers to rich nations from poor nations; instead, optimal emissions reductions can then be achieved with no sac­ rifice to current citizens by transfers from future compatriots with no transfers between nations needed. So, the upshot is that Posner, Weisbach, and Sunstein’s version of Efficiency Without Sacrifice seems to face a crippling dilemma: we’ve just seen that their view that poor nations must make transfer payments to rich nations falls apart if we endorse large intertemporal transfers as realistic, because then rich nations can compensate their current citizens in a way that makes everyone better off than under business as usual, and we previously saw that if we don’t endorse large intertemporal transfers as realistic, then their view relies on the idea that optimal global emissions reductions can be conjoined with redistribution of national out­ puts over our lifetime in a way that leads to a pareto improvement over the no policy status quo—which is not possible according to mainstream analyses.

4.  Broome’s Version of Efficiency Without Sacrifice: Intergenerational Transfers The previous section suggests that Posner, Weisbach, and Sunstein’s substantive view is not supported by a good argument even if one endorses the basic premises and logic of the Realist Argument. The argument against their view is that either it depends on the assumption that intratemporal pareto-­improving transfers are possible within our generation—which is not possible according to mainstream models since everyone in our generation is a loser from optimal emissions reduc­ tions—or else if intertemporal transfers are assumed to be feasible, then contrary

8  Russia and Canada come to mind as possible examples of nations for which this may not be true, although some commentators would probably argue that even those nations would ultimately face losses under no policy that overwhelm the cost of compliance with an optimal global tax. If there are indeed a small number of these ‘invulnerable’ nations, we can then adopt the Posner, Weisbach, and Sunstein trick of side payments to bribe them into compliance—but because the number of these invulnerable nations is at least small and arguably nonexistent, there is at least no longer any good argument at all from the premises of the realist argument above for the distinctive conclusion that rich nations must in general be paid off by large transfers from poor nations.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Political Realism, Feasibility, and Collective Action  329 to the distinguishing characteristic of their view it would in fact be better and feasible to achieve a pareto improvement without transfers from poor to rich. In this section I discuss the more promising view of John Broome, who ex­pli­ cit­ly develops his version of Efficiency Without Sacrifice in terms of transfers across time. Broome’s basic idea is that the best feasible outcome can be achieved by intertemporal transfers as described above from the future to the present that ensure that current people are made no worse off than under business as usual even while making optimal emissions reductions. In more detail, Broome’s idea is that the fully welfare optimal climate policy would require something like a cap and trade scheme with permits allocated so as to maximize welfare—and such a welfare optimal distribution of permits would involve distributing them entirely to the world’s poorest people, which would imply a massive transfer of wealth from rich to poor, which is why it would vio­ late the Realist Feasibility Constraint. Further, even if we imagine a less redistri­ butional scheme to achieve the same optimal emissions reductions but with permits allocated in a way that bracketed redistributive aims, e.g. allocated to nations in proportion to economic output, current people would still be worse off than in the no policy status quo (again, as implied by the mainstream models mentioned in the previous section), and thus even this policy would violate the Realist Feasibility Constraint. In light of that, Broome’s thought is that by using intergenerational transfers, we are able to compensate people now for the cost of making the optimal level of emissions reductions by transfers to them from future people (compatriots?) who will still be net beneficiaries due to their even larger benefit from reduced future climate damages, thereby making the desired level of emissions reductions feasible by the lights of the Realist Feasibility Constraint.9 Broome assumes that Efficiency Without Sacrifice is welfare inferior to Efficiency With Sacrifice and also that it is unjust, because he assumes that the former but not the latter would involve rich emitters being bribed to make emis­ sions reductions by others who are not large emitters. Here is Broome: Efficiency without sacrifice has the further, serious demerit that it is unjust. . . . Under efficiency without sacrifice emitters are paid to reduce their emissions by the receivers. Receivers in effect bribe emitters not to harm them. This benefits both emitters and receivers, but only relative to the initial unjust state of busi­ ness as usual. Efficiency without sacrifice perpetuates the injustice.10

9  Note that Broome assumes intergenerational transfers are feasible, which is the assumption that in the previous section we saw is sufficient to undermine the Posner, Sunstein, Weisbach argument that the best feasible outcome consistent with the Realist Feasibility Constraint must involve poor nations compensating rich nations. 10  Broome (2012: 46).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

330  Mark Budolfson However, in tension with Broome’s claim that Efficiency Without Sacrifice is worse and involves injustice, an initial quibble with Broome is that Efficiency Without Sacrifice could actually be welfare and justice superior to Efficiency With Sacrifice—and could be dramatically superior. To see why, note that Efficiency With Sacrifice assumes that transfers (in effect) happen via permit allocation and the resulting trading, and that these transfers are entirely between contemporar­ ies (crucially, are intratemporal). Efficiency Without Sacrifice, in contrast, assumes that transfers are happening from future to the present—and in all main­ stream economic models future people will be significantly richer than their con­ temporaries now. So, if it is really feasible to make those transfers across time in the way that Broome assumes, then the costs of climate change could be shifted to people in the future who are much richer than their counterparts now. There is then at least in principle no reason why the result could not be a large welfare improvement over Efficiency With Sacrifice if the latter must involve transfers between contemporaries, as Broome seems to assume. The good news for Broome is that if his version of Efficiency Without Sacrifice is indeed welfare- and justice superior to Efficiency With Sacrifice, that would only mean that Broome’s view was even more important as a focus for climate policy. Broome briefly considers and rejects the possibility that Efficiency Without Sacrifice could be an improvement over Efficiency With Sacrifice in this way, but his analysis ignores the distributional issues that tell most strongly in favor of the possibility that Efficiency Without Sacrifice is superior. In his analysis, Broome quite rightly highlights the fact that if intergenerational transfer is accomplished with, say, a lower savings rate, this will lead to forgone consumption later, and that this could lead to welfare losses in the future that are larger than the welfare gains to people now who benefit from the transfer; Broome then goes on to claim that cost–benefit analyses show that it is actually the case that the resulting welfare losses would outweigh the gains. But this is simply a matter of invoking cost–benefit analyses that are inadequate for the question at hand, since they ignore the distri­ butional issues that matter—namely, whether transfers could shift the burden of mitigation entirely away from the poorest among the contemporary poor and even other relatively poor people now onto only the richest of the much richer future people who will exist. Cost-­benefit analyses of the sort that Broome invokes completely ignore this distributional issue, and indeed cannot comprehend it since they look at only how an average person does at each time period. A key mistake that causes Broome to ignore the importance of distributional implications is his assumption that emissions reductions only impose costs on the rich.11 If this assumption were true, then his argument would go through that 11  I take leading mainstream analyses to be the results of e.g. DICE, FUND, PAGE, which are the models used in the US social cost of carbon estimates, where PAGE was also used in the Stern Review (Stern 2006), and where DICE and RICE are Nordhaus’s models. See Nordhaus (2015) for an overview.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Political Realism, Feasibility, and Collective Action  331 Efficiency Without Sacrifice is worse, since then there would be no cost of ­mitigation to the very poor. However, this assumption is mistaken. To see where Broome makes this assumption, consider the key passage on these issues in Climate Matters: The difference between Efficiency Without Sacrifice and Efficiency With Sacrifice is the distribution of resources between people. Since emitters are mainly the current rich, whereas receivers are mainly the poor and future gen­ erations, the current rich are better off in Efficiency Without Sacrifice, whereas the poor and future generations are better off in Efficiency With Sacrifice.12

The crucial assumption here relevant to the issue of whether Efficiency Without Sacrifice could be better than Efficiency With Sacrifice is the assumption that the current poor would not pay any important part of the cost of emissions reduc­ tions—this is crucial to the inference that “the [current] poor . . . are better off in Efficiency With Sacrifice”. However, that assumption is simply false according to mainstream economic analyses, where, again, the most important feature of the situation is that both rich and poor in our lifetimes suffer losses relative to business as usual if we make large emissions reductions, because a large carbon price retards the entire economy, and that is bad for the current poor, especially the current poor in developing countries. Again, that is the most important feature of our situ­ ation according to standard models: as noted above, if we are going to rely on a single global price on emissions, then we have to choose the least bad among seriously regrettable tradeoffs between people now and people in the future, keeping in mind that many people now who will be seriously harmed by carbon prices are desperately poor. That is the essence of the problem according to the economists conducting the standard modeling of climate change costs and benefits, and it is also the essence of the problem according to representatives of many of the world’s poor in our actual international climate negotiations. (The fact that representatives of a few nations like Tuvalu have a different view is not an objection to this, and neither is the fact that many philosophers have a different impression.) The upshot is that Broome’s claim is problematic that Efficiency Without Sacrifice is worse than Efficiency With Sacrifice: his argument relies on the assumption that only the rich pay for mitigation, and that is inconsistent with the essence of the climate change problem according to the mainstream economic literature that Broome otherwise quotes with approval. When distributional considerations are taken into account, I conjecture that Efficiency Without Sacrifice in the form that Broome has in mind would actually lead to a welfare- and justice improvement over Efficiency With Sacrifice—but

12  Broome (2012: 45).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

332  Mark Budolfson that is merely a conjecture. Getting to the bottom of this requires figuring out the answer to a very complicated economic question, and one that requires careful economic modeling to answer, and is not something that can be decided on the basis of existing economic modeling that is clearly inadequate for the task, as existing models ignore many of the distributional impacts of both climate dam­ ages and mitigation cost.13 To see the intuition behind this conjecture, suppose that the distribution of mitigation cost is regressive, where mitigation cost is understood to include all near-­term forgone consumption under optimal emissions reductions. If so, then existing cost–benefit analysis models underestimate the welfare loss imposed by emissions reductions paid for by current people (because they implicitly assume that mitigation cost is distributed proportional to consumption), which means that they underestimate the welfare gain that would result from having some future richer people pick up the tab instead as would be the case under Efficiency Without Sacrifice as compared to Efficiency With Sacrifice. Insofar as the inter­ generational transfers that Broome has in mind can also be made so that their incidence is progressive among the future people who pick up the tab, then that would be a further welfare-­improving dimension that is not taken into account in existing cost–benefit analyses. When I exercise my best judgment about the actual distributional facts, including under Broome’s intended intergenerational transfer scheme, I judge that both of the distributional considerations here would add welfare to Efficiency Without Sacrifice over the estimates of standard models— and would add enough welfare to make it a welfare improvement over Efficiency With Sacrifice. I take it that there is no downside for Broome’s view on this particular point, since if what I’ve conjectured is correct, that would mean that the view has even more desirable properties than has previously been acknowledged, since it might allow for a welfare improvement over even the policy that is welfare optimal assuming no transfers. It would also at least mitigate the worry that Efficiency Without Sacrifice amounts to an injustice—and, depending on how the costs and benefits play out, it could entirely remove that worry.14 Taking a step back, part of what is going on here is that by assuming the possi­ bility of intergenerational transfers, Broome has moved outside the box of the standard literature on climate economics, which assumes that there can be no transfers except via the instruments of climate policy (i.e. the standard assump­ tion is that the only transfers allowed are the small-­scale transfers that happen implicitly via carbon prices, and perhaps also—although this is almost never investigated in the standard welfare economics literature—via permit allocation). 13  For work that highlights the importance of the distribution of damages and mitigation cost, see Dennig et al. (2015) and Budolfson et al. (2017). 14  See Broome (2012) on the injustice of efficiency without sacrifice, quoted above.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Political Realism, Feasibility, and Collective Action  333 Once we are outside of that box of the standard literature, we can then consider non-­ climate-­ instrument-­ related transfers from rich to poor that are welfare improving, and if we are able to assume such transfers in large measure, then it is easy to describe conjunctions of those transfers with a climate policy achieving optimal emissions reductions that amount to a vast welfare improvement over any climate policy considered in isolation, including the climate policy that is welfare optimal assuming that no such transfers are possible. And the climate policy that is welfare optimal assuming that no such transfers are possible is sim­ ply Efficiency With Sacrifice, as described above (if we follow, as Broome does, other standard modeling conventions in describing that view, such as a har­mon­ ized global carbon price). So, if we allow large transfers outside the box, we can achieve a large welfare improvement over Efficiency With Sacrifice; and since Broome’s Efficiency Without Sacrifice assumes that some large transfers are feas­ ible, it is then easy to see how it could be a welfare improvement over Efficiency With Sacrifice.15

5.  Is Broome’s Proposal Feasible? How Realistic Is It? All of the preceding emphasizes why it is very important whether the intergenera­ tional transfers that Broome has in mind are genuinely feasible (in addition to being theoretically possible). If they are not feasible, then the entire discussion is irrelevant to what we should actually do assuming the Realist Feasibility Constraint. I leave it to others to analyze the feasibility of intergenerational transfers. To be honest, I don’t really understand in any detail how they are supposed to work in practice. But I look forward to seeing the results, and I am hopeful they could work—it is an important and exciting idea, including because it could improve both the welfare and justice properties of a realistic climate policy, as described in the previous section. In the meantime, in this section I’ll focus on a number of further problems that seem to emerge, especially from a realist perspective. Most importantly, even if some forms of intergenerational transfers are feas­ ible, it is important to wonder what the distributional implications are of the (potentially idiosyncratic) subset of transfers that are feasible in addition to merely theoretically possible. The big worry here is that insofar as intergenera­ tional transfers are feasible, it may be that only regressive transfers are feasible, with the consequence that the non-­climate welfare cost of making those transfers could be larger than the climate-­related welfare gain from making them. For 15  Note that in the last section of this chapter, I argue that if we assume that these intergenerational transfers are possible, and if we believe that a political strategy I describe in a later section is feasible, then an even better and more fair response to climate change is feasible.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

334  Mark Budolfson example, if the only feasible way to make large intergenerational transfers is via a political bargain that somehow bakes in funding cuts to schools, pensions, and investment for the future poor and vulnerable, then this would be a major problem. If one adopts a realist stance, then presumably one should take this worry par­ ticularly seriously, as it seems all-­too realistic to imagine a public policy initiative that aspires to be a progressive version of Efficiency Without Sacrifice being hijacked and devolving into a welfare-­destroying compromise deal that achieves a carbon tax at the price of cuts in effective entitlement programs. Or more darkly, one can imagine that the bargain that emerges is a fraudulent carbon pricing scheme that only serves to enrich sophisticated investment banks and thus accomplishes no good, but succeeds in harming the economy for non-­elites, and is purchased at the price of very large sacrifices of entitlement programs—and also kills millions of poor people in the coming decades through food price spikes and other more indirect ways of killing poor people, and so on. The positive ‘flip side’ of this worry is that insofar as progressive intergenera­ tional transfers are indeed possible, then there is an opportunity to promote wel­ fare by analyzing them, and then choosing the form of intergenerational transfers that is welfare optimal. In light of these important questions, what could happen next, then, is some­ thing similar to what is happening in the climate economics modeling commu­ nity of estimating the effects of combining a carbon tax with (intratemporal) ‘revenue recycling’ that uses the revenue from a carbon tax in a socially useful way, such as by refunding it in a progressive way or uses it to fund tax reform of distortionary taxes. The point of this modeling exercise is, first, to evaluate how close we can get to a ‘double dividend’ whereby a carbon tax combined with other structural reforms immediately also yields a net gain for aggregate economic out­ put, and, second, to evaluate what the net distributional effect of such policy com­ bin­ations would be for different socio-­economic groups (e.g. income deciles). The results suggest that we probably cannot get a double dividend, but we can reduce the cost to the economy with such policy combinations (vs. a carbon tax only). Perhaps more importantly for considerations of both wellbeing and feasibility, the results indicate that some of these policy combinations such as equal per capita rebate of the revenues from a carbon tax immediately make something like three quarters of citizens net beneficiaries of the combined policy, including those in all the lowest income deciles.16 One question for Broome is why he doesn’t also add these intratemporal trans­ fers to the overall package of measures he recommends. The view of many of those who are doing and promoting this modeling in policy circles is that these 16  For an overview of this research, see Metcalf (2018); see also Barron et al. (2018); Sterner (2012); Mathur and Morris (2014); Goulder et al. (2019); Beck et al. (2015).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Political Realism, Feasibility, and Collective Action  335 intratemporal transfer measures are indeed a crucial part of the best way forward. In any event, a similarly rigorous investigation of intergenerational transfers is needed to make Broome’s proposal most policy relevant, and to evaluate whether there ultimately really is any tradeoff between feasibility, on the one hand, and welfare and justice on the other. The preceding indicates that it is crucial to Broome’s argument that the inter­ generational transfers he has in mind are both feasible and welfare-­improving over business as usual, even when their effects outside the box of the climate problem are taken properly into account. But at this point, putting on a more hardheaded realist hat, we could wonder, darkly, whether in practice those transfers would even be welfare improving over business as usual. Putting on a hardheaded realist hat, notwithstanding everything above, one might not see why we should expect any complex climate interventions to be welfare improving in practice, even if we agree that their effects would be wonderful in theory were we to assume, completely unrealistically, that there would be perfect implementation by perfectly benevolent agents. Many hardheaded realists are apt to have a dark view according to which any complex grand scheme by welfare economists is certain to be hijacked by special interests. So, from that perspective, even if the proposal would result in a large improvement in theory, the actual expected effect of this kind of grand-­level, complex policy might be to make things worse than the current trajectory of increasingly enlightened self-­interested action by nations (which is better than a no policy business as usual). Given the pervasive phenomenon of government failure (on a dark view, this is typically the other horn of the regulation dilemma to market failure), it is difficult to understand why this sort of problem is not discussed in this literature. Presumably, any climate policy discussion that ­purports to be discussing realist climate policy must explicitly engage with these darker worries. With perhaps something like these darker considerations in mind, some econo­mists see virtue in a simple transparent policy: perhaps the best current contender is a carbon tax with all revenues rebated on an equal per capita basis, of the sort that is currently the focus of many modeling exercises noted above. We might say that this is an example of a maximally realistic policy portfolio that includes intra temporal transfers and aims at optimal policy: it allows us to com­ bine an optimal carbon price with ‘outside the standard box’ transfers in a way that may be both feasible and not-­easily-­subject-­to-­capture, and along both of those dimensions appears to do about as well as we can hope if we are going to add intratemporal transfers to climate policy. A key practical question for a view like Broome’s is whether there is a similarly realistic inter generational transfer policy to add to this—i.e. something as simple and as transparent, that could avoid being perverted by special interests while accomplishing the intended intertemporal transfer. If (and perhaps only if) there

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

336  Mark Budolfson is, then realists should enthusiastically endorse it as central to the maximally ­realistic policy portfolio that aims at an optimal policy response to climate change.

6.  Why the Realist Argument for Efficiency Without Sacrifice Is Invalid In the preceding discussion, I accepted the Realist Argument on its own terms, and directly evaluated the substance of the conclusions that it is taken to support. But now it is time to take a step back and note that there is a fundamental prob­ lem with Realist Argument: it is invalid, and the nature of its invalidity points the way towards what seems to be a more ethical response than Efficiency Without Sacrifice that is consistent with both the Realist Feasibility Constraint and the premise about our Circumstances. Here again is the Realist Argument: Realist Feasibility Constraint: Nations act only in the interests of their current citizens, so a response to climate change is infeasible if it requires a nation to act contrary to the interests of its current citizens. Circumstances: An optimal response to climate change requires substantial emissions reductions from the no policy status quo from many rich nations, and the current citizens of those rich nations would be made much worse off relative to business as usual by such reductions unless they were compensated for mak­ ing them. Therefore, given the Realist Feasibility Constraint and our Circumstances, the best feasible response is Efficiency Without Sacrifice, which involves re­du­cing emissions to a level that is optimal, while compensating current people, includ­ ing citizens of rich nations, for the cost of making such reductions, thereby ensuring that no one including the rich has to make any sacrifices, and that the costs are instead borne by those who are vulnerable to climate damages, because only in this way can substantial emissions reductions become feasible.

The argument is invalid because it overlooks the fact that what it is in the interest of a nation to do can change depending on the actions of other nations—in par­ ticular, even if nations are as self-­interested as the Realist Feasibility Constraint assumes, other nations can change what it is in their interest to do via threats, sanctions, and other measures, and can therefore make it in their interest to com­ ply with international standards that it would not have been in their interest to comply with in the absence of those measures. As a simple illustration of how this creates trouble for the argument, imagine an international climate treaty that requires signatories to impose globally optimal emissions taxes within their territory and empowers compliant nations

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Political Realism, Feasibility, and Collective Action  337 to impose tariffs and other penalties on non-­compliant nations as retaliation for non-­compliance. Given the existence of such a treaty, even if high-­emitting nations would be made worse off than business as usual by compliance, they might be made even worse off by non-­compliance if there are many signatories to the treaty the collective sanctions of which make the penalties for non-­compliance larger than the costs of compliance. If that is the case, then the Realist Feasibility Constraint entails that high-­emitting nations would comply and reduce emissions even though doing so would make them worse off relative to business as usual.17 Because this is consistent with the truth of Circumstances, this shows, first, that the conclusion that Efficiency Without Sacrifice is the best feasible option does not obviously follow from the Realist Feasibility Constraint and Circumstances, and, second, that that conclusion is false if there is any response analogous to the sim­ plistic example just described that is both feasible and ethically superior to Efficiency Without Sacrifice. Of course, in the real world the simplistic response just described is arguably not feasible. That is because if sanctions are threatened in the simplistic way just described, it is common knowledge that they are likely to be met with a re­tali­ atory trade sanctions and/or non-­trade measures that would make the costs of following through on them unacceptably high, thereby preventing the threat of such sanctions from being credible in the first place, thereby preventing such a response from getting off the ground. This is arguably the problem with simple proposals involving trade sanctions from commentators such as Joseph Stiglitz and William Nordhaus.

7.  Feasibility Wedges and a Meta-­Architecture for Global Agreement Nonetheless, despite the problems with the simple proposals involving trade sanctions, there are a number of feasibility enhancing measures that enable the creation of a feasible analogous regime that could realistically succeed in chan­ ging what it is in the interest of nations to do over time, ultimately making it in the interest of rich nations to reduce emissions even beyond the point at which they are made worse off than the business as usual status quo. Although none of these ‘feasibility wedges’ are individually sufficient to guarantee the success of such an ethically superior treaty, in conjunction they may make the prospect

17  Such a structure is implicit in the proposal of Stiglitz (2008), and Nordhaus (2015). Nordhaus provides an empirically informed model of how tariff sanctions, if they could be credibly threatened and if they would have no further effects outside the modeled interactions, could change what carbon price it is in the interest of nations to impose.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

338  Mark Budolfson of  success sufficiently high to make such a response a better bet for humanity.18 In what follows, I sketch some examples of such a response that in­corp­or­ates a number of feasibility enhancing features both in its meta-­architecture and strategy. The goal is to show how these measures can make progressive intergenerational transfers more feasible, which is important for the best and most fair response to climate change we can realistically hope for. So, imagine a well-­chosen climate treaty architecture that illustrates a feasi­ bility wedge at the level of its basic structure by combining three individually familiar components: first, a cap and trade scheme among signatories to the treaty, where the cap decreases each year along a path that is insensitive to the number of signatories to the treaty; second, an undemanding initial cap that ensures that when the treaty initially enters into force, no nation has to make costly emissions reductions in the short run; third, a right granted to compliant signatory nations to impose a duty on imports from non-­compliant nations.19 Ideally, the magnitude of this duty would be large, as could be feasible if one im­agines the WTO appellate body faced with the actual choice of deciding whether a reasonable regime of this kind with a large duty was permissible under, say, Article XX of the GATT (which can be read as a catch-­all exemption for any tariff that is inherently reasonable given the values that a nation endorses if it doesn’t otherwise discriminate between particular nations for any of the familiar more technical reasons). In that scenario, the appellate body would be under enormous pressure to decide that it was indeed permissible. This would be for many reasons, including that deciding otherwise could threaten the long­term survival of the WTO in its current form and the careers of the bureaucrats on the appellate body.20 If well-­designed, an initial set of nations would initially join such a treaty for a variety of reasons: in many cases, because the treaty would not require emissions reductions early on, but would only require reductions gradually later as the cap decreased and the largest emitters joined the treaty, at which point the cost to early signatories would in some cases tend to be offset by payments from

18  My notion of feasibility wedges is inspired to some extent by the notion of stabilization wedges in Pacala and Socolow (2004). 19  Because my goal in this chapter is merely to identify some feasibility-­enhancing mechanisms by which various emissions reductions architectures could realistically be implemented in as ethical a way as is practically possible given the constraint of feasibility, I wish to remain agnostic on the ques­ tion of what specific metrics, cost–benefit analysis, and architecture ought to be used. Depending on those factors, such a treaty would also include an allocation scheme for emissions permits, any num­ ber of which would be consistent with the suggestions in this chapter. Such a treaty might also include a mechanism to redistribute to developing nations some of the duties collected, in order to secure the universal compliance of developing nations and to offset the more ethically significant costs of the treaty to developing nations. 20  The literature on the practicalities of WTO permissibility seems to completely ignore this sort of ‘extra-­legal’ consideration, which from a realist perspective might be the most important con­sid­er­ ation of all in making a judgment of feasibility.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Political Realism, Feasibility, and Collective Action  339 later-­joining nations in exchange for emissions permits, and in other cases those later costs would not be relevant to the short-­run political calculations of ­politicians in democratic nations making the initial decision whether to join; in other cases, because it might be to the advantage of some initial signatories to have the power to impose unreciprocated duties on some imports from initial non-­signatories; and in other cases, for the sort of complex reasons that have led many regions such as the EU, Australia, British Columbia, California, and others to in­de­pen­dent­ly enact such policies. As a result, the initial set of signatories might be comparable to the set of nations that ratified the Kyoto Protocol. From this starting point, the idea is that the sanctions imposed by that initial critical mass of nations would set in motion a chain reaction that would make it in the interest of an increasing number of nations to join over time. Such a chain reaction would be driven by the fact that for each nation that joined the treaty, the cost would increase for each remaining nation that had not joined, because each additional cooperating nation means an additional nation imposing duties on imports from non-­cooperating nations. In addition, as more nations joined the treaty and the cap decreased over time, the cap and trade scheme would become increasingly effective, thereby increasing the emissions differential between signatories and non-­signatories, thereby increas­ ing the magnitude of each individual duty imposed. The idea is that over several decades, more and more nations would gradually join, until eventually the costs of not joining would be so high that even the most recalcitrant nations would ultimately find it in their interest to join rather than continue to hold out. Of even greater importance, once nations joined such a treaty, it would never be in their interest to pull out, at least if the treaty is successfully designed with the self-­enforcing structure imagined above, because given that structure the costs of pulling out always outweigh the costs of staying in after the point at which it is initially in a particular nation’s interest to join. This could help to solve the most serious problem for climate treaties, which is the problem of securing not merely initial ratification of the treaty, but long-­run compliance—a problem that is not­ably not fully solved by the Posner, Weisbach, and Sunstein view, given the stra­tegic incentives that would always exist under that proposal to defect in order to negotiate a better side payment.21 Of course, the realistic worry remains that strategic retaliatory measures by hostile powerful nations could scuttle even this sort of more sophisticated pro­ posal. With that in mind, such a treaty should be introduced under only the most 21  For real-­world examples and game-­theoretic analysis of international environmental treaties that have a structure analogous to the cascade to universal self-­enforcement structure described here, see Barrett (2003: ch. 9). A treaty is self-­enforcing in the relevant sense relative to a set of nations if and only if it is both individually and collectively rational to maintain agreement to the treaty from the point of view of all of those nations.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

340  Mark Budolfson favorable conditions that can be realistically expected in order to further raise its probability of success. Toward that end, additional feasibility wedges can be iden­ tified that represent the most favorable conditions that can realistically be expected, which represent an especially important place for contributions from (among other) climate negotiators and actual practitioners of international nego­ tiation and political strategy. As a few examples of additional feasibility wedges to enhance the probability of getting such a ‘meta-­architecture’ in place, consider: (1) The treaty could be introduced early in the term of a US President who sup­ ports the treaty, enabling him or her to sign the treaty and publicly endorse its permissibility under all international laws and treaties, and to use his or her power and influence to ensure that international court decisions establish as precedent that the treaty is permissible. This could help institutionalize the per­ missibility of the treaty in a way that cannot be reversed by anything short of dramatic (and hence unlikely) power politics. The US President could do all of this unilaterally even if the US Senate is initially disposed to reject the treaty unanimously, and even if powerful nations such as China oppose the treaty. (2) The treaty could be designed so that upon introduction, it is quickly ratified by a substantial proportion of developed nations, as well as a substantial propor­ tion of developing nations, thus enhancing its perceived legitimacy and, more generally, further institutionalizing it. (3) The treaty could be only one part of a policy portfolio that also includes all emissions reduction measures that are in the interest of individual nations to impose unilaterally, such as technical regulations to realize negative net cost emissions reductions, health co-­benefits, subsidies for research and develop­ ment, and perhaps even geoengineering. This will reduce the magnitude of the costs that must be imposed by the treaty in order for the overall policy portfolio to be effective, thereby reducing incentive for hostile nations to invest in stra­ tegic retaliatory measures to scuttle the treaty. (4) The cascade to self-­enforcing structure described here is also consistent with other complementary incentive schemes to encourage compliance. For example, one promising addition would be for duties from all nations to be held by a sin­ gle global administrator until the end of each year, at which point each nation’s proceeds would be disbursed only if that nation complied with the treaty’s pro­ visions in the previous year; duties could then be subtracted from the accounts of non-­compliant signatory nations based on their degree of non-­compliance, with the proceeds distributed to compliant signatories. Other ideas might include a bonus for initial signatories, and many others. (5) The treaty could be designed so that upon introduction, it tends to require only emissions reductions that nations would find feasible even without the treaty. For example, consider the ‘tax and dividend’ approach described several

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Political Realism, Feasibility, and Collective Action  341 sections above that conjoins a carbon tax with an intragenerational transfer equal to equal per capita refund of all of the proceeds of the carbon tax. Perhaps such a policy is not generally feasible if it involves the carbon tax level that would be globally optimal; nonetheless, for some n% of the optimal tax level, it could be feasible. The treaty could be designed for that upon introduction, it requires on average only a carbon price equal to n% of the globally optimal level, for whatever the largest n is such that n tends to be feasible even without such a treaty. Then, sympathetic nations (or national leaders) could feasibly enact the n% tax and dividend policy, and then sell compliance with the treaty as in the self-­interest of their nation, on the grounds that joining the treaty would be an improvement over a situation in which they are making the same emissions reductions but are not members of the treaty (which would be true only in the short run, but perhaps their audience tends to perceive only with the short run, in which case joining the treaty would be in their perceived self-­interest even if not in their long-­run self-­interest).

The preceding are intended merely as initial examples of feasibility wedges that can be added to the meta-­architecture for a global agreement described here in order to make an ethical collective action more feasible. For example, a treaty with the basic structure described above makes it easier to satisfy conditions (1) and (2), because the structure of such a treaty makes it likely that it would actually be effective, which means that nations will recognize that the costs of the treaty are likely to be non-­futile, including any political costs to leaders who support the treaty—which are already mitigated by the fact that the treaty shifts the costs of emissions reductions into the future, beyond the short-­term time horizon of most political leaders. This increases the likelihood that (1) will be satisfied relative to any particular US President, and increases the likelihood of (2) being satisfied for the various reasons described above. In this context, it is worth noting that even the transparently flawed and ineffective Kyoto Protocol was ratified by an impressive set of nations and signed by US President Bill Clinton. As a result, it is realistic to think that a US President might well be willing to expend the political capital necessary to satisfy condition (1) relative to a treaty far superior to Kyoto, such as the sort of treaty aimed at here involving a cascade to universal self-­enforcement that would also predictably involve satisfaction of (2).22

22  For example, if (1) is satisfied by a supportive US President and (2) is satisfied by an initial coali­ tion that includes roughly the nations that ratified Kyoto, then it is arguably unrealistic to think that other nations hostile to the treaty would have the power and influence to convince international courts to rule against both the US administration and that coalition regarding the permissibility of the treaty under WTO rules, which together with the resulting institutionalized legitimacy of the treaty would substantially reduce the probability that subsequent retaliatory measures could scuttle the treaty.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

342  Mark Budolfson By recognizing this emerging strategic dynamic and leveraging it as fully as possible, including by using the tools of feasibility wedges and intra- and inter­ temporal transfers described above, we might hope to move things in the right direction as quickly as possible even within the constraints of realism. With the preceding outline in hand, a climate treaty that exploits the feasibility wedges above and the particular cascade to universal self-­enforcement design structure outlined above, together with additional tools such as intra- and inter­ generational transfers is a promising route to an ethically superior climate treaty that is feasible even in light of the Realist Feasibility Constraint and Circumstances. The challenge of crafting such a treaty is then a more well-­defined engineering problem that can be addressed by experts on the relevant legal issues, economic mechanisms, dispositions of nations, and political strategy. The primary constraint on such a treaty is that it must predictably lead to universal ratification and com­ pliance in the long run via the cascade of economic incentives described above. Subject to that constraint, the treaty should be designed to distribute the costs of emissions reductions as fairly as possible, which will include much more fair dis­ tributions than would be available if there were a genuine feasibility constraint that current citizens of rich nations can be made no worse off than under the no policy status quo. If this engineering project succeeds, it will result in a treaty-­based response that is a far more ethical bet for humanity than the response intended under Efficiency Without Sacrifice and other contemporary ‘realist’ proposals. Of course, even when all of the pieces of the puzzle above are combined, there is no way of establishing that this kind of more ethical response would ultimately succeed with anything near 100% certainty—but as noted above, there are analo­ gous and perhaps more serious doubts about the success of other realist proposals including even the best versions of Efficiency Without Sacrifice. But in general, it would be surprising if the best feasible way forward was obviously feasible. Instead, we should expect the best feasible way forward to appear infeasible to many, as did the best feasible way forward at the height of US–Soviet tensions, and at many other points in world history after which a better solution than seemed feasible at the outset emerged from a combination of skilled diplomacy, good fortune, and bold leadership—and hopefully a desire for the most ethical feasible solution to this challenge.

References Aldy, J., W. Pizer, M. Tavoni, L. A, Reis, K. Akimoto et al. (2016). “Economic Tools to Promote Transparency and Comparability in the Paris Agreement.” Nature Climate Change 6: 1000–1004. Aldy, J. and R.  Stavins (eds.) (2007). Architectures for Agreement: Addressing Global Climate Change in the Post-Kyoto World. Cambridge: Cambridge University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Political Realism, Feasibility, and Collective Action  343 Anthoff, D. (2011). “Optimal Global Dynamic Carbon Abatement.” Mimeo. Anthoff, D. and R. S. Tol (2012) “Schelling’s Conjecture on Climate and Development: A Test,” in Climate Change and Common Sense: Essays in Honour of Tom Schelling, ed. R. Hahn and A. Ulph. Oxford: Oxford University Press, 260–274. Barrett, S. (2003). Environment and Statecraft: The Strategy of Environmental TreatyMaking. Oxford: Oxford University Press. Barrett, S. (2010). Why Cooperate? The Incentive to Supply Global Public Goods. Oxford: Oxford University Press. Barron, A. R., A. W. Fawcett, M. A. C. Hafstead, J. R. McFarland, and A. C. Morris (2018). “Policy Insights from the EMF 32 Study on U.S.  Carbon Tax Scenarios.” Climate Change Economics 9: 1–47. Beck, M., N.  Rivers, R.  Wigle, and H.  Yonezawa (2015). “Carbon Tax and Revenue Recycling: Impacts on Households in British Columbia.” Resource and Energy Economics 41: 40–69. Böhringer, C., E. J. Balistreri, and T. F. Rutherford (eds.) (2012). “The Role of Border Carbon Adjustment in Unilateral Climate Policy: Results from EMF 29.” Energy Economics 34, Supplement 2 (December): S95–S250. Brennan, G. (2009). “Climate Change: A Rational Choice Politics View.” Australian Journal of Agricultural and Resource Economics 53: 309–326. Broome. J. (2012). Climate Matters: Ethics in a Warming World. New York: W. W. Norton. Broome. J. (2013). “The Public and Private Morality of Climate Change.” The Tanner Lectures on Human Values 32: 3–20. Budolfson, M.  B. (2012). “Collective Action, Climate Change, and the Ethical Significance of Futility.” PhD dissertation, Princeton University. Budolfson, M. B. (2012). “Should the US and other Nations Make Unilateral Emissions Reductions? How to be a Leader Without being a Sucker.” Budolfson, M. and F. Dennig (2020). “Optimal Global Climate Policy and Regional Carbon Prices,” in Handbook on the Economics of Climate Change, ed. G.  Chichilnisky and A.  Rezai. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing, 224–238. Budolfson, M., F. Dennig, M. Fleurbaey, A. Siebert, and R. H. Socolow (2017). “The Comparative Importance for Optimal Climate Policy of Discounting, Inequalities, and Catastrophes.” Climatic Change 145(3–4): 481–494. Chichilnisky, G. and G.  Heal (1994). “Who Should Abate Carbon Emissions?” Economics Letters 44: 443–449. Dennig, F., M.  B.  Budolfson, M.  Fleurbaey, A.  Siebert, and R.  H.  Socolow (2015). “Inequality, Climate Impacts on the Future Poor, and Carbon Prices.” Proceedings of the National Academy of Sciences 112: 15827–15832. Foley, D. (2009). “The Economic Fundamentals of Global Warming,” in Twenty-First Century Macroeconomics: Responding to the Climate Challenge, ed. J. M. Harris and

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

344  Mark Budolfson N.  R.  Goodwin. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing, 115–126. Goulder, L.  H., M.  A.  C.  Hafstead, G.  R.  Kim, and X.  Long (2019). “Impacts of a Carbon Tax across US Household Income Groups: What Are the Equity-Efficiency Trade-Offs?” Journal of Public Economics 175: 44–64. Gromet, D., H. Kunreuther, and R. Larrick (2013). “Political Ideology Affects EnergyEfficiency Attitudes and Choices.” Proceedings of the National Academy of Sciences 110: 9314–9319. Kartha, S., P.  Baer, T.  Athanasiou, and E.  Kemp-Benedict (2010). “The Right to Development in a Climate Constrained World: The Greenhouse Development Rights Framework,” in Der Klimawandel, ed. M.  Voss. Wiesbaden: Verlag für Sozialwissenschaften, 205–226. Keohane, R. O. and D. G. Victor (2011). “The Regime Complex for Climate Change.” Perspectives on Politics 9: 7–23. Keohane, R. O. and D. G. Victor (2016). “Cooperation and Discord in Global Climate Policy.” Nature Climate Change 6: 570–575. Mathur, A. and A.  C.  Morris (2014). “Distributional Effects of a Carbon Tax in Broader US Fiscal Reform.” Energy Policy 66: 326–334. Metcalf, G. E. (2018). Paying for Pollution: Why a Carbon Tax is Good for America. Oxford: Oxford University Press. Nordhaus, W.  D. (2010). “Economic Aspects of Global Warming in a PostCopenhagen Environment.” Proceedings of the National Academy of Sciences 107: 11721–11726. Nordhaus, W. D. (2013). The Climate Casino: Risk, Uncertainty, and Economics for a Warming World. New Haven, CT: Yale University Press. Nordhaus, W.  D. (2015). “Climate Clubs: Overcoming Free-Riding in International Climate Policy.” American Economic Review 105: 1339–1370. Nordhaus, W. D. and J. Boyer (2000). Warming the World: Economic Models of Global Warming. Cambridge, MA: MIT Press. Pacala, S. and R. Socolow (2004). “Stabilization Wedges: Solving the Climate Problem for the Next 50 Years with Current Technologies.” Science 305: 968–972. Pauwelyn, J. (2007). “U.S. Federal Climate Policy and Competitiveness Concerns: The Limits and Options of International Trade Law.” Working Paper, Nicholas Institute for Environmental Policy Solutions, Duke University. Posner, E. and D. Weisbach (2010). Climate Change Justice. Princeton, NJ: Princeton University Press. Posner, E. and C. Sunstein (2008). “Climate Change Justice.” Georgetown Law Journal 96: 1565–1612. Rendall, M. (2015). “Carbon Leakage and the Argument from No Difference.” Environmental Values 24: 535–552.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Political Realism, Feasibility, and Collective Action  345 Sabel, C. F. and D. G. Victor (2017). “Governing Global Problems Under Uncertainty: Making Bottom-Up Climate Policy Work.” Climatic Change 144: 15–27. Schelling, T. (1960). The Strategy of Conflict. Cambridge, MA: Harvard University Press. Schelling, T. C. (1995). “Intergenerational Discounting.” Energy Policy 23: 395–401. Steele, K. (2021). “Pareto Improvements and Feasible Climate Solutions.” in this volume. Stern, N. (2006). The Economics of Climate Change: The Stern Review. Cambridge: Cambridge University Press. Sterner, T. (2012). Fuel Taxes and the Poor: The Distributional Effects of Gasoline Taxation and Their Implications for Climate Policy. Abingdon: Routledge. Stewart, R. and J.  Wiener (2003). Reconstructing Climate Policy: Beyond Kyoto. Washington, DC: AEI Press. Stiglitz, J. (2008). “A New Agenda for Global Warming” in The Economists’ Voice: Top Ecomomists Take on Today’s Problems, ed. J. E. Stiglitz, A. S. Edlin, and J. B. DeLong. New York: Columbia University Press, 22–27. Victor, D. (2017). “Three Dimensional Climate Clubs: Implications for Climate Cooperation and the G20.” Geneva: International Center for Trade and Sustainable Development. Wagner, G. and M. Weitzman (2016). Climate Shock: The Economic Consequences of a Hotter Planet. Princeton, NJ: Princeton University Press. Wiener, J. (2007). “Incentives and Meta-Architecture”, in Architectures for Agreement: Addressing Global Climate Change in the Post-Kyoto World, ed. J.  Aldy and R. Stavins. Cambridge: Cambridge University Press, 67–80.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

15

Pareto Improvements and Feasible Climate Solutions Katie Steele

1. Introduction In view of the considerable suffering that climate change threatens in the medium- and long-­term future, the global response to this problem has been woefully inadequate. This very concerning state of affairs has led many to reflect on the nature of the problem and the prospects for solving it. Accordingly, the notion of political feasibility has become prominent in debate about who can be counted on to bear the costs of climate change—the costs of mitigation and adaptation. While it is all very well arguing over what are better and worse distributions of these costs, real progress can only be achieved through action, which is constrained by what is feasible. Or so the rhetoric might go. Indeed, it is hard to deny that, ultimately, it is important that we actually advance on the status quo— that climate-­related suffering is reduced by some amount, the more the better, even if this falls far short of a morally optimal state of affairs. A specific principle or maxim has been pitched as the answer to the call for  ‘feasible’ climate solutions—International Paretianism (IP)—which basically re­com­mends framing international negotiations on problems like climate change in terms of mutual ‘national interest’. The idea is that, as far as feasibility goes, any proposal for determining responsibilities for tackling climate change must advance the interests of all states, or at least not frustrate their interests. That is, a genuine climate solution must provide a Pareto Improvement with respect to the interests of states: No state is made ‘worse-­off ’, and at least one state is ‘better off ’ than they fare under the status quo. One might question whether states really are the key international players in climate negotiations, but let us assume that this feature of IP is plausible. It is also ambiguous as to what counts as a state’s ‘self or national interest’, but let us assume it is an appropriate aggregate of the welfare of present and future citizens (more about this later). As such, the costs of a mitigation strategy for present citizens may be counterbalanced by the benefits for future citizens, such that there is a net benefit in terms of national interest.

Katie Steele, Pareto Improvements and Feasible Climate Solutions In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Katie Steele. DOI: 10.1093/oso/9780198796282.003.0016

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  347 Advocates argue that IP is uniquely placed with respect to political feasibility in the international setting (see, especially, Eric Posner and David Weisbach’s 2010 book Climate Change Justice). They claim that the only feasible proposals for significantly reducing the harms of climate change are IP proposals, and thus treaties intended to actually reduce suffering associated with climate change ought to satisfy IP criteria. While IP is not a radically new approach to inter­ nation­al agreements, it is nonetheless provocative in climate justice circles.1 Discussions of climate justice have traditionally focused rather on historical responsibility and duties of reparation, or simply duties of assistance, that are owed by rich, high-­emitting states.2 IP, on the other hand, redirects attention to what is good for rich, high-­emitting states (as well as what is good for other states). Thus IP, as an approach to climate change, may seem rather shocking. In spite of this, there is certainly something compelling, hopeful even, about IP, once we focus on political feasibility. IP is not about states recognizing their duties and stepping up to make whatever demanding reparations for climate change are morally required of them. It is rather about promoting a climate solution that the relevant actors would have no reason to resist and may have material interest in pursuing. We would be fools to let an overly aspirational treaty prevent easy agreement on a morally inferior, but nonetheless mutually beneficial, one. Or so the thinking might go. Political feasibility is certainly an important consideration in the debate about climate change. But once we look more closely at the details of what feasibility stands for, its relationship with furthering self-­interest, let alone the national interest of a state, is not as persuasive as first appears. In order to arrive at a substantial global response to climate change, IP must appeal to a notion of national interest that is rather revisionary given the existing apparent motivations of state actors. Furthermore, closer inspection of the form of the cooperative dilemma associated with climate change reveals that the pursuit of national interest alone may turn out to be self-­defeating. This is more worrisome for the feasibility argument for IP than its advocates suggest. Or so this chapter will argue. More generally, the aim here is to shed light on International Paretianism, the climate change pre­dica­ment, and the meaning and role of ‘feasibility’ in ethical-­political debate.

1  Note, however, that some philosophers have also claimed that it may be productive, at least in the short term, to focus on climate solutions that amount to Pareto improvements on the status quo; not­ ably Broome (2010); see also Schokkaert and Eyckmans (1998). The focus here is on the work of Posner and Weisbach because they make stronger, more definitive claims about feasibility in the inter­ nation­al setting. Moreover, Broome takes individual persons, rather than nation states, to be the basic agents that are subject to climate-­related costs and benefits, and so does not defend International Paretianism per se. 2  Typically, the focus is on the costs of mitigation, assuming that an appropriate global mitigation target has been decided. The question is how to divide up the costs of meeting this target; a variety of moral perspectives lead to the conclusion that the rich high-­emitting states should bear the bulk of the costs.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

348  Katie Steele Section 2 begins with the latter conceptual issue. Section 3 turns to the climate predicament itself—the kind of Pareto improvement that may be in the offing, and explores whether the achievement of such a Pareto improvement does in fact depend solely on ‘self-­interest’. Section 4 reflects on the implications of these considerations for the overall feasibility claims made by IP proponents.

2.  The Concept of ‘Political Feasibility’ As suggested above, Posner and Weisbach (2010) largely rest their case for an ­IP-­style climate treaty on the fact that only this style of climate treaty is ‘politically feasible’.3 But what is the precise meaning and significance of that term? As suggested above, the reason to focus on feasibility at all stems from an interest in actual social change, in ‘getting things done’. So ‘feasibility’ is a property that is invoked in the interests of actual change. Beyond that, however, the proper usage of the term is not obvious. And yet the plausibility of the IP proponents’ claim— the tight relationship between feasibility and Pareto improvements—depends on these details. An initial concern is that the concept is inherently vague and so the IP proponents’ claim will resist any definitive analysis. The very target of the concept— what supposedly has the property of ‘feasibility’—seems to vary in general talk and debate, let alone the rules for determining whether/to what extent the property is present. We talk about whether a goal, say, of swimming 50m in less than half a minute, is feasible; about whether change, say, towards a less racist society, is feasible; about whether a program, say, a new teaching curriculum, will feasibly improve students’ learning outcomes; about whether it is feasible that I will not procrastinate while doing some task. Some of these examples seem to be about whether actions will realistically be performed (e.g., work without pro­cras­tin­ ation), and others about whether goals will realistically be achieved (e.g., the fast swim or a less racist society) perhaps in addition to or else conditional on some action being performed (e.g., the improvement of students’ learning outcomes given reforms to the teaching curriculum).

3  A qualification must be made here: Posner and Weisbach (2010) also argue that a climate treaty should be concerned only with advancing mutual self-­interest because anything more ambitious in terms of rich states paying would amount to trying to solve other problems of global distributive just­ ice in a climate treaty, which would not be an efficient way to solve these other important problems. So they argue for IP on efficiency grounds (in the broader scheme of distributive justice) as well as feasibility grounds. It is not at all obvious that the optimal IP climate solution would in fact yield the amount of climate change mitigation that a benevolent global planner would choose, which is an aspect of the efficiency argument. Later in Section 3.1 some further remarks are made on this point. But it is beyond the scope of this chapter to discuss the efficiency argument in detail. See, however, Frisch (2012), Baer (2013), Jamieson (2013), and Shue (2013), for critiques of Posner and Weisbach’s claims regarding the ethical optimality, so to speak, of IP.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  349 The further concern is that there are apparently no set criteria for determining whether the intended target—a goal or action or both—counts as feasible, i.e., how plausible its realization must be and what is the basis for such an assessment. Perhaps we should accept that ‘feasibility’ is simply a smokescreen for differing substantial views about what are worthwhile pursuits. In the political context, the question might boil down to old disputes about what are viable prospects for social change. If IP proponents were simply staking a ‘realist’ position regarding international relations, say, then here again their claims would seem to resist any searching analysis. Quite simply: those sympathetic to the ‘realist’ view (roughly that ‘might is right’) in the international arena will be sympathetic to the IP view, unlike those of more ‘idealist’ persuasion. There is something to the above concerns, but they are a bit quick. We will see that analyzing the IP proponents’ claims about feasibility and climate change is more interesting and fruitful than first meets the eye. To begin with, it is useful to consult the growing philosophical literature on feasibility. While there is some divergence amongst the prominent definitions in the literature (and moreover some suggest that ‘feasibility’ may play multiple normative roles and so accordingly have multiple meanings) there are nonetheless a number of points of agreement amongst feasibility scholars. These points of agreement will be taken as ‘fixed points’ in our analysis. Beyond that, the strategy here is to fashion a notion of feasibility that is charitable to the IP proponents’ position. It should be a useful concept for evaluating climate change proposals and flexible enough to potentially vindicate the IP proponents’ claim regarding a strong relationship between feasibility and Pareto improvements with respect to climate change. This more substantial issue is introduced in Section  2.2 and considered in more detail in the remainder of the chapter.

2.1  A Working Definition of ‘Political Feasibility’ Despite there being a range of views about the finer details of the concept of ‘political feasibility’, there is reasonable agreement about the basics. In particular, there is broad agreement (also in line with the discussion thus far) on the general function of assessments of feasibility: They bear on directives for action and, roughly speaking, concern the possibility of success. In the spirit of ‘ought implies can’, what an agent ought to do is somehow sensitive to what is feasible (see Brennan and Southwood  2007; Gilabert and Lawford-­Smith  2012; Lawford-­Smith  2013; Southwood 2016; Wiens 2015).4 Note that this accords with the examples given 4  There are many further subtleties concerning this phrase that I will not explicitly address in this discussion—concerning the way and extent to which ‘what one ought to do is sensitive to what is feasible’. Southwood (2016), for instance, discusses different notions of ‘ought’ that have differing logical

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

350  Katie Steele above of how the term is variously employed in debate. All these examples relate to directives for action, although, as noted, the precise nature of the target seems to vary—the feasibility assessment applies sometimes to an act, sometimes to a goal or state of affairs to be acted upon, and sometimes to a combination of the two. (We will return to this diversity shortly.) It follows that feasibility assessments pertain to a given decision-­making context. All agree that feasibility has to do with whether some target may be plausibly realized, but, as mentioned, there is disagreement about the type of target. Perhaps the most natural target is an action that an agent might consider performing. This is the line taken by Brennan and Southwood (2007) and Southwood (2016). We can interpret these authors as holding that the ‘feasible acts’ are akin to what decision theorists might refer to as the ‘available acts’ for some decision-­maker, or in other words, the ‘choice set’. It follows that feasibility, understood in this way, is a binary matter: Candidate acts are either feasible (i.e., in the choice set), or not, depending on worldly constraints and whether the agent can ‘bring him/herself to perform the act in question’.5 The motivational ­constraint here is that the agent must be able to decide, if it follows from their psychological attitudes, upon the act in question, to initiate the act if it is chosen, and also to follow through in performing the act.6 So the agent must be sufficiently likely to actually go ahead and perform the act, conditional on being psychologically disposed, in some sense, to do so. Others also endorse a binary notion of feasibility that is a property of acts. There are differences, however, in emphasis, and in the kind of ‘acts’ that are taken to be at issue. Wiens (2015), for instance, attends much more closely to how mo­tiv­ation­al and resource constraints together affect what acts/options in the political domain are sufficiently likely to be realized and are thus ‘feasible’. In so doing, the acts he considers are rather more complicated ones involving multiple agents. There is not such an obvious connection between feasibility and a decision-­maker’s ‘choice set’ (to use the decision theory language introduced above). Gilabert and Lawford-­Smith (2012) also extend the target of feasibility assessments beyond simple acts; they include a goal as part of their account, thus proposing a four-­place predicate for feasibility: On their view, feasibility pertains to a particular agent performing a specified act in order to achieve some goal in a given context. Again, some of the acts (paired with goals) they consider are rather

relationships with his stated notion of ‘feasibility’. The point here is simply that the notion of feasibility should be defined in such a way that it plays some guiding role in determining right action. 5  This is, roughly, the phrase used by Southwood (2016: 11), who attributes it to Estlund (2011). Note that even if probabilities are employed in reasoning about what is feasible, on this account ­ul­tim­ate­ly the judgment must be binary because arguably it does not make sense for an act to be only partially in an agent’s choice set. 6  As per Southwood (2016: section II).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  351 complicated ones. (They also introduce more and less minimalist feasibility assessments, presumably to cater for the term having multiple roles in ethical ­discourse; one is a binary notion and depends on ‘hard’ logical-­metaphysical constraints, while the latter is a graded notion and depends on further ‘soft’ empirical details.7) To the extent that feasibility scholars have different targets in mind, they may be talking past each other. To avoid ambiguity, it helps to introduce a new term for the kinds of ‘complex acts’ referred to above: Let us call them projects or ­multi-­stage plans. This seems to best characterize the target for feasibility assessments in the climate debate. (The account developed here need not be the only useful notion of feasibility.) Climate treaties or proposals are more like projects or multi-­stage plans than simple acts that a decision-­maker may or may not choose to perform. Indeed, climate treaties not only concern a course of action that spans some considerable length of time; they also involve more than one agent (with the further complication that these agents are themselves groups). The feasibility (or predicted success) of such a project surely comes in degrees.8 But given projects are not the sort of thing a single agent may simply decide to perform, does feasibility still concern directives for action and pertain to an agent’s deliberations? Yes. For starters, it is hard to make sense of the feasibility of a project or multi-­stage plan, complex as this series of acts may be, absent the perspective of a deliberating agent and a specific way of initiating the project (even if these matters of perspective are not always made explicit in our everyday talk). Second, there are good reasons for a decision-­maker to assess the feasibility or likelihood of success of a project, were it initiated in a specific way. The most obvious reason is that the decision-­maker is considering whether it is worth doing the initiation herself. This clearly depends on the likelihood that, were she to take this action, the project would eventually be successfully implemented. That is, the full implementation of the project is one possible consequence of the act of initiation, and thus the likelihood of this consequence—the feasibility of the project were it initiated in the given way—is relevant to the agent’s deliberations about whether the project ought to be initiated in that way.9 It is worth clarifying the project account of feasibility before moving on to the more substantial questions concerning self-­interest and Pareto improvements. We see that projects are of concern to a deliberating agent. The difference with basic acts is that, at the point of choice, the agent can merely initiate the project.

7 While Southwood (2016) focuses on different logical roles for ‘feasibility’, depending on the moral concept at issue, Gilabert and Lawford-­Smith seem to focus rather on different definitions of ‘feasibility’, depending on the moral concept at issue. 8  As per Gilabert and Lawford-­Smith’s graded notion of feasibility that takes into account ‘soft’ constraints. 9  By the same token, the nature and likelihood of the project not being successfully realized is also relevant to the agent’s deliberations.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

352  Katie Steele Subsequent steps in the completion of the project are not directly under the agent’s present control; they are the province of other agents or perhaps the ‘future selves’ of the initiating agent. So projects are rather complex and there is room for a variety of eventualities once they are initiated. One of these eventualities or possible consequences is the full implementation or successful realization of the project. The likelihood of this outcome is the feasibility of the project. Note that we refer to projects or multi-­stage plans in diverse ways and so many feasibility statements could conceivably refer to projects. (Consider the examples given earlier.) Projects are sometimes described as if they were single actions, albeit complex ones that would take a number of steps to implement over some period of time. Consider the ‘act’ of me working without procrastinating, or, at the grander scale, the ‘act’ of Australia reducing its greenhouse gas (GHG) emissions by 26% on 2005 levels by 2030.10 Note that the latter example of Australia reducing its emissions may equally be regarded a goal. Indeed, projects are often described in the language of goals, for instance, ‘the goal of limiting climate change to 2 degrees C rise in average global temperature’, or ‘the achievement of universal access to high-­speed internet’. Moreover, projects may be more or less detailed. The goal of limiting climate change to 2 degrees C may be stated in such a way that the precise means to this end are specified, or else rather left open.11 But however they are referred to in casual talk, the feasibility of a project, on the account given here, depends on the way it would be initiated by the relevant agent and what the whole project amounts to, or what counts as successful implementation. A further issue is how we should think about the feasibility of projects involving multiple agents. One may refer loosely to ‘group projects’ but the idea here is that any project must somehow be initiated, and this concerns the choice of a single agent or decision-­maker.12 The choice concerns whether to perform the initiating act. Whether the project will succeed then depends, amongst other things, on the predicted actions of the other agents who are relevant to the success of the project. In other words, the behavior of these other agents, from the point of view of the decision-­maker, are simply aspects of the world, like whether or not it will rain, that bear on the consequences of the initiating act. Note that the decision-­maker may relate to the wider group in a variety of ways: He/she may be 10  As agreed to in Paris in 2015. 11 Compare the following two projects regarding the mitigation of climate change: (i) Reduce greenhouse gas (GHG) emissions by 26%; (ii) Implement extensive program of carbon-­capture-­and-­ storage so as to reduce GHG emissions by 26%. The latter project is obviously more detailed in terms of the steps to be taken, and as such, its feasibility may be higher or lower than the more open-­ended project. 12  Admittedly, a group may count as a single agent if it is sufficiently unified in the relevant ways (as per List and Pettit 2011). But the conditions for agency are relatively difficult to satisfy, and one would be hard pressed to argue that the groups of nation states that concern us here count as single agents (even if each nation state separately counts roughly as a single agent).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  353 a member of the group who, for the purposes at hand, is regarded the deliberating agent. Or else he/she may be an onlooker to the group who is nonetheless interested in the initiation of the group project. To see the distinction, consider the assessment of a proposed international climate treaty by (i) a representative of a participating state and (ii) a UN official charged with brokering the deal. These agents may justifiably have differing assessments of the feasibility of the treaty, due to differences in how they would initiate the project. So feasibility is the probability that a project or multi-­stage plan would be successfully realized if initiated in a particular way by some agent. One further important question concerns the nature of this probability judgment. How exactly should we interpret the conditional probability in question? Is it some agent’s credence that a project would be successfully realized if initiated in a particular way, say, the credence of the decision-­maker in question? Or is it supposed to be an  objective assessment? Perhaps it is best considered an objective epistemic ­prob­abil­ity—the credence that the decision-­maker should ideally have, if appropriately informed, regarding whether the project in question would be successful if initiated in the specified way.13

2.2  Feasibility and Self-­Interest The characterization of ‘feasibility’ above clearly does not directly concern ­self-­interest, let alone Pareto improvements (which we return to in later sections). Feasibility is not a property of a value function per se; it is not a measure of the extent to which a theory or account of value is in line with self-­interest. This point is worth elaborating. It is generally agreed that feasibility is not pertinent to the evaluation or ranking of states of affairs.14 Thus if IP were understood as a theory of value (with Pareto improvements on the status quo ranked higher than states of affairs that are not Pareto improvements), then feasibility would not be a relevant consideration in assessing IP. More generally, when the question is one of evaluation—say, how we should evaluate different distributions of the costs of responding to climate change—feasibility is not relevant. Note that much discussion about the ethics of climate change concerns the aforementioned question of evaluation. It is only once we start talking about directives for action that feasibility becomes relevant, and as discussed, it concerns the likelihood that certain outcomes/events will come about, as opposed to the value of these outcomes.

13  Of course, this raises further questions (which, again, come back to the precise function of the feasibility assessment) as to what it means for the decision-­ maker to be ‘ideally/appropriately informed’. The broad account given here, however, is sufficient for the purposes of this discussion. 14  Wiens (2015: 461) explicitly makes the distinction between directives for action and purely evaluative ethical claims, and notes that feasibility is pertinent to the former as opposed to the latter.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

354  Katie Steele Presumably when proponents say that ‘IP is the most (or only) feasible approach to climate change’, they are not making a purely evaluative statement but rather something in line with our working definition of feasibility above: That climate treaties designed to achieve a Pareto improvement on the status quo are likely to succeed in a way that climate treaties designed to involve sacrifices, are not. (IP proponents in fact tend to speak of feasibility as a binary matter, but as per the discussion above, a graded assessment is arguably more appropriate.) This is supposedly because the agents involved will only play their part in a treaty project if it furthers their self-­interest. If true, this would amount to a more indirect relationship between feasibility and self-­interest. It is an empirical claim about what reliably motivates the agents that climate projects depend upon. Let us not forget that, throughout our discussion, the function of feasibility assessments is to aid a decision-­maker. One thing to note is that, if we focused just on whether it would be possible for a decision-­maker to initiate a project in a specified way (whether this is a feasible act à la Southwood and co.), the as­so­ci­ ation with self-­interest would be tenuous. Recall why a simple act may not be a viable choice for an agent: If he/she would suffer some kind of irrationality or weakness of will that would ultimately prevent the performance of the act, even if he/she preferred it and tried to choose it. This does not let the agent off the hook easily when it comes to onerous acts involving self-­sacrifice. In most cases, agents who act selfishly do not do so because it was not viable (or feasible, by the lights of Southwood and co.) for them to do otherwise. More likely a range of acts were viable, and yet the agent preferred the selfish one. To give an example, it is surely perfectly viable for me to set up a regular online donation to a charity. That is, if this were what I preferred to do, I would succeed in doing it. If I do not set up the regular donation, it is rather because this is not in fact what I prefer to do, given my beliefs and values. It may thus be perfectly viable for a decision-­maker to initiate both more and less demanding climate treaties. But some of these treaty projects may be more likely to be fully realized than others, i.e., some of these treaty projects may be more feasible than others. This is the sense of feasibility—one concerning the success of a complex project—that was argued above to be most apt for our discussion here. It is easy to see why this notion of feasibility is important for a decision-­maker—it concerns how likely are the planned consequences of an initiating act—and moreover, why feasibility, thus understood, may be associated with self-­interest. It all depends on the context and the relevant empirical facts. The likelihood that a complex project will be realized, if initiated in a specified way, depends on whether other agents the project depends on (perhaps even ‘future selves’ of the decision-­maker) will play their part. The substantial question then is: Should a decision-­maker be more confident that a project or multi-­stage plan she initiates would be realized to the extent that others are expected merely to act on their self-­ interest? Common wisdom

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  355 suggests the answer is ‘yes’: Arguably, people more reliably act on self-­interest than on other motivations. In that case, a project in which others are expected to act on self-­interest as opposed to other motivations would be more likely to succeed, and would consequently be more feasible. But this picture of motivation may of course be overly simplistic. Moreover, it surely depends on the type of agent involved, whether an individual or a group entity such as a state. For now, let us simply note that the association between feasibility and self-­interested mo­tiv­ations is an empirical matter, and one that is sensitive to context.

3.  International Paretianism and Climate Change Putting aside, for now, the relationship between self-­interest and feasibility, let us examine what the climate predicament looks like if cashed out in terms of the self-­interest of the nation state actors. The first thing to note is that the ­self-­interest of nation states—being large sprawling groups extending over time and space—is not a straightforward matter. As noted in Section  1, here we assume that the ‘national interest’ is an appropriate aggregate of the welfare of present and future citizens (leaving open the finer details, although presumably there will be considerable, if not equal, weight to future citizens). That is, we assume that there is some appropriate measure of the national interest of a state, as per the self-­interest of an individual.15 The initial aim here, in Section  3.1, is to canvas plausible models of the climate-­change predicament when posed in terms of the national interest of state actors. Whether or not state leaders are actually motivated to pursue the national interest, as roughly defined here, is a further issue that will be discussed in Section  3.2, along with other motivational concerns about climate treaties that aim for Pareto improvements. That is, the game models here do not necessarily track states’ existing motivations (or rather, their apparent motivations/ values given their choices). The models are rather intended to track (in terms of the outcomes and how they are evaluated) the stipulated national interest of the respective actors.

15  Note that Posner and Weisbach (2010) do not consistently treat the ‘national interest’ this way in their defense of the IP approach to climate change. What they mean by a nation’s interest seems to vary, depending on whether they are discussing the feasibility merits or rather the moral merits of IP. When it comes to feasibility, they seem to allow future welfare, for instance, to be weighed in the ‘national interest’ just as much as current citizens see fit (see, e.g., the discussion of the importance of respecting a state’s sovereignty with respect to its own future in the discussion of the asteroid analogy on p. 77). In many other places, however, they emphasize the ethical significance of the mitigation that would come from an IP treaty, suggesting a more ethically robust notion of ‘national interest’ as per our discussion in this chapter (see especially their comments on p. 177 to this effect).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

356  Katie Steele

3.1  Prospects for IP Climate Treaties This section considers what an IP climate deal might look like by appeal to key findings in public economics regarding the usage of ‘common pool resources’. The key question is: What kind of game model plausibly fits the empirical facts and our rough definition of national interest? This is crucial for determining what sort of Pareto improvement is in the offing. Posner and Weisbach (2010: 6), for instance, suggest (in some places at least) that IP would support a global mitigation effort that is optimal (at least by utilitarian standards) and moreover substantial. We will see that this is not necessarily the case; indeed a number of factors affect the collective optimality and extent of mitigation under IP, as well as the nature of individual contributions. The clearest reason for there being an unrealized opportunity for Pareto improvement on the status quo is a failure of collective action of some sort. Environmental resources are prone to collective action problems. Consider, for example, the over-­exploitation of fisheries, deforestation, and of course climate change. The diagnosis: Many environmental resources such as those mentioned (abundant fisheries and forests, a stable climate, etc.) are common pool resources shared amongst agents who are primarily concerned with their own consumption of them. Technically speaking, common pool resources are effectively n ­ on-­ excludable, meaning that there are no barriers to anyone using the resource.16 Common pool resources have a tendency (absent any institutional arrangement) to be over-­exploited, because no one in particular has the responsibility/ability/ motivation to provide or maintain the resource given that others cannot be excluded from using/spoiling it. Climate stability is surely a classic example of a global common pool resource.17 It is commonly assumed that the collective action problem takes the form of a Prisoners’ Dilemma of international magnitude, where the ‘cooperative solution’ would be a clear Pareto improvement over the ‘non-­cooperative solution’. But this need not be the structure of the game, even assuming that the state actors pursue their ‘national interest’. The Prisoners’ Dilemma, considered particularly problematic when it comes to cooperation for mutual benefit, can be contrasted with ‘mere’ coordination problems. Moreover, even when it comes to Prisoners’ 16  Some environmental resources may also be non-­rivalrous (at least up to some threshold), meaning that the resource is not depleted (up to the threshold) by additional users. Goods that are both non-­excludable and non-­rivalrous are referred to as public goods. (Strictly speaking, it is a matter of degree as to whether a good has these properties.) While the literature on environmental dilemmas often refers to public goods, it is generally only the property of non-­excludability that is at issue, in which case, better simply to refer to common pool resources. 17  Note that there are various ways to depict climate change as involving a common pool resource. Here we refer to the resource of ‘climate stability’, as per Ostrom (2009). Others refer to the ‘carbon sink’, or the Earth’s ability to absorb greenhouse gas emissions, as the common pool resource, rather than a stable climate per se. The difference is not a substantive one.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  357 Dilemmas, there is variation in the size of the Pareto improvement in question. In what follows, these different game scenarios are discussed in more detail: Section 3.1.1 briefly considers the coordination-­game account of climate change, after which Section 3.1.2 turns to the Prisoners’ Dilemma account and the variety of forms it may take.18

3.1.1  Climate Change as a Coordination Game Some suggest that, due to the presence of a dangerous threshold of emissions that would bring catastrophe to all, swamping other impacts, climate change may be best conceived as a coordination game (see Barrett 2011 for discussion).19 Here the cooperative solution simply requires coordination: If all states were confident that the others were pursuing a common project for emissions abatement that just meets the threshold for avoiding catastrophe, then it would be in the interests of each to pursue this project as well. In other words, the presence of a dangerous threshold would, ironically, be good news for collective action, because there would exist arrangements for emissions abatement that are stable in the sense that no actor would have an incentive to defect. These sorts of joint strategies that resist unilateral defection are referred to in game theory as Nash equilibria; they are commonly regarded the ‘solutions’ to the game. The presence of a dangerous climate-­change threshold is unlikely to be as rosy as first appears, however, even from the perspective of collective action. To begin with, there is the problem of too many Nash equilibria. There would be countless projects or joint strategies involving just enough emissions abatement to avert catastrophe; some of these would involve heroic abatement efforts on the part of any given actor with little abatement from others, whereas some of them would involve very little abatement on the part of the actor in question and much greater efforts from others. It would be no trivial task for the actors to settle on one particular stable joint strategy—a tough bargain. As such, it might be risky for any given actor to pursue a particular joint strategy for averting catastrophe, given there is no guarantee that others would also opt for the same equilibrium strategy. The riskless Nash equilibrium would be for all to do little and suffer catastrophe without additionally engaging in costly and potentially futile abatement.20

18  The models of this section follow the classic work of Barrett (1990) and Carraro and Siniscalco (1993) on international protection of the environment. Note that Gardiner (2001: section VII) has a broadly similar discussion of the alternative ways to conceive of the intra-­generational problem of environmental protection (primarily climate change); a key difference, however, is that, in the first instance at least, Gardiner focuses on individual persons as the actors rather than nation states. 19  Typically the game is simplified such that there is negligible benefit to emissions abatement that is short/in excess of the designated threshold. 20  This is to suggest that the coordination problem has the form of the so-­called ‘Stag Hunt’ (see Skyrms 2004 for extensive discussion of this game). Others suggest the climate problem has the form of the roughly similar ‘Battle of the Sexes’ game (as discussed in Gardiner 2001).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

358  Katie Steele An even more crucial issue, however, is that while climate change plausibly involves a danger threshold, there is uncertainty within the scientific community about its location. (The critical amount of 2 degrees C warming has gained ­traction in international debate, but apart from any other complications, there remains uncertainty about what likelihood for this temperature increase should be treated as the threshold, and there is furthermore uncertainty about the stock of atmospheric GHGs that corresponds to the chosen likelihood.) According to Barrett and Dannenberg (2012), this sort of uncertainty would effectively turn the prima facie coordination game into the more difficult Prisoners’ Dilemma (PD) game. For this reason, our main focus here will be the PD account of climate change. Section 3.1.2 considers the variety of forms, when it comes to the magnitude of the effect (so to speak), that the climate PD could take.

3.1.2  Climate Change as a Prisoners’ Dilemma It helps to begin with the simplest characterization of how climate stability may give rise to a Prisoners’ Dilemma. In our simple model, there are N relevant actors that affect the climate by (potentially) emitting greenhouse gases (GHGs).21 Assume the N nations have equivalent circumstances and so the game is entirely symmetric (we return to this assumption later): Each actor can reduce or abate emissions, relative to the status quo or ‘business as usual’ (BAU), at a constant cost of c and a constant benefit of b per unit of abatement. Given that climate stability is a common pool resource, the costs of emissions abatement are private (borne just by the actor doing the abatement) but the benefits of this abatement accrue to all actors. So if each of N nations abates one unit of emissions, they each receive a benefit of Nb. The classic Prisoners’ Dilemma results from the following pattern of costs and benefits: Nb > c > b Since, for each actor, the individual cost, c, of their own personal abatement effort is greater than the benefit, b, they would receive from this effort, there is no incentive to act. Whatever others do, it is better for the individual to do nothing. And yet, all individuals would benefit if they pulled together, and therein lies the dilemma: If all individuals were to similarly engage in emissions abatement, the benefit to each and every actor, Nb, surpasses their respective individual costs, c. To use more formal language: In the PD game just described, the non­cooperative solution, where each actor chooses their level of emissions abatement based on their own interests, holding fixed what others do, amounts to zero total abatement and so zero total net benefit. This is because, for each actor, regardless 21  Even though there are other ways to hasten/mitigate climate change (e.g. (de)forestation), GHG emissions play a key role and thus are our focus here.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  359 of what others do, their own marginal contribution to the climate through ­abatement brings less private benefit b than the private cost c. This is the Nash equilibrium for the PD game—the solution that is stable in that no actor has an incentive to defect. The fully cooperative solution, on the other hand, is what brings max­imum benefit to each actor; it is also optimal for the group in the aggregate welfare sense. Here it involves all actors pursuing maximum abatement x, which brings (Nb—c)x net benefit to each actor. Note, however, that this Pareto improvement is not a (stable) Nash equilibrium as all actors have an incentive to defect and not abate at all. A more realistic model has cost and benefit functions that are not constant but rather depend on levels of emissions abatement.22 Assume i = 1,…, N sym­met­ ric­al actors as before. Let the total emissions abatement, Q, be the sum of all individual abatement efforts; so Q = ∑ N qi , with qi being the abatement of each actor. Assume that the marginal benefit of a unit of abatement for each actor, Bi, depends on total abatement as follows:



 Q2 Bi= (Q ) b  aQ − 2 

 N 

where a and b are positive parameters. This is a concave function with respect to global emissions abatement, which is to say that there are diminishing marginal returns for each additional unit of global abatement. The cost of abatement for each actor, Ci, is a convex function of the actor’s own abatement levels:

Ci ( qi ) = cqi2 / 2

This amounts to increasing marginal costs of abatement: the more a nation individually abates, the more costly the next unit of abatement. As per the previous model, the non-­cooperative (Nash equilibrium) solution for non-­constant cost/benefit functions is determined according to abatement levels that individual actors would choose, holding fixed the abatement of others. The solution, corresponding to total abatement Q0, is such that the marginal abatement cost for each individual equals the marginal abatement benefit for that individual alone. The fully cooperative solution, on the other hand, corresponding to total abatement QC, is such that the marginal abatement cost for each 22  This is actually considered most realistic for a pollutant that does not accumulate in the atmosphere, unlike GHGs. Barrett (1994) notes that Nordhaus (1990) employs a logarithmic cost function that depends on percentage emissions abatement, to model costs associated with GHG emissions abatement in the US. The precise forms of the cost and benefit functions, however, do not affect our (more general) discussion here.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

360  Katie Steele individual is equal to the global marginal abatement benefit. That is, QC ­maximizes the total benefit to the group, measured as the sum of benefits to individuals minus the sum of costs. The abatement and the net benefits for each actor under full cooperation are greater than the abatement and the net benefits for each actor under the non-­cooperative solution; the difference, QC − Q0, being the gain from full cooperation. The size of this gain depends on the ratio c/b (the cost factor to the benefit factor) and the size of c (see Barrett 1994: 880–881). When c is small and b is large, the cooperative gain is small because actors already benefit significantly from unilateral or non-­cooperative abatement. When c is large and b is small, the co­opera­tive gain is small because it makes little sense to abate, whether uni­lat­ eral­ly or cooperatively. The biggest gains from full cooperation, both with respect to total abatement and net benefits, are when c is comparable to b and both are large. So we see that if the international climate predicament were best modeled as a Prisoners’ Dilemma (PD), the size of the Pareto improvement (let alone whether or not it is achievable) is yet a further question. The model above well illustrates this point. For instance, in one of the PD scenarios mentioned just above—where the individual costs of emissions abatement are small and the benefits large—little is gained from a cooperative climate deal per se. Of course, such a model, if plaus­ ible, may nonetheless inspire change in revealing that the international community falls short even of the non-­cooperative solution with respect to GHG emissions abatement; perhaps states need to start acting on their ‘national interest’, so to speak, rather than overcome a further collective action problem.23 Some have indeed argued that it is not well enough appreciated that the benefits a single state receives from its own emissions abatement does to a large extent surpass the costs. Just because climate stability is a common-­pool resource does not automatically mean that it is against a state’s national interest to unilaterally do anything about it. There may well be indirect national benefits (or, in other words, lesser national costs than appreciated) associated with emissions abatement, such that the difference in abatement under the cooperative and the non-­cooperative solutions is small. Ostrom (2009), for instance, emphasizes local gains (or ­co-­benefits) associated with emissions abatement, including savings in energy costs and health benefits from lower pollution (cf. Maslin and Austin  2012 on energy security and air quality).24

23  If this were the case, the supposed motivational power of the ‘national interest’ that is implicit in the defense of IP may be undermined, because one reason for states not currently pursuing their national interest and achieving the non-­ cooperative solution is that this is politically difficult. Section 3.2 will revisit this general issue of how easy it is for states to pursue their national interest. 24  Others focus on the dynamics of the problem, and argue that the costs of emissions abatement decrease rather than increase, the greater the abatement effort to date (for various accounts of this phenomenon, see, for instance, Victor 2011; Heal and Kunreuther 2011; Tavoni 2013). The basic idea

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  361 The other PD scenario where the gains of cooperation are slight, is when the cost factor, c, for emissions abatement is extremely large compared with the bene­ fit factor, b. In this case, there is little reason, whether non-­cooperatively or co­opera­tively, to engage in emissions abatement. This kind of model would only be plausible if, for each state, (i) the BAU climate change path was relatively innocuous and/or (ii) the welfare of future citizens was largely irrelevant to the ‘national interest’. The first scenario is not very likely, given our best science. There is much uncertainty, possibly irresolvable, about how particular regions of the world will be affected by any particular rise in average global temperature, but nonetheless it is most plausible that the expected (in the mathematical sense) consequences are negative for all, compared to a future with a more stable climate.25 The second reason is a non-­starter in the context of our discussion here. It is true that, the more the future is discounted, when it comes to a state’s ‘national interest’, the less benefit and the more cost associated with emissions abatement. But a heavily discounted future contravenes one of the premises of our discussion, namely that the ‘national interest’ of a state is some appropriate aggregate of the welfare of present and future citizens. So it is important to note that the size of the cooperative gain associated with a PD can vary quite significantly. Thinking about these possibilities helps to sharpen one’s views about the climate predicament that we face (based on ‘national interest’ as roughly defined here). There is reason to think that the status quo falls short even of a non-­cooperative climate solution, that things would be different if states really acted in their national interest. On top of that, there is reason to think that the cost of GHG emissions abatement for individual actors is high, and the gains from cooperation also high. That is, it is not at all implausible that the climate predicament involves a Prisoners’ Dilemma where the gain from cooperation would be substantial and thus tragic if not achieved. (It would ar­gu­ ably be more convenient if climate change gave rise to a ‘mere’ coordination game, but this seems wishful thinking.26) A further interesting question is whether the fully cooperative solution to the climate PD would effectively be what a global planner would recommend as the optimal mitigation response. This is a topic for is that new partnerships and energy economies associated with emissions abatement would rapidly reach some critical mass whereby it is less costly to be in than out. 25  See Frigg et al. (2013) on the uncertainties surrounding predictions of regional climate change. For the multi-­region models of climate economics investigated in Nordhaus and Boyer (2000), it is indeed the case that the expected consequences of BAU climate change for all regions of the world are negative, compared to selected mitigation options. Note too that the mitigation options are costly for the present generation in all regions of the world, compared to BAU. This makes clear that mitigating climate change is better than BAU (on these models at least) only because significant weight is given to the welfare of future generations (enough to counterbalance present costs). 26 Posner and Weisbach (2010) often seem to assume, through their choice of analogies, that ­climate change is a coordination game. They do, however, also discuss the possibility of a Prisoners’ Dilemma scenario (2010: 181).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

362  Katie Steele further investigation, but the answer surely depends on whether the total welfare calculations of the global planner align with the total welfare calculations that underwrite the national interest of the respective states.27 With the character of the Pareto improvement in mind, let us now consider whether a climate treaty that aims for this outcome is really achievable just on the back of ordinary ‘­self-­interested’ motivations.

3.2  Does ‘Self-­Interest’ Suffice? If the climate-­change predicament were plausibly modeled as a game showing a large mutual gain for the state actors if only they could somehow rise to the co­ opera­ tive task, then this would seem to be good news for International Paretianism. After all, IP is all about climate proposals that are feasible in virtue of appealing only to the cause of self-­interest. And surely we can count on actors to merely pursue their self-­interest (give or take some misconceptions they might have about what is really in their own self-­interest). Even if we grant this simple rule of thumb about motivation, however, there is reason to doubt whether the Pareto improvements suggested by the Prisoners’ Dilemma models canvassed above are achievable just on this basis. To begin with, we should be cautious in treating the ‘national interest’ of states akin to the self-­interest of individuals. And secondly, even if states may be persuaded to pursue their national interest, as defined here, this alone does not suffice for overcoming a Prisoners’ Dilemma. In what follows, these two issues are discussed in turn. Many would observe that the analogy between the self-­interest of an individual agent, both in explaining and guiding behavior, and the ‘self-­interest’ of a group agent, is far from tight. The group agents here are nation states. Do states nat­ur­ al­ly pursue their ‘national interest’, especially as understood here to involve due concern for the welfare of both present and future citizens? This seems a rather heroic assumption, and yet an important one for IP, because the appeal to ‘self­interest’ has force precisely because it is supposedly motivating in a straightforward way; something that an agent would quickly recognize as worth pursuing. The obvious worry is that state leaders rather have perverse incentives, when it comes to their domestic and international negotiations, having little to do with the pursuit of the ‘national interest’. Moreover, even if democratic institutions provided some assurance that state leaders were responsive to the welfare of their

27  For instance, for a utilitarian global planner—as in the analyses of Stern (2007) and Nordhaus (1990)—the fully cooperative PD solution would only line up with that of the global planner if each state’s national interest were also a utilitarian aggregate across all citizens in the present and future. This question is relevant to the ‘efficiency’ argument for IP that was briefly mentioned at the start of Section 2 (see, especially, note 2).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  363 citizens, it is a further stretch to think that this would somehow incorporate the welfare of future citizens.28 While there is a big question mark as to the motivating force of enlightened ‘national interest’ in international negotiations, let us nonetheless proceed on the basis that this is at least a persuasive reason for state action.29 As noted already, it would no doubt be a significant international achievement if states were persuaded to act in their national interest, thus achieving the non-­cooperative game solution. As far as securing the further mutual gains associated with the co­opera­ tive game solution—the promised Pareto improvement—there are, however, further motivational worries. The problem is that projects conforming to IP may require a lot more than the pursuit of national interest to succeed, once initiated. Here the lessons of game theory are important. The notorious problem with Prisoners’ Dilemmas is that the fully cooperative solution is not a Nash equilibrium, and worse still, the greater the gains from cooperation, the greater the incentive for individual actors to free ride on the efforts of others, if the project were initiated. So even though the Pareto solution does further all actors’ interests relative to the status quo, more than self-­interested motivations are required to successfully implement this solution. Actors pursuing self-­interest alone will try to free ride on the efforts of others.30 To use the language of feasibility, if states really were purely concerned with their national interest and the game has a PD structure, the Pareto solution is not feasible. In many contexts a PD game analysis serves as a plea to change the incentive structure of the game, perhaps via punishments for free riding. But this is ar­gu­ ably not an option in the international setting, given the lack of an overarching authority that can facilitate such a change.31 Economists have otherwise explored the possibilities that may arise from coalition agreements involving only some of the actors (here states) in the game. Solutions involving coalitions may be better than the Nash equilibrium and yet also dynamically stable in the sense that no actor included in the coalition has an incentive to leave, and no actor excluded 28  In response to this point, Posner and Weisbach (2010) might be tempted to loosen the notion of ‘national interest’ so that it is more in line with the existing apparent motivations of states. But, as noted already in note 13, this would conflict with their claims about IP solving the climate pre­dica­ ment; the climate solution they have in mind depends on an account of national interest that gives due weight to the welfare of both present and future citizens. 29  Arguably, people are more naturally disposed to care for their own descendants as compared to contemporaries in distant parts of the world (cf. comment to this effect in Posner and Weisbach 2010: 142). 30  Posner and Weisbach (2010) do acknowledge this point (see, esp., p. 181), but they do not explore it in detail and nor do they give it appropriate emphasis. 31  Although one might argue that a higher authority can at least be approximated in the inter­ nation­al setting via multiple agreements, whereby punishments for non-­cooperation in one domain (e.g. environmental protection) would take the form of sanctions in another domain (e.g. trade). See, for instance, Barrett (2008) for discussion. Nordhaus (2015) makes a similar point—that a group of states or ‘climate club’ can effectively impose external trade-­related penalties on states that do not cooperate with respect to emissions abatement.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

364  Katie Steele from the coalition has an incentive to join. Hence these sorts of agreements are dubbed ‘self-­enforcing’ within the public economics literature (see Carraro and Siniscalco 1993; Barrett 1994, 2005). Unfortunately, the cooperative possibilities associated with ‘self-­ enforcing’ agreements are rather limited, as the aforementioned authors have shown. By way of elaboration: A stable or ‘self-­enforcing’ coalition arrangement is one where the individual benefit to all members within the coalition is greater than it would be if they unilaterally defected (since the remaining coalition members would accordingly lower their own cooperative efforts, in this case emissions abatement), and moreover, the individual benefit to all those outside the coalition is greater than it would be if they were to become members (since the extra gains from further cooperation would not outweigh the individual costs).32 Of course, it may be the case that all states would prefer to be outside the coalition, thus shouldering less of the total emissions abatement, than inside it. States within the coalition cooperate to achieve the greatest total benefit for the coalition, while the remaining states determine their abatement levels according to what is individually optimal, holding fixed the abatement of others. Once a stable coalition has formed, however, no state can benefit from unilateral change. Across a wide variety of assumptions regarding the precise shape of the benefit and cost functions underlying a Prisoners’ Dilemma game, it turns out that little is achieved over and above the non-­cooperative solution by stable coalition arrangements (as shown in Barrett  1994). The first game described above in Section  3.1.2, for instance, involving constant marginal cost and benefit functions, unsurprisingly does not support any self-­enforcing coalition agreement. The second game, involving a concave benefit function and a convex cost ­function, supports only a limited advantage in coalition formation. For instance, when the number of stakeholders, N, equals 100, and the cost to benefit parameter ratio, c/b, equals one (so there is potentially much to gain from cooperation), the self-­ enforcing agreement involves only three states; these signatories abate ­con­sid­er­ably more than they would under the non-­cooperative arrangement, while ­non-­signatories abate just a little less. The large number of non-­signatories means that the free-­riding effect is substantial and so little is gained over and above the non-­cooperative solution (see Barrett  1994: 882).33 In fact, to the extent that a coalition in this class of cases has a large number of signatories, there is ­little to be gained from cooperation (c/b is either very large or very small). Similar results apply to other functional forms: generally speaking, where there is much co­opera­tive gain at stake, only coalitions with two or

32 The notion of a stable coalition is attributed to earlier work on cartels by d’Aspremont et al. (1983). 33  Note that these results are obtained via simulation. For other cost and benefit functional forms, analytic results can be derived.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  365 three signatories are stable, and thus not much cooperative gain can actually be realized in a ‘self-­enforcing’ manner.34

4.  The Feasibility of IP Climate Deals The assumption that states pursue their national interest may yield a PD climate game with a sizable Pareto improvement, but we see that ‘self-­interest’ alone may not get us very far in realizing that Pareto improvement. IP proponents would thus be better served by a more nuanced story regarding the association between feasibility and self-­interest. In fact, proponents do try to sell IP on moral ‘other-­regarding’ grounds as well as ‘self-­interest’ grounds. Throughout their book, Posner and Weisbach, for instance, appeal to the happy confluence of gains in national interest and moral gains under an IP climate treaty. This suggests a hidden commitment to the moral standing of a project being relevant to its mo­tiv­ation­al appeal and thus feasibility, even if the moral gains are presented as if they were a mere by-­product of the pursuit of self-­interest. Recall too that the pursuit of ‘self-­interest’ in the context of IP is already morally demanding. It is hardly the norm for state leaders to act in the interests of present and future citizens, suitable balanced. So IP proponents appeal in various ways to moral reasons for climate action, without necessarily advertising this fact. Posner and Weisbach make an even larger concession regarding the need to appeal to moral reasons for climate action. They state that motivations beyond ‘self-­interest’ are required to secure Pareto improvements where states have an incentive to free ride (2010: 169): A key for a climate treaty is what we have called International Paretianism— nations must believe that they are better off with a treaty than without. But the obligation to achieve a broad, deep, and enforceable treaty imposes a serious ethical duty on rich and poor nations alike—the obligation to cooperate. In our view, it is unethical for a nation to refuse to join a climate treaty in order to ­free-­ride off of others.

While it has an air of authority, this statement of ethical duties has no clear justification.35 The question, when it comes to the feasibility merits of IP, is whether a 34  One might wonder whether the situation changes if we allow asymmetry, i.e., when cost and benefit functions for the N states are not identical. There has been some investigation of this case in the literature. While Barrett (1997) derives similarly pessimistic results for heterogeneous as for homogeneous parties to an agreement, McGinty (2007), on the other hand, finds that, assuming transfers are permitted within the coalition, heterogeneity can significantly increase the percentage gain from full cooperation, even when this gain is substantial. It seems that further investigation is required to establish the robustness of this result. 35  As acknowledged by Posner and Weisbach (2010: 180), although they do not allow that this lack of justification severely undermines the feasibility case for IP.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

366  Katie Steele climate treaty relying on this extra commitment to secure a Pareto improvement would be more or less motivationally compelling, for all actors concerned, as a treaty relying just on the pursuit of national interest. Let us briefly consider this question. Refraining from free riding involves a sacrifice in national interest. For this to be motivationally compelling, it would need to be offset by some other, say moral, gain. The question is then: Does an IP climate deal have the required moral merit? Does it make an obvious step forward in moral terms? On certain ways of calculating national interest and the global good, the IP climate solution may well accord with the optimal amount of global climate change mitigation. But note that this is consistent with the present generation of poor countries paying deeply for this mitigation so as to ensure climate stability for their descendants. (The alternative, of course, would be for other nations to pay more for the same mitigation effort.) So we see that, in solving one global problem—climate stability—another is exacerbated, namely, inequalities (not to mention the failure to meet historical responsibilities). Does it make sense for nations on the losing side of this equation to accept an imperative to participate in an IP climate deal rather than hold out for an alternative mitigation proposal that better promotes equality? It is far from clear that IP has the upper hand here, in terms of being motivationally compelling.36 More precisely, the problem for IP proponents is that, once we allow for other moral motivations to weigh against the pursuit of national interest, the feasibility case for IP is much shakier. This is exacerbated when the gains from emissions abatement are not symmetric and when states have more complex moral concerns (say, they care about equality and/or historical responsibility). In this case, states may have very different views about the moral merit of an IP climate treaty, which may well affect the overall appeal of the proposal for them. For rich, high-­emitting states, an IP treaty is good for national interest and it may look pretty good morally speaking as well, when one compares to the business-­as-­usual baseline; thus rich states may have both national interest and moral reasons to commit to an IP treaty. But poor states may rather see an IP treaty as morally deficient, compared to other possible treaties that would better promote equality and compensation for past wrongdoing. In other words, they may see little reason to treat business-­as-­usual as the moral baseline, and thus have little moral reason to commit to an IP treaty. In that case, there would be little reason for them to refrain from free riding or defecting from such a treaty. Let us then conclude on a cautionary note. While concern for the feasibility of climate treaties is important and timely, there is no simplistic way to make such assessments. Feasibility is simply the likelihood, from the deliberator’s perspective, that a proposed project or multi-­stage plan will be successfully implemented, 36  See Brennan and Sayre-­McCord (2016) for discussion of how moral facts can affect the feasibility of a political proposal.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  367 once initiated in a specific way in a given context. When it comes to climate treaties, the deliberator (whether a member state or an outside official) must assess whether other actors, namely states, will be motivated to follow through on their part in the project, if it were initiated as well as possible by the deliberator. The more reliable these motivations, the more likely the project would succeed, i.e., the greater its feasibility. We must conclude then that it is an open empirical question as to whether states’ motivations would likely align with an IP climate treaty; indeed it is plausible that states’ motivations would not be so aligned, given that self-­interest alone may be self-­defeating, and states may well perceive the moral situation rather differently.37

References Baer, P. (2013). “Who Should Pay for Climate Change? ‘Not Me’.” Chicago Journal of International Law 13(2): 507–525. Barrett, S. (1990). “The Problem of Global Environmental Protection.” Oxford Review of Economic Policy 6(1): 68–79. Barrett, S. (1994). “Self-Enforcing International Environmental Agreements.” Oxford Economic Papers 46: 878–894. Barrett, S. (1997). “Heterogeneous International Environmental Agreements,” in International Environmental Negotiations: Strategic Policy Issues, ed. C.  Carraro. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing, 9–25. Barrett, S. (2005). “The Theory of International Environmental Agreements,” in Handbook of Environmental Economics Volume 3, ed. K.-G. Mäler and J. R. Vincent. Amsterdam: Elsevier, 1457–1516. Barrett, S. (2008). “Climate Treaties and the Imperative of Enforcement.” Oxford Review of Economic Policy 24(2): 239–258. Barrett, S. (2011). “Avoiding Disastrous Climate Change is Possible But Not Inevitable.” Proceedings of the National Academy of Sciences 108(29): 11733–11734. Barrett, S. and A.  Dannenberg (2012). “Climate Negotiations Under Scientific Uncertainty.” Proceedings of the National Academy of Sciences 109(43): 17372–17376. Brennan, G. and G. Sayre-McCord (2016). “Do Normative Facts Matter . . . to What Is Feasible?” Social Philosophy and Policy 33(1–2): 434–456. Brennan, G. and N.  Southwood (2007). “Feasibility in Action and Attitude,” In Hommage à Wlodek: Philosophical Papers Dedicated to Wlodek Rabinowicz, ed. T. Ronnow-Rasmussen, B. Petersson, J. Jonefsson, and D. Egonsson. Online publication only: https://www.fil.lu.se/hommageawlodek. 37 Many thanks to Christian Barry, Richard Bradley, Fergus Green, Jeremy Moss, Nicholas Southwood, Kai Speikermann, and all three editors of this collection for very helpful comments on earlier drafts of this chapter.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

368  Katie Steele Broome, J. (2010). “The Most Important Thing about Climate Change,” in Public Policy: Why Ethics Matters, ed. J. Boston, A. Bradstock and D. Eng. Canberra: ANU E Press, 101–116. Carraro, C. and D.  Siniscalco (1993). “Strategies for the International Protection of the Environment.” Journal of Public Economics 52: 309–328. D’Aspremont, C., A. Jacquemin, J. Jaskold Gabszeweiz and J. A. Weymark (1983). “On the Stability of Collusive Price Leadership.” Canadian Journal of Economics 16: 17–25. Estlund, D. (2011). “Human Nature and the Limits (if Any) of Political Philosophy.” Philosophy & Public Affairs 39: 207–237. Frigg, R., L. A. Smith, and D. A. Stainforth (2013). “The Myopia of Imperfect Climate Models: The Case of UKCP09.” Philosophy of Science 80: 886–897. Frisch, M. (2012). “Climate Change Justice.” Philosophy & Public Affairs 40(3): 225–253. Gardiner, S.  M. (2001). “The Real Tragedy of the Commons.” Philosophy & Public Affairs 30(4): 387–416. Gilabert, P. and H.  Lawford-Smith (2012). “Political Feasibility: A Conceptual Exploration.” Political Studies 60: 809–825. Heal, G. and H. Kunreuther (2011). “Tipping Climate Negotiations.” Working Paper 16954, National Bureau of Economic Research. Jamieson, D. (2013). “Climate Change, Consequentialism, and the Road Ahead.” Chicago Journal of International Law 13(2): 439–468. Lawford-Smith, H. (2013). “Understanding Political Feasibility.” Journal of Political Philosophy 21: 243–259. List, C. and P.  Pettit (2011). Group Agency: The Possibility, Design, and Status of Corporate Agents. Oxford: Oxford University Press. McGinty, M. (2007). “International Environmental Agreements Among Asymmetric Nations.” Oxford Economic Papers 59(1): 45–62. Maslin, M. and P. Austin (2012). “Climate Models at Their Limit?” Nature 486: 183–184. Nordhaus, W. D. (1990). “To Slow or Not to Slow: The Economics of the Greenhouse Effect.” Mimeo, Department of Economics, Yale University. Nordhaus, W.  D. (2015). “Climate Clubs: Overcoming Free-Riding in International Climate Policy.” American Economic Review 105(4): 1339–1370. Nordhaus, W. D. and J. Boyer (2000). Warming the World: Economic Models of Global Warming. Cambridge, MA: MIT Press. Ostrom, E. (2009). “A Polycentric Approach for Coping with Climate Change.” Working Paper 5095, World Bank. Posner, E.  A. and D.  Weisbach (2010). Climate Change Justice. Princeton, NJ: Princeton University Press. Schokkaert, E. and J. Eyckmans (1998). “Greenhouse Negotiations and the Mirage of Partial Justice,” in Global Environmental Economics, ed. M. H. Dore and T. D. Mount. Oxford: Basil Blackwell, 193–217.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Pareto Improvements and Feasible Climate Solutions  369 Shue, H. (2013). “Climate Hope: Implementing the Exit Strategy.” Chicago Journal of International Law 13(2): 381–402. Skyrms, B. (2004). The Stag Hunt and the Evolution of Social Structure. Cambridge: Cambridge University Press. Southwood, N. (2016). “Does ‘Ought’ Imply ‘Feasible’?” Philosophy & Public Affairs 44(1): 7–45. Stern, N. (2007). The Economics of Climate Change: The Stern Review. Cambridge: Cambridge University Press. Tavoni, A. (2013). “Building Up Cooperation.” Nature Climate Change 3: 782–783. Victor, D. G. (2011). Global Warming Gridlock: Creating More Effective Strategies for Protecting the Planet. Cambridge: Cambridge University Press. Wiens, D. (2015). “Political Ideals and the Feasibility Frontier.” Economics and Philosophy 31: 447–477.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

16

Climate Change, Liberalism, and the Public/Private Distinction Dale Jamieson and Marcello Di Paola

In an earlier paper (Jamieson and Di Paola 2016) we explored the ways in which anthropogenic climate change and more generally the Anthropocene—this new epoch in which no earthly place, form, entity, process, or system escapes the reach of human activity—puts pressure on some central categories and concepts of lib­ eral democratic theory. In this chapter we focus on one particular distinction that is at the heart of liberal theory: that between the public and the private (hence­ forth ‘the Distinction’). Our claim is that climate change puts pressure on the Distinction in ways that are difficult for liberals to relieve. Our purpose is not to write an obituary for liberalism, but to articulate some emerging problems and to scout some possible solutions. Liberalism has shown itself to be remarkably re­sili­ ent and it may well succeed in rising to these new challenges, even if it is difficult from here to see exactly how. The chapter unfolds as follows. In Section 1 we present some basic facts about climate change and explain the notion of the Anthropocene. In Section  2 we introduce the Distinction and discuss ways in which it has figured in liberal the­ ory. In Section  3 we show how the distinction comes under pressure in the Anthropocene. In Section 4 we consider some options for relieving the pressure. Finally, in Section 5, we draw some conclusions.

1.  Climate Change and the Anthropocene In the cosmically instantaneous time of two centuries, fossil-­fueled human ac­tiv­ ities have warmed the Earth’s surface by an average of almost one degree Celsius. This is causing polar sea ice to melt, sea levels to rise, more frequent extreme meteorological events such as floods, hurricanes and droughts, and widespread eco-­systemic disruptions. Further warming poses threats to economic prosperity, water and food security, public health, local and global political stability, and other fundamental dimensions of organized human life. Despite occasional promising developments global carbon emissions have risen almost every year Dale Jamieson and Marcello Di Paola, Climate Change, Liberalism, and the Public/Private Distinction In: Philosophy and Climate Change. Edited by: Mark Budolfson, Tristram McPherson, and David Plunkett, Oxford University Press (2021). © Dale Jamieson and Marcello Di Paola. DOI: 10.1093/oso/9780198796282.003.0017

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  371 since 1992, when the nations of the Earth first committed themselves to the “stabilization of greenhouse gas concentrations in the atmosphere at a level that would prevent dangerous anthropogenic interference with the climate system.”1 The fact is that the global community still largely lacks the institutions, the technologies, and the will to decisively bend this curve. Despite its magnitude and gravity, climate change is but one part of a larger picture. Just as a fever can be seen as a sign of systemic crisis in a human body, so the rise in global temperature can be seen as a sign of transformations that are so vast and deep that they may merit the label, “epochal.”2 Because these trans­ form­ ations are primarily driven by human action, the term, ‘Anthropocene’, coined by limnologist Eugene Stoermer and popularized by chemist Paul Crutzen, is t­yp­ic­al­ly used to refer to this proposed epoch.3 In a jointly published paper, Crutzen and Stoermer (2000) claimed that humanity had become the main driver of Earth’s biological and ecological evolution. Hibbard et al. (2006) noted that the biological and geological changes initiated by human action have been subject to a “Great Acceleration” that began around 1950 and is ongoing. In 2011 Steffen et al. summarized the situation in the fol­ lowing way: In addition to the carbon cycle, humans are (i) significantly altering several other biogeochemical or element cycles, such as nitrogen, phosphorus and ­sulphur, that are fundamental to life on the Earth; (ii) strongly modifying the terrestrial water cycle by intercepting river flow from uplands to the sea and, through land-­cover change, altering the water vapour flow from the land to the atmosphere; and (iii) likely driving the sixth major extinction event in Earth history. Taken together, these trends are strong evidence that human­ kind, our own species, has become so large and active that it now rivals some

1  Framework Convention on Climate Change, Article 2, available at https://unfccc.int/resource/ docs/convkp/conveng.pdf. For global CO2 emissions, visit http://cdiac.ess-­dive.lbl.gov/ftp/ndp030/ global.1751_2014.ems; and https://www.iea.org/geco/emissions/. 2  In 2016 the Anthropocene Working Group of the International Commission on Stratigraphy (ICS), the authoritative body that sets global standards for categorizing the history of the Earth, re­com­mend­ed that the ICS recognize a new geological epoch that began around 1950 when plastics and elemental aluminum became ubiquitous, plutonium from atmospheric nuclear testing first became visible in sediments, and a spike in fly ash residue from the high-­temperature combustion of coal and oil became apparent (see https://www.theguardian.com/environment/2016/aug/29/declare-­ anthropocene-­epoch-­experts-­urge-­geological-­congress-­human-­impact-­earth). On the stratigraphic evidence for the Anthropocene, see Zalasiewicz et al. (2019); for skepticism about the Anthropocene as a geological epoch, see Finney and Edwards (2016) (Finney was Chair of the ICS at the time of writing this paper). 3  This term has also given rise to passionate controversies, often reflecting the fault lines between the “two cultures” (Snow 1959) of the sciences and humanities. Some of these controversies are dis­ cussed in Hamilton et al. (2015). For our own interventions in these debates see Jamieson and Di Paola (2016), Di Paola and Jamieson (2018), Di Paola (2017), Jamieson and Nadzam (2015), and Jamieson (2017).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

372  Dale Jamieson and Marcello Di Paola of the great forces of Nature in its impact on the functioning of the Earth system.  (Steffen et al. 2011: 843)

The development, growth, and diffusion of powerful technologies have been important in enabling the high levels of production, consumption, and popula­ tion that have transformed the planet and produced the Anthropocene. Human population, which for most of human history never exceeded more than a few million, has increased by an order of magnitude in the last three centuries. Today 7.8 billion people are organized in highly complex systems bound together by oil and gas pipelines, electrical wires, air travel, fiber optic cables, satellite connec­ tions, and cyber links. These systems empower humans in both positive and negative ways. We can save a child in a faraway land by pledging an online contribution with a few clicks; but as we do this with our computers, we tap into globally sourced energy and materials, thus activating, reinforcing, and promoting the emission of greenhouse gases that will remain in the atmosphere for centuries, etching the human face ever more deeply on the planet. The accumulation of such apparently trivial, localized, individually innocuous acts as using computers, taking hot showers, driving cars, heating homes, invest­ ing here or there, eating this or that, and having children can alter fundamental planetary systems in ways that have global consequences which are locally actual­ ized. So together, but with no unitary plan, we change the climate, drive species to extinction, and acidify the oceans—thus harming humans and non-­ human nature in faraway times and places.

2.  Liberalism and the Public/Private Distinction Many of the behaviors that contribute to climate change and other systemic global problems of the Anthropocene are not only trivial, localized, and individually innocuous, but also generally regarded as private. Yet today, these apparently pri­ vate behaviors have public consequences, however indirect, across spatial, tem­ poral, and genetic boundaries. It is this that brings the Distinction under suspicion in the Anthropocene. Through the centuries, the Distinction has been used in various ways to help mark boundaries as diverse as those between general vs. particular interests (e.g. Rousseau 1762/1985); the administrative state vs. the market economy (e.g. Smith  1776/2007); the realm of the ethical, the external, the objective vs. the realm of the natural, the internal, the subjective (e.g. Hegel 1820/1991); the par­ ticipatory self-­determination of citizens vs. their withdrawal from civic affairs (e.g. Tocqueville 1835/2008; Dewey  1927/1954; Arendt  1958; Habermas 1962/1989); the world of sociability vs. that of domesticity (e.g. Sennett  1977;

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  373 Ariès and Duby 1992); the administrative state and the market economy vs. the family (e.g. Pateman 1985, 1989); and privacy vs. exposure (e.g. Nagel 1998).4 The roots of the Distinction can be traced at least to Aristotle’s distinction between the domain of politics (polis) and the domain of the household (oikos).5 The Romans contributed to the development of the Distinction by articulating a system of private law that provided the foundations for much of modern European contract and property law. The Distinction began to take modern form with the emergence of nation-­states in the sixteenth and seventeenth centuries, which led to the configuration of a distinctly public realm and, at the same time, a reaction against the claims of monarchs or parliaments to unconstrained power. During this period popular claims to private spheres insulated from state juris­ diction became increasingly powerful. Religious conscience and rights to prop­ erty (which natural law theorists such as John Locke took as their main concerns) were seen as central to the private domain (Horwitz 1982: 1423). The Distinction took on special depth and meaning with the rise of liberalism in the eighteenth and nineteenth centuries.6 Thomas Nagel characterizes liberal­ ism in the following way: Liberalism is the conjunction of two ideals. The first is that of individual liberty: liberty of thought, speech, religion, and political action; freedom from govern­ ment interference with privacy, personal life, and the exercise of individual in­clin­ation. The second ideal is that of a democratic society controlled by its citi­ zens and serving their needs, in which inequalities of political and economic power and social position are not excessive.  (Nagel 1975: 136)

Most contemporary liberals embrace both ideals, justifying them on broadly con­ sequentialist or Kantian grounds. Libertarians endorse only the first ideal, some­ times providing a natural rights foundation rather than justifying it on consequentialist or Kantian grounds.7 The first ideal specifically invokes privacy, and a distinction between the public and private is clearly implicated in most of the domains that Nagel mentions: speech, religion, personal life, the exercise of individual inclination, and so on. Realizing the second ideal may also require observing the Distinction, but making out this claim would require further 4  This list of (related but categorically different) pair-­wise oppositions is far from exhaustive. Two books that map much of the terrain are Benn and Gaus (1983), and Weintraub and Kumar (1997). 5  Arendt (1958). Generally, on the history of the Distinction see Horwitz (1982), Bobbio (1989), Geuss (2001), and Mahajan (2009). 6  Geuss (2001: 1–5) credits Humboldt and Constant as particularly important in developing the Distinction within the context of liberal theory. 7  The classic natural rights defense of libertarianism is in Nozick (1974). Consequentialist defenses of libertarianism have been most influential in the work of economists such as Hayek (1960). There are Kantian libertarians though they are not easy to find (but visit https://www.libertarianism.org/ publications/essays/kantian-­case-­libertarianism).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

374  Dale Jamieson and Marcello Di Paola argument. In any case it is clear that all forms of liberalism endorse the first ideal and with it the Distinction. The boundaries between the public and the private became entrenched in lib­ eral societies by reciprocal empowerment. On the one hand liberals called for making people’s lives, consciences, and property more securely private. On the other hand, any government entrusted with the task of overseeing and regulating a complex liberal society had to be especially powerful. Fairly enforcing the rule of law often required more rather than less involvement with citizens’ private lives in managing files, issuing certificates and permits, keeping records and accounts, and generally tending to the bureaucracy of a modern state. The tension between a robust private sphere and the pervasive role of public authorities in keeping track of people’s lives remains an object of ongoing negotiation and a source of political division in liberal societies.8 While the Distinction can be drawn in different ways and has taken different forms at various historical moments, it is hard to imagine any recognizable form of liberalism without the Distinction playing an important role. As Stanley Benn writes: The conception of privacy is closely bound to the liberal ideal. The totalitarian claims that everything a person does has significance for society at large . . . The public or political universe is all inclusive.  (Benn 1988: 268)9

Similarly, Judith Shklar remarks: [Liberalism] must reject only those political doctrines that do not recognize any difference between the spheres of the personal and the public . . . Because of the primacy of toleration as the irreducible limit on public agents, liberals must always draw such a line . . . [T]his must under no circumstances be ignored or forgotten.  (Shklar 1989: 24)

The Distinction runs through such familiar liberal themes as the value of indi­ vidual liberty, which must be protected from unjustified interference from others; the rule of law, which separates public law (constitutional, criminal and 8  The rise of social media and other privately owned digital platforms that gather data about their users and reconstruct their propensities with a precision that may be politically pernicious complicates this negotiation significantly. The issue of what is private and what is public in the cyber-­cloud that we all now inhabit opens another set of challenges to the Distinction, which we do not address here. 9  As DeCew (2015) points out, “Privacy can refer to a sphere separate from government, a domain inappropriate for governmental interference, forbidden views and knowledge, solitude, or restricted access, to list just a few”—in other words it can refer to the private domain, and it is clear from context that this is how Benn is using the term here. The American tradition of legal philosophy that descends from Warren and Brandeis (1890) tends to have a narrower view of privacy, viewing it as a single interest or bundle of related interests.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  375 regulatory) from private law (the law of contracts, torts, transactions and com­ merce); human rights, which hinge on the recognition that individuals have inde­ pendent moral status that governments are required to respect; secularism, which (among other things) distinguishes public political debate from private religious belief; and capitalism, the economic system whose development is associated with the rise of liberalism and which depends on the recognition and enforce­ ment of private property rights. Despite its ubiquity and obvious importance to liberals, it is surprisingly diffi­ cult to find a “locus classicus” of the Distinction, or even a clear statement and explicit defense. Indeed, often it has been liberalism’s critics who have been most forthcoming about the importance of the Distinction to liberalism. Karl Marx regarded the Distinction to be at the center of liberal consciousness and an important source of alienation, since it requires us simultaneously to view our­ selves and others both as citizens and antagonistic private actors: man leads a double life . . . [I]n the political community he regards himself as a communal being; but in civil society he is active as a private individual, treats other men as means, and becomes the plaything of alien powers. (Marx 1843/1978: 28; emphasis in text)10

Some feminist critics have seen the Distinction as both at the center of liberalism and also deeply entwined with feminism itself. Carole Pateman writes: The dichotomy between the private and the public is central to almost two centuries of feminist writing and political struggle . . . Although some feminists treat the dichotomy as a universal trans-­historical, and trans-­cultural feature of human existence, feminist criticism is primarily directed at the separation and opposition between the public and private spheres in liberal theory and practice.  (Pateman 1983: 281)11

Although liberal theorists seldom explicate or explicitly defend the Distinction, they often seem to presuppose some version of it. It is in the background of Robert Nozick’s (1974) and Ronald Dworkin’s (1977) discussions of the strength (or priority) of individual rights versus other goods such as equality or general welfare. It is present in Gerald Cohen’s (1995, 2000) discussions of the extent to which people’s personal decisions should be governed by principles of justice. It is central to the debate between H. L. A. Hart and Patrick Devlin about the

10  For discussions of the Distinction from a broadly Marxist perspective see the essays collected in Mouffe (1996). 11  See also Lloyd (1984).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

376  Dale Jamieson and Marcello Di Paola extent to which the state can legitimately enforce community norms and v­ alues (Hart 1963). In Political Liberalism John Rawls actively distances himself from the Distinction. He embraces the “public/non-­public” distinction but insists that this “is not the distinction between the public and the private” (1993/2005: 220). Yet as Charney (1998) has noted, Political Liberalism is innerved by contrasts that seem ultimately to hinge on some version of the Distinction: a public conception of justice vs. “comprehensive doctrines”; political values vs. those that apply to “all of life”; and modes of reasoning that apply to public deliberation (i.e. “public rea­ son”) vs. those that animate “the background culture” and pertain to individuals, families and what Rawls elsewhere (1971: 520–529) described as “social unions” (such as churches, universities, and clubs). Perhaps Rawls’s rejection of the Distinction can be explained in the following way.12 Rawls is interested in reasons, and he recognizes different kinds of reasons each governed by its own rules, guided by its own distinctive values, and appro­ priate to its own domain. For Rawls, there is a domain of constitutional essentials, matters of basic justice, institutions and institutionalized roles and procedures; and another domain of individual preferences and voluntary associations. Public reason is appropriate to the first domain, and what Rawls calls “non-­public” ­reasons are appropriate to the second. Public reason requires that citizens deliber­ ate in the public political forum in terms of reasons that can be shared by all free and equal citizens, engaged in a fair system of cooperation.13 However, in the domain of individual beliefs, preferences, and associational and market choices, justificatory reasoning can be more sectarian and narrow. Praying, for example, can be justified by non-­public reasons that are afforded by a comprehensive doc­ trine (in this case a religion), which provides a reservoir of reasons sourced from a background tradition that is common to its members. For Rawls it is important to recognize that all reasons, even non-­public reasons, are “social.” He writes (1993/2005: 220) that “there is no such thing as ‘private reason’” and that “these [non-­public] reasons are social, and certainly not private.”14 The source of Rawls’s resistance to the Distinction may lie in his fear of conflating non-­public reasons with non-­social reasons. 12  This paragraph has been informed by correspondence with Samuel Freeman, Gerald Gaus, and Leif Wenar. Of course, they are not responsible for what we say. 13  This is part of what Rawls calls “the duty of civility” (1993/2005: 217). The duty of civility is a non-­legal moral duty citizens have to “be able to explain to one another on [. . .] fundamental ques­ tions how the principles and policies they advocate and vote for can be supported by the political val­ ues of public reason. This duty also involves a willingness to listen to others and a fair-­mindedness in deciding when accommodations to their view should reasonably be made” (Rawls 1993/2005: 217). 14  Although he does not mention Wittgenstein in this regard, we speculate that Rawls’s rejection of private reason is related to Wittgenstein’s rejection of private language. Recent scholarship has empha­ sized Wittgenstein’s influence on Rawls’s thought (see, e.g. the symposium on Rawls’s archival papers in the Journal of the History of Ideas 78(2) (April, 2017)).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  377 However, Rawls’s insistence on the social nature of reasons does not justify the rejection of the Distinction, at least as some have drawn it. John Dewey (1927/1954: 13), for example, regarded the Distinction as fundamental, yet emphatically stated that “the distinction between the public and private is thus in no sense equivalent to the distinction between the individual and social.” It is hard to see why Rawls could not have endorsed this view and thus the Distinction.15 Rawls’s focus on reasons is exemplary of the way in which many liberals tend to read the Distinction.16 They distinguish the public and private domains by identifying reasons appropriate to each. However, there is an obvious alternative to this focus on reasons as the hallmark of the Distinction. Many of the (few) philosophers who explicitly embrace (or discuss) the Distinction focus on actions rather than reasons, and look to Mill as the god­ father of the Distinction.17 In On Liberty Mill writes that: The only part of the conduct of any one, for which he is amenable to society, is that which concerns others. In the part which merely concerns himself, his inde­ pendence is, of right, absolute. Over himself, over his own body and mind, the individual is sovereign.  (Mill 1859/1978: 9)

He formulates this idea in what has come to be known as the Harm Principle (HP): The object of this essay is to assert one very simple principle . . . the only purpose for which power can be rightfully exercised over any member of a civilized com­ munity, against his will, is to prevent harm to others.  (Mill 1859/1978: 9)

Dewey echoes the HP in his characterization of the Distinction: . . . the consequences [of an action] are of two kinds, those which affect the per­ sons directly who are engaged in the transaction, and those which affect others beyond those immediately concerned. In this distinction we find the germ of the distinction between the private and the public.  (Dewey 1927/1954: 12)18

15  It is perhaps especially ironic in that Political Liberalism began life as the 1980 Dewey Lectures at Columbia. For Rawls’s engagement with Dewey, see Botti (2017). 16  The focus on reasons in contemporary moral and political philosophy (rather than e.g. linguistic or other metaphysical constructions) is a large and complicated topic that we cannot discuss here. 17  For example, Geuss (2001: 81); Gaus (1996: 171); and Rorty (1989). See also DeCew’s (2015) Stanford Encyclopedia of Philosophy entry on “Privacy” (https://plato.stanford.edu/entries/privacy/). 18  Notice that Dewey’s language of affecting is broader than Mill’s language of harming. However, Mill also recognized that actions can have positive “spillover” effects that can justify state action (e.g. the provision of free public education); see note 21 below.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

378  Dale Jamieson and Marcello Di Paola Rorty, who writes that “Mill’s On Liberty provides all the ethical instruction you need” (1998: 24) and that Mill was the “last word” in Western political theory (1989: 36), regarded the Distinction as the beating heart of liberalism. Mill was concerned with actions rather than domain-­appropriate reasons. As a rough and ready guide, we can take the HP as suggesting that actions count as private as long as they do not harm others. By contrast, actions that harm o ­ thers are public and thus may (at least prima facie) be liable to sanctioning and regulation. Following Mill, we can thus think of the private domain as constituted by actions that cause no harm to others, and the public domain as constituted by those actions that do cause harm to others. There are many complications, of course. Despite his sometimes categorical language, it is clear that Mill did not think that the HP provided a sufficient con­ dition for the rightful exercise of power over an individual (cf. Mill 1859/1978: 73). Nor did Mill think that the HP specified even a necessary condition.19 Moreover, as we have noted, Mill’s project was importantly different from that of liberal theorists such as Rawls: Mill was concerned with actions rather than domain-­appropriate reasons. Still, whatever we think about these (and other) questions, Mill’s basic insight remains central to liberalism: It is surely a mark of a liberal state that it largely keeps its nose out of its citizens’ harmless behavior, whether it involves counting blades of grass or collecting comic books. For this reason and because of its influence, despite the complications and dangling details, we will take Mill’s insight as a valuable guide to a liberal understanding of the Distinction. To summarize: Liberals aspire to a world in which governments respect the personal beliefs and choices of individuals, the life of the family, and voluntary exchanges between consenting adults. While some degree of public intervention into the private sphere is permitted when it is required for collective security (e.g. national defense) or solves an urgent coordination problem (e.g. traffic regu­la­ tion), the threshold for such intrusion is high. On one particularly influential view, articulated by Mill, public intervention into the lives, choices, opinions, actions, behaviors, and practices of individuals is only justified when it prevents harm to others. When it comes to believing in a god, eating meat, having chil­ dren, or investing in property, most liberals would take the old 1920s blues song as normative: “Ain’t nobody’s business if I do.”20

19  See Brink (2013: 183–187), and note 21 below. 20  For Bessie Smith’s 1923 recording, visit https://www.youtube.com/watch?v=2cngx_KKiWE. For a particularly awesome version, see Freddie King’s at https://www.youtube.com/watch?v=FILEQ8guI1c.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  379

3.  The Distinction Under Pressure In this section we suggest that climate change (the harbinger of the Anthropocene) threatens the viability of the Distinction, both by intensifying old tensions and by bringing new pressures to bear. Technological development and the spatial collapse that it engenders has been a constant pressure on the Distinction, which the Anthropocene intensifies. The world in which we now live, where even our most mundane everyday behaviors such as using a computer and flipping a light switch have global ripples, puts pressures on the Distinction that seem qualitatively different in scale and ubiquity. These intensifying tensions and novel pressures provoke questions about whether liberalism can survive the Anthropocene; if so, in what form; and what may be the role of the Distinction in this new epoch. The central challenge stems from the fact that when it comes to climate change and other problems of the Anthropocene, actions that are traditionally regarded as private—such as using computers, flipping light switches, taking hot showers, eating this or that, driving cars, investing here or there, and having children— contribute to disrupting planetary systems. These disruptions will harm not only those who contribute to them (generally not in proportion to their contributions), but also inhabitants of distant nations, future generations, and other species. For this reason it might be thought that actions such as taking a hot shower should be regarded as public, at least by Millian lights, and thus subject to regulation. One way of making this argument appeals to the theory of externalities as it has been developed in economic thought. The basic insight on which this appeal rests goes back to Mill, who noticed the phenomenon of positive externalities, to Sidgwick who discussed negative externalities in detail, and to Pigou who developed the general concept in a more systematic (if ambiguous) way.21 The observation on which the theory rests is that the benefits and costs of the production and consumption of a good or service are not always confined to the producers and consumers of the good or service. Sometimes they are “external­ ized” to third parties who are not parties to the transaction. For example, suppose that a widget factory produces pollution that soils Mr. Smith’s clothes when he hangs them out to dry. The cost to Mr. Smith of his soiled clothes in not reflected

21  Mill’s argument for the public provision of education was based in part on his recognition of the positive externalities that education provides to society as a whole (see his 1848 Principles of Political Economy with some of their Applications to Social Philosophy, Book V, Chapter XI (“Of the Grounds and Limits of the Laisser-­faire or Non-­Interference Principle”)). Sidgwick discussed the negative externalities related to the depletion of natural resources, the diversion of waterways, and the neglect of the interests of future generations in his 1883 The Principles of Political Economy, Book 3, Chapter 2 (“The System of Natural Liberty Considered in Relation to Production”) and Chapter 4 (“Important Cases of Government Interference to Promote Production”). Pigou’s classic treatment is in his 1920 The Economics of Welfare, Part 2, Chapter 9 (“Divergences Between Marginal Social Net Product and Marginal Private Net Product”). Interestingly, none of these figures actually used the term ‘externality’. The earliest use of the term that we have been able to find is in Bator (1958: 351).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

380  Dale Jamieson and Marcello Di Paola in either the widget manufacturer’s costs or in the price that consumers pay when they buy a widget. Mr. Smith is effectively contributing to the production and consumption of widgets without capturing the benefits. This distorts investment and consumption decisions and leads to suboptimal societal outcomes, in this case the oversupply and overconsumption of widgets. The solution is to “internal­ ize” the cost of Mr. Smith’s soiled clothes so that it is borne by the producers and consumers of widgets. There are various ways of doing this, such as regulating pollution levels from the widget factory, taxing widgets to reflect the cost of the pollution, or granting Mr. Smith property rights which would entitle him to com­ pensation for his soiled clothes. If we implement some such solution, the theory goes, we will create incentives to reduce pollution to the “optimal” level at which further reductions are more costly than the benefits realized from the reductions. This theory may be used to suggest that actions that are typically regarded as private but contribute to climate change and other problems of the Anthropocene should be regarded as public, since they produce negative externalities. This is the rationale for carbon taxes, cap and trade, regulations reflecting the social cost of carbon and so on. On this view the problem is not with the Distinction, but that we have failed to adopt policies that adequately police the boundary between the public and the private. This is an elegant theory that even works in practice in some cases. However, it has some problems of its own, and when it comes to actions that contribute to climate change and other problems of the Anthropocene, its theoretical in­firm­ ities and practical frailty are very much on display. The first problem is that rather than providing a foundation for the Distinction, the theory of externalities actually requires something like the Distinction in order to identify externalities in the first place. This is because almost everything we do affects others, often adversely, but we do not regard all of these effects as externalities. When I walk down the street my fashion choices sometimes appall other people. Whether I bike or drive affects traffic flows which affect everyone on the road. My opinions expressed in the morning newspaper make some ­people nauseous. But none of these effects are ordinarily regarded as externalities. While this is not the place to offer a full account of which effects count as externalities and which do not, we can say as a first approximation that for an effect to count as an externality, it must be considered significant and must result from actions that the actor does not have a presumptive right to undertake. Consider the following example.22 Kelly has a taste for death metal music which is not shared by Kelly’s neighbors. Their response to the music spilling out of Kelly’s house ranges from the irritated to the offended. Some neighbors are merely annoyed, while others would claim to be harmed. The jurisdiction in

22  This and the following paragraph reprise arguments found in Jamieson (2002: ch. 17, section 3).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  381 which Kelly lives, like most jurisdictions in the United States, allows Kelly to prod­uce sounds of whatever kind up to a particular decibel level, sustained over a particular period of time, as measured at particular locations. As long as these thresholds have not been exceeded, the sounds produced by Kelly that reach the ears of the neighbors are not an externality, no matter how aggrieved the neigh­ bors may be and how much they may hate Kelly’s taste in music.23 What matters in this case is the objective characteristics of the sounds rather than how they are experienced by hearers. Our prevailing norms protect Kelly’s right to produce sounds that make other people miserable, so long as Kelly observes some ob­ject­ ive boundaries. It is natural to explain this by saying that what counts as an exter­ nality presupposes something like a private sphere within which an individual is entitled to act. In this case (and others), something like the Distinction is concep­ tually and epistemologically prior to the identification and characterization of a negative externality. Many other examples can be given. While all presuppose something like the Distinction, what is striking is how different their shapes can be, which reveals how deeply contextual the idea of an externality is. For example, if we regard the noise that Kelly is producing as resulting from a socially beneficial (or even “nor­ mal”) activity, we may permit greater noise exposures than if we do not, especially if it occurs at a socially sanctioned time (i.e., the times that “decent” people are normally awake). If Kelly were renovating the house, mowing the lawn, or using a weed whacker or leaf blower, we might well permit much greater sound exposure than in the case of Kelly playing death metal music. Indeed, the neighbors might feel much less affronted by damaging noise levels coming from these beneficial (or “normal”) activities than they do from music that they hate, even if it is at a lower decibel level. Renovation and lawn work maintain or improve property val­ ues. The neighbors’ toleration of the negative effects of these activities is reminis­ cent of the old saw that the smell of pollution is “the smell of money.”24 Like the boundaries of the Distinction itself, what we regard as a negative externality is malleable. In many communities today the smell of pollution is no longer experienced as the smell of money.25 Second-­hand smoke was once seen as 23  It might be objected that the effects of Kelly’s music are externalities but not actionable externali­ ties. This response provides no way out since something like the Distinction would then be needed for distinguishing externalities from actionable externalities. 24  What this case shows, it might be said, is that the positive externalities of weed whacking are at least equal to the negative externalities, so weed whacking is accepted and we do not treat it as impos­ ing externalities. Fully addressing this claim would take us too far afield, but suffice it to say that in this case the normative content supplied by something like the Distinction is being imported into the evaluation of the economic consequences of the act in question, and this carries more than a whiff of adhocery. Moreover, we don’t ordinarily treat these as cases in which positive and negative externali­ ties are balanced; we treat them instead as cases in which the theory of externalities does not apply because we are within our rights to operate a weed whacker, at least within certain hours. 25 See, e.g. https://www.desmoinesregister.com/story/opinion/readers/2016/11/11/smell-­ money­isnt-­funny-­anymore/93451068/.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

382  Dale Jamieson and Marcello Di Paola an unremarkable consequence of smoking; it is now seen as a negative externality. In most jurisdictions, Kelly’s dancing naked in Kelly’s apartment with the window open “ain’t nobody’s business,” even if Kelly’s prudish neighbor is outraged.26 But this was not always so in the United States or Europe, and may not always be true in the future. To summarize, externalities are not simply third-­party effects nor even third-­party effects of a particular kind. Which third-­party effects are externalities is complex and dynamic, and depends on prevailing social norms. Rather than providing a foundation for the Distinction, the theory of externalities relies on many of the same resources as the Distinction. This point seems to be tacitly understood by many people, though it eludes some academics and policy makers. In some countries there is deep resistance to climate policies that are rationalized by the theory of externalities. Even where these policies have been implemented (e.g. the European Union), they do not reflect the full costs of emitting carbon.27 Even when such policies are accepted, they are typically viewed as just another tax rather than as part of the price of the commodity or service. There is a further (and even deeper) conceptual difficulty with applying the theory of externalities to actions that contribute to climate change. Actions such as taking hot showers, eating this or that, driving cars, investing here or there, and having children do not in themselves cause damages. Vast, complex, multi-­scalar physical and social systems mediate between my hot shower and the deaths and degradation that result from climate change. The emissions produced by my hot shower accumulate with those of the other billion hot showers in the world, travel through space–time, disperse into the workings and feedbacks of various physical and chemical systems operating at different scales, and at no point will the emis­ sions provoked by my shower cause any particular flood, drought, or hurricane, much less the loss of life and property. My hot shower contributes to these harms but it does not cause them. This is because of the mass of variables that intervene between my hot shower and the harms; the complex relations between the vari­ ous elements in the behavioral/physical systems that are involved; the massively different scales at which my hot shower, climate change, and the ensuing harms occur; and how all of this interacts with the concept of causation.28 26  Virtually all liberal jurisdictions recognize the right of individuals to be naked in their own houses but limit it in various ways, sometimes taking intent to be important and other times reconfig­ uring the Distinction in such a way that under certain conditions acts performed on one’s own prop­ erty are regarded as public. There are interesting issues here that we cannot explore in detail. 27  It should also be recognized that economists disagree on the social cost of carbon by an order of magnitude or more. For recent discussions see, e.g., Hansel et al. (2020), Kent et al. (2019), and Scovronick, et al. (2017). 28  This is a controversial claim put forward in Jamieson (2014a: ch. 5) and earlier in Sinnott-­ Armstrong (2005). A full explication and defense would have to provide precise understandings of such notions as contribution and causation, taking us well beyond the bounds of this chapter.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  383 Our prevailing conception of causation limits the relations that can obtain between different levels of organization—for example, between micro-­level pro­ cesses such as carbon fluxes and macro-­level events such as human migrations. Causation, as we currently conceive it, does not nimbly leap over so many levels of organization. In this respect the situation is analogous to the case in which someone might claim that a causal relation exists between some quantum flux and my desire to read Finnegan’s Wake. Some relation may exist, but not a causal relation: our concept of causation just does not work that way. “A quantum flux made me do it” does not figure in the explanation of human action. However, concepts change and speakers often engage in “persuasive definition” (Stevenson 1938) in order to get us to establish new relations between things and to see exist­ ing relations in different ways. Rather than being thought of as making sober claims to truth, those who claim level-­leaping causal relations between my taking a shower and some individual person being forced to migrate by climate change are better thought of as engaging in a normative enterprise. They are trying to get us to see things in a new way: to see a causal relation where before we could not see one. Perhaps they will succeed, and we will revise our concept of causation in the way they prefer. Or perhaps we will jettison the concept of caus­ation ­altogether. At this point nobody knows. What we can say here and now is that what is true of my emissions is true of everyone else’s emissions as well. The non-­linear, multi-­level causal rollercoaster that goes from emissions to climate change and from climate change back to harms and damages does not allow specific instances of the latter to be causally imputed (even in part) to any particular agent. Since my hot shower contributes to climate change but does not cause it, perhaps it should count as private after all and not subject to regulation.29 One way of thinking about cases in which apparently private actions contrib­ ute to harmful public consequences is to think of these actions as having two lives.30 They have an episodic life, occurring when these actions are performed, which is harmless and thus private (e.g. me taking a hot shower). But such actions also have a systemic life: the moment I turn on the hot water, an entire “infra­ structure of provision”31 that presides over the global procurement and distribu­ tion of energy is activated, whose workings require the exploitation of limited resources, cause habitat disruptions of various magnitudes, and pump greenhouse gases into the atmosphere. The infrastructure of the Anthropocene—the global energy system, the global food system, the global transport system, etc.—is arranged in ways that engender spatiotemporally and genetically unbound harms and damages. When I turn on the hot water, start a car or flip a light switch I  become implicated in these infrastructures and attendant global networks of 29  This should not be taken to imply that we cannot hold agents morally or legally responsible for their contributions on other grounds. For discussion in the context of climate change, see Lazerus et al. (2020). For a more general discussion, see Jamieson (2016). Thanks to Michael Morck for prompting this note. 30  Di Paola (2017: 24-­25). 31  See Southerton et al. (2004); Van Vliet et al. (2005).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

384  Dale Jamieson and Marcello Di Paola eco-­altering financial interests, political agreements and avenues of cultural reinforcement, and contribute to their harmful outcomes. This systemic life of our actions is as real as their episodic life: climate change would not occur with­ out all these cars being driven, hot showers taken, switches being flipped and other relevantly similar actions being performed. By performing these episodic actions that also have a systemic life, we partake in, and further propel, a ­globalized, fossil fueled system that changes climate and thus engenders harms and damages to people and ecosystems across space and time. From this perspec­ tive we seem to be left with an irreducible pluralism: some actions can be seen as both episodic and systemic, and so as both private and public. It is not easy to see how this coheres with the Distinction and its centrality to liberalism. Climate change and the Anthropocene also put pressure on the Distinction by marginalizing or even undermining its traditional justifications. According to Mill, the justification of the Distinction rests on utilitarian grounds: Utility [is] the ultimate appeal on all ethical questions; but it must be utility in the largest sense, grounded on the permanent interests of man as a progressive being.  (Mill 1859/1978: 10)

Mill thought that free ranging harmless individuality satisfies this criterion since it leads to a rich variety of opinions and ways of life, which in turn open a wide spectrum of possibilities for societal development, encourage the search for knowledge, refine institutions, and sustain lifestyle experimentation. In addition to benefitting individuals, society at large benefits from individuals doing their own (harmless) thing (Mill 1859/1978: 67–69). However, in the Anthropocene, for most citizens of liberal democracies, doing one’s own thing has come largely to mean heavy consuming. Today, affluent citi­ zens are free to choose, from an ever increasing range of possibilities, products that maximally satisfy whatever their consumer preferences may be. But food, clothing, vehicles, energy, and many other consumer goods and services, as cur­ rently provided by globalized supply chains, are typically obtained through pro­ cesses that deplete natural resources, pollute environments, and produce greenhouse gas emissions. Such unfettered consumer behavior may contribute to the immediate well-­being of individual consumers, but that was not what Mill’s conception of privateness was primarily meant to protect. Mill’s private sphere was one of opinions and practices relating to “experiments in living” that would promote the general good of mankind. It was not a space of unrestrained con­ sumption for its own sake. In fact, pursuing diversity- and innovation-­enhancing “experiments in living” in the Anthropocene, particularly in market-­enthusiastic liberal democracies, may mean not consuming what is most insistently offered by markets: it may instead mean growing one’s own food, recycling clothes and materials, sharing transportation, and so on.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  385 Since most private acts of consumption that contribute to climate change and other problems of the Anthropocene are unlikely to be diversity- and innovation-­enhancing in the ways valued by Mill, it is doubtful that these every­ day consumer behaviors are worthy of protection on Millian grounds.32 Surely they are less ­worthy of protection than diversity- and innovation-­enhancing pri­ vate “experiments in living.”33 Indeed, a private sphere that protects actions that have the kinds of systemic consequences that we have described may not only fail to provide what Mill regarded as the most important benefits of a protected pri­ vate sphere, but may undermine the conditions that have made possible human life as we know it. Respect theorists in the Kantian tradition ground the protection of a private sphere on the value of autonomy and the dignity of our individual capacities for practical reason. For them the Distinction can be seen as both reflecting the value of autonomy and contributing to the maintenance and development of our cap­aci­ties for practical reason. However, in the Anthropocene, by protecting the heavy consuming that threatens to disrupt the planetary systems on which our familiar ways of life depend, the Distinction may contribute to undermining the conditions required for practical reason’s concrete expression in liberal political communities.34 Rawls (1993/2005) thought that “the fact of reasonable pluralism” was a “permanent feature of the public culture of a democracy” (p. 36) and, while it functioned as a bastion against authoritarianism and institutionalized oppression (p. 37), it also represented the greatest threat to the stability of liberal communi­ ties, since comprehensive doctrines can be “conflicting and irreconcilable – and what’s more, reasonable” (p. 36). Rawls’s greatest concern was that diversity and discordance in beliefs could lead to stability-­threatening diverse and discordant behaviors. However, the behaviors that drive the problems of the Anthropocene are not diverse and discordant, but mostly quite uniform; and rather than grounded in significant beliefs about how to live, they largely express consumer preferences that are baked into the practices of everyday life. Whatever our religious, cultural, and philosophical persuasions, we all cook and eat; and even 32  Worries have been voiced many times over, and from many different quarters, that the consum­ erist culture that dominates the globalized world of the Anthropocene has powerful homogenizing (rather than diversifying) effects on people’s choices and lifestyles—thus depleting the private sphere of much that is most important and exciting about it. See, e.g., Marcuse (1964); Ewen (1976); Ewen and Ewen (1982); and Dawson (2004). For a pugnacious version of this argument, according to which anti-­democratic forces have used consumerism to hollow out public life by colonizing the private sphere through market channels, see N.  Chomsky’s 2017 online interview at https://www.youtube. com/watch?v=GYoKRS_eWZY. 33  The point made in this paragraph applies to industrialized liberal democracies especially. In the case of developing countries, consumption may more closely track human development and Mill’s cherished “experiments in living.” 34  In particular, destabilizing these planetary systems may threaten liberal democracy. For further discussion, see Di Paola and Jamieson (2018).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

386  Dale Jamieson and Marcello Di Paola back-­to-­nature en­vir­on­men­tal activists have children. In the Anthropocene even these basic human activities are enabled by and require the workings of global infrastructures of provision that deplete, pollute, and emit. It is agreement in these widely shared, relatively belief-­independent behaviors—not diversity and discordance in beliefs leading to diverse and discordant behaviors—that is at the root of the problems of the Anthropocene. It is the shape of everyday life, with its material workings as presently arranged, that is remaking the planet. In an emissions-­free world the disagreements in belief and discordant behaviors that Rawls was concerned about would never have undermined planetary systems in the way we are now experiencing. In sum, climate change and other problems of the Anthropocene put pressure on the Distinction in two important ways. First, they exacerbate the double life of apparently innocent private episodic actions as their systemic dimensions become increasingly prominent and threatening. Second, actions that contribute to climate change and the Anthropocene are often not the kinds of actions whose privateness is particularly worth defending by liberal lights. They do not enhance diversity in desirable ways, and because they are typically relatively belief-­independent acts of consumption they are not exactly paradigms of what we might hope for from the deliverances of practical reason. Respecting the private sphere may protect the dignity of practical reason, but in protecting unbridled consumption it may also lead to the disruption of fundamental planetary systems on which human life depends. Liberalism is worth defending, but it may seem that at least in its present forms it is made for a different world than that of the Anthropocene.

4.  Pressure Drop? We have claimed that in the Anthropocene, actions such as taking a hot shower have an episodic life that is innocuous and apparently private, yet they also have a systemic life that contributes to harms across spatial, temporal, and genetic boundaries and thus apparently public. This systemic life is enabled by the work­ ings of globalized infrastructures of provision powered by fossil fueled technolo­ gies. Most of us living in liberal societies forget or ignore that many of our actions have this systemic life, and we strenuously defend our jurisdiction over what we buy, what we eat, what we invest in, and how many children we have. This seems to present us with the following dilemma. If we focus on the sys­ temic life of actions, then each of these actions contributes to spatiotemporally unbound harms and therefore counts, at least by Millian lights, as public. If we focus instead on the episodic life of actions, then each of these actions is discretely harmless, so therefore count, at least by Millian lights, as private. Seeing these actions as public conjures up the specter of a liberal state regulating hot showers,

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  387 driving cars, heating homes, eating meat, having babies and other behaviors ­generally regarded as beyond its reach. Seeing these actions as private, on the other hand, suggests that we should permit everyday behaviors that are known to contribute to harms and damages of planetary proportions. Let us see first whether the horns of this dilemma can be dulled. One way of releasing the pressure that climate change and the Anthropocene put on the Distinction is to regard every action with a systemic life as public and open to regulation. Matters of personal hygiene, diet, housing, investment, mobility, and reproduction would thus be seen as matters of public decision-­making. This would dissolve the dilemma, but amount to the abandonment of liberalism as we know it: rather than “nobody’s business if I do,” it would be “everybody’s business if I do.” Moreover, there is no guarantee that such an approach would succeed in preventing behaviors that contribute to climate change and the Anthropocene. Illiberal states sometimes succeed in suppressing behavior that a liberal state would consider private (e.g. China’s censorship of the internet), but in other cases illiberal states can’t even control behavior that clearly causes harm to others (e.g. China’s failure to prohibit fireworks during Spring Festival).35 It might be objected that we have exaggerated what it would mean to regard every action with a systemic life as public and open to regulation. Public inter­ vention occurs on a spectrum, ranging from relatively unobtrusive gasoline taxes, for example, to coercive restrictions on reproduction. True enough, but it should be recognized that in some countries even modest interventions are regarded as unacceptable governmental overreach.36 Moreover, the kinds and scale of the interventions that would be required to significantly reduce greenhouse gas emis­ sions are much more severe than paying a few cents more per gallon for gasoline. They would include, for example, having fewer children, living car-­free, avoiding airplane travel, and moving towards a plant-­ based diet (Wynes and Nicholas 2017). Even on further reflection, the dilemma remains intact and the first horn very sharp. Dulling the second horn may seem more promising. We could discourage actions that contribute to harms of planetary proportion by moralizing them even while continuing to regard them as private.37 This is arguably the strategy that we pursue with some other actions. We moralize such apparently private matters as watching pornography, holding grudges, dropping out of college, having this or that sexual preference, or devoting time to lysergic acid

35  https://www.ft.com/content/371a8ff8-­e440-­11e6-­8405-­9e5580d6e5fb?mhq5j=e1. 36  We are thinking of course of the United States. Survey data suggests that for most of the last two decades Americans have thought that there was too much government regulation of business (visit https://news.gallup.com/poll/220400/americans-­views-­government-­regulation-­remain-­steady.aspx), and this has generally been reflected in decisions by the US Supreme Court. 37  This is suggested by John Broome (2012) who believes that, as a matter of private morality, we should offset all of our carbon emissions. For a critical response see Jamieson (2014b).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

388  Dale Jamieson and Marcello Di Paola experimentation. We pass moral judgment on smoking, even though it is rare for an individual first-­hand smoker (who does not live with a second-­hand smoker) to cause the suffering or death of an individual second-­hand smoker (and almost never on the spot). The moralization of actions goes beyond the extent to which they cause or contribute to harms, and can involve such attributes as fairness and reciprocity, group membership and loyalty, authority and respect, and purity and sanctity (Haidt 2012). We could, in the Anthropocene, begin to moralize showering, driving, flipping light switches, eating high on the food chain, having babies and all those other actions that have a pernicious systemic life, while continuing to accept that because of their harmless episodic dimension such actions should remain beyond the coercive power of the state.38 Much as many people have come to see smoking as unfair, disgusting, disrespectful, or callous, we might come to see hot shower­ ing as unfair to future generations, driving as impure, wasting food as disrespect­ ful of those who suffer climate-­induced desertification, and having more than one baby as callous towards humanity and the planet. Such actions would remain private and thus beyond the reach of the state, but we would take it upon our­ selves to regulate them in foro interno: we would be our own censors (with a little help from our friends). While this is one approach liberals might take, it may strike some as being con­ trary to liberalism’s characteristic spirit of tolerance. For liberalism to persist there must be a broad range of actions, practices, and ways of life about which morality is silent.39 Moreover, the transition to the moralization of such everyday behaviors would be painfully disruptive of the value and cultural systems that liberalism has contributed to or tolerated over the past two and a half centuries. Exactly how disruptive it would be is difficult to say. Would it be more like the transition to the abolition of smoking, or would it be more like China’s cultural revolution? Because climate-­changing and Anthropocene-­creating actions have an episodic as well as a systemic dimension, moralizing these actions might feel like the po­licing of our inner lives as well as our outward behaviors. To be successful, such change might require something resembling a new religion rather than the rational revision of our everyday moral practices. Perhaps the stakes are so high with climate change and the Anthropocene that such disruption would be ­justified. Yet the reference to China’s cultural revolution also serves as a reminder that such disruptions do not always succeed in achieving their goals. In sum,

38  Arguably, this is already occurring. For a general study of one such informal sanction (shame), see Jacquet (2015). 39  Of course there is such a thing as perfectionist liberalism. For discussion (but not endorsement) see Quong (2010); Nussbaum (2011).

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  389 while it is possible to try to dull the second horn of the dilemma, it is hard to see it as remaining anything other than sharp.40 There is another response to the dilemma that is in the spirit of liberalism, but at this point more of a recipe than a remedy. Throughout the modern era the Distinction has been something of a trickster, changing its shape in ways that keep it relevant, useful, and productive. Many liberals who have explicitly embraced the Distinction have recognized this, and they have shied away from seeing it as exhaustive, binary, timeless, or context-­free. For example, Shklar (1989: 24) writes of the Distinction that “[t]his is not historically a permanent or unalterable boundary [. . .].” Similarly, Benn writes that: The liberal cannot give absolute specifications . . . for what is private and what is not, because privacy is context-­relative. I do not mean that standards differ between cultures. That is also true, but it is a different kind of relativity. Within the one culture [italics in text] the same matter may count as private or not, rela­ tive to the social nexus in which it is embedded.  (Benn 1998: 268)

The power of what is on each side of the Distinction leads to the unavoidability of dilemmas, yet the plasticity of the Distinction allows us to go on. As Rorty writes: On the public side of our lives, nothing is less dubious than the worth of these freedoms. On the private side of our lives, there may be much which is equally hard to doubt . . . The existence of these two sides . . . generates dilemmas. Such dilemmas we shall always have with us, but they are never going to be resolved by appeal to some further, higher set of obligations which a philosophical tribu­ nal might discover and apply.  (Rorty 1989: 197).

Yet it is the very features of the Distinction that allow it to shape-­shift that have led some to say that we should do away with it or at least diminish it in our thought. For example, Geuss writes: There is no single clear distinction between public and private but rather a series of overlapping contrasts, and thus [. . .] the distinction between the public and the private should not be taken to have the significance often attributed to it. (Geuss 2001: 6) 40  In Chapter 6 of his far-­sighted work, Harm to Others (1984), Joel Feinberg sketches an account of “accumulative harms” in which he seems to have a view that is different from but related to the plural­ ism that we discuss. According to Feinberg, individual acts of polluting are only harmful if they are “unlawful as judged by a regulative agency applying rules for allocating permits in accordance with specific rules of fairness and efficiency” (p. 230). On his view it seems that some apparently private acts of polluting can only be harmful once they have been taken up in the public domain. Yet such acts would continue to live a double life, and harm (at least in this domain) would now be a consequence of how the Distinction is drawn rather than providing a foundation for it.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

390  Dale Jamieson and Marcello Di Paola Geuss seems to think that the Distinction rests on a shifty set of underlying properties and circumstances that do the real work, and this is obscured when we invoke the Distinction as if it had independent power. A similar argument could be made about the HP: the line between harm and offense is not to be found in Plato’s heaven, but in the messy facts of human psychology and social norms. Furthermore, as we have noted, the difference between an unpleasant byproduct of production and consumption, and the externalization of a cost, is in part a function of malleable norms that can be difficult to codify. Yet these observations do not drain these concepts of their utility. Geuss’s observations are correct but it does not follow that the abstract languages of “externality,” “harm,” “public,” or “private” are misleading or less significant in virtue of them.41 Despite such objections, the Distinction has persisted and played important roles in politics, jurisprudence, and morality. It has taken different forms at different times in various societies and contexts, generally remaining in the shadows and refusing to subject itself to constraining definitions. This may continue in the Anthropocene, with the Distinction taking on whatever form is required to do its work as part of the infrastructure of liberal society.42 The pressures of climate change and the Anthropocene may force a reshaping of the Distinction rather than its abandonment, bringing peace between the episodic and systemic dimen­ sions of our actions even in this new epoch. Perhaps we could expel from the perimeter of privateness certain kinds of consumption, but keep what Mill and Rawls thought was most worth defending: those actions and behaviors related to “experiments in living” and our capacity to autonomously form and pursue con­ ceptions of the good life. More broadly, we may come to think that some actions and behaviors that are now regarded as private are not worth defending at all, and other actions and behaviors that are worth defending episodically may still have to be expelled because of their systemic consequences for the planet. Using some such approach we could attempt to reverse the wheel of the presently un­sus­tain­ able infrastructures of provision—avoiding attendant greenhouse gas emissions, land and water depletion, biodiversity loss, and so on—while keeping what is most important and exciting about privateness—the self-­expression, diversity, and innovation that it protects and promotes. Perhaps in this way we could edit the Distinction to fit both the circumstances of the new epoch and the original justifications that powered the last two and a half centuries of liberal political theory and practice. This would be an epochal experiment indeed, requiring a significant ­revamping of the ways in which liberal societies currently understand the 41  Similar observations could be made about goodness. Philosophers may fruitlessly search for a “single clear distinction between” the good and the bad; this may lead to madness or bad philosophy, but not to the conclusion that goodness “should not be taken to have the significance often attrib­ uted to it.” 42  Cf. Kysar (2011) who makes a similar point with regard to tort law.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  391 Distinction and how they deliver goods and services to their citizens.43 It is hard to see from here exactly how the boundaries would be drawn between the public and private, on what occasions and in what circumstances, and what alternative social and economic arrangements would emerge. Even how to conduct such an experiment is not clear, nor can we be sure that the experiment would succeed. But no epochal social change was ever neatly planned in advance, and the price of not deviating from our present course may be so great as to make the experiment worth trying even without the guarantee of success.44 This may not be an ­intellectually elegant approach to the dilemma that we have posed, but it may be the only one that is pragmatically possible.

5.  Concluding Remarks In this chapter we have focused on the ways in which climate change and more generally the Anthropocene—this new epoch in which no earthly place, form, entities, processes, and systems escape the influence of human activity—put pres­ sure on one particular distinction that is at the heart of liberal theory: that between the public and the private. The Distinction has always been pivotal for liberalism, but also shifty, unstable, and rarely discussed explicitly and ana­lyt­ic­ al­ly. In the Anthropocene, liberals are now challenged, by the very material con­ ditions and ecological implications of the form of life that they have helped to create, to revivify the Distinction or let it fade into the background or disappear. The Distinction took root in our sensibilities because it offered the promise of a protected sphere in which individuals could invent and pursue their own ways of living and conceptions of the good life. The Anthropocene forces a reexamination of these roots to see whether the Distinction can find new life in radically chan­ ging circumstances, or whether it is destined to decay in soil that has been exploited, polluted, and impoverished by the unintended effects of systemic, interlocking forces and structures that dominate our economies, politics, and even private lives. It is too early to say whether the Distinction will survive the Anthropocene, but we hope to have illuminated the challenge.45

43  In an important early paper on climate change, Revelle and Suess (1957: 19–20) described our returning to the atmosphere and oceans the concentrated organic carbon stored in sedimentary rocks over hundreds of millions of years as an experiment. Perhaps this physical experiment with the planet requires a social and political experiment in response. 44  As Professor Irwin Corey was fond of reminding us, “If we don’t change direction soon, we’ll end up where we’re going” (as quoted in Jamieson and Nadzam 2015). 45  We are grateful to the editors of this volume—Mark Budolfson, Tristram McPherson, and David Plunkett—for their excellent comments on an earlier draft. We also thank the participants in a 2016 workshop at the Princeton Center for Human Values and the 2017 “All Tomorrow’s Climates” confer­ ence at LUISS University in Rome, at which some of this material was presented in earlier form.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

392  Dale Jamieson and Marcello Di Paola

References Arendt, H. (1958). The Human Condition, 2nd edition Chicago, IL: University of Chicago Press. Ariès, P. and G.  Duby (1992). A History of Private Life, 5 vols. Cambridge, MA: Harvard University Press. Bator, F.  M. (1958). “The Anatomy of Market Failure.” The Quarterly Journal of Economics 72(3): 351–379. Benn, S. I. (1988). A Theory of Freedom. Cambridge: Cambridge University Press. Benn, S.  I. and F.  F.  Gaus (eds.) (1983). Public and Private in Social Life. London: Croom Helm. Bobbio, N. (1989). Democracy and Dictatorship: The Nature and Limits of State Power, trans. P. Kennealy. Cambridge: Polity Press. Botti, D. (2017). “Rawls on Dewey before the Dewey Lectures.” Journal of the History of Ideas 78(2): 287–298. Brink, D. (2013). Mill’s Progressive Principles. Oxford: Oxford University Press. Broome, J. (2012). Climate Matters: Ethics in a Warming World. New York: W. W. Norton. Charney, E. (1998). “Political Liberalism, Deliberative Democracy, and the Public Sphere.” American Political Science Review 92(1): 97–110. Cohen, G. (1995). Self-Ownership, Freedom and Equality. Cambridge: Cambridge University Press. Cohen, G. (2000). If You’re an Egalitarian, How Come You’re So Rich? Cambridge, MA: Harvard University Press. Crutzen, P. and E. Stoermer (2000). “The Anthropocene.” Global Change Newsletter 41: 17–18. Dawson, M. (2004). The Consumer Trap. Chicago, IL: University of Illinois Press. DeCew, J. (2015). “Privacy.” The Stanford Encyclopedia of Philosophy, ed. E. N. Zalta. https://plato.stanford.edu/archives/spr2015/entries/privacy/. Dewey, J. (1927/1954). The Public and Its Problems. Athens, OH: Swallow Press Books. Di Paola, M. (2017). Ethics and Politics of the Built Environment: Gardens of the Anthropocene. London and New York: Springer International. Di Paola, M. and D.  Jamieson (2018). “Climate Change and the Challenges to Democracy.” University of Miami Law Review 72: 369–424. Dworkin, R. (1977). Taking Rights Seriously. Cambridge, MA: Harvard University Press. Ewen, S. (1976). Captains of Consciousness: Advertising and the Social Roots of Consumer Culture. New York: McGraw-Hill. Ewen, S. and E.  Ewen (1982). Channels of Desire: Mass Images and the Shaping of American Consciousness. New York: McGraw-Hill. Feinberg, J. (1984). Harm to Others. New York: Oxford University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  393 Finney, S.  C. and L.  E.  Edwards (2016). “The ‘Anthropocene’ Epoch: Scientific Decision or Political Statement?” Geological Society of America 26 (3): 4–10. Gaus, G. F. (1996). Justificatory Liberalism. Oxford: Oxford University Press. Geuss, R. (2001). Public Goods, Private Goods. Princeton, NJ: Princeton University Press. Habermas, J. (1962/1989). The Structural Transformation of the Public Sphere, trans. T. Burger and F. Lawrence. Cambridge: Polity Press. Haidt, J. (2012). The Righteous Mind: Why Good People are Divided by Politics and Religion. New York: Pantheon Books. Hamilton, C., C. Bonneuil, and F. Gemenne (eds.) (2015). The Anthropocene and the Global Environmental Crisis: Rethinking Modernity in a New Epoch. London: Routledge. Hänsel, M., A.  Moritz, D.  Drupp, J.  Johansson, F.  Nesje, C.  Azar, M.  Freeman, B. Groom, and T. Sterner. (2020). “Climate Economics Support for the UN Climate Targets.” Nature Climate Change 10(8): 781–789. Hart, H. L. A. (1963). Law, Liberty and Morality. Stanford, CA: Stanford University Press. Hayek, F.  A. (1960). The Constitution of Liberty. Chicago, IL: University of Chicago Press. Hegel, G.  W. (1820/1991). Elements of the Philosophy of Right, ed. A.  Wood, trans. H. B. Nisbet. Cambridge: Cambridge University Press. Hibbard, K., P.  Crutzen, E.  Lambin, D.  Liverman, N.  J.  Mantua, J.  R.  McNeill, B.  Messerli, and W.  Steffen (2006). “Decadal Interactions of Humans and the Environment,” in Integrated History and Future of People on Earth, ed. R. Costanza, L. Graumlich and W. Steffen. Boston, MA: MIT Press, 341–375. Horwitz, M. J. (1982). “The History of the Public/Private Distinction.” University of Pennsylvania Law Review 130(6): 1423–1428. Jacquet, J. (2015). Is Shame Necessary? New Uses for an Old Tool. New York: Allen Lane. Jamieson, D. (2002). Morality’s Progress: Essays on Humans, Other Animals, and the Rest of Nature. Oxford: Oxford University Press. Jamieson, D. (2014a). Reason in a Dark Time: Why the Struggle Against Climate Change Failed, and What It Means for Our Future. New York: Oxford University Press. Jamieson, D. (2014b). “Review of John Broome’s Climate Matters: Ethics in a Warming World.” Ethics and International Affairs 28(2). http://www.ethicsandinternationalaf­ fairs.org/2014/climate-matters-ethics-in-a-warming-world-by-john-broome/. Jamieson, D. (2016). “Responsibility and Climate Change,” Global Justice: Theory Practice Rhetoric9: 23-42. Jamieson, D. (2017). “The Anthropocene: Love It or Leave It,” in The Routledge Companion to the Environmental Humanities, ed. U. K. Heise, J. Christensen, and M. Niemann. London: Routledge, 13–20. Jamieson, D. and M.  Di Paola (2016). “Political Theory for the Anthropocene,” in Global Political Theory, ed. D. J. Held and P. Maffettone. Cambridge: Polity Press, 254–280.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

394  Dale Jamieson and Marcello Di Paola Jamieson, D. and B. Nadzam (2015). Love in the Anthropocene. New York: OR Books. Daniel, K.  R.  Litterman, and G.  Wagner. (2019). “Declining CO2 Price Paths.” Proceedings of the National Academy of Sciences 116(42): 20886–20891. Kysar, D. (2011). “What Climate Change Can Do About Tort Law.” Environmental Law 41: 1–71. Lazarus, O., S.  McDermid, J.  Jacquet. (2020). “The Climate Responsibilities of Industrial Meat and Dairy Producers.” Climatic Change. Forthcoming. Lloyd, G. (1984). The Man of Reason. Minneapolis: University of Minnesota Press. Mahajan, G. (2009). “Reconsidering the Private–Public Distinction.” Critical Review of International Social and Political Philosophy 12(2): 133–143. Marcuse, H. (1964). One-Dimensional Man. New York: Beacon Press. Marx, K. (1843/1978). On the Jewish Question, in The Marx-Engels Reader, ed. R. Tucker. New York: W. W. Norton, 26–46. Mill, J. S. (1848/1909). Principles of Political Economy with some of their Applications to Social Philosophy, 7th edition, ed. W. J. Ashley. London: Longmans, Green and Co. Mill, J. S. (1859/1978). On Liberty, ed. E. Rapaport. Cambridge, MA: Hackett Publishing. Mouffe, C. (ed.) (1996). Deconstruction and Pragmatism. London: Routledge. Nagel, T. (1975). “Libertarianism Without Foundations.” Review of Anarchy, State, and Utopia by Robert Nozick. The Yale Law Journal 85: 136–149. Nagel, T. (1998). “Concealment and Exposure.” Philosophy & Public Affairs 27(1): 3–30. Nozick, R. (1974). Anarchy, State and Utopia. Oxford: Blackwell. Nussbaum, M.  C. (2011). “Perfectionist Liberalism and Political Liberalism.” Philosophy & Public Affairs 39: 3–45. Pateman, C. (1983). “Feminist Critique of the Public/Private Dichotomy,” in Public and Private in Social Life, ed. S.  Benn and G.  F.  Gaus. London: Croom Helm, 118–140. Pateman, C. (1985). The Problem of Political Obligation: A Critique of Liberal Theory. Cambridge: Polity Press. Pateman, C. (1989). The Disorder of Women: Democracy, Feminism and Political Theory. Stanford, CA: Stanford University Press. Pigou, A. C. (1920). The Economics of Welfare. London: Macmillan. Quong, J. (2010). Liberalism Without Perfection. Oxford: Oxford University Press. Rawls, J. (1971). A Theory of Justice. Cambridge, MA: Harvard University Press. Rawls, J. (1993/2005). Political Liberalism. New York: Columbia University Press. Revelle, R. and H. Suess (1957).“Carbon Dioxide Exchange Between Atmosphere and Ocean and the Question of an Increase of Atmospheric CO2 during the Past Decades.” Tellus 9: 18–27. Rorty, R. (1989). Contingency, Irony and Solidarity. Cambridge: Cambridge University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Liberalism and the Public/Private Distinction  395 Rorty, R. (1998). “Pragmatism as Romantic Polytheism,” in The Revival of Pragmatism: New Essays on Social Thought, Law and Culture, ed. M. Dickstein. Durham, NC and London: Duke University Press, 21–36. Rousseau, J.-J. (1762/1985). The Social Contract, trans. H.  J.  Tozer. London: Swan Sonnenschein & Co. Scovronick, N., M. B. Budolfson, F. Dennig, M. Fleurbaey, A. Siebert, R. H. Socolow, D.  Spears, and F.  Wagner (2017). “Impact of Population Growth and Population Ethics on Climate Change Mitigation Policy.” Proceedings of the National Academy of Sciences 114(46): 12338–12343. Sennett, R. (1977). The Fall of Public Man. New York: W. W. Norton. Shklar, J. (1989). “The Liberalism of Fear,” in Liberalism and the Moral Life, ed. Nancy Rosenblum. Cambridge, MA: Harvard University Press, 21–38. Sidgwick, H. (1883/1901). The Principles of Political Economy, 3rd edition. London: Macmillan. Sinnott-Armstrong, W. (2005). “It’s Not My Fault: Global Warming and Individual Moral Obligations,” in Perspectives on Climate Change: Science, Economics, Politics, Ethics, ed. W. Sinnott-Armstrong and R. B. Howarth. Amsterdam: Elsevier, 285–307. Smith, A. (1776/2007). The Wealth of Nations. Amsterdam: MetaLibri. Snow, C. P. (1959). The Two Cultures. London: Cambridge University Press. Southerton, D., H. Chappells, and B. Van Vliet (eds.) (2004). Sustainable Consumption: The Implications of Changing Infrastructures of Provision. Cheltenham, UK and Northampton, MA, USA: Edward Elgar Publishing. Steffen, W., P.  Crutzen, and J.  McNeil (2011). “The Anthropocene: Conceptual and Historical Perspectives.” Philosophical Transactions of the Royal Society A 369: 842–867. Stevenson, C. (1938). “Persuasive Definitions.” Mind 47(187): 331–350. Tocqueville, A. de (1835/2008). De la Démocratie en Amérique. Paris: Flammarion. Van Vliet, B., H.  Chappells and E.  Shove (2005). Infrastructures of Consumption: Environmental Innovation in the Utilities Industries. London: Routledge. Warren, S. and L. Brandeis (1890). “The Right to Privacy.” Harvard Law Review 4(5): 193–220. Weintraub, J. and K. Kumar (eds.) (1997). Public and Private in Thought and Practice: Perspectives on a Grand Dichotomy. Chicago, IL: University of Chicago Press. Wynes, S. and K.  Nicholas (2017). “The Climate Mitigation Gap: Education and Government Recommendations Miss the Most Effective Individual Actions.” Environmental Research Letters 12: 1–9. Zalasiewicz, J., C.  N.  Waters, M.  Williams, and C.  Summerhayes (eds.) (2019). The Anthropocene as a Geological Time Unit: A Guide to the Scientific Evidence and Current Debate. New York: Cambridge University Press.

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Index Activist’s perspective  vii, xix, 252–6, 264, 268–9, 273, 275, 277 Adaptation cost  285, 291n Adaptation to climate change  xv, 1–2, 42, 50, 52–3, 62, 285, 288, 346 Adaptive preference  14, 20, 21, 28 Affective experience (quality of)  30, 33 Agency requirement  253–4, 261 Aggregationism 105 Axiology; see also theory of value  xvi–xvii, 4, 112–14, 116, 118, 120, 122–4, 127n, 128, 130–3 Aggregative axiology  131 Anderson, E.  179, 193 Animal agriculture  xv, 4, 42, 43–8, 53, 56n, 61–3, 242 Animal suffering  42–3, 50–2, 59–63, 219, 226, 237 Animals Farmed animals  43–5, 56n, 61, 219, 245 Wild animals  xv, 42–3, 48–56, 58–63 Anthropocene; see also Capitalocene  48, 58, 173, 370–3, 379–80, 384–6, 388, 390–1 Anthropogenic climate change; see also anthropogenic global warming (AGW)  4, 140, 179–80, 191, 255, 370 Anthropogenic global warming (AGW); see also anthropogenic climate change  179, 191 Aristotle  xviii, 154, 157, 160, 166–71, 373 Arrhenius, G.  vii, xvi, 111–36, 305n Asheim, G.B.  118, 126 Asymmetry Intuition  114, 115n Axiology  xvi–xvii, 4, 105, 112–33, 302 Barrett, S.  18, 339n, 357–67 Bayesianism 185 Bayesian conditionalization  179 Bayesian updating  184 Beneficence  51, 217, 221n, 249n Beneficiary  xv, 274, 329, 334 Bias  12, 96, 144, 158, 165, 186, 188, 191 Attribution bias  144 Confirmation bias  179, 181–5 Björnsson, G.  viii, xix, 252–80 Blackorby, C.  118, 127n, 128 Blame; see also self-blame  253, 255–60, 272

Blameworthy  xvii, 143, 145–7, 256–8, 260n Bostrom, N.  122, 124, 130, 134 Bounded Egalitarian Dominance Conditions 127 Bounded Mere Addition Principle  127–8 Brennan, G.  218n, 220, 250, 253, 264n, 279, 343, 349–50, 366n Broome, J.  viii, xx–xxi, 72, 73n, 78, 93, 106, 111–12, 114–15, 123, 128, 142n, 153n, 203n, 205–6, 270n, 281–91, 298n, 299n, 301n, 323, 325, 327–6, 338, 347n, 387n Budolfson, M.  xvi, xx, 1, 117n, 119, 121, 122n, 125n, 131, 141n, 149n, 204, 205n, 218n, 219n, 223n, 226n, 228n, 237n, 248n, 276n, 278n, 290n, 305n, 325n, 327n, 332n, 391n Buffer zones  263, 268–9 Business as usual scenario (BAU)  xx, 284, 296, 298–9, 303–4, 310–12, 325, 327–31, 335–7, 358, 361 Cantril Ladder  13, 30 Capitalocene  48, 58 Carbon dioxide, or CO2  179n, 182, 205n, 281, 284, 286, 288–9 Carbon emissions  32, 44, 67, 121, 179, 182, 270n, 272, 371, 387n Carbon footprint  32, 263, 269, 270n, 277 Carbon tax  67, 86n, 112, 156, 326–8, 334–5, 341, 380 Causation  28, 269, 382–3 Chrisoula, A.  xvi, 139–51 Citizenship  4, 21, 52 Civil liberties  28, 159 Climate treaties  339–40, 347, 351, 354–7 Cline, W.R.  97n, 101n, 106 Coincidence  70n, 186–90 Collective action problem; see also prisoner’s dilemma, coordination problem, and cooperation dilemma  xx, 2, 4, 356, 360 Collective capability  252, 273, 275–7 Collective harm  141n, 150n, 211–13, 221n, 265, 267n, 280n Collective impact  202–3, 211–17 Common pool resource  356, 358, 360 Communication  xvii, 4, 152–4, 156, 160–1, 165 Confirmation bias  179, 181, 183–5

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

398 Index Consequentialism, see also utilitarianism 69, 105–6, 190, 202, 219, 226n, 228n, 231, 320n, 373 Act consequentialism  229–33, 245 Indirect consequentialism  223 Rule consequentialism  230, 232–3 Consumption  10n, 11, 15, 32–3, 46, 48, 70–1, 76–9, 81–2, 84, 86, 92–9, 101, 106–7, 112, 207, 228, 303, 305, 313–14, 330, 332, 356, 379, 380, 384–6 Consumption index  70 Contribution ethics  224, 238, 247 Convenient truth  9–37 Cooperative dilemma  347 Coordination problem  356, 357n, 378 Corner solution  120–2, 131 Cost-benefit analysis  330 Cost distribution  ix, 141 Courage 169 Creation ethics  42, 53, 57–8 Creeping environmental problem  xvii, 140–2, 147–8 Cultural cognition  xviii, 178–95 Curzer, H.  157, 160 Dannenberg, A.  358 Dasgupta, P.  81, 90, 92n, 94–5, 106–7 Decarbonization  ix, 120–1 Decreasing marginal utility  23, 30, 49n, 70, 93n, 98, 101, 210 Democracy  28, 103, 159n, 252, 384–5 Democratic  10, 28–9, 35, 103, 153, 158, 166, 339, 362, 370, 373, 385n Denizenship 52 Descriptivism  xvi, 91–2, 100–2 Developed countries  xv, 10, 15, 21, 27, 29, 35, 44 Developing countries  9, 25n, 31n, 44, 116, 121, 216, 331, 365n Development  10, 27, 32, 35, 121–2 Difference-making explanation  225, 234–7, 239, 246–7 Discount rate  61, 79–80, 82–6, 97, 99, 286–9, 305, 307–8 Discounting  xvi, 3, 67–72, 76, 79–86, 90–108, 286–9, 303, 305, 308 Consumption-based discounting  91, 93–4, 100–1, 107 Impure discounting  91, 98 Intertemporal discounting  3, 95 Intratemporal discounting  95 Investment-based discounting  91, 93–4, 97, 100–1 Pure discounting  91–3, 97–9, 102, 106, 287

Disjunctive wrongs, principle of  315–16, 318 Distinction, The  370–91 Dismal Theorem, The  82, 85 Di Paola, M.  xxi, 370–1, 383n, 385n Distributive justice  3–4 Dominance  xvii, 77–8, 113, 118–19, 121–3, 127–8, 131 Donaldson, D.  118, 128 Donaldson, S.  49n, 51n, 52 Easterlin Paradox  29–32 Easterlin, R.A.  23, 29–32 Economic growth  4, 9–11, 23, 28, 31n, 32n, 33–5, 70n, 80–1, 83, 112, 122, 162, 284, 286–7, 295, 300, 303, 314 Economic inequality  29, 33–4 Economics  xvi, xx–xxi, 1, 9, 67, 69, 91, 98, 100–1, 104, 121, 235–6, 275, 281, 287, 296, 297–301, 306, 324, 326, 332, 356, 361n, 364 , 379n Climate economics  91, 98, 100, 301, 332, 334 Environmental economics  275, 298, 300 Marxist economics  235, 375 Neoclassical economics  235–6 Positive economics  80, 98, 101, 314 Welfare economics  332 Economics of climate change  xvi, 67, 69 Ecosystems  48, 99, 384 Egalitarian communitarians  180n, 182–5, 193 Egalitarian Dominance Condition; see also Bounded Egalitarian Condition  118 Egalitarianism  24, 70, 125, 180n, 182–5, 193 Ellsworth, P.  17 Emissions Threshold  203–4 Emotions  xvii, 2–3, 12–13, 17–18, 153, 157n, 158n, 230 Environmentally friendly  xvii, 140, 141, 147–8 Episodic life  383–4, 386 Equality  24, 26, 31, 75, 85, 159, 366, 375 Ethical explanation  233 Ethics  xvi, xviii, 2–4, 42, 53, 56n, 57, 59, 63n, 82n, 90, 95, 104, 111–13, 115–18, 120, 123–5, 127–8, 130–2, 189n, 219n, 223, 227, 231n, 233, 238, 247–8, 274, 305n, 308, 353 Evaluative focal points  223–4, 229–40, 246–8 Ex-ante equity  73–5, 84, 86 Ex-post equity  72n, 73–5, 79n, 82, 85–6 Existential risk  122–3 Exogenous  79–80, 90n, 302, 307 Exogenously given or exogenously determined  288, 302, 307

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Index  399 Expected benefit  204, 206–11, 219 Expected disutility  204, 211n Expected harm  56, 203–4, 206–7, 210–11, 219 Expected moral value  121 Expected utility  xviii, 72, 77n, 78n, 201–8, 210n, 211n, 215n Expected value  56, 72, 112n, 268–72, 273n, 277 Externalities  47, 298n, 379n, 380–2, 390 Extinction  48, 58, 71n, 81n, 90n, 97n, 122, 283, 304, 371–2 Fear  xvii–xviii, 17–18, 152–71 Civic fear  xvii, 154, 165–71 Fear appeal  xvii–xviii, 153–71 Fear mongering  153, 158, 161 Feasibility  xx–xxi, 2–3, 149, 192, 323–42, 346–55, 363, 365–7 Ferranna, M.  xvi, 67, 70n, 76, 93n Food ethics  223, 248n Fossil fuel  4, 67n, 112, 295n, 310–13, 384, 386 Freedom  11, 28–30, 101, 103, 159, 307, 315–18, 320, 373, 389 Procreative freedom  307, 315–17 Free riding  363, 365–6 Futility  162, 226n, 276n Future individuals; see also future people  xx, 4, 61, 70n, 71, 97–8, 114–16, 282, 287n, 296, 305n, 308, 310, 317, 323–4, 329–30, 332, 361n Future generations  xv, 5, 9–10, 35, 42, 53, 57, 69, 80–1, 83–5, 96, 155, 169–70, 298–9, 303, 308, 310, 317, 323–4, 331, 379, 388 Future people  xx, 4, 97–8, 114–16, 282, 296, 305–6, 308, 312n, 329–30, 332 Generations; see future generations and younger generations Geoengineering  4, 61, 179–80, 182–4, 191–4, 340 Geometrism 118 Geuss, R.  373n, 377n, 389–90 Gilabert, P.  349–51 Gini coefficient  24, 26n, 29 Global financial crisis  16, 28 Globally uniform valuation of mortality risk  284, 286–7 Gore, A.  153n, 155n Goulder, L. H.  92n, 94n, 99, 334n Greaves, H.  69n, 79n, 90n, 118, 121, 123, 125n Greenhouse gas emissions  xx, 44–6, 201, 269, 274, 281–2, 295–320, 384, 390 Greenhouse gases (GHGs)  90, 236, 264, 268, 284, 311, 356n, 358, 372, 383

Gross domestic product (GDP)  ix, 10n, 11, 15–16, 21, 23–9, 32, 35, 327–8 Grounding explanation  225, 229–31 Group agency  253–4, 279n, 368n Group-oriented stance  xvi, 142–4, 146–8 Happiness; see also satisfaction and well-being xv, 10–12, 14–15, 17–18, 22, 30, 32–3, 35, 210, 257, 272 Happiness ideology  11 Happiness science  11 Harm Principle (HP)  377–8, 389n Harsanyi, J.  71–2 Health  1–2, 11, 13–14, 22, 23n, 25–6, 28, 30, 34, 44, 47–8, 67n, 69, 123, 149, 155, 158n, 164, 282–3, 340, 360, 370 Hierarchical individualists  180n, 181n, 182–5, 192, 194 Hope (as an alternative to fear)  153, 161–4, 166, 168 Hope appeals  xvii–xviii, 153–4, 158n, 161–2, 164–5, 169–70 Human development  121–2, 385n Ideal economy v. non-ideal economy  94 Ideal theory  3 Idealism (about international relations)  349 Impartiality  91, 100, 102, 106 Impossibility theorem  xvi, 113, 118, 123–4, 129–33 Inconsequentialism 251n Individual incapability  253–4, 262–3, 271, 273, 276 Instinctive attraction/aversion  267–8, 271 Individualistic stance  xvi, 142–4, 146–8 Industrialization 48 Inefficacy  xix, 201–3, 209n, 218n, 222–7, 229, 231, 236–7, 239, 242–3, 247 Inefficacy Challenge  201–3, 208, 211, 216–48 Inequality  9–10, 15n, 24–6, 29, 31, 33–5, 70–1, 73–87, 94n, 101, 112, 236, 252, 366, 373 Income inequality  15n, 25n, 26, 31 Inequality aversion  70, 76–83, 94, 101 Integrated Assessment Model (IAM)  91, 112, 120, 128, 302, 305n Intergenerational ethics  90, 308 Intergenerational problem  141 International Paretianism (IP)  xxi, 346–67 Intergovernmental Panel on Climate Change (IPCC)  43, 111–12, 115–16, 123, 201n, 282n, 287n, 288n, 298n, 305n Intuition of Neutrality  114 Issue-oriented reasoning  274

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

400 Index Jamieson, D.  xxi, 63n, 106, 348n, 370–1, 380n, 382n, 383n, 385n, 387n, 391n Justice  xvi, xx, 9–10, 12, 62, 95, 105, 148, 236–7, 240, 247, 257, 270n, 274, 290, 307–9, 314–16, 320, 331, 333, 335, 375–6 Climate justice  347 Distributive justice  3–4, 327n, 348n Intergenerational justice  xx, 295–6, 305–6, 308–9, 311, 314–15, 317–20 Social justice  26, 308, 320n Kahan, D.  xviii, 178n, 180–6, 190n, 192, 193n, 194 Kaldor-Hicks efficiency  296, 299–301 Kantian  213, 215n, 217, 227, 373, 385 Kaplan, A.  105–6 Kelleher, J. P.  95, 105, 106n, 108n, 298n, 299n Kymlicka, W.  51n, 52 Koopmans, T.  95, 305 K-strategists  49, 56, 60 Large Quantity-Quality Tradeoff  129 Lawford-Smith, H.  203n, 206n, 233n, 254, 261n, 268, 277n, 349–51 Lerner, A.  103 Liberal Liberal democracy  159n, 370, 384, 385n Liberal politics  158 Liberal theory; see also Liberalism  xxi, 373n, 375, 391 Liberal societies  374, 378, 385–6, 390 Liberalism  xxi, 2, 370, 372–9, 384, 386–9, 391 Life satisfaction  13, 24, 31 Markowitz, E.  153, 158, 165 Maximin  118, 122, 126 McPherson, T.  xix, 1, 102n, 132n, 149n, 195n, 212n, 222, 223n, 228n, 233n, 237n, 290n, 391n McQueen, A.  xvii, 152, 153n, 160 Meat  46, 167, 201, 217–19, 222, 226–8, 237, 242, 289, 379, 387 Mere Addition Principle; see also Bounded Mere Addition Principle 127 Mill, J.S.  377–8, 384–5, 390 Mindset M  143, 146–9 Mintz-Woo, K.  xvi, 61, 63n, 81n, 90, 92, 94n, 104n, 106n, 108n, 205n, 219n Mitigation  xv, xix, 1–2, 42–3, 45–53, 61, 80, 81n, 99, 156n, 158n, 162–3, 179, 222–4, 229–30, 237, 243, 245–7, 305n, 323, 328, 330–2, 341, 346, 347n, 348n, 352n, 355n, 356, 358n, 361, 366

Mitigation of climate change  xv, xix, 1–2, 42–3, 45–53, 61, 80, 81n, 99, 157n, 162–3, 179, 222–4, 229–30, 237, 243, 245–7, 305n, 323, 328, 330–2, 341, 346, 356, 361, 366 Moral agent  62, 157, 253–5, 260–1, 278n–79n Moral expert  xvi, 91, 101, 103–5, 107–8 Moral expertise  xvi, 91 Moralization 388 Mortality-related harm  284–5 Mortality-related cost  285 Multi-stage plan  351–4, 366 Nagel, T.  373 Nash equilibria  357, 359, 363 Nation state  347n, 352n, 355, 357n, 362 National interest  346–7, 355–6, 360–3, 365–6 Nefsky, J.  xvii, 201, 202n, 206n, 209n, 215n, 223n, 224n, 253, 264, 265, 267, 270n Negligibility  224, 227, 242–4, 246–7 Nesse, R.  17 Ng, Y.-K.  55n, 118 Non-ideal theory  3 Non-identity problem, The  xx, 58–9, 296, 298, 309, 314–15, 317–18, 320 Nordhaus, W.  xxi, 68, 81, 90, 97, 112, 120, 153, 302n, 324–7, 330n, 337, 359n , 361n, 362n, 363n Obama administration  69n Obama, B.  162, 170 Obligation Climate Change Related obligation  203, 208, 220n, 279n Collective obligation  214, 252–5, 261–2, 268, 273, 278n, 280n Imperfect obligation  212–19 Individual obligation  viii, xviii, xix, 151n, 201, 211, 214, 221n, 251n, 252–3, 256, 258–62, 270n, 271, 273, 277, 280n, 395n Moral obligation  151n, 201–2, 221n, 251n, 252–3, 255–6, 258–62, 272n, 280n, 308, 395n Shared obligation  viii, xix, 252, 254, 260, 262–3, 272–3, 275, 276–7, 278n, 279n Ord, T.  121, 125n Pareto  xxi, 77, 101n, 128, 296, 299 Pareto efficient  299 Pareto improvement  300, 325–9, 347–9, 351, 353–7, 359–60, 362–3, 365–6 Pareto inefficient  298 Pareto optimal, see also Pareto efficiency  297 Parfit, D.  58–9, 90n, 93, 106, 113–19, 123, 124n, 149–50, 212n, 227n, 243n, 264, 266, 268, 302n, 303n

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

Index  401 PERMA  33, 35 Philanthropy  45, 50 Plunkett, D.  1, 62, 149n, 195n, 248n, 391n Policy Climate policy/ies  xvi–xvii, xxi, 67, 69–70, 80–1, 86, 91, 96–7, 111–12, 114–15, 120–1, 125–6, 130–1, 192, 272, 296, 301–2, 305–6, 308, 323, 329–30, 332–3, 335, 382 Intergenerational policy/ies  111–12 Political duties  52, 62 Population  xv–xvii, 3, 9–11, 15, 21, 22–3, 27–8, 32–5, 42–5, 48–51, 53, 55–9, 62, 67, 70, 71n, 73–4, 81–3, 86, 87, 111–32, 191n, 235, 269, 284, 286, 295n, 302–3, 305n, 307, 308, 372 Population axiology  xvii, 113–14, 118–23, 127n, 130–2 Population change  111, 115 Population ethics  xv, xvi, 3, 42, 53, 57, 82n, 111–13, 115, 116n, 118, 120, 123–5, 127–8, 130–2, 305n Population growth  81, 114, 307 Posner, E.  xxi, 71n, 305, 323, 325–8, 329n, 339, 347–8, 355n, 356, 361n, 363n, 365 Possibility theorem for bounded axiologies  128 Poverty  9, 23, 26, 98, 112, 189 Praiseworthiness  xvii, 143, 146–7 Precautionary effect  79 Preference  20, 70n, 71–2, 76, 81, 86, 95, 99, 101–3, 104n, 140–1, 306, 376 Adaptive preference  14, 20–1, 28 Consumer preference  384–5 Equity preference  86 Individual preference  70n, 72, 77–8, 81, 91, 102, 376 Mere preference  101 Policy maker preference  75 Population preference  70 Satisfaction of  301, 306 Social or collective preference  75, 91, 103 Subjective preference  296 Time preference  70n, 80, 81n, 85–6, 90–2, 97–8, 106n, 306 Prescriptivism (about discounting)  xvi, 91, 92n, 100–4, 106–7 Prioritarianism  xvi, 67, 70–2, 76, 79n, 86, 95, 125 Prisoners’ Dilemma  141, 324, 356–8, 360–4 Private  xxi, 62, 74n, 141, 358–9, 370, 372–81, 383–91 Promotion relation  227, 230 Public Public intervention  378, 387 Public/Private distinction  xxi, 371–2

Puzzle of inefficacy  viii, xix, 222–4, 229, 231, 236–7, 239, 242, 247 Quality of life  xvi, 9, 14–16, 27, 31n, 33, 35, 56, 60, 115–17, 130, 310 Quality of will  256–8 R-strategists  49, 55–6, 60 Railton, P.  xvi, 9 Ramsey rule  xvi, 68, 72, 79, 80–3, 91–7, 107 Rawls, J.  62, 94, 96n, 376–8, 385–6, 390 Realism (about international relations)  325, 342 Reparations  52, 347 Repugnant conclusion, The  59, 113–19, 123–33, 296, 302, 308 Responsibility  3, 4, 51, 156, 165–7, 272, 285, 347, 356, 366 Collective responsibility  4 Individual responsibility  3 Institutional responsibility  4 Rights / rights theory / rights theoretic approaches  50–1, 55, 57, 59, 257, 306–7, 309, 315, 317, 320, 373, 375, 380 Risk  xvi, 9, 11, 20, 44, 57, 68, 70–87, 95–6, 97n, 106n, 144, 147, 153–4, 164, 193, 211n, 215n, 252, 266, 269–70, 274, 300, 301n, 318 Mitigation of  46 Risk assessment  4 Risk aversion  70–1, 78–9, 83, 93n, 94n Risk factor  33 Uncertainty about  56, 70n, 93n Sadistic Conclusion, the  117–18, 129 Shariff, A.  153, 158, 165 Sebo, J.  xv, 42, 56, 108 Self-blame  xvii, 143–9 Self-interest  42, 324, 335–6, 341, 347–8, 351, 353–5, 362–3, 365, 367 Self-praise  xvii, 143, 145–8 Self-sustaining social patterns  238–40, 242, 247 Seligman, M.E.P.  18, 33 Shue, H.  52, 348n Sider, T. R.  118 Social cost of carbon (SCC)  69n, 88n, 109n–10n, 281–2, 284, 291n, 322n, 330n, 380, 382n Social discount factor  69 Social discount rate (SDR)  xvi, 67–9, 80–2, 92–3, 95–100, 107, 305, 307 Positive social discount rate  305, 307 Social welfare function  69–70, 72, 75–6, 78, 80, 82, 86, 92, 95, 100, 101n, 105, 123, 126, 128, 132–3

OUP CORRECTED AUTOPAGE PROOFS – FINAL, 12/02/21, SPi

402 Index Spears, D.  xvi, 111, 114n, 117n, 119, 121, 122n, 125n, 131, 305n Stanczyk, L.  xx Standard of living  9–10, 21, 23n, 29, 33, 35, 307n, 310 Steele, K.  xxi, 347 Stern, N.  81, 90, 97, 101, 106, 152–3 Stern Review  81, 90, 97, 101, 106 Sustainability (sustainable)  10–11, 99, 303, 320 Systemic life  383–4, 386–8 Tännsjö, T.  119 Theory of value  353 Tragedy of the commons  236 Uncertainty  3, 5, 57–8, 70n, 75, 80–1, 83–5, 87, 90n, 92n, 93–4, 112n, 121–2, 154, 161, 165, 207n, 289, 358, 361 Normative uncertainty  121 Utilitarianism, see also consequentialism  xvi, 50, 55, 67, 68n, 70–2, 74, 77–8, 80, 82, 85–6, 91, 94–6, 104, 106 Average utilitarianism  60, 116–22, 126–7, 129, 132 Discounted utilitarianism  xvi, 67, 80, 86, 94–6 Generalized utilitarianism  118, 125–6 Negative utilitarianism  58n Total utilitarianism  60, 115–16, 118, 120–4, 126, 128, 130, 132–3, 296, 301–3, 308 Utility; see also expected utility  xviii, 72, 74–8, 80–1, 83, 91n, 92, 95, 97–8, 100–1, 104, 106n, 118, 125, 132, 384, 390

Utility discounting  86, 91 Utility function  70, 72, 81, 83, 93 Utility monster  303, 308 Valuing the Global Mortality Consequences of Climate Change (VGMC)  282–9, 291n Veil of ignorance  72 Very Repugnant Conclusion, the  119 Walton, D.  154, 159 Weisbach, D.  xxi, 305, 323, 325–8, 329n, 339, 347–8, 355n, 356, 361n, 363n, 365 Weitzman, M.L.  71, 81–5, 90, 283, 308n Welfare; see also utility  xx, 48, 69–72, 74n, 75–80, 82–3, 86–7, 91–2, 94n, 95, 97–8, 100, 101n, 104–5, 114–19, 122–4, 126–33, 296–300, 302–7, 309, 320, 329–35, 346, 355, 359, 361–4, 376, 379n Total welfare  72, 74, 115, 297, 303n, 304, 362 Welfare economics  xx, 100, 296–300, 306, 332, 335 Welfarism  105, 305 Well-being / wellbeing Average well-being  59–60, 111, 121–2 Subjective well-being  xi, xv, 10–35 Total well-being  50–1, 55, 59–60 Williams, R. C.  93n, 94n, 99 Wishful thinking  164–5, 167, 189, 361 World Values Survey  12 Younger generations  317, 320 Zuber, S.  81n, 82n, 85, 94n, 118, 126