Ethics at the end of life : new issues and arguments [1 [edition].] 9781315727950, 1315727951

Table of contents :
Is it possible to be better off dead? / Geoffrey Scarre --
How does death harm the deceased? / Taylor W. Cyr --
The significance of an afterlife / Benjamin Mitchell-Yellin --
The severity of death / Jens Johansson --
Defining death / John K. Davis --
Autonomy, competence, and end of life / James Stacey Taylor --
Deciding for the incompetent / Eric Vogelstein --
Change of mind: an issue for advance directives / Paul T. Menzel --
Medical futility and respect for patient autonomy / Nancy S. Jecker --
Refusing lifesaving medical treatment and food and water by mouth / Paul T. Menzel --
Suicide, physician-assisted suicide, the doing-allowing distinction, and double effect / Thomas S. Huddle --
Grief and end-of-life surrogate decision-making / Michael Cholbi --
Solidarity near the end of life: the promise of relational decision-making in the care of the dying / Bruce Jennings --
Justice and the aging of the human species / Colin Farrelly.

Citation preview

Ethics at the End of Life

The 14 chapters in Ethics at the End of Life: New Issues and Arguments, all published here for the first time, focus on recent thinking in this important area, helping initiate issues and lines of argument that have not been explored previously. At the same time, a reader can use this volume to become oriented to the established questions and positions in end-of-life ethics, both because new questions are set in their context, and because most of the chapters— written by a team of experts—survey the field as well as add to it. Each chapter includes initial summaries, final conclusions, and a Related Topics section. John K. Davis is Professor of Philosophy at California State University, Fullerton. He has published articles on various aspects of end-of-life care.

Ethics at the End of Life New Issues and Arguments

Edited by John K. Davis

First published 2017 by Routledge 711 Third Avenue, New York, NY 10017 and by Routledge 2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN Routledge is an imprint of the Taylor & Francis Group, an informa business © 2017 Taylor & Francis The right of the editor to be identified as the author of the editorial material, and of the authors for their individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging in Publication Data Names: Davis, John K. (Associate Professor of Philosophy), editor. Title: Ethics at the end of life : new issues and arguments / edited by John K. Davis. Description: 1 [edition]. | New York : Routledge, 2016. | Includes bibliographical references and index. Identifiers: LCCN 2016048693 | ISBN 9781138829756 (hardback) Subjects: LCSH: Death—Moral and ethical aspects. Classification: LCC BJ1409.5 .E84 2016 | DDC 179.7—dc23 LC record available at https://lccn.loc.gov/2016048693 ISBN: 978-1-138-82975-6 (hbk) ISBN: 978-1-315-72795-0 (ebk) Typeset in Times New Roman by Apex CoVantage, LLC

Contents

List of Contributors Introduction

vii 1

J O H N K . D AV IS

PART I

The End of Life 1 Is It Possible to Be Better Off Dead?

9 11

G E O F F R E Y S CARRE

2 How Does Death Harm the Deceased?

29

TAY L O R W. C YR

3 The Significance of an Afterlife

47

B E N J A M I N M I T CHE L L - YE L L I N

4 The Severity of Death

61

J E N S J O H A N SS ON

5 Defining Death

74

J O H N K . D AV IS

PART II

Who Decides When to End Life? 6 Autonomy, Competence, and End of Life J A M E S S TA C EY TAYL OR

91 93

vi

Contents

7 Deciding for the Incompetent

108

E R I C V O G E L ST E I N

8 Change of Mind: An Issue for Advance Directives

126

PA U L T. M E N ZE L

9 Medical Futility and Respect for Patient Autonomy

138

N A N C Y S . J E CKE R

PART III

How to End Life

153

10 Refusing Lifesaving Medical Treatment and Food and Water by Mouth

155

PA U L T. M E N ZE L

11 Suicide, Physician-Assisted Suicide, the Doing-Allowing Distinction, and Double Effect

171

T H O M A S S . H UDDL E

PART IV

Other Parties and the End of Life

199

12 Grief and End-of-Life Surrogate Decision-Making

201

M I C H A E L C H OL BI

13 Solidarity Near the End of Life: The Promise of Relational Decision-Making in the Care of the Dying

218

B R U C E J E N N INGS

14 Justice and the Aging of the Human Species

235

C O L I N FA R R E L LY

Index

249

Contributors

Michael Cholbi is Professor of Philosophy at California State Polytechnic University, Pomona. He has published books on the ethics of suicide and on Kant’s ethics and is currently completing a monograph on philosophical issues related to grief and bereavement. Taylor W. Cyr is a graduate student at the University of California, Riverside, who is interested in various topics at the intersection of metaphysics and ethics, especially free will, moral responsibility, death, and immortality. John K. Davis is Professor of Philosophy at California State University, Fullerton. He has published articles on various aspects of end-of-life care. Colin Farrelly is Professor in the Department of Political Studies at Queen’s University. His publications on aging include articles in the BMJ, Biogerontology, and Bioethics. His book titled Biologically Modified Justice is forthcoming with Cambridge University Press. Thomas S. Huddle, MD, PhD, FACP, is Professor of Medicine in the Division of General Internal Medicine at the UAB School of Medicine in Birmingham, Alabama. He teaches internal medicine and clinical ethics and writes about medical education and medical ethics. Nancy S. Jecker, PhD, is Professor at the University of Washington. She is coauthor of Wrong Medicine: Doctors, Patients, and Futile Treatment, 2nd edition (Johns Hopkins Press) and coeditor of Bioethics, 3rd edition (Jones and Bartlett). Bruce Jennings is Adjunct Associate Professor at Vanderbilt University. His publications include: The Hastings Center Guidelines for Decisions on Life-Sustaining Treatment and Care Near the End of Life (2013) and Ecological Governance: Toward a New Social Contract with the Earth (2016). Jens Johansson is Associate Professor of Practical Philosophy at Uppsala University. His main research areas are the philosophy of death and personal

viii Contributors identity. He is coeditor of Oxford Handbook of Philosophy of Death (2013; with Ben Bradley and Fred Feldman). Paul T. Menzel, PhD, is Professor of Philosophy emeritus at Pacific Lutheran University. He has published widely on moral questions in health economics and health policy. His most recent writings are on end-of-life issues, including advance directives for dementia. Benjamin Mitchell-Yellin is Assistant Professor of Philosophy in the Department of Psychology and Philosophy at Sam Houston State University. Geoffrey Scarre is Professor of Philosophy at Durham University, UK. In recent years he has taught and published mainly in moral theory and applied ethics. His books include Death (2007) and On Courage (2010), and he is currently editing The Palgrave Handbook of the Philosophy of Aging. James Stacey Taylor teaches philosophy at The College of New Jersey. His books include Stakes and Kidneys (Ashgate, 2005), Practical Autonomy and Bioethics (Routledge, 2009), and Death, Posthumous Harm, and Bioethics (Routledge, 2012). Eric Vogelstein is Assistant Professor in the Department of Philosophy and the School of Nursing at Duquesne University in Pittsburgh, Pennsylvania. He specializes in bioethics and ethical theory.

Introduction John K. Davis

End-of-life ethics was among the first topics to develop when modern medical ethics got going in the early 1970s. It is well developed. It is not, however, static. This volume is a collection of new articles on various aspects of end-oflife ethics, with a focus on cutting-edge work and new issues. It is meant both to display recent thinking in end-of-life ethics and to initiate new issues and new lines of argument that have not been explored before. At the same time, a reader can use this volume to become oriented to the established questions and positions in end-of-life ethics, both because new questions are set in their context and because most of the chapters survey the field as well as adding to it. The chapters are organized by topic, and the order of the topics reflects the order in which someone deliberating about the end of life might consider those topics. It is natural to begin with questions about when and whether a patient may be better off dead (or at least has a life that is no longer worth living). Those questions are covered by the chapters in Part I, which develop connections between end-of-life ethics and the now-lively topic of philosophy of death. However, it is inevitable that we will not agree about those questions in all cases, and therefore the next set of issues concerns whose answer is decisive—that is, who gets to decide whether a patient is better off dead. Those issues are addressed by the chapters in Part II, which discuss when a patient with diminished competence may make that decision, how such decisions might be made in advance, and who makes them if the patient cannot. Once a decision has been made that a patient’s life should no longer continue, questions arise about the proper way to end it. Many people believe that letting the patient die is morally permissible, but that killing, suicide (killing oneself ), and physician-assisted suicide are not permissible. The two chapters in Part III question those constraints. Parts II and III represent the core of end-of-life ethics as it is usually understood. Part IV groups together three chapters that raise largely new issues. The first two concern the relationship between the patient and other parties, and explore the ethics of grief and solidarity between the patient and others. The final chapter discusses the farthest frontier of all: the

2

John K. Davis

possibility of slowing aging and extending life far beyond its present biological limits. Most discussions of end-of-life ethics note that patients are sometimes better off dead, or at least do not benefit from continued life. In Chapter 1 Geoffrey Scarre asks whether there is any sense in the idea that a patient could be “better off dead.” After all, if death involves nonexistence (a possibility we should not rule out), it is hard to see how nonexistence is good or bad for someone who no longer exists (how can anything be good or bad for you if there is no you?). Scarre considers what makes a life a “worth living,” whether a life has intrinsic value (and not merely instrumental value) apart from its value for the one living it, whether someone might rationally think that suffering makes his or her life more worth living, and whether such judgments are objective or subjective. Scarre concludes that some patients really are better off dead, but that judgments about this are heavily subjective in nature, and that it’s hard to generalize about which lives are worth living. Whether a patient can be better off dead can be seen as the flip side of some well-known questions about the harm of death: if death involves nonexistence, how can death be bad for the one who dies? In Chapter 2 Taylor W. Cyr surveys several aspects of that question, including the two main positions on the issue. According to the first view, the deprivation account, death is harmful to the one who dies not because death is bad for that person but because death deprives him or her of further life that would have been good for that person. In short, it is not the state of nonexistence that is bad; it is the absence of a good existence. According to the second view, Epicureanism, death is not harmful to the one who dies. Cyr discusses several arguments for and against each of these views. The traditional puzzles about the harm of death assume that we cease to exist when we die, but what if that is false? After all, many people claim to have had near-death experiences in which they briefly experienced an afterlife. Does the existence of an afterlife make a difference to questions about whether a patient can be better off dead, or whether death is ever harmful to the one who dies? Benjamin Mitchell-Yellin argues that it does not. In Chapter 3 he discusses the harm of death and the desirability of immortality, and concludes that the existence of an afterlife has less importance for those issues than one might think. When we ask whether, when, and how a patient might be better off dead, we are also raising questions about how bad death would be for that patient (i.e., the condition of being dead, not the process of dying). We tend to think that death is more tragic the earlier it comes in life, for younger people have more years of potential life to lose. This is a deprivation account of what we might call the severity of death—how bad or tragic a death is. Jens Johansson explores this issue in Chapter 4, points out that a person’s life expectancy at

Introduction 3 the time of death is not the only relevant factor in determining how bad death is for that person, and defends the deprivation account of the severity of death against two objections. When we ask whether a life should end, we must also ask when it does end. Modern medicine can keep patients alive for years in a state of permanent unconsciousness. Such patients are warm, they breathe, and they metabolize, but their consciousness and personality are forever gone. Are they still alive? Is the body alive but the patient dead, or are both dead? How can there be two kinds of death, and if there are, do we die twice? If not, when do we die? In Chapter 5 John Davis surveys and evaluates several criteria and definitions of death, and argues that death occurs when all capacity for experience is forever gone. Patients, in other words, are essentially sentient beings, and they die when they lose the capacity for sentience. There is no consensus on when life is no longer worth living, and there may never be. We can settle such questions, therefore, only by settling the question of who gets to settle the question—that is, who gets to decide when the patient should no longer be kept alive. Up to the 1960s and 1970s, doctors tended to decide what was best for the patient, sometimes without informing or consulting the patient. By the 1960s and 1970s, a series of court cases established the doctrine of informed consent. Now nothing could be done to the patient without informing the patient and getting the patient’s consent. This was matched by a consensus among medical ethicists that providers should follow a principle of respect for patient autonomy. However, we do not respect the decisions of all patients—just those who are competent and autonomous. Autonomy means “self-determination,” roughly speaking, but there is more to it than that. We need a concept of autonomy that captures the precise sense in which a patient’s desires are truly her own, and not the result of some interference or irrationality. Competence is a particularly difficult concept to pin down; merely being conscious, adult, and “not crazy” is intuitively plausible but not quite right. We need a definition and criterion of competence that capture cases where we think the patient has made an irrational choice, but do so without requiring us to second-guess whether his or her choice is a good one. In Chapter 6 James Stacey Taylor addresses these challenges, and explains how the concepts of autonomy and competence apply to end-of-life decisionmaking in general. Very often patients are deeply and permanently demented or unconscious, and cannot make decisions about whether and when to end their lives. When the patient cannot make that decision, who does? Sometimes patients use a legal document known as a living will to make decisions about their medical care in advance, including decisions to reject further care and to die. However, living wills have been controversial. Some ethicists argue that if the patient is permanently incompetent, or better yet, permanently unconscious, the patient’s

4

John K. Davis

previous intentions and values cannot now be attributed to the patient, and the patient’s past intentions and values have no moral authority (i.e., we are not ethically required to honor them). Instead, these critics argue, surrogate decisionmakers should do whatever is in the best interests of the patient as he or she is now. Eric Vogelstein defends this view in Chapter 7, arguing that we can respect the autonomy of currently incompetent patients only by respecting the desires and values the patient currently has, not the ones he used to have. He also questions the view that what is in a patient’s current best interest depends partly on the patient’s prior values. Not all patients who are incompetent near the end of life are unconscious, and there are degrees of dementia. Some patients are demented or otherwise incompetent but still capable of having desires or making decisions, at least on simple matters. It is possible for such patients to have different judgments and desires than they had earlier, when they were competent and wrote a living will. However, they may be too demented to comprehend their own advance directives or reevaluate them in a reasoned way, so they cannot quite rethink or reconsider their earlier decisions. They cannot be said to have changed their mind in the sense of consciously forming a new intention for their current medical treatment, but they do have cognitive states that at least resemble a change of mind. Should we respect their advance directives, or should we treat these as cases where the patient changed his or her mind, and respect the patient’s current values? Paul T. Menzel discusses such cases in Chapter 8, and argues that many of these cases involve a change of mind. Menzel argues for a refined concept of change of mind and contends that we should respect the patient’s current values, not the patient’s earlier choices, even if the patient is now too incompetent to understand and revise his or her earlier decision. By the 1990s the medical profession had largely come to agree that withdrawing or withholding life support is often the right decision, at least when the patient’s quality of life is very low. This produced a new problem: sometimes the patient’s surrogate decision-makers (typically family members) refuse to terminate life support, perhaps because they do not believe the prognosis, or they cannot let go, or they do not want to be responsible for a relative’s death, or they believe that all human life is sacred, or they think the patient’s quality of life is better than the providers tell them it is. In response, many medical professionals and ethicists invoke medical futility to justify refusing to provide life support in such cases. An intervention is futile in this sense when it will not achieve the goal of treatment, not because it is difficult or unlikely to succeed, but because it is simply not going to produce that result under any realistic circumstances. The idea is that, just as physicians can unilaterally refuse to provide interventions like antibiotics for a viral infection (antibiotics work on bacteria, not viruses), they can also unilaterally refuse to provide life support when it is futile to do so because doing so confers no benefit on the patient. Some people object that unilaterally refusing to provide an intervention on

Introduction 5 grounds of futility does not respect patient autonomy and gives too much power to providers. In Chapter 9 Nancy S. Jecker defends this use of the concept of futility against these objections. She argues that these objections are based on an excessively broad conception of patient autonomy, and that there is a social agreement among physicians and society that supports such refusals. Most patients who wish to die, or who no longer benefit from continued life, die when the doctors withdraw or withhold whatever is needed to keep the patient alive (life support, cardiopulmonary resuscitation, antibiotics, and so on). This is described as “letting die.” However, there are patients who are in a condition bad enough that they wish to die, but they do not require life support, CPR, antibiotics, or other interventions to stay alive. Letting them die in the usual way is not an option. Their lives can be terminated only by killing them. Many ethicists believe that killing such patients is wrong even in cases where letting die would be morally permissible if they were on life support and otherwise in the same medical condition. Everyone, however, requires food and water. Some patients receive this by tubes, others by mouth, but both groups can refuse food and water. This is a trend; it is becoming more common for patients to intentionally end their lives by refusing food and water. This need not involve the pangs of starvation and painful dehydration. It can be paced and managed with palliative care so that it is not hard on the patient, yet still leads to death. However, there is some controversy over this practice. Some ethicists oppose it because they believe that only medical treatments are morally permissible to refuse, and they think that delivering food and water, even by intubation, is not a medical treatment. Some object that withholding food and water from the patient is a kind of killing (or suicide, given that the patient requests this), and that killing (and suicide) is morally wrong. In Chapter 10 Paul T. Menzel argues that currently competent patients have a moral right to refuse food and water (even if we consider it a form of suicide, which Menzel does). However, Menzel also says it is less obvious that this right extends to cases where the decision was made in advance, when the patient was competent, and the patient is now permanently incompetent. Menzel also discusses two problematic features of advance directives that refuse food and water. First, incompetent patients usually retain a desire to eat. Second, providing food and water is basic personal care. Menzel argues that the first objection fails in cases where the advance directive clearly addresses the future patient’s desire to eat, and that the second objection fails on the grounds that if patients have a right to refuse medical care, they also have a right to refuse personal care. Menzel undermines the prohibition against killing patients by arguing that withdrawing food and water is a form of letting die. However, there are other ways to challenge that prohibition. One well-known work-around concerns patients at the end of life who are in very severe pain. If the pain is severe enough,

6

John K. Davis

the amount of morphine needed to kill the pain may suppress respiration and kill the patient too. Even so, the patient’s pain needs to be addressed, and it seems cruel to prolong the patient’s agony just because he is not on life support. Some ethicists have argued that this practice is morally permissible because the moral rightness of treating pain outweighs the moral wrongness of killing a patient. The most developed version of this argument invokes the doctrine of double effect. According to one formulation, this doctrine says that an action that has a foreseeable effect (killing the patient) is morally justified if that effect is not intended, and that effect is necessary to achieve a proportionately good effect (i.e., good enough to outweigh the wrongness of the unintended effect). Thomas S. Huddle discusses the doctrine of double effect in Chapter 11. He is skeptical that it can be made to work, primarily because he is skeptical that defenders of the doctrine can overcome the “problem of closeness.” The problem is that it is hard to draw the distinction between intended effects and unintended but foreseen effects in a way that prevents the doctrine from justifying actions that seem clearly immoral. For example, a terrorist who sets off a bomb killing dozens of civilians could argue that he foresaw but did not intend to kill them; his real intention was to weaken the government he considers his enemy and thereby liberate his own people. Even if we assume for discussion that the government he opposes really is tyrannical and unjust, and that his people are oppressed and have the right to rebel, his action seems immoral, yet the doctrine seems to justify it. Defenders of the doctrine must solve the problem of closeness. Huddle is skeptical that they have succeeded. The doctrine of double effect is supposed to justify some cases of killing without also justifying other forms of killing, such as euthanasia and physician-assisted suicide. However, some ethicists argue that euthanasia and physician-assisted suicide are morally permissible. They reject the claim that there is a morally relevant distinction between killing and letting die, and their view is now coming into the mainstream. At the time I am writing this, euthanasia and physician-assisted suicide are still illegal in most places, but euthanasia is legal in the Netherlands, Belgium, Columbia, and Luxembourg, while physician-assisted suicide is legal in the Netherlands and the American states of Washington, Oregon, Vermont, New Mexico, Montana, and California. Huddle addresses the killing/letting die distinction, and the more general doing/allowing distinction it is drawn from. He surveys a variety of arguments for and against the moral relevance of that distinction, and concludes that the doing/allowing distinction is not a good basis for objecting to physicianassisted suicide. Most writing on end-of-life ethics focuses on the person whose life is ending, but there are other affected parties to consider, particularly the patient’s family. Providers who engage in moral reasoning in the clinic routinely consider the interests and wishes of family members. Theoreticians, however, have not.

Introduction 7 But two of our contributors break new ground by considering parties other than the patient. In Chapter 12 Michael Cholbi explores the ethics of grief in the setting of surrogate decision-making about end-of-life care. He argues that grief makes it much more difficult for surrogate decision-makers (who are usually relatives of the patient) to be rational and make decisions that comply with what the patient wanted, or would have wanted had she considered the situation. He also offers some practical suggestions for dealing with this. Bruce Jennings approaches end-of-life decision-making from another angle in Chapter 13, where he uses the concept of solidarity to propose an alternative to the individualistic model of end-of-life decision-making, in which the patient makes decisions alone, to a form of shared decision-making based on concern for others and on mutuality and a relational interpretation of autonomy liberty, dignity, justice, and other more traditional concepts. The last chapter considers the ultimate end-of-life care: the possibility that we might be able to postpone the end of life far beyond present limits, not by keeping a patient alive in an elderly and senescent condition but by slowing, halting, or reversing aging itself. Until perhaps 15 years ago, this possibility was highly speculative and not widely considered plausible. Over the last few years, however, many well-known geroscientists at major universities have begun making public comments to the effect that slowing aging in humans may well be possible, and might be achieved within the coming century. Not everyone is convinced this is a good idea. Some of those who are opposed argue that extended life would not be a good life, that we might grow bored, or fail to use our time well, or that our lives would lose some of their meaning. Some people are concerned about bad social consequences, such as the possibility that the world will face a Malthusian disaster if everyone extends their life, or that a world where life extension is too expensive for everyone to get it is an intolerably unjust world. In Chapter 14 Colin Farrelly addresses the justice objection to life extension and the Malthusian concerns. Farrelly argues that justice requires accelerating the development of life extension so we get it sooner, and that failing to develop it when it is possible is an injustice, for that failure condemns billions of people to the chronic diseases of late life. Until now, it was not possible to protect people from the suffering and indignities of senescence, but now it is, so now controlling aging becomes a requirement of justice. He agrees that justice requires making it available to as many people as possible, and that we must be careful not to let it produce excessive population pressure or damage to the environment. And with Chapter 14 we have gone from asking whether anyone can ever be better off dead to asking what ways of ending life are morally permissible, to asking whether life has to end on any timetable at all, and what ways of failing to extend it are unjust. Perhaps the right to die will be eclipsed by the right not to age.

Part I

The End of Life

1

Is It Possible to Be Better Off Dead? Geoffrey Scarre

Summary Although the coherence of supposing that a person could ever be better off dead has sometimes been questioned on metaphysical grounds, I argue in favour of the common intuition that in some circumstances death could be better for a person than continued life. I proceed to defend three conclusions: (a) that generalisations about when a person may be better off dead are elusive, the answers being personal and existential and affected by how individuals see the world and their place within it; (b) that the question of whether it would be better for a person to die than live arises not only in medical contexts, where she may be suffering severe physical or mental distress or disability, but also in situations in which the prospect of prolonged life is unattractive because living longer would be incompatible with her values; and (c) that because judgements as to whether further life would be worth living are heavily dependent on the subject’s value commitments, criticism from an alternative value-standpoint of a person’s judgement that she would be better off dead is unlikely to change her mind and may beg the question against her.

Introduction: To Be, or Not to Be? Could it ever be the case that one would be better off dead? David Hume had no doubt that it could. “Age, sickness, or misfortune,” he wrote, “may render life a burden, and make it worse even than annihilation” (Hume 1903: 595). In his view, not to be is better than to be in a state of misery that there is no realistic prospect of alleviating. If one can no longer pursue the things that made one’s life worth living, or one suffers from chronic pain and hopelessness, then death may seem a lesser evil than continued life. Thus a patient in the late stages of incurable cancer or a victim of “locked-in syndrome” may well look forward to death as a “happy release” from the prison of life. But not everyone takes this view. Some people believe that even a miserable existence is better than no life at all, holding that the extinction of our conscious existence is the

12

Geoffrey Scarre

worst thing that can happen to us. In John Milton’s poem Paradise Lost, the fallen angel Belial poses the rhetorical question: . . . for who would lose, Though full of pain, this intellectual being, Those thoughts that wander through eternity, To perish rather, swallowed up and lost In the wide womb of uncreated night? (Paradise Lost, Bk. II, ll.146–150) —to which he evidently expects the answer “no one.” Samuel Johnson is one human who undoubtedly would have agreed. Finding the prospect of his own death terrifying, Johnson told his biographer Boswell that “mere existence is so much better than nothing, that one would rather exist even in pain, than not exist” (Boswell 1963: I, 212). Construed as a generalisation about everyone’s preferences, Johnson’s proposition is certainly false; many people, including Hume, would prefer to die than to live on in hopeless misery. (People vary also in regard to how much suffering they would be willing to tolerate before considering that they would be better off dead.) But Johnson may have been offering a normative rather than a descriptive claim: people ought to see any form of existence as superior to non-existence, and if they don’t, they misunderstand their own interests. This claim Hume rejected, although his statement that death may sometimes be regarded as preferable to life should not be misread as the statement that it ought then to be preferred, for Hume seems willing to allow that existential tastes may reasonably differ. Whether Hume is right about this is one of the issues to be addressed in this chapter. In order to throw light on the question of whether we would ever be better off dead (henceforth “BOD” for short), we need to think about what makes lives worth living and about how we should appraise the quality of particular lives in a world of plural and often conflicting values. These are very large themes and there is only space here to sketch some lines of relevant discussion and suggest some tentative conclusions. Following this introduction, I proceed in the second section to address briefly a familiar metaphysical objection to the idea that death, as non-existence, can ever coherently be said to be good for us or preferable to life (of any sort). In the third section I turn to the ethical issues that are at the heart of this chapter, asking what makes lives worth living and how the value of a life is related to its subject’s own desires. The fourth section furthers this discussion by considering the importance of the shape of lives, while the fifth section examines the role of personal values in making lives worthwhile. In a short concluding section, I propose that whether a person’s life is worth living (or whether she would be BOD) is a question requiring some reference to her own value system.

Is It Possible to Be Better Off Dead? 13 One connected topic raised by Hume that can receive only passing mention here is whether suicide (to which we may add voluntary euthanasia) is morally permissible where the quality of a life has fallen very low. For the everpractical Hume, if life offers the prospect of more pain than pleasure, then the sensible thing to do is to end it. Asserting that “suicide may often be consistent with interest and with our duty to ourselves,” Hume dismissed the scriptural injunction Thou shalt not kill as applying only to the killing of others, “over whose life we have no authority.” When a person has become miserable and incapable, he would be better off dead, and since the primary interest in his life is his own, he is not obliged to continue living if he does not want to (Hume 1903: 590). Although Hume sketches a rationale for a right to commit suicide where a life has become one of suffering, it does not follow from the fact, if ever it is one, that a person would be BOD that he or she would be morally entitled to commit suicide; nor does Hume claim that it does. Some people think there are compelling moral or religious reasons to desist from suicide no matter how bad existence becomes. Hume’s advocacy of suicide when life reaches a certain level of unhappiness is premised on the proposition that the life of a human being is his own, to dispose of as he pleases. According to Hume, this is because a person’s life matters primarily to him, secondarily to the society in which he lives, but not at all to the cosmos as a whole (“the life of a man is of no greater importance to the universe than that of an oyster” [1903: 590]). This claim is not beyond dispute, and dissentients might counter that human life is far more valuable than an oyster’s, not because we can make much more impact on the universe than oysters can but because our human capacities for reason, creativity, empathy and moral conduct are amongst the marvels of the cosmos, miracles (one might say) of God or nature (cf. Dworkin 1993: 82). And because a human life is a wonderful thing, it is wonderful to be the possessor of it and wrong ever deliberately to end it. It might even be claimed on this basis that a person could never truly be BOD, no matter how great her suffering. I shall address this claim shortly, in the next section. But the question of whether suicide and euthanasia are morally permissible in any circumstances, or whether it is always wrong to end the “miracle” of life, is one for discussion elsewhere in this book.

Could Non-Existence Ever Be a Good? However subjectively miserable a person’s life may have become, she remains something rare and irreplaceable, a conscious and rational self. Is it therefore better for her to remain alive, in spite of her pain? We have seen that personal responses to this question may differ; but is there in fact a “right” answer to it? It might be suggested that, even for a person who exists in a state of unrelieved misery, her cognitive and affective capacities and her consciousness of the

14

Geoffrey Scarre

world around her enrich her life more than her suffering impoverishes it. But even if she acknowledges the value-making features of her life, she may not consider their contribution great enough to make them worth retaining at the cost of further suffering. To insist at this point that she does not understand her own interests correctly would be tendentious. However, it would be equally so to assert that someone was confused who took the opposite view of his own interests, holding that the retention of conscious capacities weighed more with him than the measure of suffering he endured. The fact is that there is no compelling neutral and non-question-begging way to settle the question of whether people should prioritise the one or the other. Some writers, such as Ronald Dworkin and Jeff McMahan, have pointed out that a human life may have another dimension of value besides the personal value it has for the subject. The “impersonal” or “intrinsic” value of a life is defined by McMahan as “value that is independent of its being valued by or good for anyone, whether its possessor or others” (McMahan 2002: 330). The idea here is that the fact of your or my existence contributes in this impersonal way to the value of the universe; by existing, we make the universe a better place. This could be because we are the handiwork of God (some people believe that human life possesses a certain kind of sanctity as a result of this origin), or, from a more secular point of view, simply because of our possession of “uniquely high capacities” (McMahan 2002: 332). What we make of those capacities—what Dworkin terms the “investments” we make in our life (1993: 93)—can then affect our life’s impersonal as well as its personal value. An obvious question to ask here is why an individual should trouble herself about how her existence contributes to the value of the universe. But it would be misleading to suggest that the impersonal view of her life is simply not the subject’s business. To be sure, she may be more inclined to ask what value her life has for other people in her social world than how the preservation of her high capacities enriches the cosmos as a whole. And consequently she may be more willing to prolong a painful life for the sake of other people she cares about than because the universe will lose value by her demise. Yet the latter consideration may have some, if limited, power to move her. Alternatively, she may believe that her life has intrinsic or impersonal value because it is a gift of God. But even if life’s status as a divine gift places certain constraints on what we might do with it, that does not mean that we would not be BOD if the quality of our life should fall below a certain threshold. From the personal point of view, our life has run its course and no longer offers a sufficient return for the investments we have made in it. McMahan writes that “there is . . . no sense in the idea that the investments that were once made in a person’s life make it important to extend his life when, for whatever reasons, there is no prospect of those investments yielding further good” (2002: 46). From the personal perspective, the subject would

Is It Possible to Be Better Off Dead? 15 now be BOD. But it is also very doubtful whether from the impersonal view of the universe much is to be gained by the prolongation of a life in these circumstances, since there may be little chance of a person’s high abilities yielding further good when her life has become focused on its suffering. (In any case, the ending of an episode of suffering may plausibly be regarded as a betterment of the condition of the universe.) It is not clear why this should not be God’s view too. Gifts can go sour and become unwanted, and this can happen even with the God-given gift of life. And while God might conceivably wish people to suffer occasionally in order to expiate their sins or to strengthen their virtues, it is obscure why a benevolent deity should favour a person’s living on in a state of misery that serves no constructive purpose. As this is not an essay in theology, I shall take as a working assumption what, following Fred Feldman, has been called the “termination thesis”: namely, that at death we cease to exist as personal selves (though we may leave identifiable physical remains behind) (Feldman 1991: 205). On this assumption, there can be no positive value in (the state of ) being dead, since non-existence has nil utility-value. Perhaps the most plausible and intuitive view is that life—that is, the state of being alive—is valuable not so much in itself as instrumentally; that is, life matters because it is the underlying necessary condition of our being able to realise the goods available to human beings. Only the quick and not the dead can reason, love, act kindly and virtuously, create useful or beautiful things, rise to challenges, solve problems and experience joys and salutary sorrows. “I believe that no man ever threw away life while it was worth keeping,” said Hume (1903: 595); being alive is valuable so long as we are able to flourish, and once we are not, life loses its purpose. This view would be misconceived if life were always worth keeping, whatever its quality. But it is hard to find a persuasive reason for believing this. “But wait!” an objector might say. “Something has been forgotten here. Life might indeed become very painful, distressing or empty for an individual, but there is no need to claim there to be some mysterious objective value to human life to justify rejection of the idea that she would be BOD; for if we accept the termination thesis, then we deny that there is any she after her death to have any quality of existence, good or bad. To say of someone that she would be BOD is not analogous to saying that she would be better off moving to another address or changing her job, for whereas in those cases we envisage a different possible future existence for the subject, in the former we imagine there being no future subject. The human goods or ills in terms of which we think of ourselves as being well or badly off are necessarily good and ills of subjects; so, in the absence of a subject there can be no good or ill. Hence we should dismiss any claim that a person would be BOD as incoherent, since after death there is no longer a subject to be either better or worse off than the formerly existing person.”

16

Geoffrey Scarre

This argument will have a familiar ring to anyone conversant with the extensive literature on the question of whether death is ever truly an evil. Ancient and modern Epicureans hold that a person can be the subject of good or bad fortune only provided that he exists, and this “existence condition” has been frequently cited to warrant asserting that death is no misfortune, since before a person dies he plainly does not suffer the putative evil of death and after he has died he is no longer around to suffer it; hence there is no time at all at which he suffers it (see Epicurus 1926: 85). As David Benatar has noted, the same line of argument, if valid, also establishes that death cannot benefit a person whose life has become dreadful: “Death cannot spare anybody from anything any more than it can deprive anybody of anything” (Benatar 2006: 201). For to be dead is not to be anybody anymore. In the present discussion I shall take a short way with the objection just cited—some readers will undoubtedly think too short a way—since I wish to move on from this well-trodden territory to some less-commonly addressed, and more distinctively ethical, issues about the conditions and circumstances in which someone might reasonably be held to be BOD. Briefly, then, one might begin by questioning whether the existence condition applies in what McMahan has called the “special case” of death. Normally a person has to exist to be the subject of good or ill fortune. Yet as McMahan complains, “To insist that [death] cannot be an evil because it does not meet a condition that most if not all other evils satisfy is tantamount to ruling it out as an evil simply because it has special features” (McMahan 1993: 241); put differently, it begs the question in favour of the universal applicability of the existence condition. An Epicurean would be likely to retort that this lays the charge of petitio at the wrong door, and that it is those who deny that the existence condition applies in the special case of death who need to justify their stance. Yet McMahan’s contention that the existence condition does not obviously apply in the case of death has strong intuitive support. It is natural to think that, where death is an evil, it is an evil of a privative sort, in the sense that when it occurs, the subject is deprived of further life, where that life would have been, on the whole, good for him. Raymond Angelo Belliotti puts the point with admirable succinctness: “Death deprives us when it cuts short a life that would otherwise be an overall good” (Belliotti 2009: 138).1 And Thomas Nagel observes that “if death is an evil, it is the loss of life, rather the state of being dead, or non-existent, or unconscious, that is objectionable” (Nagel 1979: 3). Bernard Williams remarks that when we want something, we have “reason for resisting what excludes having that thing; and death certainly does that, for a very large range of things that one wants” (Williams 1973: 85). Take a subject—call him Mark—who dies in a car crash at the age of 30. Suppose that if the crash had not occurred, Mark would have enjoyed many further years of generally happy life. In this case, common sense is clear that Mark’s death at 30 is bad for him. It is true

Is It Possible to Be Better Off Dead? 17 that it would be philosophically more comfortable if this intuition could be buttressed by a rationally unassailable demonstration of why the existence condition does not apply here; and this demonstration is still awaited. But for present purposes I shall take it that the common view is correct and that death for someone in Mark’s circumstances is genuinely bad for him. And now one can imagine a different case, in which at the age of 30 Mark succumbs to a painful and incurable illness that deprives him of the capacity to pursue effectively the projects and interests that gave his life its meaning. Depressed and frustrated, Mark desires an end to his existence. He may not articulate this desire in terms of being “BOD”; he may instead say that he wants his life to cease now that it has no longer gives him satisfaction. However he puts it, he is expressing a plain preference for a state of affairs in which he dies soon (or as soon as he has put his affairs in order) over one in which he lingers for longer. If death is bad, where it is, because of what it takes away from us, superficial symmetry might suggest that when death is good for us, that must be because of what it provides us with. But this is not so. Death, being non-existence, cannot supply us with any positive goods. Where death benefits Mark, it must also be because of what it removes from him: in his case the evils of pain, frustration, depression and so on. Thus both good and bad deaths are to be characterised in terms of loss; each kind of death takes something away from us, though in the latter case the “loss” is rather a liberation than a deprivation.2

What Makes Lives Worth Living? From this point I shall take for granted that it is not incoherent to suppose that a person could be BOD. The questions now arise under what conditions a person might be BOD, and how it might be determined when these conditions obtain. To this end it is necessary to consider in what circumstances a person can be said to have a life worth living (henceforth “LWL”). There is a need to tread cautiously here, as the expression “a life worth living” has become politically charged in recent years, with some defenders of the rights of the disabled and promoters of the hospice movement contending that there is no such thing as a life that is not worth living. Read as the claim that everyone’s life matters, and that illness, disability, old age and dependency do not diminish a person’s entitlement to equal moral consideration, the denial that there are lives not worth living is unexceptionable; all human beings, we may readily grant, are first-class citizens of the kingdom of ends (as Kant termed the set of morally considerable beings). But while there are no worthless people, there are lives that acquire deeply undesirable features, often through no fault of their subjects. When Mark at 30 falls victim to serious and incurable illness, his life enters a phase in which its quality falls far below that previously enjoyed. But

18

Geoffrey Scarre

there is no decline, either sudden or gradual, in Mark’s moral status. The moral worth of a person is one thing, the subjective quality of his life another. This is scarcely a very subtle distinction; yet it is one that may sometimes be missed by well-meaning people, amongst them some who reject the moral propriety of voluntary euthanasia or its variant, assisted suicide. To insist that even voluntary euthanasia must always be wrong because everyone’s life is worth living is to conflate the high and equal value of every living human being with the quality of the lives that people actually lead, which plainly varies. To propose that a person has the moral right to end his life when it has reached an unsatisfactorily low level is not to imply that he has become outcast from the kingdom of ends, or that he is taking up space in the world that others could better occupy. In the remainder of this essay I shall therefore speak of lives that have (wholly or largely) ceased to be lives worth living without intending any negative judgements about the moral standing of their subjects. It has become customary to distinguish three major theories of well-being, or of what makes life worth living. Hedonistic theories weigh the value of a person’s life according to the extent that it is pleasant rather than painful. Some versions of hedonism follow J. S. Mill in favouring “higher” spiritual, intellectual or aesthetic pleasures over “lower” physical or mentally undemanding ones (Mill 1910: ch. 2). Desire-satisfaction accounts appraise the quality of a life in terms of its subject’s success in satisfying her (reasonable) desires, especially, though on most versions not exclusively, those desires that relate to herself, her own circle and connections, rather than to the world at large (her desire to pass a particular exam, say, as distinguished from her desire for an end to world hunger). Desire-satisfaction theories usually add that to have a good life a person needs to have a range of lively and forceful desires, to rule out the idea that someone who has very few or very weak desires is in the running to attain a high level of well-being (see Brandt 1979). Finally, objectivelist theories measure the well-being of a life by the presence or absence in it of certain objectively good things, irrespective of whether those things give pleasure or satisfy the subject’s desires. Common candidates for the list of objective goods include the possession and use of moral, intellectual and aesthetic capacities, including virtues and intelligence, beneficent agency, liberty, health, a sense of beauty and self-respect; pleasant experiences and satisfied desires may also be included in so far as their objects are worthy rather than footling (see, e.g., Griffin 1986: ch. 3). Although these theories are commonly presented as rival accounts of well-being—and they undoubtedly do not always yield the same answer to the question of whether a particular life is flourishing—there is a case for saying that each has correctly identified some salient feature of human existence that can contribute, positively or negatively, to the overall quality of a human life. A hedonistic account might rate poorly the life of a manic-depressive professor

Is It Possible to Be Better Off Dead? 19 of philosophy of extraordinary intellectual abilities and achievements whose existence scores highly on an objective-list account. But it is easy to see that the professor’s life would be better still if he threw off his depression and started to take more pleasure in things. Desire-satisfaction theories that particularly value desires to have pleasant and avoid unpleasant mental states come closest to hedonism (though the theories do not completely converge, since pleasant mental states can turn up unexpectedly, without having been desired beforehand). The lives that are most worth living normally contain high levels of subjective satisfaction (pleasures, or fulfilments of desire) as well as objective goods. Derek Parfit has proposed that the best life is that in which the subject takes his or her greatest pleasure in projects and activities of high objective value (though this leaves room for debate on which these are) (Parfit 1986: Appendix I). Even if one is not committed to a desire-satisfaction theory of well-being, one might think that one particular desire is essential if future life is to be good for a person—namely, the desire to live longer. Yet if that desire is to be rational, it requires underpinning by one or more desires that give a reason for wanting further life: as Williams puts it, there needs to be some desire that “propels him on into the future.” This desire cannot be one “that operates conditionally on his being alive, since it itself resolves the question of whether he is going to be alive”—that is, it provides him with a reason for living. Such a desire Williams terms “categorical” (1973: 86). Although, he concedes, a bare desire to remain alive may reflect an evolved instinct for self-preservation, a rational desire to that effect “has to be sustained or filled out by some desire for something else, even if it is only, at the margins, the desire that future desires of mine will be born and satisfied” (1973: 86–87). If one lacked at least one such categorical desire, one would have nothing to reply to the question “Why do you want to go on living?” Note, however, that even if one survives to satisfy that desire it is still not guaranteed, according to a hedonistic or objective-list theory, that one will be better off than if one had died beforehand. Suppose, for instance, that the categorical desire that propels one into the future is a desire to assassinate one’s greatest enemy but that after carrying out the fell deed one is racked with remorse; one might now wish that one had died sooner rather than have become a murderer. As Christopher Belshaw points out, while one might desire to live longer, “[i]t may be that the extra life available to you is one that is not worth living” (Belshaw 2009: 115). Williams’s claim that a desire to stay alive must be sustained by one or more categorical desires leaves open the question of whether someone who had no categorical desires or a desire to remain alive could benefit from any future life. It seems that although no desires propel him into the future, a satisfying future could still await him. Thus a maltreated slave who has given up wishing for release from his bondage (which seems to him as pointless as

20

Geoffrey Scarre

wishing for the moon) and who wants only to die and end his misery may, unbeknown to him, be on the verge of liberation and a much happier existence. If he dies now, he foregoes this better future. His belief that he would be BOD is justifiable by reference to his present but not to his future circumstances. This example vindicates Jeff McMahan’s claim that “it seems clear that the loss of future goods that are undesired at the time of death can contribute to the badness of death” (McMahan 2002: 182). Commenting on this claim, Belshaw suggests that the potential goods would, in order to be goods, have to be desired at some stage in the future if one were not to die beforehand (2009: 117); but this seems strained, as it is neither impossible nor improbable that a significant good may materialise out of the blue without ever being previously desired. (On an objective-list account, it would not even have to be regarded as desirable, provided that it really were good for one.) These considerations are relevant to the situation of Mark, who, struck down at 30 by a painful, debilitating and incurable disease, believes that he would be BOD. Mark may entertain the hope that one day a cure will be found for his disease; but because he has no expectation that that will happen in the foreseeable future, it would be inaccurate to characterise this as the categorical desire that he, Mark, should be cured. One does not waste time desiring the seemingly impossible or unattainable.3 Imagine, however, that contrary to Mark’s (assume well-grounded) expectation that no cure will be found for his condition any time soon, one will in fact be discovered in five years’ time. If Mark could somehow see into the future, he might well abandon his belief that he would be BOD. He might also do so if he learned that there was a 50% probability that a cure for his condition would be available in five years. But there is no uniquely rational way to perform such existential-actuarial calculations, and other people in a similar situation to Mark’s might legitimately arrive at different conclusions to his.

Contents of Lives, Shapes of Lives In judging the quality of a life, then, and in deciding whether and when a person has a LWL or, on the other hand, would be BOD, there are multiple factors that should be taken into consideration: factors concerning pleasures and pains, factors concerning desires and factors to do with the objective quality of capacities, interests, activities and attainments. But beyond all these, as Mark’s case indicates, there are important structural issues having to do with the overall organisation of our lives and how the individual phases of our existence fit together. Many recent writers have emphasized the importance of such structuring, frequently comparing it to the planning of a narrative. For example, John Martin Fischer suggests “that we think of ourselves as authors in the sense that we freely endow our lives with a special sort of meaning: an

Is It Possible to Be Better Off Dead? 21 irreducible narrative dimension of meaning” (Fischer 2009: 21). This proposition could be taken either descriptively or normatively: as a claim about how people actually regard their lives or as a proposal for how they should, or usefully could, regard them. Maybe not many people explicitly adopt the metaphor of authorship when reflecting on their lives, but any thoughtful person will occasionally assume a perspective that has similarities to that of a writer who plots a story. Assuming a “narrative view” of our lives means taking the long view, and assessing the value of individual episodes and phases by the contribution they make to the whole. This can help us to avoid aimless drifting, time-wasting or becoming at cross-purposes with ourselves, and hence to achieve meaning in our lives. However, it is easy to overestimate our abilities to predict and to control future events, though recognition of these limitations should not deter us from planning but remind us of the need to have alternatives on hand when pet schemes go awry. Common prudence dictates that we should make our significant life-choices with care, and it would be hard to be effective agents in the world if we formed no long-term goals, in accordance with our values. To adopt the narrative conception is to construct one’s life like a novel, with a continuous plot-line, rather than like a series of stand-alone vignettes. Still, it would be too bold to assert that lives lived in a more serendipitous fashion were doomed to absurdity; a person who is content to “take things as they come” may achieve a meaningful life, though in doing so she is offering precious hostages to fortune, drifting before the winds of life like a ship without a steersman. In any case, one could incorporate a carpe diem attitude to life within a narrative perspective if one believed that “seizing the day” was the best way to make the most of one’s life as a whole. None of the three theories of well-being previously distinguished is entirely satisfactory when it comes to thinking about lives holistically, where the emphasis is shifted from the ingredients of the good life to its shape. Although desires are of necessity directed on the future, and pleasures and other goods may be the objects of keen anticipation, a comprehensive account of wellbeing should acknowledge the need to select amongst our possible futures in order to attain coherence. Our lives should not be mere ragbags into which we try to pack as many good things as possible (pleasures, satisfied desires, skills and abilities, rewarding experiences and relationships and so on) but time-extended enterprises during which we grow, mature and age, undergo various rites of passage, struggle with life’s multiform challenges and attempt to mould, as far as is possible, our own destiny (for inevitably our narratives will also intersect with others that affect their course). When Mark decides to put up with present suffering for the sake of arriving at a future promised land of health, he shows how short-term and long-term reflection can diverge on the question of whether a person would be BOD. Were he predominantly concerned with his present and short-term-future state,

22

Geoffrey Scarre

and relatively uninterested in his more distant condition, he might think that the benefit tomorrow is not worth waiting for and that early death would be preferable. While this represents a prioritising of a short-term over a longerterm perspective, Mark might also reach the same conclusion on a more holistic view of his life, where he judges it better to escape an immediate future of certain suffering than to gamble on a less certain distant future that will be pain-free. There is little to be said in favour of extreme short-termism, in which the subject takes no care at all for the morrow, unless one takes a counterintuitive view of the self as being in a state of perennial flux, with rational selfconcern extending only as long as the transient self lasts. Here I shall simply take for granted (against such “episodists” as Galen Strawson; see Strawson 2004, 2007) that the common-sense view of a unitary self that endures over the course of a lifetime is correct. On this assumption, it would not be rational for Mark, unless his current pain were utterly intolerable, to claim he would be BOD if he also believed that the prospect of a cure was imminent, with the prospect of many years of healthy life beyond it. It is possible to imagine circumstances in which a lengthy period of suffering or hardship might even be looked forward to as a good, in virtue of the role it plays in a person’s life-plan. Imagine that Sir Lancelot has been living a wicked but happy life until it dawns on him that he is walking the high road to perdition unless he can make his peace with God. A holy monk advises him that the only sure way to save his soul is to spend the remainder of his days in prayer and repentance, mortifying his flesh and enduring hunger, thirst and cold. Sir Lancelot heartily dislikes his new existence and reflects that, were only his present comfort to be weighed, he would be BOD. But his greatest fear is that he will die before he has suffered enough to placate a vengeful God. This example may seem remote from contemporary experience, but one can easily recognise modern analogues in which a person prefers to endure some severe suffering rather than seek an early escape through death. This may be because, like Sir Lancelot or like the remorseful murderer Raskolnikov in Dostoevsky’s Crime and Punishment, he wishes to expiate his previous offences, or because he has unfinished business that requires his attention before he dies, or because he wants to have time for reflection on life’s important questions. It is from a narrative, or holistic, perspective that the full significance of such exercises can best be seen. The individual who in a busy, crowded life has rarely stopped to reflect on fundamental questions may, after being struck down by serious illness or disablement, value at last having the time and opportunity to do so. She may also recognise that she has other loose ends to tie up: debts, literal or figurative, to be received or discharged; forgiveness to be offered or sought; damaged relationships to be repaired; unfinished projects that require completion or delegation; and responsibilities to be transferred. She rejects the idea she could be BOD while there remains so much to do.

Is It Possible to Be Better Off Dead? 23

Judging the Value of Life Deciding whether one would be BOD can be difficult. Views of what makes a LWL vary not only from culture to culture but also often between people within the same culture. Dilemmas can also arise for individuals where a clash of values resists resolution. Would it be better to die soon or to live for several months in intense pain where that time is required to complete a project that plays a major role in one’s life narrative? Even if one happens to desire one of these options more strongly than the other, one may hesitate to declare that it is the more desirable. So far we have been looking at cases in which a person has fallen victim to a distressing medical condition of sufficient severity to raise the question of whether her life is still worth living. Such cases have of course been the main focus in the bioethical literature. However, there are other conditions than those involving illness or pain (whether physical or mental) that can cause people to find the prospect of death preferable to that of further life. A person who feels that she has achieved her main life-objectives may have little relish for living on into old age, even if that old age would be a relatively healthy and trouble-free one. If the narrative structures that gave her life its meaning have been largely completed, and her future appears in the light of a dull and tiresome appendix to her story, then she may, like Holy Simeon, be ready to pronounce her “Nunc dimittis.” Eighty-six-year-old Dutch citizen Edward Brongersma obtained legal permission in 1989 to request assisted suicide on the simple ground that he was “tired of life.” (I am indebted to Christopher Cowley for this example.) It would not be irrational for someone whose interests are centred on projects and activities practicable only in youth and middle age to reject an old age that relentlessly confronts him with a sense of what he has been but is no longer. Not for him the tranquil life of reading, tending the garden, walking the dog and waiting for the next mealtime. “If that is the life I should have,” he muses, “then I would be BOD.” Admittedly, his failure to find a way of living a satisfying life in a gentler key may argue some shortfall in imagination. But he may reject the envisaged future life on the ground that it is “just not me,” its goods costing a radical transformation of self that he is unwilling to contemplate. For a different kind of instance, in the traditional Japanese military code of bushido, a general defeated in battle, or a soldier who had let down the emperor, could not remain alive without forfeiting whatever shreds remained of his honour; to recover what he had lost he had to commit ceremonial suicide (hara-kiri). A samurai who, through cowardice, failed to take his own life would be a sorry sight in his own eyes as well as in others’. This honour code may now seem extravagant but in classical Japan there would have seemed no LWL for the once-shamed samurai. The European tradition of chivalry too

24

Geoffrey Scarre

(which survived in key respects until the Great War) could be very demanding of the knight or soldier aspiring to military honour. Reputedly, at the close of the Battle of Waterloo, the rump of Napoleon’s Old Guard chose to be shot down en masse rather than accept an invitation to surrender; in their eyes, death came before dishonour, and a life following surrender would have been a life not worth living.4 “A life not worth living”—or a life that seemed to them not worth living? Which description should be preferred? To choose to die for the defeated emperor on the field of Waterloo rather than save their lives by laying down their arms may appear an irrational choice to those of us to whom the deathbefore-dishonour code of the Guards is alien; and for anyone who finds it alien, such a choice would be irrational. But it does not follow from this that it was equally irrational from the personal perspective of the Guards. Or, to put things in a more guarded way, it was not obviously irrational for them on the field of Waterloo, given the character of their personal narratives to that date. If, against their will, they had been spared, some of them might eventually have come to be glad of the fact, having adapted their narratives to fit their changed circumstances. Even from a single individual’s perspective, the answer to the question of whether the life before him is one worth having can vary over time. Therefore the possibility that one might change one’s mind or one’s standards later should always be considered by a person who presently feels he has nothing to live for. Only if one is alive can one enjoy the luxury of second thoughts. Of course, it may sometimes be possible to identify inconsistencies or errors within a person’s own thinking, and by pointing them out, we may persuade him that, relative to his own character and values, he would not really (as he thinks) be BOD. Maybe we can persuade him that he overestimates the scale of his current problems or overlooks the likely future goods his life will hold. There can be no reasonable moral objection to such light-touch paternalism, and we might even be held to be failing in our duty of care if we neglect to provide it. Neither is there anything very problematic about inviting him to consider viewpoints alternative to his own. It would be too pessimistic to assume that where value-systems conflict, there can never be constructive dialogue between them; what Amartya Sen has called “globally sensitive questioning” can be of enormous benefit in helping societies and their members to scrutinise their local beliefs (Sen 2010: 76). The moral problems arise if we lose sight of the difference between dialogue and indoctrination. Heavy-handed paternalism that treats persons’ sincerely held values or core beliefs dismissively may succeed in persuading them to change their minds but shows little respect for their autonomy. It might be suggested that an objective-list theory of well-being could take up some of the slack here, and by identifying some fundamental human goods provide genuinely objective reasons for thinking that certain lives are worth

Is It Possible to Be Better Off Dead? 25 living even if their possessors do not think they are. So an objective list might rank (say) the possession of intellectual or creative faculties much higher than the retention of honour according to an aristocratic code like that of the samurai. By the standards of this list, the defeated samurai who commits hara-kiri in the belief that he no longer has a LWL is quite mistaken. But there are two problems with this. The first is that one person’s “objective” list may not match another person’s “objective” list (and none of us have access to God’s objective list of what is best for human beings). A great deal of intercultural conversation may have to take place before the differences can be resolved, there being rarely a shortcut to agreement. The second problem is that it is unsafe to assume that there is only one way of ranking values objectively. Taking a pluralist view of values and the good life does not mean that we should be wholly uncritical of others’ values and tolerant of the acts that flow from them; as Sen has remarked, some culturally accepted practices (e.g., female genital mutilation) are so “tyrannical and nasty,” and so restrictive of basic freedoms, that they merit objection wherever they are found (Sen 2010: 237fn). But pluralism is inconsistent with the claim that there is one unique ordering of objective values. Hence there are serious limits to the extent to which an objective-list conception of well-being can provide decisive criteria for determining whether a particular person’s life is worth living. Respecting others’ autonomy involves, amongst other things, respecting their right to make their own significant life-choices (or, in other words, to construct their own life-narratives without external interference). But what should we do where a person comes to a seemingly autonomous conclusion that she would be BOD, and subsequently seeks to end her life, on the basis of values that appall us? Take, for instance, the practice of sati, which is still reported from time to time as occurring in rural India. Does a woman who believes that she would not have a LWL if she lived on as a widow after her husband’s death have the right to have her wish for self-sacrifice respected? Should she be allowed to immolate herself on her husband’s funeral pyre, or ought others to restrain her from that grisly death in what they, but not she, may describe as “her own best interests”? How far should pluralism about values be taken? In this case it might be argued that paternalistic intervention to prevent her suicide is justified because she can be regarded as the victim of a false consciousness that impedes her autonomy, the product of a bad ideology that perceives women as merely the chattels of men. For sati might certainly be classed amongst the “tyrannical and nasty” practices that have no place in the civilised world. Still, if a woman contemplating sati cannot be persuaded to alter her outlook, then preventing her from living up to her values means enforcing on her further life of which she may be able to make no good sense, failing as it does to fit her preferred narrative. Although this may still be the right thing to do, there is a cost to be paid for doing it.

26

Geoffrey Scarre

Conclusion It is time to state briefly some conclusions. The first—and perhaps the main— point that has emerged from this enquiry into the conditions under which it may be said of a person that she would be BOD is that generalisations are elusive. That is because so many of the questions are personal and existential, their answers depending on how particular individuals see the world and their place within it. So it would be rash to propose that in every instance where a person has begun to experience continuous or chronic pain beyond a certain threshold of intensity, she can no longer have a LWL. Such may often be the case, but not where she has unfinished business whose completion she sees as worth the sacrifice of personal comfort. Second, the question of whether a person would be BOD arises not only in medical contexts, where she may be suffering severe physical or mental distress or disability, but also in circumstances in which the prospect of prolonged life is unattractive on account of the affront that living longer would do to her values. The soldier who prefers death to dishonour, or the early Christian saint who chose martyrdom rather than renege on his Christian faith, believed that a morally or spiritually sullied life would be a life not worth living; for him, violent death on the battlefield or in the amphitheatre was the lesser of two evils. Are such judgements justified? The third major conclusion is that often there is no definitive way of answering such questions. This is because the judgements depend heavily on a person’s value commitments, and where different people’s commitments clash, there may be no impartial, non-question-begging way of arbitrating between them. The Hindu widow who immolates herself on her husband’s funeral pyre performs what to Western eyes appears an appalling and unnecessary act but one that, from her perspective, is a matter of duty whose non-performance would mean perpetual dishonour. We may have no doubt she is misguided, but unless she can be brought to share that view, preventing her fatal act may leave her facing a life for which she has little relish. In the light of her present values, the appropriate end to her life’s narrative would appear to be now. Conflicts of values can also occur within a single individual, and there too they can cause perplexity. A soldier may be torn between thinking that running away from the enemy to save his life would make him unworthy of it, while wondering, like Shakespeare’s Falstaff, what use honour would be to him if he were dead. That soldier faces an existential question that he needs to answer in constructing his life narrative. If we are lucky, we will never have to consider the question of whether we would be BOD. But it could happen to any of us. If it does, we have to hope that we shall feel sufficiently in control of our own narrative to know how to answer it.

Is It Possible to Be Better Off Dead? 27

Related Topics Chapter 2, “How Does Death Harm the Deceased?” by Taylor W. Cyr Chapter 3, “The Significance of an Afterlife” by Benjamin Mitchell-Yellin Chapter 4, “The Severity of Death” by Jens Johansson

Notes 1. Belliotti continues, with equal succinctness: “Death can be seen as a benefit when it mercifully terminates a life that would otherwise be an overall detriment” (2009: 138). 2. One could complicate this formula by saying that death in Mark’s case liberates him from the deprivation of goods that his illness has caused him. Even so, it is only the lesser of two evils; it would be better still for Mark’s condition to be cured and his goods restored to him. 3. This proposition might be questioned. One might say, “I would like to go to the moon” where one has no expectation that one will ever have the chance to do so. Does this express a real desire? Note that it has a quite different feel from the statement “I want to go to New York this summer,” which expresses a desire that one recognises is realisable. Maybe one could say that the former expresses a “thin” and the latter a “thick” desire. Yet the fact that there is something paradoxical about saying, “I desire to [have] X, but I know that I never shall [have] X,” suggests that such “thin” desires would be mislabelled as such, and are not genuine desires but rather of the genus of “pipe dreams.” 4. The Guards’ defiance at Waterloo was said to have been expressed by the proud words of their leader, General Cambronne: “La Garde meurt, elle ne se rend pas.” Regardless of whether Cambronne actually said those words, the eyewitness sources agree that the Old Guard at Waterloo made a highly courageous and costly stand (later to be immortalised by Victor Hugo in Les Misérables).

References Belliotti, R.A. (2009) Roman Philosophy and the Good Life, Lanham, MD: Lexington Books. Belshaw, C. (2009) Annihilation: The Sense and Significance of Death, Stocksfield: Acumen. Benatar, D. (2006) Better Never to Have Been: The Harm of Coming into Existence, New York: Oxford University Press. Boswell, J. (1963) The Life of Samuel Johnson (2 vols.), London: Dent’s Everyman’s Library. Brandt, R. (1979) A Theory of the Good and the Right, Oxford: Clarendon Press. Dworkin, R. (1993) Life’s Dominion, New York: Alfred A. Knopf. Epicurus (1926) The Extant Fragments, C. Bailey (ed.), Oxford: Clarendon Press. Feldman, F. (1991) “Some Puzzles about the Evils of Death,” Philosophical Review 100(2): 205–227. Fischer, J.M. (2009) Our Stories: Essays on Life, Death, and Free Will, New York: Oxford University Press. Griffin, J. (1986) Well-Being: Its Meaning, Measurement and Moral Importance, Oxford: Clarendon Press.

28

Geoffrey Scarre

Hume, D. (1903) “Of Suicide,” in The Essays of David Hume, London: Grant Richards, 585–604. McMahan, J. (1993) “Death and the Value of Life,” in J.M. Fischer (ed.) The Metaphysics of Death, Stanford, CA: Stanford University Press, 231–66. McMahan, J. (2002) The Ethics of Killing: Problems at the Margins of Life, Oxford: Oxford University Press. Mill, J.S. (1910) “Utilitarianism,” in J.S. Mill (ed.) Utilitarianism, on Liberty, Representative Government, London: Everyman’s Library, 1–60. Nagel, T. (1979) “Death,” in T. Nagel (ed.) Mortal Questions, Cambridge: Cambridge University Press, 1–10. Parfit, D. (1986) Reasons and Persons, Oxford: Clarendon Press. Sen, A. (2010) The Idea of Justice, London: Penguin. Strawson, G. (2004) “Against Narrativity,” Ratio 17(4): 428–452. Strawson, G. (2007) “Episodic Ethics,” in D.D. Hutto (ed.) Narrative and Understanding Persons, Cambridge: Cambridge University Press, 85–116. Williams, B. (1973) “The Makropulos Case: Reflections on the Tedium of Immortality,” in B. Williams (ed.) Problems of the Self, Cambridge: Cambridge University Press, 82–110.

Further Reading Bradley, B. (2009) Well-Being and Death, Oxford: Clarendon Press. Feinberg, J. (1984) The Moral Limits of the Criminal Law, Volume One: Harm to Others, Oxford: Oxford University Press. Luper, S. (2009) The Philosophy of Death, Oxford: Oxford University Press. Pitcher, G. (1984) “The Misfortunes of the Dead,” American Philosophical Quarterly 21(2): 183–188. Taylor, J.S. (2012) Death, Posthumous Harm, and Bioethics, New York and Abingdon: Routledge. Velleman, J.D. (2003) “Narrative Explanation,” Philosophical Review 112(1): 1–26.

2

How Does Death Harm the Deceased? Taylor W. Cyr

Summary The most popular philosophical account of how death can harm (or be bad for) the deceased is the deprivation account, according to which death is bad insofar as it deprives the deceased of goods that would have been enjoyed by that person had the person not died. In this chapter, the author surveys four main challenges to the deprivation account: the No-Harm-Done Argument, the No-Subject Argument, the Timing Argument, and the Symmetry Argument. These challenges are often raised by Epicureans, who (following Epicurus) claim that death cannot harm the deceased, and each challenge is addressed in Thomas Nagel’s classic essay, “Death,” which has been very influential on recent developments in the literature on the philosophy of death. The author of this chapter summarizes some of these recent developments as the challenges are considered.

Introduction At the beginning of his great and influential essay, “Death,” Thomas Nagel writes, “If death is the unequivocal and permanent end of our existence, the question arises whether it is a bad thing to die” (Nagel 1970: 73). Most of us do in fact believe that death can, at least in some cases, be bad for the one who dies. In other words, most of us think that death can harm the deceased. (Note: I take the expressions the badness of death, on the one hand, and the harm of death, on the other hand, to be roughly equivalent, so I will use these and related expressions interchangeably.) Now, of course, one reason that some of us believe that death can be bad is that some of us believe that death is not the end; some of us believe in an afterlife, where some of those who have died will experience punishments involving suffering (which, we can all agree, would be a bad thing). Others of us share Nagel’s assumption that death is the permanent end of our existence, not something we will experience but rather an experiential blank. For the purposes of this chapter, let’s assume that Nagel is

30

Taylor W. Cyr

right about this and that death really is the end. (If you don’t share this assumption, you can imagine this chapter as exploring a hypothetical scenario: what would you say about the harm of death if you didn’t believe in an afterlife?) Nevertheless, most of us want to say that death is one of the greatest evils that can befall a person. As it turns out, it is quite puzzling to explain how death could harm the deceased, given that the deceased no longer exists, and there are interesting and difficult challenges to the belief that it is even possible for death to harm the deceased. Before we continue, however, three clarifications are in order. First, when we talk about the harm or badness of death, we are talking about a state of nonexistence—the state that follows a person’s life and that begins when the person has ceased living. We must not confuse death with dying, the latter of which is what happens during the end of a person’s life, before the person has died. It is not very puzzling how dying can be bad for a person who is experiencing it, since dying all too often includes suffering. Second, the philosophical issues considered here are concerned with the possibility of death’s badness for the deceased and not for, say, the person’s loved ones who continue to exist after the person’s death. Here again, it is not very puzzling how a person’s death can be bad for that person’s loved ones, since they will very likely experience suffering as a result. Lastly, in discussing how death harms the deceased, we are going to be concerned with how death can harm the deceased. In other words, our concern is whether it is even possible for death to harm the deceased. Thus, even if we decide that death can be bad for the one who dies, we are not thereby committed to the claim that death is always bad—that it is bad in every case. It might turn out that death always harms the deceased, but perhaps not. Our concern is how death could harm the deceased. So, given that death is an experiential blank, can death be bad for the one who dies? There are two main answers to this question in the contemporary philosophical literature on the philosophy of death. The first says that it is possible for death to be bad for the one who dies and that it is bad insofar as it deprives the deceased of goods that would have been enjoyed by that person had the person not died. For obvious reasons, this view is called the deprivation account, and it has several defenders (see, e.g., Feldman 1992; Bradley 2009; Fischer 2009), though the details of their accounts vary. If you are like me and find yourself with the belief that death can, at least in some cases, be bad for the one who dies, then the deprivation account may strike you as obviously true, or at least as very plausible. According to the deprivation account, death harms the deceased by taking away opportunities to continue enjoying the goods of life. This might happen by depriving the deceased of potential pleasures, such as the pleasures of eating good food or of listening to music, or it might happen by depriving the deceased of opportunities for certain desires to be satisfied, such as the desire to write a book or to see one’s grandchildren

How Does Death Harm the Deceased?

31

graduate college. In any case, according to the deprivation account, death can harm the deceased, and it does so by depriving her of various goods. The second main answer to the question of whether death can be bad is called Epicureanism, named after Epicurus, an ancient Greek philosopher, and this view denies, as Epicurus himself did, that death can be bad for the one who dies. On this view, there is no case of death that harms the deceased. Again, if you are like me and find it plausible that death can be bad, then Epicureanism might strike you as obviously false, or at least as very implausible. But it turns out that there are several Epicurean challenges to the deprivation account—arguments for the conclusion that death is never bad for the one who dies. Indeed, in the essay mentioned earlier, Nagel suggests four potential challenges to the deprivation account (Nagel actually counts them as three, but the second involves two distinct challenges), and the remainder of this chapter will consider both how those four challenges could be regimented into arguments for Epicureanism and how the deprivation account could respond to each. The first three (the “No-Harm-Done Argument,” the “No-Subject Argument,” and the “Timing Argument”) are suggested by Epicurus himself, and the fourth (the “Symmetry Argument”), which has recently generated a lively discussion, is suggested by Lucretius, an ancient Roman philosopher who was himself an Epicurean. As we consider each argument, I will highlight contributions to this debate from Nagel’s classic essay, but I will also discuss, especially in connection with the Symmetry Argument, some more recent developments.

The No-Harm-Done Argument The first Epicurean challenge that Nagel mentions concerns whether anything can be bad for someone without being experienced as bad by that person (Nagel 1970: 75–76). Epicurus puts the challenge this way in his Letter to Menoeceus: “Accustom yourself to the belief that death is nothing to us. For all good and evil lie in sensation, whereas death is the absence of sensation . . . Therefore that most frightful of evils, death, is nothing to us” (Long and Sedley 1987: 149–150). On Epicurus’s own view, in order for a thing to be evil (or to count as bad for a person), that thing must involve some unpleasant sensation—it must involve some pain. But since, on Epicurus’s view, death doesn’t involve any pain, he concludes that death cannot be bad for the one who dies. We could put Epicurus’s argument, the No-Harm-Done Argument, like this (for discussion of a similar argument, see Feldman 1992: ch. 8): 1. 2. 3.

Everything that is bad for a person consists in sensation. A person who has died cannot have any sensations. Therefore, death cannot be bad for the one who dies.

32

Taylor W. Cyr

The No-Harm-Done Argument is valid, which is to say that the conclusion (3) follows from the premises (1)–(2). The question we need to ask, then, is: Are the premises true? Given our assumption about death being a permanent experiential blank, we are committed to the truth of (2); you can’t have sensations if you don’t exist. So it looks like the deprivation account must somehow deny (1), and this is exactly what Nagel does. Imagine the following scenario: Unbeknownst to you, your closest friends have recently spread lies about you to everyone else you know, including the members of your family and your other close friends, thereby betraying your friendship. Imagine, further, that they have been very careful in spreading the lies behind your back, making sure that you never find out about the lies and that you never notice that anyone treats you differently as a result of having heard the lies about you. Do you think that, in such a case, it would be a bad thing for you that you had been betrayed, even though you won’t ever find out about the betrayal or suffer as a result? Most people, I think, would say that this betrayal would be bad for the one betrayed; if so, however, then the premise (1) must be false, for we’ve found a case in which something is bad for a person without consisting of sensation. In other words, if it can be bad for you that your friends spread lies behind your back without you ever finding out, then something can be bad for you even though you never have any unpleasant sensations—you never find out about the lies and thus never feel any pain. Nagel argues that such cases are indeed bad (Nagel 1970: 76–78), and thus he would say that premise (1) must be false; there can be bads (or harms) that befall us without our experiencing them as bad. It looks, then, as though the deprivation account has an adequate response to the No-HarmDone Argument. Now, of course, one response on behalf of Epicureanism is to deny that this case of betrayal really is bad for the one who is betrayed but never finds out about it. I do not find this response very plausible, but I suppose that Epicureans could dig in their heels on this point. It is worth noting, though, that this would be a very striking result, for there are many cases in which we take something to be bad for someone who never experiences the thing as bad, such as when a person’s right to privacy is violated without that person’s knowledge (Nozick 1999). (For a discussion of this and similar cases, see Fischer 2009: ch. 1; and also Fischer 2014). For this reason, I don’t find this response very plausible. A different Epicurean response, however, would be to argue that the betrayal case is not relevantly similar to the case of death. An Epicurean might argue as follows. In the betrayal case, it is at least possible for the one betrayed to discover what’s happened and thus experience the badness of being betrayed; but in the case of death, it is not even possible for the deceased to have any experience of the supposed badness of death; thus (the Epicurean

How Does Death Harm the Deceased?

33

concludes), even if the betrayal really is bad for the one betrayed, it doesn’t follow that death can be bad for the deceased. How might someone defending the deprivation account respond? One way is to construct another case—one in which a person is harmed even though it isn’t possible for the person to experience the harm. John Martin Fischer aims to do exactly this in his modified version of the betrayal case: You are betrayed behind your back by people who you thought were good friends, and you never actually find out about this or have any bad experiences as a result of the betrayal. But now suppose that these friends were (very) worried that you might find out about the betrayal. In order to guard against this possibility, they arrange for White to watch over you. His task is to prevent you ever from finding out about the betrayal. So, for example, if one of the individuals who betrayed you should decide to tell you about it, White can prevent him from succeeding: White can do whatever is required to prevent the information from getting to you. (Fischer 2009: 39) In the modified version of the betrayal case, it is not possible (in the relevant sense) for you to find out about the betrayal because White ensures that this information never reaches you. Still, it certainly seems like this case of betrayal is bad for you, despite the fact that it isn’t possible for you to experience the betrayal as bad. And this case is more similar to the case of death; in both the modified betrayal case and the case of death, it is not possible for the person to experience the relevant items (betrayal, on the one hand, and death, on the other) as bad. But, the proponent of the deprivation approach could say, if we accept that the betrayal is bad for you, despite the impossibility that you experience it as bad, then we should say that death can be bad for the deceased, despite the impossibility that the deceased experiences it as bad. Again, an Epicurean could bite the bullet and deny that this case of betrayal is really bad for the one betrayed, but this is very counterintuitive and would imply, as I mentioned earlier, that many cases in which we ordinarily assume that something is bad for someone (e.g., certain violations of privacy) would turn out to be different than we thought. An alternative Epicurean strategy is to maintain that there is some relevant difference between the case of death and the other cases considered (such that, while the latter can be bad for the person involved, the former cannot). One way to continue pursuing this strategy is to focus on the fact that, while there is a possible subject of the harm in the betrayal cases, there is no possible subject of the alleged harm in the case of death. This leads us to the next challenge.

34

Taylor W. Cyr

The No-Subject Argument The second challenge is also inspired by Epicurus’s Letter to Menoeceus: Therefore that most frightful of evils, death, is nothing to us, seeing that when we exist death is not present, and when death is present we do not exist. Thus it is nothing to either the living or the dead, seeing that the former do not have it and the latter no longer exist. (Long and Sedley 1987: 150) The passage just quoted is very suggestive and might be read in a number of ways, but there are two interpretations worth considering here (these two interpretations are also discussed in Kagan 2012: ch. 10): first, Epicurus is emphasizing that there is no subject of the harm of death, and thus that death can’t be bad (this is the challenge considered in this section); second, Epicurus is emphasizing that there is no time at which death harms anyone, and thus that death can’t be bad (this is the challenge considered in the next section). On the first reading of the passage, Epicurus is claiming that a person must exist in order for anything to be bad for that person. Martha Nussbaum, who has written much on Epicurus and Lucretius and their arguments, has formulated this type of Epicurean argument, the No-Subject Argument, as follows: 1.

2. 3.

An event can be good or bad for someone only if, at the time when the event is present, that person exists as a subject of at least possible experience, so that it is at least possible that the person experiences the event. The time after a person dies [that is, death] is a time at which that person does not exist as a subject of possible experience. Hence the condition of being dead [that is, death] is not bad for that person. (Nussbaum 1994: 201–202)

The No-Subject Argument, just like the No-Harm-Done Argument that we’ve already considered, is valid; so, in order to respond to the argument, we need to show which premise is false. As with the No-Harm-Done Argument, premise (2) of the No-Subject Argument follows from our assumption about death being a permanent experiential blank; we’re assuming that death is the end, so once the deceased is in fact deceased, she no longer exists as the subject of possible experience. Thus, to avoid the Epicurean conclusion, the deprivation account must deny (1). One way to motivate the denial of (1) is to appeal to examples like the ones considered in the previous section (involving betrayal) and to argue that death

How Does Death Harm the Deceased?

35

is not relevantly different from such cases even though there is no subject of possible experience in the case of death. Consider the following example introduced by Nagel: Suppose an intelligent person receives a brain injury that reduces him to the mental condition of a contented infant, and that such desires as remain to him can be satisfied by a custodian, so that he is free from care. Such a development would be widely regarded as a severe misfortune, not only for his friends and relations, or for society, but also, and primarily, for the person himself. This does not mean that a contented infant is unfortunate. The intelligent adult who has been reduced to this condition is the subject of the misfortune. He is the one we pity, though of course he does not mind his condition—there is some doubt, in fact, whether he can be said to exist any longer. (Nagel 1970: 77) As with the betrayal cases, it looks as though this case shows that something can be bad for a person even if that person cannot experience the thing as bad. The new case, however, is not very different from the case of death; the intelligent adult who is reduced to the mental state of an infant, even if he still exists (which, as Nagel notes, may be doubted), has apparently been harmed in a way that he can no longer appreciate. And since this case is not very different from the case of death, it looks as though the proponent of the deprivation account can maintain that a person need not exist as a subject of possible experience in order to be harmed, which is to say that premise (1) is false. But, it may be argued, a problem with this new example, as well as the betrayal cases, is that they remain importantly different from the case of death insofar as they involve subjects of possible experience, whereas the deceased is not the subject of possible experience. Indeed, Nussbaum argues that, in the case of the adult whose mental capacity is reduced, it is hard not to feel that the continued existence of the damaged person, who is continuous with and very plausibly identical with the former adult, gives the argument that the adult has suffered a loss at least a part of its force. (Nussbaum 1994: 206) According to Nussbaum, we simply can’t extrapolate from cases involving a subject (like betrayals or cases of mental loss) to the case of death, for it is precisely this feature of death (that there is no subject when it has occurred) that prevents it from being a harm. We could put this Epicurean defense of the No-Subject Argument as follows: the strategy for denying premise (1) that we

36

Taylor W. Cyr

are considering begins by appealing to a non-death example of alleged harm and then argues that the case of death is not relevantly different (and so it can harm the deceased); but if the strategy appeals to any example of alleged harm other than death, then there must be a subject of possible experience in that example (since it isn’t a case of death); and if there is a subject, then the case is importantly different from the case of death (in which there is no subject); so this strategy can’t establish that death harms the deceased despite there being no subject. One way to respond on behalf of the deprivation account is to consider examples in which it is intuitive that, though a person has died (and thus is no longer the subject of possible experiences), something done after that person’s death can nevertheless harm him or her. Imagine that, after a famous author’s death, everyone comes to believe (falsely) that her books were written by her brother, who died in Mexico long before the books were published (Nagel 1970: 76). Or, to take a case from Nussbaum’s more recent work, in which she makes some concessions concerning such cases, Suppose that Virgil’s Aeneid had been burned at his death (as, in fact, he is said to have requested). Then the posthumous story of Virgil’s importance for the whole history of literature, art, and thought would have been completely different. It is not implausible, I think, to view these posthumous events as altering the significance of Virgil’s life of striving; that life is now rightly seen as a fantastic success, and some of that achievement consists in altering the way millions of people think about life. (Nussbaum 2013: 33) As Nussbaum notes, it is not implausible that the significance of Virgil’s life would have been affected by the burning of the Aeneid at his death. Moreover, it is not implausible that this would have been bad for Virgil, despite the fact that he would never have experienced anything differently had it happened. Similarly, if an author posthumously loses her fame as a result of widespread false belief about the authorship of her books, this would be bad for that author, despite her nonexistence after her death. If it is true that these subject-less cases involve harms, then premise (1) of the No-Subject Argument is false. At this point, a convinced Epicurean would, of course, deny that there really are any posthumous harms in these cases. From the Epicurean point of view, it makes no difference to Virgil what happens after he dies, for he isn’t around to be affected by it! And this leaves us roughly where we started: the Epicurean and the proponent of the deprivation approach simply disagree about the various examples (including the case of death), and it appears that neither side can convince the other to change positions. (This is an instance of what Fischer calls a “dialectical stalemate.” For a nice discussion of this dialectical

How Does Death Harm the Deceased?

37

situation, see Fischer 2009: ch. 7.) How should we proceed from here? One way to attempt to make progress would be to consider more examples with the hope of discovering one that will do the trick, convincing the Epicurean that premise (1) of the No-Subject Argument is false. Another way forward is to move on, evaluating other arguments for Epicureanism. Let’s proceed in the latter way.

The Timing Argument In the previous section, I mentioned that there were two ways to read a certain passage from Epicurus. Here is the passage: Therefore that most frightful of evils, death, is nothing to us, seeing that when we exist death is not present, and when death is present we do not exist. Thus it is nothing to either the living or the dead, seeing that the former do not have it and the latter no longer exist. (Long and Sedley 1987: 150) We considered the first reading (with an emphasis on there being no subject of death) in the previous section; the second reading has it that Epicurus is emphasizing that there is no time at which death harms anyone, and thus that death can’t be bad. Following Jens Johansson, we might formulate Epicurus’s argument, the Timing Argument, as follows: 1. 2. 3.

Anything that is bad for a person is bad for her at a time. There is no time at which death is bad for the person who dies. Hence, death is not bad for the person who dies. (Johansson 2013b: 255)

As with the previous arguments, the Timing Argument is valid. In this case, however, one can resist either premise without giving up any of the assumptions with which we began. On the one hand, one can deny premise (1), claiming instead that death can be bad for a person without being bad for her at a time. This view has been dubbed “atemporalism” (Johansson 2013b). On the other hand, one can deny premise (2), claiming instead that there is a time at which death can be bad for the deceased. Since there are several candidate times for the badness of death, there are several possible ways to reject premise (2) of the Timing Argument. Following Steven Luper, we can identify and label the possible views on the timing of the badness of a person’s death as follows: “eternalism” says that death is bad at all times; “subsequentism” says that death is bad after the person has died; “concurrentism” says that death is bad at the time of death (where death is the first moment of the person’s

38

Taylor W. Cyr

nonexistence); and “priorism” says that death is bad before the person dies (Luper 2009). Yet another possible view, endorsed by Nagel, is “indefinitism,” according to which the time of the badness of death is indefinite (Luper 2009). Since I don’t have the space to evaluate each of these possible responses to the argument here, I will instead focus on Nagel’s claims concerning the timing of the harm of death. (For a thorough discussion of the other positions, see Johansson 2013b.) Recall the case we considered in the previous section involving an intelligent adult who is (tragically) reduced to the mental state of an infant. Concerning this case, Nagel says, This case should convince us that it is arbitrary to restrict the goods and evils that can befall a man to nonrelational properties ascribable to him at particular times. As it stands, that restriction excludes not only such cases of gross degeneration, but also a good deal of what is important about success and failure, and other features of a life that have the character of processes. I believe we can go further, however. There are goods and evils which are irreducibly relational; they are features of the relations between a person, with spatial and temporal boundaries of the usual sort, and circumstances which may not coincide with him either in space or in time. A man’s life includes much that does not transpire within the boundaries of his body and his mind, and what happens to him can include much that does not take place within the boundaries of his life. These boundaries are commonly crossed by the misfortunes of being deceived, or despised, or betrayed. (If this is correct, there is a simple account of what is wrong with breaking a deathbed promise. It is an injury to the dead man. For certain purposes it is possible to regard time as just another type of distance.) (Nagel 1970: 77–78) It is relatively simple, according to Nagel, to determine the boundaries, both spatial and temporal, of a human being. Still, Nagel thinks, much of a person’s life, including what happens to the person, lies outside of those boundaries. Much of Virgil’s life, for example, depended on his Aeneid not being burned upon his death; in fact, the event of Virgil’s wishes being ignored (concerning the burning of his Aeneid ) is something that happened to Virgil, despite the fact that this event occurred outside the temporal boundaries of Virgil’s life. As Nagel puts it, we can think of cases in which some later event affects the deceased (either by harming or benefitting the deceased) as instances of temporal distance. Just as I can be harmed by an event that occurs spatially distant from me (e.g., when my bicycle is stolen while I am away), so too, on Nagel’s view, I can be harmed by an event that occurs at some temporal distance from me.

How Does Death Harm the Deceased?

39

Talk of temporal distance might suggest subsequentism about the timing of death’s badness (the view that death harms the deceased after her death), but Nagel’s view is that the harm does not have so specific a temporal location even if both the person harmed and the event that brings about the harm (death) do have specific temporal locations. (For an endorsement of subsequentism that makes reference to temporal distance, though, see Fischer 2009: ch. 3.) According to Nagel, When a man dies we are left with his corpse, and while a corpse can suffer the kind of mishap that may occur to an article of furniture, it is not a suitable object for pity. The man, however, is. He has lost his life, and if he had not died, he would have continued to live it, and to possess whatever good there is in living . . . [A]lthough the spatial and temporal locations of the individual who suffered the loss are clear enough, the misfortune itself cannot be so easily located. One must be content just to state that his life is over and there will never be any more of it. That fact, rather than his past or present condition, constitutes his misfortune, if it is one. Nevertheless if there is a loss, someone must suffer it, and he must have existence and specific spatial and temporal location even if the loss itself does not. (Nagel 1970: 78) The harm (or misfortune) of death, on Nagel’s view, need not have a specific spatiotemporal location in order to be a genuine harm to the deceased. In other words, we may not be able to pinpoint the exact temporal location of death’s harm to the deceased, but we can nevertheless maintain that death is a harm to the deceased despite the fact that the event of death (which results in the harm) is temporally distant from the deceased. Thus, on Nagel’s view, premise (2) of the Timing Argument is false; even if the timing of the harm of death is indeterminate, it does not follow that there is no time at which death harms the deceased.

The Symmetry Argument Even if each of the challenges that we’ve considered can be met, however, the one that remains to be considered is perhaps the most difficult challenge for the deprivation account. This last challenge is suggested by Lucretius in his De Rerum Natura: Look back now and consider how the bygone ages of eternity that elapsed before our birth were nothing to us. Here, then, is a mirror in which nature shows us the time to come after our death. Do you see anything fearful in

40

Taylor W. Cyr it? Do you perceive anything grim? Does it not appear more peaceful than the deepest sleep? (DRN 972–977, trans. Martin Ferguson Smith)

According to Lucretius, there is a certain symmetry between the time before we exist (which we can call our prenatal, or pre-vital, nonexistence) and the time after we exist (which we can call our posthumous nonexistence, or simply death); in both cases, we do not exist. But, Lucretius argues, since there is nothing grim about our prenatal nonexistence, and since our posthumous nonexistence is not relevantly different, we should have the same view of our posthumous nonexistence. (Note that, while Lucretius seems primarily concerned with whether we should fear death, I have been setting aside questions about the fear of death in this chapter to focus on whether death can be bad. These sets of questions are related, to be sure, and answers to one set of questions may be relevant to the other set, but we can keep them separate for our purposes here and consider a version of Lucretius’s argument that focuses on death’s badness.) We could articulate Lucretius’s argument, the Symmetry Argument, as follows: 1. 2. 3.

Prenatal nonexistence is not bad for the person who comes into existence. There is no relevant difference (with respect to badness) between prenatal nonexistence and posthumous nonexistence (death). Thus, death is not bad for the one who dies.

On my view, the Symmetry Argument is the most interesting and challenging weapon in the Epicurean arsenal. Both of the premises are very plausible, and the inference from those premises to the conclusion appears valid. How might a proponent of the deprivation account respond to the Symmetry Argument? One possible response is to reject premise (1), thereby accepting the counterintuitive position that prenatal nonexistence is bad for the person who comes into existence. Fred Feldman commends this approach: There are, after all, two ways in which we can rectify the apparently irrational emotional asymmetry. On the one hand, we can follow Lucretius and cease viewing early death as a bad thing for [the deceased]. On the other hand, we can at least try to start viewing late birth as a bad thing. My suggestion is that in the present case, the latter course would be preferable. (Feldman 1991: 223) If one accepts premise (2) of the Symmetry Argument, one could nevertheless reject the Epicurean conclusion that death can’t harm the deceased. The catch, of course, is that doing so while maintaining (2) requires accepting, as

How Does Death Harm the Deceased?

41

Feldman does, that our prenatal nonexistence is bad for us. On this approach, one denies both premise (1) and the conclusion (3) of the Symmetry Argument. While I don’t have a knockdown argument against this approach, I find it difficult to accept that my own prenatal nonexistence is bad for me. Hence, I’m motivated to look elsewhere for a response to the argument. On my view, a more promising response to the Symmetry Argument is to argue that there is some asymmetry overlooked by the argument and thus that, in fact, premise (2) is false. In particular, a proponent of the deprivation account could argue that death deprives us of something which prenatal nonexistence does not. This is how Nagel responds to the Symmetry Argument: It is true that both the time before a man’s birth and the time after his death are times when he does not exist. But the time after his death is time of which his death deprives him. It is time in which, had he not died then, he would be alive . . . But we cannot say that the time prior to a man’s birth is time in which he would have lived had he been born not then but earlier. For aside from the brief margin permitted by premature labor, he could not have been born earlier: anyone born substantially earlier than he was would have been someone else. Therefore the time prior to his birth is not time in which his subsequent birth prevents him from living. (Nagel 1970: 79) Nagel’s suggestion is that, while death can deprive the deceased of goods (and thereby harm the deceased), prenatal nonexistence does not deprive anyone of any goods, since anyone who was born earlier enough to receive any extra goods would not be the same person. On Nagel’s view, then, there is an important asymmetry between prenatal and posthumous nonexistence; if prenatal nonexistence were to end earlier (i.e., if one came into existence earlier), it would be a different person that came into existence, whereas if posthumous nonexistence were to begin later (i.e., if a person were to die later), it would be the same person who would be able to enjoy more goods of life. Thus, Nagel thinks, one can accept premise (1) of the Symmetry Argument (which says that prenatal nonexistence is not bad) without accepting the Epicurean conclusion that death can’t be bad for the deceased. A common objection to Nagel’s suggestion is that it depends on the claim that it is not possible for a person to come into existence earlier than that person actually comes into existence. This apparently relies on a strict version of what is sometimes called an “essentiality of origins” thesis, according to which it is essential to a person that she develops from a particular fertilized egg that comes into existence at a particular time—any other fertilized egg, or the same fertilized egg that came into existence at any other time, would necessarily be a different person, on this view. I find this strict essentiality of origins thesis

42

Taylor W. Cyr

incredibly implausible, and (perhaps unsurprisingly) it is very controversial. Consider a case suggested by Fischer: a fertilized egg is frozen and thawed out at a later time (many years later, say). For the individual who develops from this fertilized egg, it was possible that she be born earlier, since she could have been thawed out sooner (Fischer 2009: 65–66). Nagel himself confesses to being troubled by his response to the Symmetry Argument, partly because of this kind of worry. When his essay was reprinted in his book Mortal Questions, he added a footnote to suggest a different response: I suspect that [a response to the Symmetry Argument] requires a general treatment of the difference between past and future in our attitudes toward our own lives. Our attitudes toward past and future pain are very different, for example. Derek Parfit’s writings on this topic have revealed its difficulty to me. (Nagel 1979: 9) Rather than attempting to show that it isn’t possible to be born earlier, as he had suggested at first, Nagel’s second suggestion, inspired by Parfit, is that we look for an asymmetry in our attitudes toward past and future events in order to explain how the Symmetry Argument fails. So, Nagel has provided two suggestions for responding to the Symmetry Argument, the first of which, as he acknowledges, faces some difficulties, and the second of which he leaves to be developed. In the remainder of this chapter, I want to explore a recent development of each of Nagel’s two suggestions. The first suggestion, according to which it isn’t possible to be born earlier, is defended (after some modification) by Frederik Kaufman (Kaufman 1999). The second suggestion, according to which there is an important asymmetry in our attitudes toward the past and future, is developed and defended by Anthony Brueckner and John Martin Fischer (Brueckner and Fischer 1986; Fischer and Brueckner 2012). Let’s consider each in turn. While Kaufman endorses the criticisms of Nagel’s first suggestion discussed earlier, he nevertheless claims that there is a sense in which it is not possible for us to be born substantially earlier than we are in fact born. To see why this is the case, Kaufman thinks, we need to distinguish between two senses of the term person (Kaufman 1999: 11–12). On the one hand, person might refer to a person in the “thin” sense, which picks out the metaphysical essence of some human being, whatever that may be (perhaps a certain soul, or a brain, or some such). On the other hand, person might refer to a person in the “thick” sense, which picks out one’s subjective sense of self, including one’s psychological features (e.g., memories and character traits). Crucially, a person in the “thin” sense could have lived a very different life and thus been a different person in the “thick” sense; if someone were to have a very different history, with the

How Does Death Harm the Deceased?

43

result that she would have a very different psychological makeup, she would not be the same person in the “thick” sense. How is the distinction between “thin” and “thick” persons relevant to the Symmetry Argument? Well, on Kaufman’s view, the reason that death can harm the deceased is that it can deprive a person, construed in the “thick” sense, of various goods; but, according to Kaufman, it isn’t possible for prenatal nonexistence to deprive a person, construed in the “thick” sense, of any goods (Kaufman 1999: 12–14). If someone were to be born earlier, the result would be a very different “thick” person with a new history, different memories, and perhaps even an unfamiliar character. (This is the sense in which Kaufman thinks that it isn’t possible to be born substantially earlier than we are in fact born.) By contrast, a “thick” person can always, as it were, get “thicker” by accruing new psychological features without a loss to what was already part of her “thick” self. So, on Kaufman’s view, prenatal nonexistence is not bad for us (construed as “thick” persons), but death can be bad for us insofar as it deprives us (again, construed as “thick” persons) of various goods. Thus, according to Kaufman, premise (2) of the Symmetry Argument is false—there is an asymmetry concerning the badness of the prenatal and posthumous nonexistence of “thick” persons. Before we turn to Nagel’s second suggestion, as developed by Brueckner and Fischer, let me mention just one difficulty for Kaufman’s version of Nagel’s first suggestion. Kaufman’s proposal can succeed only if the only deprivations that matter to us are deprivations of goods for the “thick” persons that we currently are. Kaufman thinks that these are, in fact, the deprivations we care about, and this is why he thinks that only posthumous (and not prenatal) deprivations can be bad for us. But it is not clear that Kaufman is right about this, as it is a very commonplace occurrence for a person to regret a past deprivation that would have resulted, had it occurred, in a different “thick” person, with various changes to that person’s psychological features. Some children, to take just one example, are deprived of a good education, and this can be bad for the children so deprived even if they would have been different “thick” persons had they received a better education. (For discussion of similar cases, see Fischer 2009: ch. 5.) For this reason, I am skeptical that it is merely our “thick” selves that we care about when thinking of past and future deprivations. Perhaps we also care about whether our “thin” selves are associated with particular “thick” selves. An alternative response to the Symmetry Argument takes Nagel’s second suggestion and attempts to show that some asymmetry in our attitudes toward past and future events justifies our view that death, but not prenatal nonexistence, can be bad for us. Nagel notes that Parfit’s writings on this subject revealed its difficulty to him, and it turns out that Parfit’s writings also played a part in Brueckner and Fischer’s development of Nagel’s suggestion. On Parfit’s view, we prefer to have our sufferings in our past rather than in our

44

Taylor W. Cyr

future, and it is not obviously irrational for us to regard our past and future sufferings asymmetrically. Brueckner and Fischer argue that, while Parfit is right about the asymmetry in our attitudes toward past and future events, the asymmetry concerning sufferings is not sufficient to provide an adequate response to the Symmetry Argument—in particular, they worry that Parfit’s approach will apply only to bads that are experienced as bad (like suffering), which will not apply to the case of death (Brueckner and Fischer 1986: 216). Instead, they argue, we need to focus on our attitudes toward past and future goods (Brueckner and Fischer 1986: 219). In their first article on the subject, Brueckner and Fischer’s thesis was that “[d]eath deprives us of something we care about, whereas prenatal nonexistence deprives us of something to which we are indifferent” (Brueckner and Fischer 1986: 219). They give the following thought experiment, inspired by some of Parfit’s examples (see Parfit 1984), to highlight our actual asymmetric attitudes toward past and future goods (e.g., pleasures): Imagine that you are in some hospital to test a drug. The drug induces intense pleasure for an hour followed by amnesia. You awaken and ask the nurse about your situation. She says that either you tried the drug yesterday (and had an hour of pleasure) or you will try the drug tomorrow (and will have an hour of pleasure). While she checks on your status, it is clear that you prefer to have the pleasure tomorrow. (Brueckner and Fischer 1986: 218–219) As this case shows, we prefer to have pleasures in our future rather than in our past. According to Brueckner and Fischer, this is why we regard death as a harm to the deceased but do not regard prenatal nonexistence as bad for anyone; death can deprive us of future goods, whereas prenatal nonexistence deprives us of goods to which we are indifferent. So far, however, Lucretius (and other Epicureans) can agree with Brueckner and Fischer. It may be the case, Lucretius could say, that we actually have a preference for future goods over past goods, but that doesn’t show that it is rational for us to have such preference patterns. And if it isn’t rational for us to have such preference patterns, then it looks like the Symmetry Argument is untouched by the Brueckner/Fischer approach. In response to this and related worries (some raised by Feldman 2011; Johansson 2013a, 2014), Fischer and Brueckner have added a “rationality component” to their view: “Although we originally put our point in terms of what we took to be people’s actual preference patterns, we should have put it in terms of the rationality of such patterns of preference” (Fischer and Brueckner 2014: 3; see also Fischer and Brueckner 2012). Rather than relying on the asymmetry of the preference patterns with which we happen to find ourselves, Fischer and Brueckner maintain that these asymmetrical patterns are actually rational for us to have and thus that it is

How Does Death Harm the Deceased?

45

rational for us to regard death as a harm to the deceased (since the deceased is deprived of future goods that it is rational to care about) without regarding prenatal nonexistence as a similar harm (since it deprives us of something it is not rational to care about). And this allows Fischer and Brueckner to deny premise (2) of the Symmetry Argument—even though both prenatal and posthumous nonexistence can deprive a person of goods, only posthumous nonexistence (i.e., death) can deprive a person of goods that it is rational to care about, and thus death, but not prenatal nonexistence, can harm the deceased. Despite the apparent complexity in their approach, I am inclined to think that Fischer and Brueckner’s account best captures our ordinary way of thinking about the way that death can harm the deceased.

Conclusion As I hope it is clear from this chapter, it is actually quite challenging to defend the very natural view that it is possible for death to harm the deceased, assuming that death is the end. Given the challenges, one starts to see the attractions of the Epicurean view that it isn’t possible for death to harm the deceased. Still, as we have seen, there are ways to resist the Epicurean arguments, and I have surveyed Nagel’s influential responses to each of the Epicurean challenges he considered in his classic essay on the topic. Whether such resistance is ultimately successful remains, in my view, one of the most interesting debates in contemporary philosophy.

Related Topics Chapter 1, “Is It Possible to Be Better Off Dead?” by Geoffrey Scarre Chapter 3, “The Significance of an Afterlife” by Benjamin Mitchell-Yellin Chapter 4, “The Severity of Death” by Jens Johansson

References Bradley, B. (2009) Well-Being and Death, New York: Oxford University Press. Brueckner, A.L. and J.M. Fischer (1986) “Why Is Death Bad?” Philosophical Studies 50: 213–221. Reprinted in J.M. Fischer (ed.) The Metaphysics of Death, Stanford: Stanford University Press, 1993: 221–229. Feldman, F. (1991) “Some Puzzles about the Evil of Death,” The Philosophical Review 100(2): 205–227. Reprinted in J.M. Fischer (ed.) The Metaphysics of Death, Stanford: Stanford University Press, 1993: 307–326. Feldman, F. (1992) Confrontations with the Reaper: A Philosophical Study of the Nature and Value of Death, New York: Oxford University Press. Feldman, F. (2011) “Brueckner and Fischer on the Evil of Death,” Philosophical Studies 162: 309–312. Fischer, J.M. (2009) Our Stories: Essays on Life, Death, and Free Will, New York: Oxford University Press.

46

Taylor W. Cyr

Fischer, J.M. (2014) “Mortal Harm,” in S. Luper (ed.) The Cambridge Companion to Life and Death, Cambridge: Cambridge University Press, 132–148. Fischer, J.M. and A.L. Brueckner (2012) “The Evil of Death and the Lucretian Symmetry: A Reply to Feldman,” Philosophical Studies 163: 783–789. Fischer, J.M. and A.L. Brueckner (2014) “Prenatal and Posthumous Non-Existence: A Reply to Johansson,” The Journal of Ethics 18: 1–9. Johansson, J. (2013a) “Past and Future Non-Existence,” The Journal of Ethics 17: 51–64. Johansson, J. (2013b) “The Timing Problem,” in B. Bradley, F. Feldman, and J. Johansson (eds.) The Oxford Handbook of Philosophy of Death, New York: Oxford University Press, 255–273. Johansson, J. (2014) “Actual and Counterfactual Attitudes: Reply to Brueckner and Fischer,” The Journal of Ethics 18: 11–18. Kagan, S. (2012) Death, New Haven: Yale University Press. Kaufman, F. (1999) “Pre-Vital and Post-Mortem Non-Existence,” American Philosophical Quarterly 36(1): 1–19. Long, A.A. and D.N. Sedley (1987) The Hellenistic Philosophers (vol. 1), Cambridge: Cambridge University Press. Luper, S. (2009) “Death,” in E.N. Zalta (ed.) The Stanford Encyclopedia of Philosophy (Winter 2014 Edition), URL = . Nagel, T. (1970) “Death,” Nous 4(1): 73–80. Reprinted in his Mortal Questions, Cambridge: Cambridge University Press, 1979: 1–10. Reprinted in J.M. Fischer (ed.) The Metaphysics of Death, Stanford: Stanford University Press, 1993: 61–69. Nozick, R. (1999) “On the Randian Argument,” in Socratic Puzzles, Cambridge: Harvard University Press, 249–264. Nussbaum, M. (1994) The Therapy of Desire: Theory and Practice in Hellenistic Ethics, Princeton: Princeton University Press. Nussbaum, M. (2013) “The Damage of Death: Incomplete Arguments and False Consolations,” in J.S. Taylor (ed.) The Metaphysics and Ethics of Death: New Essays, New York: Oxford University Press, 25–43. Parfit, D. (1984) Reasons and Persons, Oxford: Clarendon Press.

Further Reading Bradley, B., F. Feldman, and J. Johansson (eds.) (2013) The Oxford Handbook of Philosophy of Death, New York: Oxford University Press. (This is a recent collection of essays on the philosophy of death, with articles dedicated to the “Timing Argument” and “Symmetry Argument” discussed in this chapter.) Fischer, J.M. (ed.) (1993) The Metaphysics of Death, Stanford: Stanford University Press. (This is a collection of the classic papers on the philosophy of death, with several articles on the badness of death and a very helpful introduction, by Fischer, to the philosophical issues at play in the literature.) Kagan, S. (2012) Death, New Haven: Yale University Press. (This book is an excellent introduction to thinking about death; it is a very enjoyable read, and it provides some metaphysical background for questions about death’s badness.) Nagel, T. (1970) “Death,” Nous 4(1): 73–80. Reprinted in his Mortal Questions, Cambridge: Cambridge University Press, 1979: 1–10. (This short paper, which I’ve referenced through the current chapter, is at least partly responsible for getting contemporary philosophers talking about death’s badness, and it is a must-read for anyone interested in the topic.)

3

The Significance of an Afterlife Benjamin Mitchell-Yellin

Summary What role does belief in the afterlife play in our thinking about death? This chapter argues that insights from traditional philosophical discussions of the badness of death and the desirability of immortality, which operate on the assumption that there is no afterlife, remain relevant even if we assume that there is an afterlife. It turns out that believers and non-believers have more common ground on these issues than one might have thought.

The Annihilation Assumption This chapter considers the significance of an afterlife for several questions about death. Traditional treatments of these issues in the Western philosophical canon rest on the assumption that death is annihilation—one ceases to exist at the moment one dies. Call this the annihilation assumption. My experience has been that students often find the annihilation assumption unsettling. Many of my students are religious themselves or have grown up surrounded by people who believe in an afterlife. Even those without religious backgrounds are intimately familiar with the idea that people exist after their deaths, albeit in a different place or manner. It informs our cultural context. Ghost stories are a hallmark of late childhood; books and movies depict departed souls and the afterlife; we commonly talk about deceased loved ones “being in a better place.” The thought that death doesn’t mark the end may be easy to disavow in a cool, intellectual mood. But it’s not so easy to completely excise belief in an afterlife from the way we collectively think about and discuss issues surrounding death. Regardless of whether we take the afterlife seriously, its presence is surely felt in our conversations about The End. One might expect that seriously considering the significance of an afterlife would mean stepping away from the terms in which discussion of death has proceeded in certain philosophical circles. But this expectation would be mistaken. Philosophers who make the annihilation assumption have focused on

48

Benjamin Mitchell-Yellin

several concepts and distinctions that prove important also for consideration of the issues in conjunction with the assumption of an afterlife. This is not to say that everything remains the same. Death is sure to look different to one who believes in an afterlife. And as we will see, there are some philosophical puzzles that dissolve once we reject the annihilation assumption. My aim in this chapter is not to advocate for the annihilation assumption. Rather, I aim to demonstrate that the traditional philosophical conversation proves fruitful even when this assumption is rejected. I will consider two debates with ancient origins. The first has to do with the badness of death. Philosophers have long disagreed about whether death is bad for the deceased—that is, whether your death is bad for you. I trace one strain of debate back to the ancient Greek philosopher Epicurus, who affirmed the annihilation assumption. And I argue that this conversation focuses on issues that remain salient even if we abandon this assumption. I then turn to a second issue: immortality. Faced with the realization that you will one day die, you might share the reaction of the Sumerian king Gilgamesh, as recounted in his eponymous epic. You might search for any possible means of staving off demise, any way of quelling your terror at your own mortality. But as this most ancient of stories does, we can ask whether immortality would, in fact, be a good thing. Should you want to live forever? Once again, one influential strand of the philosophical debate on the matter rests on the annihilation assumption. And as with our first topic, I will argue that the lessons to be learned from this debate are relevant even when that assumption is rejected.

Is Your Death Bad for You? It seems simple to account for your sadness at a loved one’s death. Her passing marks a loss for you. No more shared laughs or hugs; no future conversations or vacations. Some of the good things in life have gone away with her. Your mourning reflects appreciation of your loss. But what about your own death? It’s common to feel sad, afraid even, at the thought of your own demise. Can these feelings be accounted for in the same way as your feelings about your friend? Many philosophical treatments of this issue have been shaped by the longago arguments of Epicurus and Lucretius. Both claimed that your own death cannot be bad for you and should not be regarded as such. And both argued for this on the assumption that you cease to exist after you die. The key elements of Epicurus’s view are present in the following argument: Accustom thyself to believe that death is nothing to us, for good and evil imply sentience, and death is the privation of all sentience; . . . Death, therefore, the most awful of evils, is nothing to us, seeing that, when we

The Significance of an Afterlife 49 are, death is not come, and when death is come, we are not. It is nothing, then, either to the living or to the dead, for with the living it is not and the dead exist no longer. (Laertius 1925: 651) Not only does Epicurus avow the annihilation assumption. His argument turns also on the experience requirement: something can be bad for one only if one can have an unpleasant experience as a result of it (Rosenbaum 1986; Fischer 1997). This yields a further plausible requirement, the existence requirement: one must exist in order to be harmed. All three of these theses may be rejected. Philosophical debate has focused on the two requirements. But as my remarks in the introduction suggest, many people find the annihilation assumption foreign, perhaps even more foreign than the experience or existence requirements. They find it difficult to suppose that they will cease to exist after they die. Lucretius, a Roman defender of Epicurus’s views, also assumes that death is annihilation and that badness requires the possibility of experience. On the basis of these Epicurean assumptions, he argues that we have just as much reason to take our death to be bad for us as we do to consider the time before we were born to be bad for us. And since we do not much care about our nonexistence prior to birth, we should display symmetrical equanimity toward our nonexistence after death. In neither case does one exist to be pained. Thus, there is nothing bad for one about either stretch of one’s nonexistence. This has come to be known as the “symmetry argument” (for a translation, see Warren 2001: 467). Philosophical discussion of Lucretius’s symmetry argument has centered on whether there is good reason to adopt asymmetrical attitudes toward one’s past and future nonexistence (Nagel 1970; Brueckner and Fischer 1986; Rosenbaum 1986; Feldman 2013; Fischer and Brueckner 2013). Once again, the annihilation assumption is left unquestioned. The experience requirement, too, is left unchallenged in the context of this debate. The focus is on whether there are relevant differences between the stretch of time during which one does not exist prior to birth and the stretch of time during which one does not exist after death. But it is not clear that the experience requirement and the existence requirement are true. Nagel’s (1970) betrayal case is provocative here. Suppose you are betrayed behind your back and never find out or experience any repercussions. It seems that this betrayal is bad for you, even though you don’t have any unpleasant experiences as a result. We may even suppose that you cannot have any unpleasant experiences as a result of this betrayal, and it still seems bad (Fischer 1997). The lesson appears to be that the experience requirement is false. And without this requirement, there seems to be no good reason to hold onto the existence requirement. Why require that one exist in order to be

50

Benjamin Mitchell-Yellin

harmed if there is no requirement that the harm have any effect on an experiencing subject? The upshot of considering the betrayal case is that two of the Epicurean theses—the experience and existence requirements—seem false. It is not, in general, true that one must experience something in order for it to be bad for one, nor is it, in general, true that one must exist in order to be harmed. Since the betrayal case is relevantly similar, but not identical, to the case of one’s own death, its ability to show these theses false does not beg the question against the Epicurean. Thus, it provides strong grounds for rejecting Epicurean arguments against the badness of death. Moreover, and important for our purposes, the betrayal case neither challenges nor relies upon the annihilation assumption. So, the lessons learned from considering the betrayal case are of interest regardless of whether one accepts the existence of an afterlife. One influential account of why death may be bad for the deceased is the deprivation theory: your death is bad for you because it deprives you of possible future goods (Nagel 1970). This account both is consistent with the annihilation assumption and accounts for a wide range of robust intuitions. On its face, death does seem to be bad for the deceased, and we take earlier death to be worse than later death. We can explain both judgments in the same manner. One’s death forecloses possible future enjoyment of the goods in life, and the earlier this occurs, the more possible goods are forfeit. The deprivation theory has been marshaled in opposition to both Epicurus and Lucretius (Nagel 1970; Brueckner and Fischer 1986; Fischer 1997; Fischer and Brueckner 2013). In opposition to Epicurus, it is claimed that the deprivation theory is explanatorily superior (Nagel 1970). Supposing betrayal is bad for you and that you would be pained if you found out about it, we might want to know which explains the other. Would finding out about it be painful because betrayal is bad? Or is betrayal bad because finding out about it would be painful? This is, essentially, an application of the so-called Euthyphro question (Plato 1981). Here the answer seems clear. One’s pained response to the revelation that one was betrayed is explained by the badness of being betrayed, not the other way around. The deprivation theory can explain the painful experiences appealed to by the Epicurean. In opposition to Lucretius, the deprivation theory seems to stand behind a plausible account of why it is rational to adopt asymmetric attitudes toward one’s postmortem and prenatal nonexistence. Consider the following thought experiment: Imagine that you are in some hospital to test a drug. The drug induces intense pleasure for an hour followed by amnesia. You awaken and ask the nurse about your situation. She says that either you tried the drug yesterday (and had an hour of pleasure) or you will try the drug tomorrow (and

The Significance of an Afterlife 51 will have an hour of pleasure). While she checks on your status, it is clear that you prefer to have the pleasure tomorrow. (Brueckner and Fischer 1986: 218–219) This case is supposed to help us see that we have a rational asymmetry in our attitudes toward past and future goods. We prefer the latter to the former. Conceiving of death as a deprivation of possible future goods, it would seem rational to care more about that than about the deprivation of possible past goods, which may be claimed to account for the badness of prenatal nonexistence. Quite aside from the (presumed) fact that we do not exist before our births or after our deaths, the two time periods are not on a par. Once again, we have a challenge that rests on an intuition about a case that is relevantly similar, but not identical, to the case of one’s death. The symmetry argued for by Lucretius does not seem to rest on principles that are generally applicable. Indeed, the asymmetry he argued against seems to be on firmer footing. And once again, the lessons we have learned from considering this case hold independently of the annihilation assumption. Thus, they should be of interest regardless of whether we believe in an afterlife. Fair enough, one might say; there are lessons to be learned from the conversation between philosophers who accept the annihilation assumption. But what if one doesn’t accept that death is annihilation? Might this make a difference to one’s assessment of Epicurus’s and Lucretius’s arguments? Even if denying this assumption may leave some things in their place, this doesn’t mean it will leave everything as it was. It doesn’t take long to see that denying the annihilation assumption may provide one with Epicurean grounds for objecting to the Epicurean conclusion that death is not bad. As Brueckner and Fischer (1986: 213) note, “[i]f one believes in an afterlife, one could explain how death (conceived of roughly as the cessation of bodily functioning) can be bad insofar as it can involve eternal torment—an indefinitely long sequence of (highly) unpleasant experiences.” Depending on how one conceives of the afterlife, and how one conceives of one’s own life in relation to what will come next, the Epicurean conclusion would seem to go to hell in a handbasket. Death is bad precisely because of impending unpleasant experiences. One need not believe in the possibility of eternal torment in hell in order to think that death may be bad for one because of what one will experience after one dies. The Myth of Er, with which Plato concludes his Republic (Plato 1992), is explicitly invoked to demonstrate the importance of living a just life by telling of the repercussions in the afterlife for those who do not. Socrates portrays the afterlife as involving punishment or reward in proportion (tenfold, to be exact) to the justice or injustice of one’s earthly conduct. This is followed by rebirth. Some time after one’s death, one has a choice about which form of

52

Benjamin Mitchell-Yellin

life to inhabit the next time around. Virtuous people, who have knowledge of the good, are able to choose wisely. Vicious people make poor choices, which they often regret. It is at this moment of choice, Socrates warns, “that a human being faces the greatest danger of all” (Plato 1992: 289). A poor choice regarding one’s next life can have repercussions for many lifetimes, as it may take many cycles of rebirth, and some luck, to alight on a rebirth that would allow one to pursue virtue and knowledge of the good. Both this Platonic myth and the Judeo-Christian-Islamic conception of hell present versions of the afterlife that provide grounds for thinking that death may be bad for one because of experiences to follow. A third conception is familiar from Buddhist teaching. The afterlife may be conceived of as painful due to the frustration of one’s desires and as involving a cycle of rebirth, albeit one that can be broken. The doctrine of karma links one’s past and future lives. Similar to the Myth of Er, those who live well in the past are reborn in better situations in the future. One main difference, however, is that correct living, on the Buddhist conception, may lead to the cycle being broken. The afterlife is not necessarily eternal. One can reach enlightenment and thereby free oneself from the cycle of life, death and rebirth. Thus, there is a sense in which one’s death will, ultimately, be a good thing, even if many bad deaths precede it. All three of these conceptions of the afterlife provide grounds for claiming that death may be bad for Epicurean reasons, specifically in terms of the experience requirement. But is this requirement really left untouched by denying the annihilation assumption? There are numerous ways in which one might fail to experience (or fail to be able to experience) something that is intuitively bad for you even if you were to continue to exist after your death. How one conceives of the afterlife is crucial here. Suppose that entrance into the afterlife cuts one off from the possibilities that would have been open to one given a lengthier earthly existence—for example, those in heaven are unable to witness what happens on earth, or we are reborn in a different and unconnected realm. Then you, once deceased, cannot be affected by the opportunities for good experiences that would have been available to you were you to have lived longer. You could not, for instance, witness your child’s wedding, which happens after your death, and long to be there to take part in it. This is consequential if the experience requirement is true. Then the fact that you couldn’t attend the wedding wouldn’t be bad for you because you can’t have any related experiences. That is assuming that the experience requirement is true. But if the experience requirement is false, as the betrayal case suggests, then one way in which your death may be bad for you, even given an afterlife disconnected from earthly possibilities, is that it deprives you of possible future goods. Intuitively, your inability to attend your child’s wedding is a loss, even if you cannot have any unpleasant experiences related to it. Thus, there is reason for those who

The Significance of an Afterlife 53 conceive of the afterlife in this manner to be concerned about the truth of the experience requirement. There are other ways of conceiving of the afterlife, such that one could have unpleasant experiences that account for the intuitive harms that may come with no longer being alive. Familiar depictions of the afterlife from various traditions present the deceased as either watching over or even interacting with the living. Thus, there is no clear answer either way as to what the status of the experience requirement is in the context of assuming that there is an afterlife. Much depends on how one conceives of the afterlife. This makes for some interesting connections between inquiry into the nature of the afterlife and inquiry into the harm of death, a line of inquiry that is largely absent from discussions based on the annihilation assumption. Consider now the existence requirement. It’s clearly not directly challenged by the assumption of an afterlife. Supposing that we continue to exist after we die does not itself threaten the assumption that existence is required for harm. Indeed, it may seem to make the requirement more secure. As we’ve seen, however, the existence requirement does seem to be challenged by the apparent falsity of the experience requirement. One lesson of the betrayal case was that there seems to be little reason to accept the existence requirement in the absence of adherence to the experience requirement. That point stands independently of the annihilation assumption. And as we just saw, the truth of the experience requirement is not secure given the supposition of an afterlife. Thus, the existence requirement is not entirely secure in this context either. Even if we deny the annihilation assumption, it’s not clear what role the continued existence of the deceased plays in the context of debates about whether her death is bad for her. It may seem as though the existence requirement simply ceases to be relevant if we assume that there is an afterlife. But that is not the case. Consider Lucretius’s symmetry argument. It doesn’t even get going if we assume posthumous existence, for on that assumption, the time period before birth is crucially different from the time period after death. And given that one accepts, as Lucretius does, that existence is required for harm, the assumption that there is an afterlife would allow for a clear-cut case in favor of the rationality of taking one’s death to be bad (or at least potentially bad), and so caring about that stretch of future time, while remaining indifferent to one’s prenatal nonexistence. Moreover, this is all compatible with accepting the grounds given by Brueckner and Fischer for having asymmetrical attitudes toward the two time periods. Their arguments are consistent with both accepting and denying the annihilation assumption. Finally, the assumption that there is an afterlife does not challenge the deprivation theory of death’s badness. Indeed, it may strengthen its appeal. Though proponents of the deprivation theory typically do not accept the experience

54

Benjamin Mitchell-Yellin

and existence requirements, there is nothing stopping one from doing so. And if we continue to exist after death, then, given at least one familiar conception of the afterlife, one can experience certain deprivations due to one’s death. You might be able to see what you’re missing at your child’s wedding. You might be both deprived and pained; death’s badness may be accounted for twice over. Moreover, positing an afterlife allows proponents of the deprivation theory to provide a simpler and more intuitive explanation of when the deceased is harmed by his death. Supposing that death is bad because it deprives one of possible future goods, when is it bad for one to be so deprived? When is one harmed by one’s death? Given the annihilation assumption, the deceased no longer exists. The deprivation theorist must claim that someone is harmed who does not exist (i.e., the deceased is harmed at the time of the deprivation after his death) (Fischer 1997), or that the harm comes before the foul (i.e., the deceased is harmed during his life and prior to the deprivation that accounts for the harm), or that the time of the harm is indeterminate (Nagel 1970). None of these answers is an entirely comfortable one. Positing an afterlife obviates the need to choose between them. The deprivation theorist can claim that the deceased is harmed at the time of the deprivation and that he exists at that postmortem time. Thus, one puzzle that arises in the context of the traditional philosophical debate would seem to dissolve if we reject the annihilation assumption.

Is Immortality Desirable? We have been considering the implications of positing an afterlife for competing views about the badness of death. So far, we have seen that though the traditional conversation assumes that death is annihilation, many of the lessons learned and views developed within this debate remain salient in the context of accepting that there is an afterlife. Let’s turn now to consider the relevance of denying the annihilation assumption to a second set of issues. Supposing that your death is bad for you, it seems reasonable to infer that longer life, eternal life even, would be good for you. But this inference is not obviously sound. In an influential essay on the topic, Bernard Williams (1973) argues that though death is reasonably regarded as bad for the one who dies, immortality would necessarily be “intolerable.” Williams’s main line of argument rests on the annihilation assumption, but he explicitly considers the difference an afterlife might make. And he concludes that it would make none. The contention that an eternal afterlife would be desirable faces the same hurdles as the contention that an eternal continuation of this life would be desirable. In both cases, Williams argues, one would necessarily become bored. The thing to do, he ultimately concludes, is to hope for the lucky happenstance of dying neither too early nor too late.

The Significance of an Afterlife 55 Williams’s argument appeals to the case of EM, which he borrows from a play-turned-opera. The story is that of a woman, Elena Makropulos, or various aliases representable by the initials “EM,” who is given an elixir of life at the age of 42 (or 37, depending on whether one follows Williams or the original author). She lives for a further 300 years, frozen at the biological age of 42 (37), before refusing to take the elixir again and dying. The trouble for EM is that she lived so long she became bored; the world no longer held any interest for her. She became completely withdrawn. And with no reason to continue living, she gave up on continued life and let herself die. As Williams analyzes the case, the trouble with EM was that her “categorical desires,” those desires that provide her with reasons to live, were extinguished. According to Williams, humans have two kinds of desires. Categorical desires settle the question whether one should go on living. By contrast, “conditional desires” represent things one wants, given that one will continue to be alive. A desire for a hot cup of coffee is something one wants on the assumption that one will be around to drink it—a conditional desire. A desire to find one’s soul mate is (plausibly) something that keeps one going day after day—a categorical desire. The fact of having categorical desires and needing them to “propel [one] on into the future” is a basic fact of being human (Williams 1973: 86). Thus, EM is not a special case in this regard. And so the explanation of her boredom—that she ran out of categorical desires to keep her engaged with life—is a potential pitfall for us all. According to Williams, it’s worse than that: immortality would be necessarily boring. EM’s case suggests that for immortality to be desirable it must be “unthinkable” that one’s categorical desires should run out of steam (Williams 1973: 95). Even if one could stay engaged past 342 years, there would come a time when one would run out of reasons to keep on living. Given long enough, every human being would eventually cease to be engaged in a life. This was EM’s fate, and it would be the fate of any human who lived too long. Indeed, Williams contends, it would be the fate even of a human who dies and finds herself in the afterlife. Insofar as she is still human, she will have reason to continue on only given categorical desires. The prospect of heaven may look attractive, but the proof remains in the pudding. Could it engage one for eternity? It will not do to answer Williams’s challenge by citing the otherworldliness of the afterlife. The question is whether you, human as you are, could continue to be happy in such a context. And that question must be settled, on Williams’s view, by appeal to your character. Will the afterlife allow for the satisfaction of your categorical desires? Two conditions must be satisfied for this to be the case. First, it must clearly be you who continues on. It must be the case either that your categorical desires now, here on earth, remain the grounds of your reasons to go on in the afterlife, or else that you can see the different

56

Benjamin Mitchell-Yellin

categorical desires you come to have as suitably related to the ones you have now. As Williams puts it, you must have a constant character. Second, you must recognize it as possible to continue satisfying your categorical desires in the afterlife. That which keeps you engaged in this life must be provided for in the afterlife as well, for eternity. These conditions have to do with you and your relation to your context. Heaven may be a wonderful place, but you have to feel at home in it. Otherwise, you would fare no better than EM. The challenges Williams’s argument presents for those who would like to argue for the desirability of immortality in the afterlife are no different from the challenges it presents for those who would like to argue for the desirability of immortality in this life. Once again, a philosophical conversation that proceeds on the basis of the annihilation assumption is applicable even when this assumption is denied. That’s not to say that these challenges cannot be met; nor is it to insist that they can. Rather, it is to say that those champions of immortality who believe that death is annihilation and those who believe in the possibility of eternal happiness in the afterlife are united in a common cause. Various philosophers have responded to Williams, defending the desirability of immortality (e.g., Fischer 1994; Bostrom and Roache 2008; Bortolotti 2010; Rosati 2013; Fischer and Mitchell-Yellin 2014). Each of them has accepted the annihilation assumption. Thus, these responses are not predicated on an afterlife. But they are instructive nonetheless. In order to bring this out in more detail, let’s consider two distinctions invoked to argue against Williams’s necessary boredom thesis. As we will see, they are relevant in the context of the afterlife as well. The first is another distinction between kinds of desires. Williams distinguishes between categorical desires and conditional desires, but we can further distinguish between two different kinds of categorical desires in terms of their contents (Fischer 1994; Fischer and Mitchell-Yellin 2014). Self-focused desires are for states, experiences or conditions of the person herself—for example, a desire to have a full belly, a desire to have sex or a desire to be healthy. Otherfocused desires are for states, experiences or conditions of individuals or things other than the person herself—for example, a desire to cure cancer, a desire to read all of the great novels or a desire to beautify the world. Given this distinction, it seems that we can more easily home in on categorical desires that appear apt to sustain one’s interest in life for eternity. Among self-focused desires, some of them seem indefinitely “repeatable” (Fischer 1994). It seems plausible that you could have great sex once a week for eternity, without losing interest. This is not to say that the only things that can provide you with reasons to continue living are repeatable, self-focused pursuits. The point is that it’s not at all obvious that such pursuits are not sufficient to propel one into the future in the manner that Williams claims categorical desires do. Moreover, it seems even clearer that other-focused desires can provide one

The Significance of an Afterlife 57 with reasons to continue living indefinitely. Consider the desire to read all of the great novels (Fischer and Mitchell-Yellin 2014). Why think that this desire would necessarily be exhausted after some period of time? First, there is the prospect of new novels being written in the time it takes to tackle one’s initial list; there is no good reason to think that the stock of novels is finite and limited. Second, there is the fact that one can read the same novel more than once, and with interest. A good book can change the way one sees things, including other good books. Thus, even if there were a finite list of great novels to be read, there seems to be no good reason to suppose that the desire to read them all will be exhausted once they have each been read once. They might hold one’s interest a second (and a third, and . . .) time through. These points do not rely on the reality of an afterlife, but they apply to it just the same. What is heaven supposed to be like, such that one might exist there blissfully for eternity? Perhaps that depends on what one wants to do with one’s time. Even if the contents of heavenly desires and pursuits are supposed to be different from the contents of earthly ones, we should still want to make sense of how they can sustain one’s interest indefinitely. And we can make some headway toward doing so by recognizing that some desires are potentially repeatable and others potentially inexhaustible. A second distinction found in the literature in response to Williams is also relevant in the context of positing an afterlife, even though it is traditionally invoked in the context of accepting the annihilation assumption. We can distinguish between two kinds of immortality (Cave 2012; Fischer and MitchellYellin 2014). Medical immortality is the type experienced (one might say, suffered) by EM. A medical immortal is immune to death by natural causes, but she can die nonetheless. In EM’s case, she stopped taking the elixir and then passed away. We might imagine also someone who has access to a medical technology that allows him to maintain a healthy biological age of 42, but which does not protect him from the mortal harm of moving buses. He will not die from a stroke, no matter how old he gets, but he still has to look both ways before crossing the street. Medical immortality is a form of radical life extension and may be contrasted with true immortality: invulnerability to death. A truly immortal person cannot die, even if hit by a bus. This form of immortality is more closely associated with the afterlife than it is with earthly immortality. As just mentioned, Williams’s discussion focused on a case of medical immortality, as does the provocative short story, “The Immortal,” by Jorge Luis Borges (Borges 1962). Both of them agree that this form of immortality is undesirable. In the end, however, the subjects of the tales on which they base their contention are released from life—EM dies after ceasing to take the elixir; Borges’s immortal dies after drinking from a river that reverses his immortality. A true immortal does not have this prospect. If a true immortal suffers the form of boredom that

58

Benjamin Mitchell-Yellin

plagues EM, there is no way out. If life ceases to engage him, he cannot end it. This seems a terrifying prospect. But it is not clear that it is of concern to those who deny the possibility of a supernatural afterlife—that is, an eternal afterlife in some other realm. Even if we were to realize the possibility of indefinitely extending our lives here on earth (or any planet, really) the eventual heat death of the universe would spell the end. Several billion years is not forever. (See Egan [1995] for another short story that considers immortality, but with reference to a billion-year time limit.) There is, then, a sense in which the distinction between medical and true immortality is even more salient in the context of denying the annihilation assumption than in the context of accepting it. It’s more reasonable to consider true immortality given the assumption that there is an afterlife (and a supernatural one, at that). Thus, this second distinction, made in the context of accepting the annihilation assumption, is especially relevant for those interested in questions about the afterlife. If we want to know whether the afterlife is something to look forward to, we do well to consider the desirability of true immortality. Or, following Williams, we might put it this way: we do well to consider whether it is unthinkable that true immortality would be undesirable.

Conclusion According to Nagel, questions about the significance of death need not be focused around questions about immortality, including in the afterlife, because “one’s attitude toward immortality must depend in part on one’s attitude toward death” (Nagel 1970: 74). I have been arguing for two weaker theses. I have tried to show that one’s attitude toward death conceived of as preceding the afterlife may profitably depend in part on one’s attitude toward death conceived of as annihilation. Similarly, I have tried to show that one’s attitude toward immortality in the afterlife may profitably depend in part on one’s attitude toward immortality in this life. In both cases, insights drawn in the context of accepting the annihilation assumption are relevant also in the context of denying it. Belief in the afterlife does not fundamentally change the terms of inquiry into matters of (eternal) life and death. To echo something I said at the beginning of this chapter, this is not to say that the afterlife is of no significance. Surely, for many people belief in the afterlife has profound implications for how they conceive of and live their lives and for how they approach their deaths. And I have pointed out some ways in which it has special significance when asking the two questions addressed here, about the badness of death and the desirability of immortality. I don’t wish to belittle belief in the afterlife or to suggest that it has no part in serious intellectual inquiry. What I do wish to accomplish is to uncover common ground for two groups of thinkers who are often seen as approaching these matters from

The Significance of an Afterlife 59 fundamentally different and irreconcilable perspectives. Questions of life and death are central to inquiry about what it is to be human. My aim has been to demonstrate that we human beings, deep thinkers that we can be, may all profitably engage in this inquiry regardless of how exactly our worldviews align.

Acknowledgments I would like to thank David Wright and John Davis for helpful comments on an earlier draft of this chapter.

Related Topics Chapter 1, “Is It Possible to Be Better Off Dead?” by Geoffrey Scarre Chapter 2, “How Does Death Harm the Deceased?” by Taylor W. Cyr Chapter 4, “The Severity of Death” by Jens Johansson

References Borges, J.L. (1962) “The Immortal,” in D.A. Yates and J.E. Irby (eds.) Labyrinths, New York: New Directions, 105–118. Bortolotti, L. (2010) “Agency, Life Extension, and the Meaning of Life,” The Monist 93: 38–56. Bostrom, N. and R. Roache (2008) “Ethical Issues in Human Enhancement,” in J. Ryberg, T. Peterson, and C. Wolf (eds.) New Waves in Applied Ethics, Basingstoke, UK: Palgrave Macmillan, 120–152. Brueckner, A. and J.M. Fischer (1986) “Why Is Death Bad?” Philosophical Studies 50: 213–221. Cave, S. (2012) Immortality. New York: Crown. Egan, G. (1995) “Learning to Be Me,” in Axiomatic, London: Millennium, 185–202. Feldman, F. (2013) “Brueckner and Fischer on the Evil of Death,” Philosophical Studies 162: 309–317. Fischer, J.M. (1994) “Why Immortality Is Not So Bad,” International Journal of Philosophical Studies 2: 257–270. Fischer, J.M. (1997) “Death, Badness, and the Impossibility of Experience,” The Journal of Ethics 1: 341–353. Fischer, J.M. and A. Brueckner (2013) “The Evil of Death and the Lucretian Symmetry: A Reply to Feldman,” Philosophical Studies 163: 783–789. Fischer, J.M. and B. Mitchell-Yellin (2014) “Immortality and Boredom,” The Journal of Ethics 18: 353–372. Laertius, D. (1925) Lives of Eminent Philosophers (vol. 1), R.D. Hick (trans.), Cambridge, MA: Harvard University Press. Nagel, T. (1970) “Death,” Nous 4: 73–80. Plato (1981) “Euthyphro,” in G.M.A. Grube (trans.) Five Dialogues, Indianapolis: Hackett, 5–22. Plato (1992) Republic, G.M.A. Grube (trans.) and C.D.C. Reeve (revised), Indianapolis: Hackett.

60

Benjamin Mitchell-Yellin

Rosati, C. (2013) “The Makropulos Case Revisited: Reflections on Immortality and Agency,” in B. Bradley, F. Feldman, and J. Johannson (eds.) The Oxford Handbook of Philosophy of Death, New York: Oxford University Press, 355–390. Rosenbaum, S.E. (1986) “How to Be Dead and Not Care: A Defense of Epicurus,” American Philosophical Quarterly 23: 217–225. Warren, J. (2001) “Lucretius, Symmetry Arguments, and Fearing Death,” Phronesis 46: 466–491. Williams, B. (1973) “The Makropulos Case: Reflections on the Tedium of Immortality,” in Problems of the Self, Cambridge, UK: Cambridge University Press, 82–100.

4

The Severity of Death Jens Johansson

Summary Assuming that death is sometimes a bad thing for the one who dies, what factor or factors determine how bad it is for the one who dies? According to the Deprivation Approach, a person’s death is bad for her to the extent that she would have been on balance intrinsically better off if it had not taken place. One common objection to this view is that it overestimates the severity of death by underestimating the significance of psychological unity. Another common objection is that the Deprivation Approach underestimates the severity of death in “preemption” cases. This chapter questions these objections, partly by focusing on various ways in which badness is, and is not, related to prudential reasons.

Introduction Just as some illnesses and injuries are worse than others, so some deaths appear to be worse than others. This is so not only for the fairly trivial reason that those deaths that are bad are worse than those deaths that are not bad: less trivially, some bad deaths seem to be worse than other bad deaths. For instance, whereas it may well be bad for an 80-year-old to die, it is likely to be even worse for a 40-year-old, and still worse for a 20-year-old. Supposing that the badness of death does indeed vary in this way, what factor or factors does it vary with? The example just given indicates that age is one such factor: in particular, the earlier one’s death occurs, the worse it is. For a number of reasons, however, things are more complicated—though, as we shall see, maybe not quite as complicated as some philosophers have suggested.

Our Topic I begin with six points of clarification.

62

Jens Johansson

First, some people undergo a much worse—in particular, a much more painful—process of dying than others. But this is irrelevant here, for our topic concerns the value of the event of death: the brief event that terminates that process (or more generally, the individual’s life). Second, a person’s death is often bad for others, such as friends and family, and it is often bad full stop (or bad “from the point of view of the universe”). However, our investigation concerns the badness of death only for the person who dies. Third, according to the “Epicurean” view of the value of death, no person’s death is in any way bad (or good) for her. The present inquiry would be completely misguided if this view were true. And maybe it is true: although Epicureanism is counterintuitive, there are arguments for it, which are well worth considering (Rosenbaum 1986; Smuts 2012; Taylor 2012). But not here: for present purposes, I am simply going to assume that Epicureanism is incorrect. Fourth, for an obvious reason, my death is going to have the same intrinsic value for me as your death is going to have for you. For surely Epicureans are right about this much: no person’s death is intrinsically good or bad—good or bad in itself—for her. If it is bad for her, it is instead extrinsically bad for her: bad for her because of what it causes or prevents. Fifth, something may be extrinsically, pro tanto bad for a person yet fail to be extrinsically, overall bad for her: it may affect her negatively in some ways but affect her positively (or neutrally) on the whole. For instance, even if a person’s death prevents her from enjoying a nice meal the following evening, it might also spare her decades of torture. Our topic concerns the degree to which a person’s death is extrinsically, overall bad for her. (I will often leave the “overall” qualification implicit.) Sixth, it is important to distinguish the badness of death from the wrongness of killing. Many believe that the moral objection to killing me is equally strong as the moral objection to killing you, no matter how much we differ in terms of age, intelligence, happiness, popularity, and so forth. There is no immediate connection between this thesis and the thesis that the badness of death does not vary from person to person, for it is far from clear that the moral objection to killing varies with the badness of death (McMahan 2002; Luper 2009: ch. 7; Hanser 2013; Ekendahl and Johansson 2015).

Age Again, it is easy to think that the badness of death varies with the victim’s age: the younger the victim, the worse her death is for her. This may indeed serve quite well as a rough guideline. For at least two reasons, however, it can be little more than that. First, the proposal seems hopeless without some sort of “typically” qualification. Arguably, for instance, the death of a healthy forty-year-old—or eighty-year-old—who would have otherwise

The Severity of Death 63 enjoyed many additional happy years is worse for its victim than is the death of a twenty-year-old, whose incurable disease would guarantee continued life to be short and only moderately enjoyable. No doubt the latter individual seems more unfortunate overall, but that does not show her death to be worse for her: on the contrary, if someone’s death is not so bad for her, then this is a pretty clear indication that she is the victim of other misfortunes. Death’s not being highly regrettable can itself be highly regrettable. Second, even aside from this first consideration, the victim’s age at the time of death can at most be evidence that her death is bad for her to a certain degree. It cannot plausibly be what makes her death bad for her to that degree. At least, there must be some more fundamental explanation of why the death of a twenty-year-old is much worse for its victim than the death of an eighty-yearold—more fundamental, that is, than the fact that the twenty-year-old is much younger than the eighty-year-old. It is such a more fundamental explanation that participants in the debate are looking for.

The Deprivation Approach The most widely accepted view of the value of death—including the degree of that value—is the “Deprivation Approach” (e.g., Feldman 1991; Bradley 2009; Luper 2009). It can be formulated as follows: The Deprivation Approach. Someone’s death is extrinsically, overall bad (good) for her if and only if, and to the extent that, her life would have been on balance intrinsically better (worse) for her if her death had not taken place. The Deprivation Approach is agnostic about what it is that makes someone’s life intrinsically good (or bad) for her—about what makes for a person’s wellbeing (or ill-being). Perhaps it is pleasure (pain); perhaps it is desire satisfaction (desire frustration); perhaps it is something else—the Deprivation Approach is compatible with any view on this matter. Strictly speaking, the Deprivation Approach is also agnostic about the value for a person of events other than her own death. However, it would be unmotivated not to extend what it says about death to other events. Unsurprisingly, then—and this will be of some relevance later—advocates of the Deprivation Approach often accept also the following, much more general view: The Generalized Deprivation Approach. An event is extrinsically, overall bad (good) for a person if and only if, and to the extent that, her life would have been on balance intrinsically better (worse) for her if the event had not taken place.

64

Jens Johansson

The Deprivation Approach explains why early death is typically worse than late death. Usually, the value difference between the actual life and the counterfactual life—the life that the person would have had, had her death not occurred—is smaller when a forty-year-old dies than when a twenty-year-old dies. Their respective counterfactual lives may well be equally good, but the forty-year-old’s actual life is likely to be better than that of the twenty-year-old (simply because it is likely to contain more pleasure, desire satisfaction, and other candidate components of well-being). The Deprivation Approach also explains why, in some less typical cases, the death of a twenty-year-old is not as bad for its victim as is the death of a forty-year-old: again, for some twentyyear-olds, continued life would have been only moderately happy. (Similarly, of course, it explains why, in some cases, the death of a twenty-year-old is not bad for her at all; but recall that we are focusing on what makes some bad deaths more severe than other bad deaths.) The Deprivation Approach may appear to require us to make a dubious comparison. Does it not ask us to compare how well off the person is at certain times after her death with how well off she would have been at those times if her death had not taken place? Such a comparison seems to presuppose the questionable idea that the person occupies a well-being level when she is dead—questionable since, provided that she ceases to exist at death, she has no being after death and hence, it seems, no well-being. Not only does it seem wrong to say that a dead person, while dead, occupies a positive or negative well-being level; she does not even seem to occupy a neutral well-being level, a well-being level of zero. In fact, however, the Deprivation Approach requires no such comparison, but only a comparison between the intrinsic value of the individual’s total actual life and the intrinsic value of her total counterfactual life (Feldman 1991). We can make this comparison without supposing that the person occupies any well-being level while dead; as it is sometimes put, we make a “life-life” comparison rather than a “life-death” comparison. There are still problems in the vicinity of this issue. Many philosophers are convinced that in order for an event to be bad for a person, there has to be a time at which she is worse off than she would have been had the event not occurred. It is difficult to find such a time in the case of death: before a person’s death, it seems too early for her to be affected by it, and again, once she dies she apparently no longer occupies any well-being level at all. However, this problem, while not insignificant, afflicts any view according to which a person’s death can be bad for her; since I am assuming that the Epicurean view is false (see second section), we can therefore ignore this challenge here (for discussion, see Bradley 2009: ch. 3; Luper 2009; Johansson 2013; Feit 2015). There are various other challenges to the Deprivation Approach, however, which we should not ignore. As I said, it has the virtue of explaining why early death is typically worse than late death. Nevertheless, it is often alleged

The Severity of Death 65 that the theory sometimes gets the severity of death wrong—in some cases by overestimating it, in others by underestimating it. I shall consider the most discussed alleged instance of the former mistake in the fifth section, and the most discussed alleged instance of the latter one in the sixth section (for some other alleged examples, see McMahan 2002; Belshaw 2013; for replies see Bradley 2007, 2009).

Psychological Unity On the Deprivation Approach, as we have seen, the only thing that matters, in determining how bad a person’s death is for her, is how much intrinsically better her total life would have been if she hadn’t died. The Deprivation Approach thus ascribes no significance to how strongly psychologically connected (in terms of sameness or continuity of beliefs, desires, character traits, etc.) the person is, at the time of death, with herself as she would have been in the future—with her “future self,” as some like to say—if she had instead continued to live. More exactly, the theory ascribes no significance to this factor except insofar as it affects the intrinsic value of the person’s life. Some might think that, all else being equal, a person’s life is intrinsically better the more psychologically unified it is: radical changes in one’s beliefs, desires, personality traits, and so forth make one’s life intrinsically worse, at least if they occur too abruptly and too frequently. As noted in the previous section, the Deprivation Approach can be combined with any view of what makes someone’s life intrinsically good (or bad) for her, including this one. It is thus open to an advocate of the Deprivation Approach to maintain that, if someone’s death deprives her of a future—perhaps a very happy future—in which she would be only weakly psychologically connected with herself, as she is when she dies, then her death is not very bad for her, since her counterfactual life is not intrinsically much better than her actual life. However, the present point is that the Deprivation Approach does not allow that psychological connections affect the degree of the value of a person’s death independently of how they affect the intrinsic value of her actual and counterfactual lives. On the Deprivation Approach, once we have determined the intrinsic values of the person’s actual life and the life that she would have had, were it not for her death, our job is done. According to some writers, this speaks against the Deprivation Approach (McMahan 2002; DeGrazia 2007; cf. Millum 2015). They contend that our job is only partly done, for although the degree of death’s badness is, indeed, partly determined by the degree to which the person’s total life would have been intrinsically better for her if she had not died, it is also partly determined by the strength of the relevant psychological connections. In particular, all else being equal, the weaker these connections—between the person as she is when she dies and herself as she would have been at those later times at which she

66

Jens Johansson

would have received various benefits, had she continued to live—the less her death is bad for her. Thus, even if the person’s life would have been intrinsically much better if she had not died, her death can still fail to be very bad for her. Although proponents of the Deprivation Approach are right that, even in cases where the relevant connections are relatively weak, it is bad for the person to be deprived of those future goods that she would have enjoyed, had she continued to live, they considerably overestimate the degree of this badness. Why, then, think that this is so? Several reasons have been suggested. First, the Deprivation Approach yields that death is typically worse for the deceased the earlier it occurs. As I have pointed out, some instances of this claim are highly plausible: for example, that the death of a twenty-year-old is normally worse for its victim than that of a forty-year-old, and still worse than that of an eighty-year-old. However, other instances are allegedly less attractive. For example, the Deprivation Approach also yields that the death of a small infant is normally much worse for its victim than that of a twenty-year-old. Critics charge that, intuitively, the reverse is true: death is normally much worse for the twenty-year-old, even though the value difference between the actual and the counterfactual lives is much greater in the infant’s case. If the reverse claim is indeed true, moreover, then this is likely precisely because the psychological connections between the infant, as she is when she dies, and herself as she would have been at later times, had she continued to live, are much weaker than those between the twenty-year-old, as she is when she dies, and herself as she would have been at later times, had she continued to live (for the simple reason that the twenty-year-old’s psychology is substantially richer than the infant’s). Second, and relatedly, Jeff McMahan claims (2002: 165) that the Deprivation Approach has the implausible result that it is extremely important to prevent someone from coming into existence, if she otherwise dies immediately after having come into existence, for on the Deprivation Approach, such an individual’s death would be very bad for her (assuming she would have otherwise had a very good life), whereas nothing bad ever happens to an individual who never exists. Thus letting this individual come into existence causes something extremely bad for her, whereas preventing her from coming into existence does nothing of the sort. In response, Ben Bradley (2009: 125) reminds us that death is only extrinsically bad. Whether we have reason to perform a certain action, he contends, depends partly on how much intrinsic value it would bring about, but it does not at all depend on whether it would bring about extrinsically bad things. Suppose, for example, that we can save someone’s life, and thereby cause her to die a much later death. No matter how extrinsically bad this later death would be, this is not in any way a reason against saving her life. Thus the Deprivation Approach does not, in fact, have the result that it is extremely important to prevent the person from coming into existence in McMahan’s

The Severity of Death 67 story. (Personally, I find Bradley’s response convincing; however, see Millum [2015] for a reply on behalf of McMahan.) Third, McMahan has put forward the following case: The Cure. Imagine that you are twenty years old and are diagnosed with a disease that, if untreated, invariably causes death (though not pain or disability) within five years. There is a treatment that reliably cures the disease but also, as a side effect, causes total retrograde amnesia and radical personality change. Long-term studies of others who have had the treatment show that they almost always go on to have long and happy lives, though these lives are informed by desires and values that differ profoundly from those that the person had prior to treatment. You can therefore reasonably expect that, if you take the treatment, you will live for roughly sixty more years, though the life you will have will be utterly discontinuous with your life as it has been. (McMahan 2002: 77) According to McMahan, it is prudentially permissible for the person—“you”— to refuse the treatment, even though this gives you an intrinsically worse total life. (Of course, those who place an enormous weight on psychological unity, in determining the intrinsic value of a life, might not agree that it does; cf. Bradley 2009: 119. But let’s not worry about this.) This indicates that, contrary to the Deprivation Approach, refusing the cure is not bad for you (i.e., not extrinsically, overall bad for you—it can, of course, still be extrinsically, pro tanto bad for you, as it deprives you of various benefits contained in the longer, psychologically less unified life; see the second section). And this, in its turn, suggests that how extrinsically bad an event is for someone is partly determined by psychological connectedness. Is this convincing? Some do not agree that it is prudentially permissible for you to refuse the treatment (Bradley 2009: 117; cf. Luper 2009: 209–210). However, even if it is, it is still debatable that it follows from this that refusing the treatment is not extrinsically bad for you (cf. Bradley 2009: 134, 146). Note first that extrinsic badness is not time-relative: although of course an extrinsically bad event, like any other event, occurs at a particular time, it is not extrinsically bad at a particular time. At any rate, this surely holds for the notion of extrinsic, overall badness that the Deprivation Approach is about— for one thing, it makes little sense to say that the intrinsic value of a total life can vary over time (Feldman 1991; Johansson 2013: 258–260). Even if there is some other notion of extrinsic value which is time-relative, this cannot be a problem for the Deprivation Approach. Because we are not here considering, then, the extrinsic value that the event of your refusing the cure has when this event occurs, as opposed to other times, it is hard to see why what you are

68

Jens Johansson

prudentially permitted to do at that time should have any crucial bearing on the event’s extrinsic value. This would perhaps be of limited significance if your prudential reasons to bring about or otherwise promote this event were always equally strong, but that does not seem to be the case. Consider, for instance, a much earlier time t, at which you are not yet strongly psychologically related to yourself as you will be from age twenty and onwards, whether you are going to receive the cure. Would you be, at t, prudentially permitted to promote—for example, to act so as to increase the likelihood of—your refusing the cure (at age twenty)? Intuitively, it does not seem so. After all, that would risk forgoing a much longer and happier life, something that could hardly be outweighed by the relatively weak psychological connections between you as you are at t and yourself as you would be after having refused the cure. So, your prudential reasons regarding the relevant event—your refusing the cure at age twenty— apparently vary over time, and those you have when the event occurs have no special standing with regard to its extrinsic value. This lends credence to the idea that it can be extrinsically bad for you to refuse the cure at age twenty even if you are prudentially permitted to do so, for prudence might allow a sort of temporal partiality that is alien to extrinsic, overall badness. Without rejecting the Deprivation Approach, then, we can accept that a person is prudentially permitted to give priority, at a certain time, to those possible “future selves” to which she would, at that time, stand in close psychological connections. Here is a related point. Plausibly, the intrinsic value for a person of an event—say, an experience of pleasure—is not something that varies from one time to another, depending on facts about psychological connectedness. For instance, even if a person is now only weakly psychologically connected to herself as she will be when she receives the experience, this does not somehow make the experience less intrinsically good for her now than at other times, at which those connections are stronger. In light of this, it would be odd and unmotivated to say that, by contrast, if an event occurs now which prevents this future experience, then the weakness of these psychological connections does automatically affect the extrinsic badness for her of that event. (Again, that would make more sense if the event’s extrinsic badness were somehow possessed by the event now but not at other times; but once again, this is not the case.) A more attractive view is that the weakness of these psychological connections only weakens her prudential reasons to prevent, at the present time, the event from occurring. These reflections might also be useful in responding to the first criticism earlier: that, contrary to the Deprivation Approach, death is normally worse for a twenty-year-old than for an infant. Those who find this judgment appealing—I confess I am not one of them—may do so because they are influenced by the more plausible claim that the twenty-year-old has more prudential reason to bring about benefits accruing to her “future selves” than does the infant, since

The Severity of Death 69 the twenty-year-old’s psychological connections to these “future selves” are much stronger. Once we separate this latter claim from the claim that an event that deprives the twenty-year-old of future benefits (e.g., death) is worse for her than an event that deprives the infant of future benefits (e.g., death) is for her, the latter claim may lose some of its intuitive force.

Preemption According to some writers, the Deprivation Approach underestimates the severity of death in “preemption” cases such as the following one, provided by McMahan: The Young Pedestrian. A young man, aged twenty, absentmindedly steps off the curb into the path of a bus and is instantly and painlessly killed. During the autopsy, it is discovered that he had a hitherto silent cerebral aneurysm that would inevitably have burst within a week if he had not been hit by the bus. And the bursting of the aneurysm would certainly have been fatal. (McMahan 2002: 117) The Deprivation Approach yields that the Young Pedestrian’s death is no serious misfortune for him: if it had not occurred, he would have died soon enough anyway from the aneurysm. McMahan and several others take this to be an unattractive result: intuitively, they say, the Young Pedestrian’s death is very bad for him. After all, it ensures that he is not going to have a happy future; the fact that the aneurysm also does so seems not to detract from its badness. (Of course, the alternative view discussed in the previous section—invoking the strength of the psychological connections between the one who dies and herself as she would have been later, were it not for her death—is no less vulnerable to this criticism.) If the Young Pedestrian’s death is, indeed, very bad for him, what makes it so? It cannot be that he would have had an intrinsically much better life had his death not occurred: he wouldn’t (but see Bradley 2009: 53–60 for some complications). Maybe it is instead that his death is causally sufficient for his not having an intrinsically much better life. But this condition seems to be satisfied even in cases where death is, intuitively, not bad at all—for example, in some cases where the deceased was seriously and incurably ill (cf. Bradley 2012: 409). Various more complex suggestions could be offered here, but instead of delving into these, let us focus on how unsatisfactory it really is to deny that the Young Pedestrian’s death is very bad for him. On closer inspection, such a denial does not seem to me particularly unattractive. Consider what the Young Pedestrian has reason to desire and do, for

70

Jens Johansson

his own sake (cf. Bradley 2009: 70). Does he have a reason to strongly desire, for his own sake, that his death—that particular event—had not occurred? It does not seem so. Does he have a prudential reason to sacrifice a substantial amount of his own well-being, if this means that the particular event of his death will not occur? Again, it does not seem so. But if the Young Pedestrian’s death were very bad for him, it is hard to see why he would not have such reasons. (In light of the previous section, this would perhaps not be so hard to see if this case involved a significant lack of psychological connectedness. But let us assume that it does not.) Also, it seems that the fact that he lacks such reasons is due to precisely the factor that the Deprivation Approach highlights: the non-occurrence of the Young Pedestrian’s death would make such a minor difference to his well-being; it would not significantly increase the intrinsic value of his life. It is difficult to see why this factor should preclude him from having these reasons—that is, reasons to strongly desire the event not to happen, and to exchange some of his well-being in return for the event’s non-occurrence— but nevertheless not preclude the Young Pedestrian’s death from being very bad for him; in other words, since these reasons are so sensitive to this factor, it would be surprising if the badness were, by contrast, so insensitive to it. A critic of the Deprivation Approach may try to support the judgment that the Young Pedestrian’s death is very bad for him in the following way. Death from the aneurysm would be a great misfortune for the Young Pedestrian (had he still been alive at that time), even by the standards of the Deprivation Approach: it would deprive him of many happy years. So, if the Deprivation Approach is true, then the Young Pedestrian’s actual death (in the bus accident) is much better for him—much less bad for him—than his death from the aneurysm would have been. Thus if he had avoided his actual death—by, for instance, staying on the sidewalk a bit longer—he would instead have walked right into another, much worse one (death from the aneurysm). And therefore, the critic might contend, defenders of the Deprivation Approach are committed to the absurd claim that it is in fact a good thing for the Young Pedestrian that he dies the early death: that it would have been bad for him to stay on the sidewalk a bit longer. (Because staying on the sidewalk a bit longer is, obviously, not the event of death, and the Deprivation Approach is strictly speaking only about the event of death, this line of thought assumes that friends of the Deprivation Approach are also friends of the Generalized Deprivation Approach—but as noted in the fourth section, they usually are, and for good reason.) This line of reasoning is mistaken, however. The claim in question is indeed absurd, but nothing commits defenders of the Deprivation Approach (or the Generalized Deprivation Approach) to it. The crucial factor for them, remember, is whether the relevant event would have made the person’s life intrinsically worse, and the Young Pedestrian’s staying on the sidewalk a little longer certainly would not have satisfied this condition—on the contrary, it would

The Severity of Death 71 have given him an enjoyable extra week to live. Of course, it would have led to an extrinsically very bad event (death from the aneurysm), but it would not thereby itself be extrinsically bad. This is not strange: after all, saving a happy person’s life often leads to a huge extrinsic evil for her—a much later death— while itself being extrinsically very good for her. Granted, it would be implausible to say that there is nothing at all in the situation that is very bad for the Young Pedestrian; it is, after all, a tragic case. However, as Fred Feldman notes in connection with a relevantly similar example (Feldman 1991), there are several states of affairs, closely related to the event of the person’s death, which friends of the Deprivation Approach can say are bad for him—for example, the state of affairs that the Young Pedestrian dies so young. Of course, this state of affairs is, precisely, a state of affairs, whereas the Deprivation Approach is about the event of death. (We cannot assume that states of affairs just are events; for instance, a common view is that events are concrete entities, whereas states of affairs are abstract entities.) However, just as it is natural for proponents of the Deprivation Approach to apply what they say about the event of death to other events (and thus accept the Generalized Deprivation Approach), so it is natural for them to apply what they say about the event of death to states of affairs—that is, to take a state of affairs to be (extrinsically, overall) bad for a person to the extent that her life would have been intrinsically better if the state of affairs had not obtained. On this view, the state of affairs that the Young Pedestrian dies so young is very bad for him, for if it had not obtained, then neither his actual death nor his preempted death from the aneurysm would have occurred, and (we can presume) he would have had many more happy years to live. Can advocates of the Deprivation Approach even say that the situation involves some event that is very bad for the Young Pedestrian? Neil Feit (2015) has suggested that, in cases such as this one, there are at least events that are together very bad for the person. What is required for this, Feit suggests, is that the person’s life would have been intrinsically much better if none of these events had occurred. Consider, for example, the Young Pedestrian’s actual death and his incurring the aneurysm, taken together. On Feit’s view, although neither of these two events is itself particularly bad for the Young Pedestrian, they are together very bad for him, since his life would have been intrinsically much better if neither of them had taken place. This is an ingenious proposal. It is debatable, however, whether it really is in harmony with the Deprivation Approach, for, independently of the Deprivation Approach, it seems reasonable to take the extrinsic value that two events, E1 and E2, together have to be the same as the extrinsic value of the conjunctive state of affairs that E1 occurs and E2 occurs—rather than the disjunctive state of affairs that E1 occurs or E2 occurs. After all, we are concerned with the extrinsic value of the combination of E1 and E2: these two events taken together. (Compare with the intrinsic

72

Jens Johansson

value that E1 and E2 together have. Surely it is more plausible to identify it with the intrinsic value of the conjunctive state of affairs than with the intrinsic value of the disjunctive state of affairs.) Hence, given what I just suggested that advocates of the Deprivation Approach should say about the extrinsic value of a state of affairs—that is, that it is bad for him to the extent that his life would have been intrinsically better if it had not obtained—the Young Pedestrian’s actual death and his incurring the aneurysm are not very bad for him, even taken together, for if not both of them had occurred, then one of them would still have occurred, and his life would then not have been intrinsically much better. In any case, in light of Feldman’s “state of affairs” solution, it is not clear that Feit’s suggestion—or, for that matter, any other solution that implies that the story involves some event or events that are very bad for the Young Pedestrian—is really needed. After all, it is not as if the “state of affairs” approach somehow renders the case any less tragic. Bad states of affairs are not to be trifled with.

Conclusion Although much more could be said about virtually everything covered in this chapter, I think this much is reasonably clear: it is not clear that the Deprivation Approach is mistaken in judging a person’s death bad for her to the extent that she would have been on balance intrinsically better off without it. The Deprivation Approach has the virtues of elegance and simplicity, and, I think, a sort of immediate compellingness. The objections considered here, at least, do not seem to me to establish that the theory either overestimates or underestimates the badness of death. Arguably, normally there is no fact of the matter as to exactly what would have happened if an event, which in fact occurs, had not occurred. And even when there is, there need be no fact of the matter as to the precise intrinsic value for a person of that counterfactual scenario. In many cases, then, we should not expect there to be any determinate answer to the question of exactly how much intrinsically better off the deceased would have been if her death had not occurred. (Often it might be determinately true that she would have been much intrinsically better off, or slightly intrinsically better off, and so forth; but there need be no precise degree such that it is determinately true that she would have been intrinsically better off to that degree.) So if the Deprivation Approach is true, there is often no fact of the matter as to precisely how bad a person’s death is for her; and presumably, sometimes there isn’t even any fact of the matter as to whether it is bad for her at all. Far from being a problem for the Deprivation Approach, however, this is a mark in its favor (though admittedly one that many rival views will share): it is highly counterintuitive that the severity of death is often an entirely precise and determinate matter.

The Severity of Death 73

Related Topics Chapter 1, “Is It Possible to Be Better Off Dead?” by Geoffrey Scarre Chapter 2, “How Does Death Harm the Deceased?” by Taylor W. Cyr Chapter 3, “The Significance of an Afterlife” by Benjamin Mitchell-Yellin

References Belshaw, C. (2013) “Death, Value, and Desire,” in B. Bradley, F. Feldman, and J. Johansson (eds.) The Oxford Handbook of Philosophy of Death, New York: Oxford University Press, 274–296. Bradley, B. (2007) “How Bad Is Death?” Canadian Journal of Philosophy 37: 111–127. Bradley, B. (2009) Well-Being and Death, Oxford: Oxford University Press. Bradley, B. (2012) “Doing Away with Harm,” Philosophy and Phenomenological Research 85: 390–412. DeGrazia, D. (2007) “The Harm of Death, Time-Relative Interests, and Abortion,” Philosophical Forum 38: 57–80. Ekendahl, K. and J. Johansson (2015) “Epicureanism, Extrinsic Badness, and Prudence,” in M. Cholbi (ed.) Immortality and the Philosophy of Death, Lanham, MD: Rowman and Littlefield, 39–52. Feit, N. (2015) “Plural Harm,” Philosophy and Phenomenological Research 90: 361–388. Feldman, F. (1991) “Some Puzzles about the Evil of Death,” Philosophical Review 100: 205–227. Hanser, M. (2013) “The Wrongness of Killing and the Badness of Death,” in B. Bradley, F. Feldman, and J. Johansson (eds.) The Oxford Handbook of Philosophy of Death, New York: Oxford University Press, 391–408. Johansson, J. (2013) “The Timing Problem,” in B. Bradley, F. Feldman, and J. Johansson (eds.) The Oxford Handbook of Philosophy of Death, New York: Oxford University Press, 255–273. Luper, S. (2009) The Philosophy of Death, Cambridge: Cambridge University Press. McMahan, J. (2002) The Ethics of Killing: Problems at the Margins of Life, New York: Oxford University Press. Millum, J. (2015) “Age and Death: A Defence of Gradualism,” Utilitas 27: 279–297. Rosenbaum, S. (1986) “How to Be Dead and Not Care: A Defense of Epicurus,” American Philosophical Quarterly 23: 217–225. Smuts, A. (2012) “Less Good But Not Bad: A Defense of Epicureanism about Death,” Pacific Philosophical Quarterly 93: 197–227. Taylor, J.S. (2012) Death, Posthumous Harm, and Bioethics, London: Routledge.

5

Defining Death John K. Davis

Summary A little over half a century ago, medical science gained the ability to keep patients alive in a state of permanent unconsciousness: a state where the mind seems to be gone, but the body is alive. This raised questions about how to define death, and which criterion of death is correct. A criterion of death is some physiological state of affairs found in all and only those patients who are dead according to whatever definition of death matches that criterion. A criterion is correct if it matches the correct definition of death. However, previous discussions of death have not always clearly distinguished criteria of death from definitions of death, and as a result, the arguments for and against various definitions have not always been clearly developed. I review three criteria of death, four definitions of death, the arguments for and against each definition, and the relationships between each definition and its matching criterion. When we distinguish definitions from criteria and articulate the arguments, we see the advantages of a higher brain criterion, which says that all permanently unconscious patients are dead. We should reject the whole brain criterion (currently the law in most jurisdictions), which says that some of them are dead and some are alive, depending on the state of their brain stems.

Introduction This handbook is about the ethics of end-of-life care. Any moral duty toward someone near the end of life comes to an end when life itself comes to an end. But when does human life end? The answer may seem obvious, but consider patients who are permanently unconscious even though their bodies are kept alive on life support. Ventilators and other forms of life support make such cases common. Some people believe that such patients are dead; once you are permanently unconscious, the real you is gone. Others believe they are alive; after all, the

Defining Death 75 bodies are—how can a patient be on life support without being alive? Some want to split the difference and say that the death of the patient is one thing, while the death of the body is another. Others protest that there is only one kind of human death, not two. Some concede that there is only one kind of death for humans, and say that death for humans is just like death for all other organisms (we are organisms, after all). Others make the same concession but say that death for humans is different, for humans have a unique capacity for rationality and self-awareness. Defining death is not as straightforward as it may seem.

Anatomy, Ventilators, and Criteria of Death To define death we must first understand some anatomy and recent medical history. The brain is the command and control center for the body. The functions of the brain are divided between the cerebral hemispheres (the higher brain) and the brain stem (the lower brain). The higher brain gives us our capacity for consciousness, sensory awareness, pain and pleasure, voluntary actions such as bodily movement, and respiration when we breathe consciously. The brain stem controls vegetative functions, such as metabolization, internal regulation of temperature, involuntary reflexes (e.g., the eyes dilating in response to light), and respiration when we do not think about breathing. Among other things, the brain stem contains the ascending reticular activating system—the portion of the brain stem that regulates the sleep-wake cycle: it puts you to sleep and wakes you up. Some writers claim the brain stem has the general task of managing the body as an organism—more on this later. The heart, lungs, and brain are interdependent. When the heart stops beating and blood stops circulating, blood no longer delivers oxygen to the rest of the body, and the brain, lungs, heart, and rest of the body soon suffer irreversible damage and quickly cease to function at all. When the lungs stop working, the blood no longer receives oxygen, and again the rest of the body—including the heart and brain—starts to break down and ceases to function. If all functions of the brain irreversibly cease, then breathing stops (because neither the higher brain nor the brain stem is regulating respiration), and again the body is deprived of oxygen—including the heart, lungs, brain, and everything else. In short, if any of these three organ systems fails, the other two fail, and soon the entire body shuts down. For many years, law and medical practice used the cardiopulmonary criterion to determine when a patient is dead: Cardiopulmonary criterion of death: a patient is dead when all heart and lung functions have irreversibly ceased. (This is sometimes referred to as the “circulatory criterion.”)

76

John K. Davis

In other words, once you have permanently stopped breathing and circulating blood, you are dead. Note that these functions must have stopped in a way that cannot be reversed (merely declining to reverse when we can easily do so does not render the patient dead). Note also that it is the functions that must stop, not the parts of the patient’s body that perform those functions. For example, if the patient’s heart and/or lungs have stopped but her blood and oxygen are kept circulating by machines, then her circulatory and respiratory functions have not stopped at all. The same is true if she has a mechanical heart. The cardiopulmonary criterion worked well for decades, but it ran into trouble in the 1950s and 1960s, after ventilators were invented and intensive care became more sophisticated. A ventilator is a breathing machine that takes over for the lungs, allowing respiratory function to continue.1 The heart can function on its own for a while even when the brain stem no longer functions at all, but it can function even longer with cardiac and blood pressure regulation— sometimes for many years. Life support became more sophisticated, and soon there were many patients who were permanently unconscious but were classified as alive under the cardiopulmonary criterion because machines took over for or supplemented their respiration and circulation. This created a serious practical problem: keeping someone in intensive care indefinitely costs a fortune. The obvious solution, of course, is to withdraw life support and let the patient die, but around that time another problem surfaced: physicians began performing organ transplants. However, there was (and still is) a shortage of organ donors. All those permanently unconscious patients on life support were an obvious source of organs (organs don’t store well for long, and they are useless unless they are removed immediately after the body dies, or sooner). Why not take organs from them? They are as good as dead anyway, it will save money, and it will save other patients who are conscious. There was only one problem: under the cardiopulmonary criterion of death, they were still alive, and taking organs amounted to killing them. One solution to both problems was to change the criterion of death and reclassify these people as dead. In 1968 a committee at Harvard Medical School that has come to be known as “The Ad Hoc Committee” proposed a new criterion of death to deal with both problems (Ad Hoc Committee 1968). A few years later President Carter appointed the U.S. President’s Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research (“the President’s Commission”) to do the same thing: propose a new criterion of death. Both the Ad Hoc Committee and the President’s Committee recommended the same new criterion (President’s Commission 1998): Whole brain criterion: a patient is dead when all functions of the whole brain, including both the higher brain and the brain stem or lower brain, have irreversibly ceased.2

Defining Death 77 Within a few years the whole brain definition became law in all fifty states and most countries (see, e.g., California Health and Safety Code, §7180). However, the whole brain criterion does not classify as dead all patients who are permanently unconscious. Patients who are permanently unconscious and whose brain stems have completely ceased to function are dead under this criterion. However, many patients who are permanently unconscious have partial or even complete brain stem function, and they are alive under this criterion. Therefore, the whole brain criterion is only a partial solution to the cases of permanent unconsciousness it was devised to handle: it handles some of those cases but not all. On top of that, it some people are inclined to say that any patient who is permanently unconscious is dead, regardless of the condition of the brain stem. They believe that permanent unconsciousness alone is necessary and sufficient for death. Those who believe this subscribe to a third criterion of death: Higher brain criterion: a patient is dead when all functions of the higher brain have irreversibly ceased. This criterion is met in all and only cases of permanent unconsciousness, even if the brain stem still functions completely. So we must choose between the whole brain criterion, which says that permanent unconsciousness is necessary but not sufficient for death, and the higher brain criterion, which says that permanent unconsciousness is both necessary and sufficient for death.3 Before we can consider the arguments for and against each criterion, however, we must first review some terminology.

Five Concepts Used in Connection with Defining Death To settle which criterion of death is correct, we must settle which definition of death is correct. “Defining death” is used loosely to mean various things: a concept of death, a criterion of death, diagnostic tests for death, or a legal definition of death, among others. I will use the phrase “definition of death” this way: a definition of death, strictly speaking, is a philosophical definition. It articulates a concept of death. That is, we may have a rough, inarticulate sense of when a patient is dead. A definition takes that idea and makes it explicit and precise. We support a definition of death by making philosophical arguments for it. The claim that it is consistent with our concept of death is one argument; there are many others. A criterion of death, by contrast, is a physiological state of affairs that doctors can detect using diagnostic techniques and devices. The point of a criterion is to enable doctors to tell whether a patient is dead according to the definition of death that that criterion matches. For example, whether a patient’s whole

78

John K. Davis

brain has irreversibly ceased to function is a physiological state of affairs doctors can detect with an EEG. If doctors detect that condition in a patient, then they know the patient is dead according to the whole brain criterion, and therefore dead according to whatever definition of death matches the whole brain criterion. A criterion matches a definition of death if and only if all and only those patients who fit that criterion are dead according to that definition. For example, the higher brain criterion tells us that a patient is dead when all functions of the higher brain have irreversibly ceased. The higher brain criterion of death matches the sentience definition of death because all and only those patients whose higher brains have irreversibly ceased to function at all will have irreversibly lost the capacity for sentience. Thus, a criterion is correct in a different way than a definition. Instead of being correct because it is supported by the best philosophical arguments, a criterion is correct because it matches the definition that is supported by the best philosophical arguments. A legal definition of death is a law that tells doctors which criterion of death to use. Legal definitions typically do not mention a definition or a diagnostic test. They leave the former to philosophers and the latter to doctors. Finally, a word about being “brain-dead.” The general public often uses this phrase as if it were a criterion or definition of death, but it’s used in too many ways to be useful. Sometimes it is used to refer to the whole brain criterion or the higher brain criterion—it’s not clear which, and that’s part of the problem. Sometimes it’s used to suggest that a patient might have a dead brain but a living body, as when someone suggests we should take a patient off life support because he is “brain-dead.” However, it is controversial whether the brain—or the person—can be dead before the entire body is dead. I think this is possible, but to avoid begging that question, and to avoid ambiguity, I’ll avoid the phrase “brain-dead.” So here is how we will proceed. First, we consider the main definitions of death and the philosophical arguments for and against each of them. Then we ask which criterion of death fits each definition, and consider the merits and demerits of each criterion. For practical and historical reasons, the literature on defining death tends to ask which criterion is correct, but this puts the cart before the horse. We need to start with definitions.

Four Definitions of Death We will consider four definitions of death: the integrated function definition, the auto-integration definition, the personhood definition, and the sentience definition. These terms are not in the literature on defining death; I had to coin them because they are usually not clearly distinguished from the criteria they match with, and are typically not named. (They are sometimes spoken of as the

Defining Death 79 rationale or argument for this or that criterion.) I want to bring the definitions into the light in their own right, and name them so we can state them clearly and identify their supporting arguments. These four definitions fall into two groups. The first and second identify the patient as an organism, and ask when the organism has died. The third and fourth identify the patient as what we might call a being (an entity capable of awareness) and ask when that being has died. The Integrated Function Definition When the President’s Commission argued for the whole brain criterion, the Commission spoke of “integrated function” (President’s Commission 1998: 124–125). The Commission was focused primarily on criteria and only secondarily on definitions, so it did not specify the definition it believed the whole brain criterion matched, but here is one possible definition they might have had in mind: Integrated function definition: a patient is dead when the patient’s functions are no longer integrated and directed in a coordinated way. This makes some sense; the word “organism” suggests something that is organized as a unit (e.g., a clump of algae is not an organism), and is not merely a collection of parts that happen to be held together somehow. (That’s why a clump of algae or a coral reef is not an organism.) Moreover, the emphasis on organizing and directing is consistent with the fact that the brain directs the rest of the body. The President’s Commission and nearly everyone else who has considered the issue believe that it is the brain stem that manages, coordinates, and integrates the parts and functions of a human being into a single, coordinated organism. It is possible that the Commission held the integrated function definition of death in mind when they recommended the whole brain criterion. Notice that the integrated function definition is focused on the function of integrating and directing the body; by its terms, it does not require that the organism itself perform that function. Therefore, according to this definition, the integration and coordination of a human organism need not be performed by the brain stem. That means that a patient whose brain stem has completely ceased to function is still alive if we have technology that can substitute for the functions of the brain stem—and we do, more or less (with life support, drugs, round-the-clock intensive care, and so on). But this means that none of the patients who are permanently unconscious are dead—not even the ones whose brain stems completely ceased to function— so long as they are on life support. Since they are still alive, we cannot remove the life support, so they must be put on life support (to prevent their death) and kept there (to prevent their death). This makes the integrated function

80

John K. Davis

definition useless as a means to solve the two problems that motivated the shift from a cardiopulmonary criterion to a whole brain criterion: too many permanently unconscious bodies on life support, and not enough organs available for transplant. The integrated function definition says all those bodies are still alive. The Auto-Integration Definition But perhaps the integrated function definition wasn’t stated properly. Perhaps we can save it by amending it into this: Auto-integration definition of death: A patient is dead when the patient’s functions are no longer integrated and directed in a coordinated way by the patient’s own body. On this definition, an organism is dead even if something external to the body, such as hospital equipment and a team of medical providers, organizes its parts and functions. (Presumably they organize the parts by keeping all the parts within the parameters that other parts require—e.g., the right temperature.) This definition does a better job of solving the problems with permanently unconscious patients: now at least the patients whose brain stems have completely ceased to function are dead, and that was the intent behind the definition in the first place. This definition, or something close to it, is likely to be the definition the Commission had in mind, for it is consistent with the whole brain criterion: if the brain stem completely ceases to function, then the body is no longer integrating its own parts and their functions. As we noted earlier, this classifies only some permanently unconscious patients as dead (those whose brain stems are completely nonfunctional), but at least this takes care of part of the population of permanently unconscious patients. Definitions of death are philosophical definitions, and they must be supported by philosophical arguments. Perhaps the best or at least most welldeveloped argument for this definition is what I’ll call the argument from usage and tradition. This argument begins with the fact that people have concepts long before they have definitions of those concepts. When we develop a definition of a concept we already have, our definition needs to fit that concept— otherwise the definition is not accurate. As James L. Bernat puts it, “[D]eath is a nontechnical word that is and has been used broadly and correctly by the public. Any formal attempt to define [death] should strive to capture this ordinary, nontechnical meaning” (Bernat 1998: 15; see also Bernat et al. 1981). The President’s Commission made a related point, saying that its mandate as a political commission was not to change the definition of death, but merely to recommend a new criterion for the concept we already had, one adapted to the

Defining Death 81 situations created when patients are on ventilators (President’s Commission 1998: 124, 130). So what is the traditional concept of death? In an influential early paper on the definition of death, Bernat and his colleagues Charles M. Culver and Bernard Gert argued that the traditional concept of death is permanent cessation of function of the organism as a whole. As Bernat later explained, this refers to that set of vital functions of integration, control, and behavior that are greater than the sum of the parts of the organism, and that operate in response to demands from the organism’s internal and external milieu to supports its life and to maintain its health. (Bernat 1998: 17) The President’s Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research (“President’s Commission”) agreed with Bernat that a definition of death should match the traditional concept of death, and offered a similar account of the traditional concept: “[T]he traditional means of diagnosing death actually detected an irreversible cessation of integrated functioning among the interdependent bodily systems” of the heart, lungs, and brain (President’s Commission 1998: 124). In effect, both the Commission and Bernat claimed that the traditional concept of death is that we die when we cease to be organisms. But how do we know that this is the traditional concept of death? The argument must be that this concept fits our traditional judgments about death better than other concepts or definitions. Does it? It is hard to say. If there is a traditional concept of death, it evolved long before ventilators were invented. Until ventilators and other forms of life support came along, we never had to apply our concept of death to patients who were permanently unconscious but whose bodies were still breathing and warm to the touch. In a world without ventilators, any patient who needed a ventilator would soon die from oxygen deprivation under any definition or concept of death. Because we never faced such cases, our concept of death may well be silent about them. The integrated function and auto-integration definitions are both consistent with these pre-ventilator cases: the patient has no integrated function and also cannot integrate herself. Later we will discuss the personhood definition, which says you die when you cease to be a person; pre-ventilator patients are not persons. After that, we will consider the sentience definition, which says you are dead when you lose the capacity for sensation or awareness of any kind; these patients lack that too. In short, our judgments about death before life support technology was developed are consistent with all the definitions of death, and none of them are more consistent with those judgments than any other definition. Our traditional concept of death was simply this: you are dead when

82

John K. Davis

everything shuts down permanently. That concept said nothing about cases where some functions or organs shut down permanently and others do not. We needed a new definition of death for such cases precisely because we did not know what to say about them. Another argument for the auto-integration definition is that the auto-integration definition is consistent with our judgments about embryos and fetuses, while the personhood and sentience definitions are not. We tend to think that we come into existence either at conception or somewhere during gestation. Some scientists believe that we do not develop the capacity for sensation (sentience) until seven or eight weeks at the earliest, and perhaps as late as twenty-six weeks (Derbyshire 2006). However, we exist before that stage. Because we are neither sentient nor persons when we come into existence, we are not dead just because we irreversibly lose the capacity for sentience or personhood. There is a problem with this argument. First, the auto-integration definition says we die when we irreversibly lose the capacity to regulate our own vegetative functions. If so, then we do not come into existence until we gain that capacity. However, fetuses do not regulate their own vegetative functions until around eight to twelve weeks; until then the mother’s body regulates the fetuses’ vegetative functions. Defenders of the auto-integration definition would have to claim that we come into existence only around eight to twelve weeks after conception. This is an intellectually consistent view, but many people will find it hard to accept. Moreover, the auto-integration definition faces some objections. Here is an example: imagine that scientists have accidentally created a virus that destroys brain stems but leaves higher brains intact, and that a patient has contracted the virus. Her brain stem has permanently ceased to function but her higher brain functions normally. Her integrated function is not handled by her body. It is handled by machines connected to her body in some way, so she is dead according to the auto-integration definition (she has integration but not the auto part). Clearly she is alive, so the auto-integration definition is false. The Commission came close to considering cases like this when it pondered the objection that patients whose brain stems had permanently ceased to function would still be alive on the whole brain criterion, for the organizing function of the brain stem was handled by life support and a medical team. The Commission dismissed this objection on the grounds that such a body is merely a “perfused corpse”: it does not yawn, its eyes do not track the light, its pupils do not dilate, and it does not react to pain or reflex stimulation (President’s Commission 1998: 128). This is a poor response. First, it does not show that it’s a metaphysical impossibility to have a working higher brain and a nonworking brain stem (the virus was imaginary but conceivable). The fact that such cases do not occur now does not entitle us to brush them aside. Second, it is not clear that such

Defining Death 83 bodies are merely “perfused corpses,” even given current medical technology. Why does the appearance of such bodies, and their ability to yawn or dilate their pupils, disqualify them from being alive? The term “corpse” begs the question when we are considering whether their functions are integrated in the way said to be necessary for life. It seems we cannot ignore the need to include the function of awareness and cognition as part of the human organism. Perhaps we should revise the autointegration definition to include it. Bernat eventually did, claiming in a later article that there are three categories of critical functions of the human organism as a whole: (1) breathing and circulation, (2) integrating functions that ensure homeostasis (keeping the internal conditions of the body—e.g., temperature— stable and constant), and (3) consciousness. Bernat says that death requires loss of all three categories of critical functions (Bernat 1998: 17). Let’s follow his lead and consider revising the auto-integration definition of death to include higher brain functions: a patient is dead when the patient’s own body (specifically the brain) irreversibly ceases to organize its parts and functions, including all higher brain functions. In other words, you are dead when your brain stem and your higher brain no longer function at all. There are two ways to specify this: Both are necessary for life: Both brain stem functions and higher brain functions are necessary for being alive, and irreversible loss of either is sufficient for being dead. Either is sufficient for life: Either brain stem functions or higher brain functions are sufficient for being alive, and irreversible loss of both is necessary for being dead. These two versions have very different implications. Bernat seems to have the first version in mind when he says that consciousness and other functions are each necessary for being alive—if you lose any one of them, you are dead. However, because it makes higher brain functions necessary for life, the first version implies that all patients who are permanently unconscious would be dead. (It also implies that anencephalic infants are stillborn, for they are born without higher brains.) This greatly increases the number of dead people (to put it crudely), for not only are all patients with permanent unconsciousness and no brain stem function dead, but so are all patients who are permanently unconscious but still have at least some brain stem function left. Many people will have no problem with this implication, but defenders of the auto-integration definition cannot accept it, for they are typically defenders of the whole brain criterion as well; they are already committed to saying that patients with partial brain stem function are not dead. Therefore, the first version goes much further than they may want to go. That alone is not a reason to reject this version, but

84

John K. Davis

it does indicate that defenders of the auto-integration definition cannot make use of it. They must prefer the second version. However, the second version, which says that neither set of brain functions is necessary for being alive, implies that neither the vegetative functions nor the higher brain functions are essential for (necessary properties of ) a human organism. This is a peculiar implication. If neither the functions of our bodies as organisms nor the functions of the our higher brains are necessary for being a human, what is? The whole point of a definition of death is to identify something so essential to being human that, when we lose it, we lose life itself, regardless of whether some organic part of us may still be alive. We seem to have lost any rationale for the definition itself. (As we will later see when we discuss criteria of death, many writers have objected to the whole brain criterion on more or less this ground—that it combines a higher brain criterion with a brain stem criterion without any rationale for doing so.) Which Criterion Matches the Auto-Integration Definition There is another problem with the auto-integrated definition (and the integrated function definition too, for that matter): it matches the whole brain criterion, and the whole brain criterion has a problem of its own. Note that the whole brain criterion requires that all functions in both the lower and the higher brain have irreversibly ceased. Only the brain stem is involved in the homeostatic integration Bernat and the President’s Commission proposed as the core of a definition, but the Commission required irreversible cessation of all capacity for consciousness and sensation as well, even though they have no role in integrating the vegetative functions of the body, and that is the version that has been legislated into existence. This means that there is an inconsistency in the whole brain criterion: if it is based on the integrative function concept, then why does the criterion include the whole brain—that is, why require that all functions of the upper brain, where consciousness happens, have irreversibly ceased as well? After all, it’s the brain stem that integrates the vegetative functions, not the upper brain. The obvious answer is that we want a criterion of death that matches our intuition that loss of consciousness and sensation is part of death (we don’t want patients who are conscious but dead). However, that is not an integrated function concept; it is a concept based on consciousness and sensation, or perhaps a mix of the two. The Personhood Definition of Death Many philosophers prefer a definition of death that focuses on whether the person still exists. John Harris says that persons are what matter morally, so

Defining Death 85 we should not focus on the death of the organism. On this understanding, your body might remain alive even after you have died, provided being a person is essential to you (Harris 1985: 241). Shelly Kagan holds the same view: after you cease being a person, your body might remain alive but you will be dead (Kagan 2012: 173–174, 177). What matters, for Kagan, is whether you are still a person, not whether you are still alive, but, perhaps because that is what matters, he contends that death should be defined in terms of personhood (Kagan 2012: 178; see also McMahan 2012: 292). This point of view indicates another definition of death: Personhood definition: a patient is dead when he or she suffers irreversible loss of the capacity for personhood. According to the personhood definition of death, all permanently unconscious patients are dead. Not all humans are persons. Philosophers use “person” to refer to beings who have rationality, self-awareness, and perhaps other characteristics, such as a sense of themselves existing over time and the ability to communicate, among other things—roughly speaking, a being with a mental life comparable to that of a normal human in early childhood and later. So far as we can tell, no animals are persons in this sense, though nothing hangs on this. However, not all humans are persons in the way that philosophers use that word. Infants are not persons in this sense; we become persons sometime during early childhood and cease to be persons when we become deeply demented. Which criterion matches this definition of death? That is not clear. Certainly not the whole brain criterion, which says that people whose brain stem functions (at least in part) are still alive. However, the higher brain criterion does not quite match it either. The higher brain criterion says a patient is dead only when all functions of the higher brain have irreversibly ceased. The problem is that there are patients who are not persons yet still have partial higher brain function. For example, patients in deep dementia have lost some higher brain function, for they are no longer capable of rationality or self-awareness. They are capable merely of sensation, and perhaps very rudimentary awareness—like an infant. They are alive according to the higher brain criterion, but they are not persons. A criterion for the personhood definition of death would have to identify the specific regions of the higher brain necessary for rationality and self-awareness, or perhaps specify the degree of higher brain tissue necessary for rationality and self-awareness. In short, the personhood definition requires a new criterion of death that specified which parts and how much of the higher brain are necessary for personhood. This is difficult but not impossible in principle. What is the argument for the personhood definition of death? One argument for the personhood definition is that persons are what matter morally, so we

86

John K. Davis

should focus on the death of the person, not the death of the organism (Harris 1985: 241; Kagan 2012: 173–174, 177–178; see also McMahan 2012: 292). Of course it may be true that persons are morally more important than sentient beings or organisms that are not persons. (I happen to think this is true.) However, that shows only that the death of a person is morally worse than the death of a sentient being or a nonsentient organism. We cannot assume that we are identical with whatever is morally most important. Moreover, the death of a nonperson is tragic too; we do not want a definition of death that permits us to deny life support to a baby or demented elderly patient on the grounds that they are not persons. They are morally important enough to include in our definition of death. Another argument is that we are essentially persons, that being persons is a necessary property for us, and if we lose it, we cease to exist. We die when we stop being persons. Are we essentially persons? If we are, certain things follow. First, when I was a baby I did not yet exist (the baby did but I didn’t). I was not a person, so the baby preceded my birth, so to speak. Second, if I become demented later in life, I will cease to exist and be succeeded by another entity—a nonperson who looks a lot like me and bears my name and social security number. We can, of course, accept these implications and thereby defend the personhood definition, though to do that sincerely you must really believe you came into existence a couple of years after your birth. Few of us believe that, but perhaps we can change our beliefs on that point. Still, asking what we essentially are may be the wrong approach to defining death. The Sentience Definition of Death Here is a definition of death that takes another approach: Sentience definition: a patient is dead when that patient irreversibly loses the capacity to experience sensation or awareness of any kind, even basic awareness of pain. Sentience is the ability to experience feeling or sensation. Being sentient does not require the capacity for rationality or self-awareness. Normal humans are sentient; so are higher animals and many lower ones. According to this definition, patients who are permanently unconscious are dead regardless of the condition of their brain stem, but patients in deep dementia are still alive. This definition probably matches the higher brain criterion, assuming that a partly functioning higher brain necessarily makes it possible to have some degree of awareness. (If that assumption is false, then the higher brain criterion would have to be revised and made specific to the tissues or degree of damage necessary to eliminate the capacity for sentience.)

Defining Death 87 The argument for the sentience definition of death begins with a concession: there is more than one definition of death, and there should be. To die is simply to cease existing. Many things can die: empires, political movements, hope. Also cells, organs, organisms, sentient beings, and persons. There is no single correct definition of death for all purposes—which definition is correct depends on what kind of entity we are interested in. Humans are many things: organisms, sentient beings (almost always), and persons (at least most of the time). Which of these are we interested in? That depends on the context. For example, in educational and many other social contexts we are most interested in whether a person still exists. But we are talking about death in a medical context. In medical contexts the concept of death is typically invoked when we think about whether we still have a duty to the patient to provide medical care. Notice I did not say, “a duty to provide medical care to the patient.” We might have a duty to provide medical care to the patient but not owe that duty to the patient; we might, for example, owe a duty to the family to keep the patient on life support a while longer so they can come to terms with the death of a loved one. I am talking about a duty to the patient to provide medical care to that patient. When do we no longer have such a duty to the patient? We might say that a patient is dead when we are morally permitted to let her die, but that is too broad. There are situations where it is morally right to let someone die because continued life is not in that person’s interest, even though she is alive and may live much longer. When a patient is dead and when we may let that patient die are different questions. It is better to say we have no medical duty to the patient when the patient is no longer a patient. This sounds circular, but it isn’t. To be patient is to endure something one might prefer not to, such as a delay. To be a patient is to be a candidate for or under the care of someone providing medical services. These concepts are related; the words “patient” and “patience” come from the Latin verb pati: to suffer. We say that doctors, nurses, and veterinarians have patients, but that coroners, undertakers, and car mechanics don’t. It is essential to being a patient that the being is at least capable of suffering or enduring something it would avoid if it could. That is why a living organism that cannot experience sensation or awareness of any kind cannot be a patient. For example, a marine biologist keeping a coral colony alive, or a gardener working to save a flower bush, normally would not speak of the coral or the bush as a patient, even if they were using drugs and surgery to keep the coral or the bush alive. Sentient beings can be patients, but a living body that permanently lacks awareness of any kind would not be, even if it were human, and even if we had other reasons for wanting to keep it alive (as a mere body). It is true that we use the word “patient” when we speak of permanently unconscious humans, but we really shouldn’t. Using the word that way is a carryover from situations where humans are sentient, as they usually are, and perhaps also a courtesy to the family of the “patient.”

88

John K. Davis

A being cannot be a patient unless it has at least the capacity for sentience, for if it has no capacity for awareness, it cannot be helped or harmed in any way. A person may die before the patient does, so to speak, but medical duties survive the loss of personhood. They do not survive the loss of patienthood. A critic might object that a living being can be helped or harmed even if that being is not sentient. For example, you can harm a tree by pruning it too aggressively, or make it better off by watering it. My response is to distinguish between two kinds of welfare, and therefore between two kinds of helping and harming. First, there is welfare in the sense of making an entity better or worse off relative to what that entity prefers or seeks. Second, there is welfare in the sense of making an entity better or worse off relative to something else, such as a natural state of flourishing, or suitability for human purposes. We can help or harm sentient beings in both senses, but when we help human patients, even deeply demented ones, we typically help in the first sense. We help or harm that patient by giving the patient better or worse sensations, or prolonging or shortening the time during which the patient can have sensations (i.e., prolonging or shortening his life). Of course, there is a kind of welfare that is external to the patient’s body, so to speak. For example, there are unexperienced harms, events, or states that are harmful to you even if you never become aware of them, or suffer any conscious unpleasantness. Your friend might talk about you behind your back, or your spouse might have an affair. Similarly, there are unexperienced benefits; when your friend talks about you behind your back, she might be saying nice things about you. However, it is hard to think of any unexperienced benefits or harms for beings that are incapable of having experiences. For example, it might be morally wrong to make fun of an ugly anencephalic baby, but it would not be wrong because it violates a duty to that baby (it is wrong for other reasons). I am not saying that it’s impossible for nonsentient beings to have unexperienced benefits or harms, only that I can’t think of any examples of this. In the absence of any examples, the burden of argument lies with those who oppose the sentience definition of death. My conclusion is that the sentience definition of death is correct, and the other three definitions are incorrect. Which Criterion Matches the Sentience Definition of Death The sentience definition matches the higher brain criterion: Higher brain criterion: a patient is dead when all functions of the higher brain have irreversibly ceased. This criterion is met in all and only cases of permanent unconsciousness, even if the brain stem still functions. As noted earlier, no state or country has adopted this criterion yet.

Defining Death 89 The higher brain criterion does not match the personhood definition. A patient can lose personhood but still retain enough higher brain function to be sentient; the patient would be alive according to the higher brain criterion even though the patient is not a person. This is important because implying that the higher brain criterion correlates with the personhood definition leaves the higher brain criterion open to spurious objections that adopting a higher brain criterion means we will classify deeply senile elders as dead when they are merely demented. The higher brain criterion implies no such thing. Another version of this objection is that the higher brain criterion will lead us down a slippery slope, and in the end we will start declaring deeply demented patients to be dead: a grandfather suffering from Alzheimer’s will be declared dead while he still has some cognitive function left. This is a bad objection. First, even if there is a slippery slope here, that does not show that the higher brain criterion, properly understood, implies that demented people are dead. At most it shows that we are at risk of mistakenly thinking it does. (By the way, that is not really a slippery slope; that is a concern about whether doctors can apply this criterion accurately in practice. Given that doctors are scientifically trained, it is not likely they will have that problem, even if laypeople do.) Second, to make a slippery slope objection, one must show that declaring death when all higher brain functions are irreversibly lost will eventually lead us to declare death when only some of them are lost. That claim, in turn, requires an argument that either we will eventually change the law to another criterion, or that we will cease to apply the criterion accurately. There is no reason to fear that either of these things will happen. Unlike the whole brain criterion, the higher brain criterion is unproblematic and not subject to any crippling objections.

Conclusion Our ability to keep patients alive in conditions that, a few decades ago, would have meant certain and immediate death has required changes in our concept and definition of death. Society and the medical community have moved from the traditional cardiopulmonary criterion of death to a whole brain criterion of death, but without fully understanding the rival concepts and definitions of death that underlie those criteria. Once we understand the definitions of death and the arguments for and against them, it becomes clear that we need to move from the whole brain criterion of death to a higher brain criterion of death.

Related Topics Chapter 1, “Is It Possible to Be Better Off Dead?” by Geoffrey Scarre Chapter 7, “Deciding for the Incompetent” by Eric Vogelstein

90

John K. Davis

Notes 1. Ventilators are sometimes called “respirators.” In addition to ventilators, cardiac pacemakers, drugs for blood pressure, and intensive care can help keep cardiac and respiratory functions going. 2. Notice that the whole brain criterion says a human body is not dead until “all” functions of the whole brain have irreversibly ceased. However, we do not apply the criterion that strictly in practice. There are a number of functions that continue for a time after patients are pronounced dead: neuroendocrine function and some cerebral electrical activity, for example. (Bernat acknowledged this, and later revised his proposal to limit it to “critical” functions; Bernat 1998: 17.) The whole brain criterion probably needs to be revised to state which brain functions are essential. In principle this should not be an insurmountable problem, but it does tell us that the criterion has not been fully spelled out. 3. One other criterion deserves mention: some people say death occurs when the soul leaves the body. This may well be true, but it makes a bad criterion of death because this event (if it occurs) does not appear to be a physiological state of affairs, so we have no diagnostic tools to detect it directly. We might detect it indirectly if it occurs under certain physiological conditions, but we cannot use that as a criterion without first knowing precisely which conditions those are. Presumably this criterion would be supported by the sentience or personhood definitions (depending on the cognitive state of the soul, I guess).

References Ad Hoc Committee (1968) “A Definition of Irreversible Coma: Report of the Ad Hoc Committee of the Harvard Medical School to Examine the Definition of Brain Death,” JAMA 205(6): 85–88. Bernat, J.L. (1998) “A Defense of the Whole-Brain Concept of Death,” Hastings Center Report 28(2): 14–23, 15. Bernat, J.L., C.M. Culver, and B. Gert (1981) “On the Definition and Criterion of Death,” Annals of Internal Medicine 94(3): 389–394. California Health and Safety Code, §7180. Derbyshire, S.W.G. (2006) “Can Fetuses Feel Pain?” British Medical Journal 332(7546): 909–912. Harris, J. (1985) The Value of Life: An Introduction to Medical Ethics, New York: Routledge. Kagan, S. (2012) Death, New Haven: Yale University Press. McMahan, J. (2012) “Death, Brain Death, and Persistent Vegetative State,” in H. Kuhse and P. Singer (eds.) A Companion to Bioethics (2nd Edition), Malden, MA: WileyBlackwell, 286–298. President’s Commission (1998) “Defining Death: Medical, Legal, and Ethical Issues in the Definition of Death,” in A.R. Jonsen, R.M. Veatch, and L. Walters (eds.) Source Book in Bioethics, Washington, DC: Georgetown University Press, 118–142.

Further Reading Gervais, K.G. (1986) Redefining Death, New Haven: Yale University Press. (An influential overview of the topic.)

Part II

Who Decides When to End Life?

6

Autonomy, Competence, and End of Life James Stacey Taylor

Summary Bioethicists generally agree that (1) patients should be able to decide for themselves which medical treatments they will accept, and which they will refuse; (2) only patients who are competent to make decisions for themselves should have their decisions respected in this way; and (3) this is required out of respect for patient autonomy. Given the centrality in contemporary bioethics of the concepts of “competence” and “autonomy,” it is important to understand what conditions must be met for a patient to be competent, or to be autonomous, and what relationships hold between these attributes. The aim of this chapter is to provide an outline of these issues, with special attention being paid to their relevance for end-of-life decision-making. Bioethicists generally agree that patients should be able to decide for themselves which medical treatments they will accept, and which they will refuse; they also agree that patients should be able to request treatments that they believe would best serve their interests. Moreover, bioethicists also generally agree that only patients who are competent to make decisions for themselves should have their decisions respected in this way; they also generally agree that this is required out of respect for patient autonomy. The concepts of “competence” and “autonomy” are accordingly central to contemporary bioethics. It is thus important to understand what conditions must be met for a patient to be competent, or to be autonomous, and what relationships hold between these attributes. The aim of this chapter is to provide an outline of these issues, with special attention being paid to their relevance for end-of-life decision-making. This chapter consists of three sections. In the first two sections the related concepts of competence and autonomy will be discussed and their importance for ethical issues concerning patients’ decision-making at the end of life outlined. In the third section the ethical implications of valuing such decisionmaking will be addressed, focusing on the issues of voluntary euthanasia, physician-assisted suicide, and advance directives.

94

James Stacey Taylor

Competence Patients who have a variety of treatment options at the end of their lives are faced with a series of decisions. Most obviously, they are faced with the decision of whether they wish to utilize the life-extending treatments that they have access to. While at first sight it might appear that the answer to this question would always be in the affirmative—for who would wish to die when she could live?—it should be recognized that many persons are concerned about the quality of their lives as well as their duration. A person might thus choose not to pursue the means of extending her life if she judges that her quality of life would not justify this. For a person to make such decisions for herself and to have these decisions respected by others, she should be competent to decide; if she is not competent then her decisions should carry no weight, and her treatment decisions should be made by another. This raises the question of the conditions that must be met for a person to be competent to decide.1 As Beauchamp and Childress note, the core meaning of “competence” is “the ability to perform a task” (Beauchamp and Childress 2013: 115). Since persons can be competent to perform certain tasks but not others whether a person will be competent to make a decision will depend on the type of decision that is being made. A person might, for example, be competent to decide whether she would prefer to remain in her own home during the last stages of a terminal illness or enter a hospice, but not be competent to make decisions concerning the particular regime of medications that she should be prescribed, since this would require specialized training. A person’s competence to make certain decisions will also vary across time. A person might, for example, be competent to judge whether she should remain in her own home or enter a hospice during the early stages of Alzheimer’s disease, but lose this competence as the disease progresses. Similarly, a professional ornithologist might fail to be competent to identify a species of bird during a period of temporary amnesia, but recover her competence in this area once she recovers her memory. This pair of examples highlights a feature of competence that is often overlooked in discussions of this concept in bioethics: that the question of whether a person is competent to make decisions can arise in contexts where her decisions are about purely descriptive matters as well as in contexts where her decisions will be partially informed by normative considerations. (The terms “descriptive competence” and “normative competence” will be used here to identify these two different forms of competence, although they are not commonly used in the philosophical literature on this issue.) To assess whether a person is descriptively competent one needs to assess whether her judgements about the issues that she is deciding on (e.g., what species a particular bird is) are likely to be objectively correct. Similarly, to assess whether a person is normatively competent one needs to assess whether her judgements about the

Autonomy, Competence, and End of Life 95 issues that she is deciding on (e.g., whether she should remain at home or enter a hospice) are likely to achieve her intended ends. A person is normatively competent to make a particular decision if her making it is likely to lead to success, where this success is measured against the aim of her decision. Conversely, a person will be incompetent to make a particular decision if it is unlikely to achieve the ends that she envisages achieving when she makes it. (Since the focus in bioethics is typically upon a person’s normative competence, it is this that will be focus of discussion from now on.) Rather than attempt to assess in each case whether a person is competent to make decisions concerning her end-of-life medical care, healthcare professionals instead assume that persons are competent in the relevant sense unless there is reason to believe otherwise (i.e., she should be assumed to be descriptively competent if she is deciding a descriptive issue, normatively competent if she is deciding a normative issue, and both if she is deciding an issue that has both descriptive and normative elements). If, for example, a person made decisions that led to her harming herself (e.g., by cutting herself repeatedly and deliberately) the background assumption that she would not want to self-harm would lead healthcare professionals to the initial belief that her decisions in this sphere were not likely to lead to outcomes that the person in question was aiming to secure, and hence that she was not competent to make the decisions in question. This should, in turn, lead them to assess whether she was in fact incompetent to make the decisions in question. How, then, can one assess whether a person is normatively competent to make a particular decision? At first sight such a determination would require only a simple hypothetical test: one would first determine the ends that the decision was made to secure, and then one would determine if the decision in question would be likely to secure them. If this answer is negative, then the person would not be competent to make this decision. But this simple hypothetical test should be rejected for two reasons. First, the question of whether a person is competent to make a particular decision is not necessarily only a question about whether she is engaging in appropriate means-end reasoning; it is also a question about the appropriateness of the ends that she is pursuing. A person might, for example, choose a course of action (such as self-mutilation) that would be justified given his religious beliefs, but where the lack of evidential support for those beliefs combined with the seriousness of the definite harm that he inflicted upon himself would lead him to be deemed incompetent in this sphere of decision-making (Beauchamp and Childress 2013: 117). Second, it should be recognized that this test is not purely descriptive, for it requires that one make a judgement concerning whether the decision in question was “likely” to secure the desired ends. This is a normative question (for how likely must be the desired result be for it to be sufficiently likely to qualify the decision in question as being competently made?) and hence the standards that would apply to it must be identified and defended.

96

James Stacey Taylor

Assessing whether a person is competent to make decisions concerning her end-of-life treatment in cases where this is in doubt is thus doubly normative: one must assess whether her ends are those that a competent person would pursue, and then, if the answer to this first question is affirmative, assess whether she is able to identify those courses of action that are sufficiently likely to achieve these (approved) ends. Such assessments are clearly difficult to achieve. It is not clear how one could defensibly adjudicate between ends that no competent person could endorse and those that might simply be unusual. Nor is it clear how one could determine when a course of action would be so unlikely to secure the desired end that a decision to pursue it would indicate that the person who made that decision was not competent to decide how to achieve the end in question. Owing to these difficulties a sliding-scale approach is often endorsed for determining when a patient should be deemed incompetent (Grisso and Appelbaum 1998: 139). (Note that this sliding-scale approach does not necessarily hold that a person’s competence to make particular decisions varies according to the degree of risk, but instead it holds that the degree of evidence that is needed to ascribe incompetence to a person varies with the level of risk; Beauchamp and Childress 2013: 120.) The greater the risks to the patient that accompany a particular decision the more stringent the criteria for ascribing competence to a person for a particular decision should be. In such a case the bar for a person’s ends to be considered those of an incompetent person is set low, as is the bar that must be overcome for a person’s proposed course of action to secure her ends to be considered that of an incompetent person. By contrast, if the risks to the patient that accompany a particular decision are small, the weaker the standards for ascribing competence to a person for a particular decision should be. In such a case the bar that must be overcome for a person’s ends to be considered those of an incompetent person would be set high, as would the bar that must be overcome for a person’s proposed course of action to secure her ends to be considered those of an incompetent person. To illustrate the forgoing considerations consider two cases: first, that of a person who has bodily identity integrity disorder (BIID), a condition that is characterized by the desire to amputate a healthy limb, and, second, that of a person who is unhappy with the shape of her nose, and who wishes to undergo cosmetic surgery to alter it. There would in both cases be a need to assess whether the persons concerned were descriptively competent to decide to undergo the respective surgeries that they were considering. To do this it would need to be assessed whether they understood the empirical elements of their surgeries, such as the expected surgical outcomes and the possible side effects. Their normative competence would also need to be assessed: Would these surgeries be likely to achieve the ends that they were aiming to achieve through them? For this to be assessed the ends that each person was aiming at would have to be identified. For example, the person with BIID might have

Autonomy, Competence, and End of Life 97 as his sole end the removal of a limb that he had never considered to be “his,” while the person considering cosmetic surgery on her nose might believe that having this would improve her marriage or make her more professionally successful. Clearly, the closer the persons’ ends are to the medical ends of their surgery the easier it will be to assess their normative competence with respect to their view that the course of treatment that they desire would achieve their ends. Given this, it might be easier to assess whether the person with BIID was normatively competent with respect to his view that his course of treatment would secure his ends than it would be to assess the normative competence of the person who desires cosmetic surgery on her nose. However, this is not the only assessment of normative competence that must be made: it must also be assessed whether the ends that the persons are aiming at are those that a competent person would pursue. And here it would be easier to assess the ends of the woman who desired cosmetic surgery than assess the (unusual) end of the person with BIID. Finally, given that there are greater risks associated with amputation than with minor cosmetic surgery, the bar for the person with BIID to be considered incompetent to decide to undergo amputation would be lower than that which must be met for the person considering undergoing minor cosmetic surgery to be considered incompetent to decide to undergo it.

Autonomy Yet even if a person is deemed competent to make a decision concerning her end-of-life treatment this would not necessarily suffice to ground the claim that there is a moral duty to respect her decision. It is widely held within contemporary medical ethics that there is a moral duty to respect persons’ autonomy—and hence to respect those decisions that she is autonomous with respect to—but not necessarily to respect those decisions that she competently made. While the reason for this differential ethical response to the value of competence and the value of autonomy has received almost no attention in the philosophical literature, it can be justified once the concept of autonomy has been clarified. The term “autonomy” stems from two Greek roots: “autos,” meaning “self,” and “nomos,” meaning “rule,” or “law.” To be autonomous, then, is to be selfruled. But this simple characterization of autonomy obscures a multitude of complexities. Persons sometimes believe that some of their desires are not truly their own; that they are “alien” to them in some way. A person who is trying to quit smoking, for example, might say that her desire to smoke is not really “hers,” but merely a desire that operates within her and that she wished that she did not have. A dominant line of research in contemporary autonomy theory addresses this phenomenon, aiming to provide an account of what conditions must be met for a person’s desires to be truly her own (i.e., those that she is autonomous with respect to, and whose satisfaction would create value for

98

James Stacey Taylor

her) as opposed to those that she is alienated from. If we understand a person’s self as being (at least partially) constituted by the set of desires, values, preferences, and other pro-attitudes that are truly her own, then this line of autonomy research could be understood as providing an account of the “self ” that rules when a person is autonomous. (Let us call this the “value-creating self ” approach to autonomy.) To hold that a person is autonomous if she is self-ruled, then, while correct, is theoretically unhelpful, for it obscures the fact that (at least for persons engaged in this line of research into autonomy) the question of which desires constitute the ruling self is precisely what is at issue. And just as the question of what conditions must be met to identify the “autos” (self ) in autonomy is controversial, so too is the concept of “nomos.” While the dominant line of research in contemporary autonomy theory focuses on providing an analysis of what conditions must be met for a person to be autonomous with respect to her desires, an alternative line focuses on understanding autonomy as a political, rather than a metaphysical, concept. This line of research (the “political” approach to autonomy) focuses on the question of when a person can properly be said to be ruling herself, rather than being ruled by others. Despite their differences both of these current research programs into the nature of autonomy accept that persons are autonomous when their decisions and consequent actions are governed by desires that are truly their own. These analyses of personal autonomy thus stand in stark contrast to the understanding of autonomy that was made famous by Immanuel Kant. For Kant, autonomy is “the property of the will by which it is a law to itself (independently of all properties of the objects of volition)” (Kant 2002: 58). On this understanding of autonomy a person’s will is autonomous if it imposes upon itself the moral law (it is “a law to itself ”) where the source of this law is not from any property “of the objects of volition”—that is, it does not stem from any of the agent’s desires, preferences, or other pro-attitudes. An autonomous will is thus one that is self-governed in accordance with pure rationality. On this view, then, autonomous agents should all be motivated to act in the same way insofar as they are acting as autonomous agents. Yet while Kant’s view of autonomy has been tremendously influential in moral philosophy in many discussions of end-of-life issues in which the concept of autonomy plays a central role, it is not the impersonal Kantian conception of this concept that is at issue but that of personal autonomy, “whereby a person is autonomous with respect to her desires, actions, or character to the extent that they originate in some way from her motivational set” (Taylor 2005: 1). The reason for this focus on personal, rather than Kantian, autonomy in discussions of end-of-life issues is simple. Many of these issues focus on the question of whether a person’s decisions concerning her treatment at the end of her life should be respected by her caregivers. As was outlined earlier, for this question to be answered in the affirmative a person should be competent

Autonomy, Competence, and End of Life 99 with respect to the decisions that she is making, with this competence being assessed (in part) against the likelihood that the course of action that she decided upon would realize her intended ends. The reason for respecting a person’s decisions that she makes concerning end-of-life issues is thus grounded in a concern that her individual ends be realized. As will be outlined ahead, a similar reason grounds the concern that a person’s autonomy be respected. Since it is personal, and not Kantian, autonomy that is focused on the desires of the individual agent, it is the conception of personal autonomy that has come to be the focus of discussions of end-of-life issues that draw upon this concept. Interestingly, although they differ in their focus, both the value-creating and the political approaches to personal autonomy agree on the primary justification for respecting a person’s competent and autonomous decisions: that doing so is likely to be the best way of protecting and promoting her well-being. The link between the value-creating approach to analyzing autonomy and the protection or promotion of her well-being should be clear: insofar as a person is autonomous if she is acting to satisfy those pro-attitudes that are truly her own, and thus whose satisfaction would create value for her, then she will be acting to protect or promote her well-being. A person with BIID, for example, would, if the strong desire to amputate his left leg that existed in his motivational set was truly his own, promote his well-being by having it amputated. Similarly, if one endorsed the political approach to analyzing personal autonomy, one could hold that since it is likely that a person would know better than others what acts would secure her well-being a concern for the protection and promotion of individual well-being would justify allowing persons to rule themselves rather than be ruled by others. This approach could be supplemented by the claim that the person with the greatest incentive to secure a person’s well-being would typically be the person herself—and so if we are morally concerned with protecting and promoting personal well-being we should (given the epistemic claim earlier) prima facie allow persons to rule themselves, rather than be subject to the rule of others. On this view the example of the person with BIID would be especially pertinent, for his desire to amputate his leg is not widely held—and so given his privileged access to his unusual desire to do this he would be better placed than other people to decide which course of action would promote his well-being. Moreover, given that he would bear (most of ) the costs of the decision to amputate if this turned out to be the wrong decision for him to make, he would be the person most motivated to determine if this was the correct decision for him to make. Frankfurt’s and Christman’s Analyses of Autonomy Yet despite the widespread agreement that personal, rather than Kantian, autonomy should be the focus of those debates concerning end-of-life issues

100

James Stacey Taylor

that are grounded on a concern with autonomy, there is little agreement over what conditions must be met for a person to be autonomous. One of the early accounts of autonomy that address the question of what conditions must be met for a person’s desires to be truly her own, for her to be autonomous with respect to them rather than be alienated from them, was developed by Harry Frankfurt in the early 1970s (Frankfurt 1971: 10).2 For Frankfurt, a person is autonomous with respect to one of her effective first-order desires (e.g., to put in place a “do not resuscitate” [DNR] order) if she volitionally endorses it: if she both wants to have this desire and also wants it to be effective in leading her to act. This approach to analyzing what it is for a person to be autonomous with respect to her effective first-order desires is an ahistorical approach, for it does not impose any restrictions on the origins of the desires that it would accept persons as being autonomous with respect to. This aspect of Frankfurt’s analysis is widely held to be problematic, for it leaves open the possibility that a person could be manipulated into having a particular desire that she would pretheoretically be held to be alienated from. A person might, for example, be hypnotized into both having an effective first-order desire to have a DNR order and volitionally endorsing that desire. On Frankfurt’s view this person would be autonomous with respect to that desire for it met the ahistorical conditions that were both necessary and sufficient for this, but such a desire would be one that would be paradigmatically alien to the person in question. In response to this concern John Christman has developed a historical approach to analyzing personal autonomy. For Christman, an agent P is autonomous with respect to a desire D at time t if and only if i) P did not resist the development of D (prior to t) when attending to this process of development, or P would not have resisted that development had P attended to the process; ii) The lack of resistance to the development of D (prior to t) did not take place (or would not have) under the influence of factors that inhibit self-reflection; iii) The self-reflection involved in condition i is minimally rational and involves no self-deception; And iv) The agent is minimally rational with respect to D at t (where minimal rationality demands that an agent experience no manifest conflicts of desires or beliefs that significantly affect the agent’s behavior and that are not subsumed under some otherwise rational plan of action). (Christman 1993: 288)

Autonomy, Competence, and End of Life 101 Yet while Christman’s analysis of what it is for a person to be personally autonomous with respect to a desire D at t avoids the problem that beset Frankfurt’s earlier historical analysis, it seems to fail to provide either necessary or sufficient conditions for a person to be autonomous with respect to her effective desires. Consider a child who is beaten by his mother to motivate him to learn to play the piano, and who as a result comes to love piano playing for its own sake when an adult. However, despite his love of piano playing the adult repudiated the means by which the desire to play the piano was inculcated into him. Despite this, as an adult this person is still intuitively fully autonomous with respect to this desire; it is legitimately one of the elements of the motivational set that is (at least partially) constitutive of his “self.” Since this is so, Christman’s analysis cannot provide necessary conditions for a person to be autonomous with respect to his effective first-order desires, for this person would be autonomous with respect to his desire to play the piano despite failing to meet Christman’s conditions for this. Moreover, even if all of Christman’s conditions for a desire to be one that a person is autonomous with respect to a desire D are met, a person could still fail to be autonomous with respect to it. Meeting these conditions thus does not suffice for a desire to be one that a person is autonomous with respect to. Consider, for example, a slavish person who wishes to subjugate his will to that of another, such that the other person will determine which desires he has and acts on. If this person is successful in this aim, then the desires that he has as a result of the desire of his controller that he have them will meet conditions (i)–(iv). Yet they will be desires that he is paradigmatically heteronomous—other-directed—with respect to (Taylor 2005: 11). Christman’s account could be defended if one balks at the characterizations of these desires as being those that the agents were (respectively) autonomous with respect to, and heteronomous with respect to. Thus, one might balk at the characterization of the man’s desire to play the piano in the first of these examples as being one that he was autonomous with respect to in adulthood on the grounds that this desire was inculcated in him by another without his consent. One might also balk at the characterization of the slavish person’s desires as being those that he was heteronomous with respect to on the grounds that he wanted his effective desires to have the origins that these had (i.e., being determined by another), and was not concerned about their objects—and so their satisfaction would have value for him. And if one rejects these ascriptions of autonomy and heteronomy, then one will not find these objections to Christman’s analysis of autonomy persuasive. Autonomy, Respect for Autonomy, and Competence Although it might be tempting to view this discussion of Frankfurt’s and Christman’s analyses of autonomy as being only of parochial interest to autonomy

102

James Stacey Taylor

theorists, this would be mistaken. This discussion can be drawn upon to illustrate how the two different approaches to analyzing personal autonomy that were outlined earlier (the “value-creating self ” approach and the “political” approach) focus on different aspects of this conception of autonomy. Illustrating the different foci of these approaches to analyzing autonomy is useful for one concerned with the role that autonomy should play in decisions that persons make at the end of their lives, for it can be used to show both what conditions should be met for a person’s decision in this context to be worthy of respect and why such respect would be owed to a decision that meets these conditions. This discussion can also be used to elucidate the relationship between competence and autonomy—a relationship that both is relevant to questions concerning the status of decisions made at the end of life and has as yet been underexplored in the philosophical literature. What, then, underlies the different characterizations of the man’s desire to play the piano, and the slavish person’s desires whose objects stemmed from the will of another, as being those that these persons either were, or were not, autonomous with respect to? One will hold that the man who desired to play the piano for its own sake after this desire was inculcated into him through beating and the slavish person whose desires merely followed the will of another were autonomous with respect to these desires if one considers the question of whether a person is autonomous with respect to a particular desire to focus on whether the satisfaction of a person’s desire could ex ante be expected to create value for her. On this view the satisfaction of a desire that one was autonomous with respect to would be ex ante expected to create value for one, whereas the satisfaction of a desire that one was alienated from would not. (This is the “value-creating self ” approach to autonomy.) Alternatively, if one holds that these persons are heteronomous with respect to these desires, then one will consider the question of whether a person is autonomous with respect to her desires to be primarily concerned with the question of whether it was she, or another agent, who had decided that she should possess them. (This is the “political” approach to autonomy.) Recognizing that the ascription of autonomy to a person with respect to her desires depends on one’s approach to analyzing autonomy might lead one to despair of uniformly using this concept to determine when a person’s desires concerning her end-of-life treatment should be respected. But such despair would be misplaced, for three reasons. First, it might be the case that one of these analyses of autonomy was either more theoretically justified than the other and hence should be drawn upon in discussions of patients’ end-of-life decisions. Second, it might turn out that one of these approaches is not concerned with autonomy at all, but with a related concept—such as, for example, the question of whether a person’s desire was authentically hers rather than whether she was autonomous with respect to it. Finally, it might be the case

Autonomy, Competence, and End of Life 103 that both approaches to analyzing autonomy would give the same normative results in clinical cases (i.e., their proponents would concur over what should be done) and so their differences would be merely theoretical. Addressing the first two issues would take us too far afield. Fortunately, it is likely that the third point will hold. This can be illustrated by noting that the proponents of both of these approaches to autonomy would offer the same normative responses to the cases of the piano-playing adult and the man with the slavish desires. In the case of the piano-playing adult both approaches to autonomy would hold that he should not have had the desire to play the piano inculcated into him in the way that it was. The value-creation approach would come to this conclusion on the grounds that since he would repudiate the means by which he came to have this desire, the value to him of satisfying it does not outweigh the diminution in his well-being that he endured to acquire it. Thus, although he would wish to satisfy this desire now that he has it and so it would be value-creating for him—and hence a desire that he was autonomous with respect to—he would have preferred not to have had it inculcated into him. The political approach would hold that since he was not self-determining with respect to it, it should not have been inculcated into him. In the case of the person with slavish desires the value-creation approach would say that even though the objects of his desires were determined by another person, since he would secure value through satisfying them his decision to have them and satisfy them should be respected. The political approach would also hold that his decision to satisfy them should be respected, for although it would hold that he was heteronomous with respect to them it would also hold that this heteronomy was autonomously chosen, and so the desires that flow from it should be respected, where this respect was derivative from respecting his autonomous decision to possess them. Both approaches would thus hold that respect for autonomy would prima facie oppose manipulating persons. Respect for autonomy would not require that the decision that a person has been manipulated into making is one that should be respected. Thus, since a person could be manipulated into making a competent decision the mere fact that a person’s decision is competent is not sufficient for it to carry any moral weight. (Othello, for example, made decisions that were competent while he was manipulated by Iago; his ends were those of a competent person, and he was able to identify those courses of action that were sufficiently likely to achieve them.) For a person’s decision to be one that carries moral weight, then, it should be one that she was competent to make and one that she was autonomous with respect to. Thus, a person’s decision would not carry moral weight if she was competent to make it but was not autonomous with respect to it, and nor would it carry moral weight if she was autonomous with respect to it but lacked competence to make it. This last point might at first sight appear puzzling—how could a person be autonomous

104

James Stacey Taylor

with respect to a decision that she was not competent to make? But there is nothing odd about this possibility. To see this note that a person who mines iron pyrites thinking them to be gold would not be competent to identify minerals, but he would be autonomous with respect to his decision if (a) it originated in his motivational set, and not that of another agent (on the self-determination approach) or (b) he ex ante expected that it would create value for him (on the value-creation approach).

Decisions at the End of Life Recognizing that a person must be both competent to make a decision and autonomous with respect to it for it to carry moral weight is important for two reasons. First, it makes clear the relationship that holds between competence and autonomy—a relationship that has in the past been theoretically underdeveloped. Second, if there is an assumption that a person is autonomous with respect to her decisions unless there is evidence to the contrary, then if a person’s competent decision was made under conditions that did not compromise her autonomy, there would be prima facie reason to believe that it carries moral weight, and hence to respect it. The burden of proof would thus lie with those who would thwart her decisions. This has important implications for end-of-life issues. Consider, for example, the case of Re B (adult: refusal of treatment). In this case a woman (referred to as Mrs. B) decided after suffering a hemorrhage of the spinal column that, should she suffer a life-threatening condition or become permanently mentally impaired or unconscious, she would not want any life-sustaining treatment she was receiving to be continued. She subsequently became tetraplegic and was placed on a ventilator, which she repeatedly asked to be switched off—a request that her doctors refused. Mrs. B took her case to court, where it was decided in her favor, after it was determined that she was competent to make this decision, and had made it under conditions that did not compromise her autonomy. Of course, one might challenge the court’s decision here, holding that a person who had decided to end her life could not be normatively competent with respect to this decision since this is not an end that a competent person would pursue. In response to this, one could invoke the underlying rationale outlined earlier for respecting autonomy (whether one endorses the value-creation approach or the political approach) that such a decision could be one that a competent person could endorse if she had come to believe that her life would contain a negative amount of well-being. (Note that given the grave and irrevocable risks of a mistaken decision here the bar for determining that the person in question was not competent to make this decision would be set very low.) Given this, it would seem that a decision made by a patient to have her life ended through the act of another or herself should carry similar moral

Autonomy, Competence, and End of Life 105 weight—that requests for active voluntary euthanasia and physician-assisted suicide should also be prima facie acceded to. (The caveat here is intended to recognize that the persons to whom these requests are made might have the right to refuse them.) But this is often not the case. The distinction here is justified by appeal to an alleged moral difference between letting someone die and actively killing them; it is held that the latter is immoral, and hence requests to perform it (or to aid someone in performing it) should not be respected. However, as noted earlier, the burden of proof in such cases rests with those who would thwart the decisions of persons competent to make them and who did so under conditions that did not compromise their autonomy. Thus, unless it can be shown that there is a clear moral difference between killing and letting die, and that killing persons who request this is morally wrong, the moral weight of end-of-life decisions made by persons under the conditions outlined earlier should support acceding to their requests actively to die. Yet it is not always clear that a person’s decisions concerning her end-of-life treatment that were made while she was competent and under conditions that do not compromise her autonomy should carry moral weight. Consider a person who has made a living will concerning how she is to be treated medically if she ever becomes incompetent. Some years later she does indeed become incompetent. Should she be treated in accord with her living will, or not? One argument in favor of the view that she should be so treated holds that, just as respect for a person’s autonomy requires that prima facie she be treated in accord with her interests, so too should her “surviving interests”—those whose objects concern events that might transpire after she has ceased to be an autonomous agent—carry moral weight. Thus, on this view respect for personal autonomy requires that a person’s preferences over how her surviving interests be treated should also be respected (Buchanan and Brock 1990: 162–163). The view that a person’s “precedent autonomy” should be respected faces a serious objection: it is possible that a person’s interests and preferences might undergo drastic changes, perhaps as a result of a debilitating medical condition, such as Alzheimer’s disease. In such cases not only might the original interests that the person had when she made the advance directive no longer survive, but also following them might be against the interests of the person to whom the advance directive now applies (Dresser and Robertson 1989). The initial response to this objection is to note that to say that a preference exists is not to say that its possessor currently exists. Instead, to say that a preference exists at a time T is merely to say that it was possessed by a being (possibly prior to T) and that its object exists at T. A person who died in 2015 might have a preference about which team would win the 2016 World Series; since this preference would be fulfilled or thwarted by the events in 2016 it would exist at 2016, even if its bearer did not.3 It should then be noted that if a person has two conflicting preferences and a “resolution preference” that favors one of

106

James Stacey Taylor

them respect for her autonomy would require that the resolution preference be given moral weight.4 If a person has become incompetent, it might be the case that she formed a resolution preference adjudicating between her (anticipated) current preferences and those that she held earlier concerning her diminished state. And since in her diminished state she would have formed no conflicting resolution preference, her formerly formed resolution preference would be the only one in existence. (It would exist as it could be thwarted or fulfilled.) Thus, since respect for autonomy requires that a person’s resolution preferences carry prima facie moral weight, absent any reason to the contrary respect for autonomy requires that a person’s advance directives should be followed if they were made when the person was competent and deciding under conditions that did not compromise her autonomy.

Conclusion There is significant debate over both the conditions that must be met for a person to be deemed competent to make certain decisions and those that must be met for her to be autonomous with respect to them. Despite these theoretical difficulties, however, there is widespread agreement among bioethicists that persons’ competent decisions that were made under conditions in which their autonomy was not compromised carry moral weight and prima facie should be respected. This respect extends to the decisions that persons make about the preferences that they wish to be respected in the future, where these are expressed through advance directives.

Acknowledgments I thank John K. Davis for his extremely helpful comments on an earlier version of this chapter.

Related Topics Chapter 7, “Deciding for the Incompetent” by Eric Vogelstein Chapter 8, “Change of Mind: An Issue for Advance Directives” by Paul T. Menzel

Notes 1. The following discussion will focus on the notion of competence as this is used in bioethics. The legal account of competence is similar, although it includes a communicative element. As Alec Buchanan notes, “Legal competence is specific to the task at hand. It requires the mental capacities to reason and deliberate, hold appropriate values and goals, appreciate one’s circumstances, understand information one is given, and communicate a choice” (Buchanan 2004: 415).

Autonomy, Competence, and End of Life 107 2. For criticisms of understanding Frankfurt’s view as a view of autonomy, see Taylor (2009: ch. 3). 3. For arguments that such thwarting would harm the person who is now dead, see Pitcher (1984). For responses to this argument see Taylor (2012: 13–18). 4. This and the argument that follows ahead follow the argument of Davis (2009).

References Beauchamp, T.L. and J.F. Childress (2013) Principles of Biomedical Ethics (7th Edition), New York: Oxford University Press. Buchanan, A. (2004) “Mental Capacity, Legal Competence, and Consent to Treatment,” Journal of the Royal Society of Medicine 97(9): 415–420. Buchanan, A.E. and D.W. Brock (1990) Deciding for Others: The Ethics of Surrogate Decision Making, Cambridge: Cambridge University Press. Christman, J. (1993) “Defending Historical Autonomy: A Reply to Professor Mele,” Canadian Journal of Philosophy 23: 281–289. Davis, J.K. (2009) “Precedent Autonomy, Advance Directives, and End-of-Life Care,” in B. Steinbock (ed.) The Oxford Handbook of Bioethics, New York: Oxford University Press, 349–374. Dresser, R.S. and J.A. Robertson (1989) “Quality of Life and Non-Treatment Decisions for Incompetent Patients: A Critique of the Orthodox Approach,” Law, Medicine and Health Care 17(3): 234–244. Frankfurt, H.G. (1971) “Freedom of the Will and the Concept of a Person,” The Journal of Philosophy 68(1): 5–20. Grisso, T. and P.S. Appelbaum (1998) Assessing Competence to Consent to Treatment: A Guide for Physicians and Other Health Professionals, New York: Oxford University Press. Kant, I. (2002) Groundwork for the Metaphysics of Morals, A.W. Wood (ed. and trans.), New Haven: Yale University Press. Pitcher, G. (1984) “The Misfortunes of the Dead,” American Philosophical Quarterly 21(2): 183–188. Taylor, J.S. (2005) “Introduction,” in J.S. Taylor (ed.) Personal Autonomy: New Essays on Personal Autonomy and Its Role in Contemporary Moral Philosophy, New York: Cambridge University Press, 1–32. Taylor, J.S. (2009) Practical Autonomy and Bioethics, New York: Routledge. Taylor, J.S. (2012) Death, Posthumous Harm, and Bioethics, New York: Routledge.

7

Deciding for the Incompetent Eric Vogelstein

Summary This chapter discusses the moral framework for surrogate decision-making for incompetent medical patients. The chapter focuses on the question of how we can respect the autonomy of those who are no longer competent to make such decisions. The standard counterfactual account of how to respect the autonomy of the incompetent is evaluated, along with accounts that ground respect for autonomy on the patient’s most recent desires and values (regardless of whether the patient still possesses those desires and values) as well as accounts according to which only the current desires and values of the patient matter. Other issues discussed include whether questions of personal identity should affect decisionmaking for incompetent patients, whether a patient’s former desires and values factor in to the patient’s current well-being, and whether an incompetent patient’s “investment interests” should guide decisions on the patient’s behalf.

Introduction Medical patients, under ordinary circumstances, are free to make their own choices about how to be treated (or not to be treated). A cancer patient, for example, is free either to accept or to refuse life-extending chemotherapy, based, for example, on their own view of how to weigh length of life versus quality of life. Or consider the following case: Ozzie is diagnosed with a disease that is easily cured with a blood transfusion, but fatal without a transfusion. Ozzie is a devout Jehovah’s Witness, and thus believes that accepting blood transfusions is contrary to God’s commands; he therefore refuses the transfusion, knowing full well that he will die. The widely accepted view is that Ozzie ought to be allowed to refuse the transfusion, regardless of whether others view that decision as wrongheaded

Deciding for the Incompetent 109 or seriously harmful to Ozzie; the law in the United States and most other developed nations follows suit, giving Ozzie the legal right to refuse treatment. When patients are unable to make their own choices, however (due, e.g., to dementia, delirium, unconsciousness, or mental illness), someone else must make medical decisions on their behalf. In those cases, various questions arise about how such decisions ought to be made. In particular, we want to know who should make those decisions—who should be the surrogate decision-maker— as well as how surrogates should go about making the relevant decisions—that is, on which criteria they should base their decisions. Standard practice has it that patients, if they wish, may designate a particular person to be their surrogate should they become unable to make decisions (such a person is sometimes called a “proxy” or “agent”). The official legal designation or document is usually called a medical power of attorney or a power of attorney for health care. If such a designation has not been made, then one’s “next of kin” becomes the surrogate; individual states have particular orderings for who is to be the surrogate—the “top three” are usually (1) spouse, (2) parents, and (3) adult children. But even if it is clear who is or ought to be the surrogate decision-maker, there is still the question of how that person ought to go about making the often momentous, life-and-death decisions he or she will be faced with. Presumably, the surrogate should not be allowed to make any decision he or she wants—but how should the surrogate decide? For example, should the surrogate’s decision be based purely on what’s in the patient’s best interests—that is, what best promotes the patient’s welfare or well-being? That might mean, for example, giving Ozzie the transfusion if he is unable to make his own decision—after all, it is plausible that continuing to live would be in Ozzie’s best interests. Or should the decision be based on the patient’s desires and values, even when they conflict with the patient’s welfare? Acting in accordance with Ozzie’s values would mean not giving him the transfusion, and allowing him to die. According to the standard view, surrogates should try to glean what the patient would want if she were competent, and act accordingly (e.g., via instructions for treatment that the patient provided prior to becoming incompetent—either verbally or documented in a “living will”—or based on an inference from what the surrogate knows of the person’s values, the so-called substituted judgment principle). But is that standard view correct, or does it suffer from philosophical or ethical problems? If the standard view is problematic, then what is the correct principle for surrogate decision-making? These are some of the questions that will be addressed in this chapter. A good place to begin answering these questions is with the moral principle of respect for autonomy and its application in the medical domain. Respect for the patient’s autonomy, or self-determination, is a guiding principle in medical ethics. According to widely accepted views, patients have a strong

110

Eric Vogelstein

prima facie right to make their own medical decisions. That right, however, is conditional upon the patient’s decision-making competence (roughly, having a certain degree of understanding and rationality), in the following way: a decision made while incompetent is considered to be non-autonomous, and thus the principle of respect for autonomy does not require us to honor such a choice. At the same time, it is still considered possible to respect or violate the autonomy of some incompetent medical patients—specifically, those who have been competent in the past. (Those who have never been competent, on the other hand, have no autonomy to either respect or violate—in that case, the widely accepted view is that such patients should be treated based on what best promotes their well-being.) There is some controversy, however, about how we respect the autonomy of the incompetent, and thus how, on the prevailing view, surrogate decision-makers ought to decide for the incompetent. There is debate, that is, about the theory that correctly explains how to respect incompetent patients’ autonomy. And theory matters, because different theories will license different actions—the decision that is made for a particular incompetent patient may hinge on which theoretical framework is applied to the case. At no time is this more important than at the end of life, when decisions about forgoing and discontinuing life-sustaining treatment are prevalent. In this chapter, I survey and evaluate various theoretical approaches to deciding for incompetent medical patients. I shall not address the issue of the moral authority of surrogate decision-makers—that is, the question of who may or should make decisions on behalf of the incompetent and on what moral ground their decision-making authority rests (for discussion of such issues see, e.g., Buchanan and Brock 1989; Wrigley 2015). Instead, I shall focus on the proper criterion for surrogate decision-making—that is, the basis on which ethically sound decisions for incompetent patients should be made, by whoever has the moral authority to do so. First, I will evaluate the standard criterion for surrogate decision-making: the counterfactual account. I shall survey various objections to a counterfactual criterion for surrogate decision-making, and ultimately suggest that a counterfactual account will not be fully adequate to capture what we want from surrogate decision-makers. Then, I consider the merits and drawbacks of alternative theories. I will suggest that on the most plausible theory, respecting the autonomy of incompetent patients requires simply that we honor the desires and values that the patient currently has, not the desires and values that the patient used to have, but no longer does. As we shall see, this view has implications for certain cases of severe dementia and vegetative states, and more generally for any incompetent patients who have lost the desires and values that they had when competent. Finally, I address the extent to which promotion of a person’s well-being, as opposed to respect for autonomy, should drive decisions for the incompetent.

Deciding for the Incompetent 111

The Counterfactual View According to a fairly standard view in medical ethics and the law, when a patient becomes incompetent to make her own medical decisions, such decisions should be based on what the patient would choose in those circumstances were she competent to decide (Pearlman 2008; Jonsen et al. 2010; Beauchamp and Childress 2013). This counterfactual criterion is thought to capture the idea that a patient’s autonomy is morally relevant and applicable to how she should be treated even after she is no longer competent to make medical choices. In such cases, according to this view, we respect a patient’s autonomy by honoring her counterfactual wishes—that is, what her wishes for treatment would be if she were competent to make the relevant choice.1 For example, suppose that Ozzie is unconscious when brought into the hospital, and is not expected to regain consciousness unless his illness is cured via the blood transfusion. Ozzie’s surrogate decision-maker must now decide whether Ozzie is to be transfused. On the counterfactual view, Ozzie’s surrogate must base that decision on what Ozzie would choose if he were competent to make the decision. And Ozzie, it seems clear, would not accept the transfusion if he were competent, and thus his surrogate should not authorize the transfusion (doing so would violate Ozzie’s autonomy). Most other patients, however, would accept the transfusion, if competent, and thus in those cases surrogates should authorize the transfusion (doing so would be the way to respect the autonomy of those patients). The counterfactual view has roots in United States case law. For example, the Supreme Court of New Jersey has ruled that “the goal of decision-making for incompetent patients should be to determine and effectuate, insofar as possible, the decision that the patient would have made if competent” (Conroy 1985: 1229). The counterfactual criterion was also invoked by that court in the seminal case of Karen Quinlan (Quinlan 1976). Similarly, the Supreme Court of Massachusetts has stated, A decision whether to withhold medical treatment from a mentally incompetent person should conform as closely as possible to the decision which would be made by the incompetent person, if that person were competent, but taking into account his present and future incompetency as one of the factors which would necessarily enter into the decision-making process. (Saikewicz 1977: 728) A trial court ruled, and the Court of Appeals of New York affirmed, that “[t]here is no doubt that this court is obligated to ascertain, if possible, the choice which [the patient], now incompetent, would make if he became competent for a moment and consciously faced his afflictions” (Eichner 1981: 210). Notably,

112

Eric Vogelstein

however, some judicial rulings conflate the following distinct criteria: (1) what the patient would want, if she were competent; (2) what the patient wanted, prior to becoming incompetent; and (3) what the patient currently wants (e.g., Westchester County Medical Center 1988; Cruzan 1990). Here, we will be sure to keep separate those similar yet importantly different concepts. The counterfactual view has been attacked on various fronts. For example, it has been argued that the view gives us the wrong results in cases in which the patient’s wishes are conditioned on being incompetent (Wierenga 1983; Barnbaum 1999; Phillips and Wendler 2015). Suppose that one’s wishes are as follows: if and only if one becomes permanently unconscious, one is to have no lifesaving treatment for serious illness. If one then becomes permanently unconscious, and thus permanently incompetent, applying the counterfactual criterion gives us the wrong results: if one were competent, and thus conscious, one would not refuse lifesaving treatment because one would not meet one’s own criterion for that refusal: permanent unconsciousness. But, we think, the patient’s desire not to have lifesaving treatment should be honored, as doing so is the way to respect her autonomy. Therefore, the counterfactual criterion, at least as traditionally conceived, cannot be correct. One response to this objection is to specify the relevant counterfactual condition not as being competent per se, but rather as being momentarily competent, keeping fixed all other aspects of one’s illness (Wierenga 1983). The question would then be: if one were to be competent for an instant, knowing that one will immediately return to the condition one indeed has, what sort of treatment would one want? This suggestion, however, faces a problem analogous to the previous one: it might be that one’s desire for some treatment or lack thereof is conditioned on the fact that one has not momentarily become competent—if one were to become competent, even for an instant, one would want something else. Again, in such cases the counterfactual view gives us the wrong results: it implies that the way to respect the patient’s autonomy is to thwart his wishes (Wierenga 1983; Nagasawa 2008; Wrigley 2011). A different attempt to save the counterfactual criterion appeals to a particular specification of the consequent of the relevant counterfactual: “one would choose x.” That phrase is ambiguous, because it does not specify whether the target of one’s counterfactual choice—that is, the person for whom one would choose—is oneself in the counterfactual world or in the actual world; and the problem for the counterfactual view arises only in the former case. So instead of imagining what the patient, if competent, would choose for his competent self in the counterfactual world—a test that gives us the wrong results in cases in which the patient would choose something other than what he actually wants—we should imagine what the patient would choose for himself as he is in the actual world (Barnbaum 1999). That way, the patient’s choice will not be about a situation in which he does not actually find himself—his choice is

Deciding for the Incompetent 113 now about himself as he actually is, incompetence and all, and thus, unlike the original “inter-world” interpretation of the relevant counterfactual, the new “intra-world” interpretation gives us the right results in the sort of case at issue. A problem with the intra-world interpretation, however, is based on the suggestion that “oneself ” in the counterfactual world is not actually oneself; rather, it is one’s counterpart (Lewis 1986). Likewise, a person in the actual world is herself a counterpart of her own counterpart—that is, A, a person in the actual world, is the counterpart of B, A’s counterpart in the counterfactual world. Thus, it may be argued, when we imagine what an incompetent patient would, if competent, choose for herself in the actual world, we are imagining what one’s counterpart in the counterfactual world chooses not for herself, but rather for someone who is in many ways just like her, but is not in fact her— we are imagining her choosing for her counterpart. But a person’s choices for her counterpart might not, and do not normally, reflect the sort of concern a person has for herself; rather, one’s choice for one’s counterpart might diverge substantially from one’s wishes for oneself. Thus, one’s counterpart’s wishes for oneself might not reflect one’s actual wishes for oneself, in which case the counterfactual criterion will yield the intuitively wrong results—it may license acting contrary to a patient’s wishes on the basis of the distinct wishes of her counterpart (Wrigley 2011). A further problem with the counterfactual view in general (be it intra-world or inter-world) involves cases in which a patient would change his mind if competent. Imagine an incompetent patient whose wishes for treatment were made clear prior to becoming incompetent. That in itself does not preclude the possibility that if the patient were to become competent again, his wishes would change, and thus he would make a choice that is not consistent with his wishes prior to incompetence. In such cases, the counterfactual criterion (on any interpretation) gives us the wrong results: it tells us, implausibly, that the way to respect a person’s autonomy is to ignore the person’s actual desires in favor of desires that are merely hypothetical (Wierenga 1983). Consider, for example, the following case: Wendy, due to severe anoxic brain damage, is in a permanent vegetative state. Prior to becoming incompetent, Wendy was adamant that she wanted her life preserved should she ever be in such a state—her religion, she says, requires it. Wendy, however, has a rare neurological quirk such that if she were to miraculously become competent, she would immediately change her mind—she would disavow her religion and choose (competently) to have life support removed from her vegetative self. In this case, it seems, we respect Wendy’s autonomy not by honoring her counterfactual choice, but her actual wishes—after all, Wendy has not changed her

114

Eric Vogelstein

mind, and the mere fact that she would do so were she to become competent seems irrelevant. In response, one might suggest limiting the counterfactual scenario to situations in which one maintains one’s prior values and desires—thus the relevant criterion is not what one would choose, were one competent, but rather what one would choose, were one competent and were one to maintain one’s prior values and desires. We could even specify the counterfactual scenario as being indexed to the last time at which one was competent, thus ensuring that one’s values in the counterfactual scenario correspond to one’s actual prior values. In other words, the counterfactual scenario, instead of being what the patient would choose were one to be competent and to maintain one’s prior values, could be described as what the patient would have chosen at the last point in time at which the patient was competent, with regard to the situation in which one actually finds oneself (Wrigley 2011). In that case, in making decisions for incompetent patients we would ask, when the patient was last competent, if she had made a choice about this situation (the patient’s actual condition), what would she have chosen for her future self ? This would give us the right results in Wendy’s case: if Wendy, when last competent, had made a choice about her current situation—being in a permanent vegetative state—she would have chosen to have life support continued. Thus, Wendy’s actual values, not the values she would have in some other possible world, determine how to respect her autonomy. Furthermore, this revised formulation of the counterfactual view has the virtue of avoiding the problem generated by the intra-world interpretation, because it requires only an inter-world interpretation: we’re now imagining what Wendy would have chosen for her actual future self, not what she would have chosen for her counterpart. This revision, however, is not sufficient, due to the fact that one might decide in a way that thwarts one’s own values and desires (Johansson and Broström 2011; Phillips and Wendler 2015). One’s choice in the moment might reflect irrationality or weakness of will, and thus the choice one would have made might not be the right way to serve one’s values and desires. For example, we might know that someone usually makes foolish decisions he later comes to regret and that contradict his own better judgment, or that someone is normally led astray by certain cognitive biases or faulty reasoning. In such cases, we might conclude that if the person were competent, he would make a poor choice with respect to his own underlying values and desires—that is, he would choose something that contravenes his ultimate ends. But we do not respect that person’s autonomy by acceding to the mistaken judgment he would likely make; rather, we respect his autonomy by honoring the underlying values and ultimate aims he has, even if he would, if competent, unwittingly make a decision that undermines those values and aims. Suppose, for example, that Wendy, when competent, was inclined to opt for ineffective homeopathic treatment whenever she had an illness—not because

Deciding for the Incompetent 115 she had a philosophical objection to modern medicine, but simply because she believed, falsely, that homeopathic remedies are more effective than medication. Now, suppose that Wendy, in her vegetative state, has contracted pneumonia, and if she is not given antibiotics immediately, she will die. In this case, it is reasonable to believe that Wendy, when competent, would have chosen homeopathic care instead of antibiotics if asked to decide about her current condition. Recall, however, that what Wendy ultimately wanted was to live as long as possible; thus the choice she would have made would not be the way to achieve her own ultimate goal. It seems that in such a case, we respect Wendy’s autonomy by honoring her ultimate aim to live as long as possible—and thus by treating her with antibiotics—rather than by doing what she would have chosen, which would have been based on a mistaken idea of what would in fact achieve her ultimate aim. Of course, we could amend the counterfactual account in response to this problem, such that the counterfactual condition includes that one chooses in concert with one’s ultimate values and desires. The new criterion would then be that we should decide for the incompetent based on what that person, when last competent, would have chosen for herself in her actual, current situation, were she to have chosen in concert with her ultimate values and desires. But at this point, we are in danger of losing the distinctiveness of the counterfactual view. In fact, this latest revision is equivalent to a view that simply invokes the patient’s prior values and desires without reference to any counterfactual scenario. The account could be put simply as follows: in order to respect the autonomy of an incompetent patient, we must treat that patient in concert with the values and desires the person had prior to becoming incompetent. Let us call this the prior values view. (Beauchamp and Childress 2013 call this the “pure autonomy” view; Phillips and Wendler 2015 term a similar view the “endorsed life” approach.) Thus, the problem with the counterfactual view, amended as it ought to be, is that the counterfactual element is thereby rendered superfluous, and therefore the view itself becomes overly complicated and provides no insight beyond the prior values view itself. The prior values view, then, would be the right theory—we respect the autonomy of the incompetent simply by acting in the way that best reflects and coheres with the values and desires the patient had when she was last competent.

The Prior Values View and Precedent Autonomy The prior values view is prima facie quite plausible. It captures the idea that we respect the autonomy of the incompetent by acting in the way that best coheres with what the person most wanted, without invoking the superfluous machinery of counterfactual scenarios. However, the prior values view is not uncontroversial. One reason is that the view is committed to the validity of precedent

116

Eric Vogelstein

autonomy, the idea that we can respect a person’s autonomy by honoring her former values—that is, the values one used to have but no longer does. This is relevant because it is possible that a patient loses the values and desires that he had prior to becoming incompetent. Indeed, some medical conditions that cause incompetence plausibly cause the annihilation of the values and desires one had prior to such conditions, such as severe dementia and vegetative state; on the prior values view, those former values and desires nevertheless determine how to respect the patient’s autonomy. The acceptance of precedent autonomy is controversial. It stands to reason that in order to respect a person’s autonomy, one must honor only the desires and values that the person has, not the desires and values that the person used to have. If I used to be a Red Sox fan and thus used to want to attend their games, but now I don’t care about the Red Sox, my autonomy is not respected by giving me tickets to a Red Sox game—after all, I no longer desire to attend the game, and the fact that I did have such a desire seems irrelevant to how to respect my autonomy now. Thus, there is a strong prima facie case to be made against precedent autonomy. It can be argued, however, that appearances mislead us here. The argument begins with the observation that ordinary, paradigmatic cases of someone losing a desire or value are cases in which that desire or value is replaced (these are simply cases of a person changing her mind—i.e., desiring or valuing something else instead). If I don’t want to go the Red Sox game, for example, I will want to do something else instead—for example, work or watch TV. Furthermore, the plausible way to understand the doctrine of precedent autonomy is not as the idea that any former desire or value contributes to a person’s autonomy, but rather as the notion that the most recent desire or value a person has, within a group of desires or values that would conflict with one another, contributes to a person’s autonomy, regardless of whether one currently possesses that desire or value (Vogelstein 2016). Thus, when we ask ourselves whether former desires or values contribute to autonomy—that is, whether precedent autonomy is valid—we implicitly imagine cases in which one’s previous desire or value is replaced with a new desire or value that conflicts with it, and thus, because it is more recent, overrides it. It is no wonder, therefore, that we are inclined to accept prima facie that former desires and values cannot contribute to a person’s autonomy: in the usual sort of case, they do not. But it is an overgeneralization to then conclude that precedent autonomy is invalid when one’s most recent desires and values have not been replaced, as plausibly occurs in the sort of medical conditions at issue—for example, certain cases of severe dementia and vegetative state. Thus, on this line, the prima facie case against precedent autonomy is flawed (Davis 2002, 2007). But the prima facie case is only one case against precedent autonomy. Another argument relies on our intuitive reaction to cases in which one’s prior

Deciding for the Incompetent 117 desires or values have not been replaced. For example, suppose that Jon used to want exactly one gray hair, and that he wanted this very badly—he viewed having a single gray hair as a profound mark of distinction. Now, Jon has no desires at all about his gray hairs; perhaps Jon no longer understands what gray hair is. Suppose that Rachel has the ability to give Jon a single gray hair. Would Rachel be respecting Jon’s autonomy by exercising that ability now? According to the argument under consideration, the intuitive answer is “no”— Jon’s autonomy is not respected by Rachel giving Jon the gray hair he used to want, and it is not thwarted by her not doing so. But Jon’s desire for a single gray hair is not only former but also most recent—it has not been replaced by any desire that conflicts with it. Thus, it seems, we have a counterexample to the doctrine of precedent autonomy: that view would have it that former, most recent desires and values always contribute to a person’s autonomy, and thus that they command respect under the principle of respect for autonomy, but here we have a case in which a former, most recent desire fails to do just that (Vogelstein 2016). This argument will of course be controversial; but the rejection of precedent autonomy is a view worth taking seriously. And if precedent autonomy is indeed invalid—that is, if a desire or value must currently be possessed by an agent in order to command respect under the principle of respect for autonomy—then instead of the prior values view, it seems, we ought to adopt the current values view: that we respect the autonomy of the incompetent (and everyone else) by acting in concert with their current desires and values. This would have serious implications for how we ought to treat incompetent medical patients whose conditions appear to have destroyed their prior desires and values. For example, suppose that the kind of neurological damage that has resulted in Wendy’s vegetative condition has also destroyed her religious values and her desire to be kept alive at all costs. In that case, on the current values view, honoring Wendy’s former desire to be kept alive will not be a way of respecting her autonomy, and thus the strong moral imperative to keep Wendy alive may no longer exist (which will depend on other morally relevant factors present in the case, such as Wendy’s best interests). It might be objected that the current values view is absurd on its face, since the current desires and values of the incompetent are precisely what do not command respect—the patient, after all, is currently incompetent. Of course, incompetently held desires and values do not command respect under the principle of respect for autonomy; but patients who are incompetent to make medical decisions might nevertheless have desires and values that are competently held even if they are unable to decide based upon them or to become consciously aware of them. In explaining how this is so, the distinction between dispositional and occurrent mental states or attitudes is vital. Roughly, a dispositional state or attitude is one that is unconscious—that is, not present in

118

Eric Vogelstein

awareness—or is not currently operative—that is, it is not currently causing its characteristic mental or physical effects. An occurrent state or attitude, on the other hand, is a state or attitude of which one is aware or one that is causing its characteristic effects. An occurrent desire, for instance, is one that is motivating an agent, or causing the agent’s attention to be directed in certain ways, or causing certain affective responses to stimuli (e.g., pleasure). A dispositional desire is one that lacks these features, but is present nonetheless (e.g., during deep sleep). Thus, incompetent patients might have a slew of dispositional desires and values—ones that are competently held (because they were acquired prior to becoming incompetent)—and on the current values view, those desires and values would determine how to respect that patient’s autonomy. It should also be noted that in a broad and important range of cases, the current values view gives us the same results as the prior values view. That is because common conditions that make a person incompetent, such as unconsciousness, delirium, and moderate dementia, do not destroy a person’s values, but may simply render those values dispositional. Therefore, in such cases, what the person valued prior to becoming incompetent will also be what the patient currently values.

Personal Identity In this section, I highlight a type of case in which it might be problematic to use an incompetent person’s former desires and values in making treatment decisions for that person even if the prior values view is true. The worry at issue is that an incompetent patient whose mental states and attitudes have changed dramatically compared with when he was last competent, such as might occur with cases of severe dementia, is no longer the same person as he was prior to incompetence. Suppose, for example, that Jane was an accomplished philosophy professor, was deeply interested in politics and committed to social justice, was a feminist activist and writer, enjoyed classical music and fine dining, loved her husband, children, and grandchildren, was an ardent Red Sox fan, practiced Judaism, and deeply valued her independence, her intellectual ability, her dignity, and her legacy, both personally and professionally. Now, Jane has severe dementia. She remembers almost nothing of her prior life and no longer understands most of the things she once valued and to which she had devoted so much of her energy, including her career and life’s work, politics, feminism, social justice, her family, and her religious traditions. She no longer enjoys classical music and fine food, and does not know who the Red Sox are. Furthermore, she appears not to understand the concepts of dignity and legacy, and thus seems to have lost the desires related to those aspects of her life. Still, Jane’s life is not a bad one, overall—Jane is able to enjoy simple pleasures,

Deciding for the Incompetent 119 such as watching television and interacting with her caretakers and family (even if she doesn’t know who they are), and she is not emotionally distraught or in physical pain. In other words, Jane’s life is “worth living”—that is, it is in Jane’s best interests to be alive rather than dead, taking into account her current quality of life. Now, suppose that Jane’s prior wishes were that if she were ever to be in a state of severe dementia, she would want no lifesaving treatment of any kind, should she develop a life-threatening illness. Her desire was based on the values to which she was so strongly committed, most importantly her dignity (which she perceived as being incompatible with severe dementia) and her legacy. If Jane develops an easily curable but otherwise fatal disease, should we honor her desire not to be treated and thus allow her to die, even if Jane’s life is otherwise worth living? We know that on the current values view, honoring Jane’s wish for no treatment is not a way of respecting her autonomy, because (it seems clear) she no longer has such a wish. But what about Jane’s autonomy on the prior values view? It might seem that on that view, Jane’s autonomy is respected by honoring her prior wishes for no treatment. But, it might be argued, because Jane is so different from the person who did not want treatment, she is now, quite literally, a different person. And since one person’s values and desires do not determine how to respect the autonomy of a different person, the prior values and desires of Jane would fail to apply to the “Jane” who currently exists. On this view, honoring her desire for no treatment is not a way of respecting Jane’s autonomy, because for all intents and purposes, Jane no longer exists. This view is controversial for several reasons. First, it relies on a particular theory of personal identity: the so-called psychological theory, according to which the persistence conditions of a person are determined by one’s psychological features (as opposed, e.g., to one’s physical or biological features), such as psychological continuity and similarity (Locke 1690/1979; Parfit 1971). Some philosophers reject this theory of personal identity (cf. Olson 1997). Furthermore, even on a psychological theory it is an open question just when a person’s psychology has been sufficiently altered such that a new person exists. And it might be that the criterion is so strict—that is, that the degree of psychological difference and discontinuity that must be achieved in order for someone to count as a different person is so great—that any psychological change great enough to change one’s identity would render one a nonperson. In that case, it could not be claimed that one person’s prior wishes might be applied to a different person, but only that a person’s wishes might be applied to a being who is not a person—and the latter situation, it seems, is compatible with the moral permissibility of honoring those wishes, at least in some cases, given that the rights and interests of nonpersons carry significantly less weight than those of persons (Buchanan 1988; Buchanan and Brock 1989). And there are reasons for thinking that the criterion is indeed that strict. In

120

Eric Vogelstein

particular, if all it takes to render one a different person are sufficiently significant changes within one’s psychological makeup, then we ought to drastically alter our view of a slew of social and institutional practices that we think are entirely appropriate—for instance, practices that involve the maintenance, through such changes, of property rights, debts, and other obligations (both those one owes and those one is owed), as well as criminal and civil liability— practices that would be morally illegitimate given that the kinds of psychological changes at issue render one an entirely different person (Buchanan 1988; Buchanan and Brock 1989; Olson 1997).

Well-Being, Former Values, and Investment Interests Thus far, we have focused on the question of how to respect incompetent patients’ autonomy, even when they lack autonomous decision-making capacity at that time. Most broadly, the key to respecting the autonomy of the incompetent is to honor their desires and values, whether their most recent desires values (including former desires and values), or simply their current desires and values (including dispositional desires and values). But regardless of what the correct criterion is, in order to put that criterion into practice one must have evidence of what the patient’s desires and values were, or are, as it applies to the treatment decisions at issue. Sometimes, that evidence comes in the form of specific instructions about what to do that the patient issued prior to becoming incompetent; sometimes it comes in the form of general values the patient previously espoused that can be applied to the case at hand. In many cases, however, that sort of evidence is lacking—there is no particular evidence of how the patient’s desires and values would cohere with a given treatment decision. In that case, the standard view is to act based on the patient’s well-being, taking into account those aspects of well-being that would apply to most people, such as pleasure and the absence of pain, mobility and independence, cognitive and communicative ability, and so forth. The theoretical basis for the focus on patient well-being in such instances will differ depending on whether the prior values or the current values view is true. On the prior values view, promoting the well-being of patients when we do not know how their unique values and desires would translate into a treatment decision can be thought of as a way of respecting the patient’s autonomy, given what we know, since it can be presumed that the patient indeed cared about her own welfare (even if that desire has been lost). On the current values view, promoting the patient’s well-being might be a way of respecting the patient’s autonomy, given that she indeed still cares about her well-being (even if only dispositionally); but if the patient’s desire to promote her well-being no longer exists, even dispositionally, then promotion of the patient’s wellbeing can be viewed as honoring the “next best” moral value after respect for

Deciding for the Incompetent 121 autonomy. In either case, on the current values view, we act ethically by promoting the patient’s well-being in such instances. A controversy arises, however, regarding whether an incompetent patient’s former values can determine what is in her best interests—that is, what best promotes her well-being. According to some ethicists, even if a patient’s medical condition has destroyed the values she once had, what is best for such a person might be to act in accordance with her prior values. These ethicists distinguish between so-called critical interests and experiential interests, where critical interests are determined by what satisfies one’s values (including former values), while experiential interests are determined by positive and negative experiences—for example, pain and pleasure (Dworkin 1993; Davis 2007). Thus, on this view, even if a patient has no experiential interests at all (e.g., because she is permanently unconscious), things can still be good or bad for her if they cohere or fail to cohere with her former values. This is not the same as accepting precedent autonomy; rather, it is to accept a kind of precedent well-being. That said, if critical interests are weighted more heavily than experiential interests, morally speaking, then the issue of precedent autonomy, as well as debate over the current values view and the prior values view, may be largely irrelevant, at least in terms of its practical upshot: in either case, we ought morally to treat incompetent patients in concert with their most recent desires and values, including desires and values they no longer possess. This is particularly relevant to cases of severe dementia like Jane’s, in which (1) the patient when competent wanted not to live in such a condition, (2) it appears that the patient no longer possesses such a desire, and (3) the patient presently maintains a level of well-being sufficient to make her life worth living—that is, better than no life at all. If critical interests indeed exist, then satisfying Jane’s prior desire not to live in such a state would contribute to her well-being; and if critical interests outweigh experiential interests, then doing so would best promote her well-being. Such a view would allow us to justify withholding lifesaving treatment from Jane even if doing so is not a way of respecting her autonomy. It will be controversial, however, whether a patient’s critical interests indeed outweigh her experiential interests in such cases. Moreover, it is highly controversial whether critical interests exist in the first place; notably, many of the most prominent theories of well-being flatly rule them out. For example, critical interests are incompatible with hedonism—the view that well-being consists in the balance of pleasure over pain, broadly speaking—as well as theories according to which well-being consists in overall life satisfaction. It is also incompatible with those desire-satisfaction theories that require a desire to be satisfied while it exists in order for its satisfaction to contribute to one’s well-being. Still, one might think that the existence of critical interests is so intuitively plausible that it serves as a counterexample to such theories.

122

Eric Vogelstein

One strategy for those who wish to defend the existence of critical interests is to rely on the premise that people can be harmed—that is, their well-being can be decreased (compared with what it otherwise would have been)—by acting contrary to what a person formerly valued. This point is often made in the context of posthumous harm, but applies to any case in which a person’s values have been destroyed—for example, by brain damage. Examples of the sort of harm at issue include damaging someone’s reputation or legacy, and thwarting a person’s projects or goals (Feinberg 1984). These are cases in which a person valued things that extend beyond the life of those values themselves, or even the life of the person. The claim is that in such cases, it is intuitive that preventing the realization of the objects of those values harms the person whose values they were. If that is correct, then such values would ground critical interests. In response, it can be claimed that we may explain away the intuition that supports the sort of harm in question by noting that the kind of act at issue might involve wronging a person without harming that person. For example, it certainly seems reasonable that one wrongs a deceased person by smearing her reputation, or by preventing her life’s work from coming to fruition—but wronging does not entail harming. According to the response under consideration, the fact that we wrong such a person, not that we harm them, is what we’re tracking when we intuit (mistakenly) that such a person is harmed. Indeed, it is easy to conflate wronging and harming, because wronging a person so often does involve harming him. (One might claim that if an incompetent patient can indeed be wronged by thwarting his former values, then it doesn’t matter whether that is due to harm or something else—in either case, we should honor those values. But in fact it does matter, because of the possibility that thwarting the incompetent patient’s former values might promote that person’s well-being. The issue, then, becomes one of priority—i.e., which consideration is more morally important. I address this in more detail ahead.) Of course, one way of wronging a person without harming him is to violate his autonomy—and some might claim that that is the sort of wrong done to the dead by thwarting the realization of that which they formerly valued. But recall that the import of the possibility of critical interests, as it pertains to deciding for the incompetent, is that critical interests would provide a justification for honoring the former values of the incompetent even if we reject precedent autonomy. Thus, the opponent of critical interests, in this context, rejects the view that an incompetent patient’s former values determine how we ought to treat her, whether via precedent autonomy or critical interests. Such a person therefore cannot appeal to respect for autonomy in explaining how it is that we can wrong the dead without harming them. One way of understanding how we might wrong the dead or the incompetent without either violating their autonomy or harming them is to countenance what Davis (2006) calls investment interests. Ordinarily, the term “interests” in ethics refers to factors that contribute to a person’s well-being. To say that

Deciding for the Incompetent 123 one’s best interests are served by some action, for example, is to say that one’s well-being is most strongly promoted by such action. Davis suggests, however, that a person’s former values can determine a different sort of morally relevant interest: investment interests. These are interests the honoring of which might not contribute to a person’s well-being, but on the basis of which we have prima facie moral obligations to the interest-holder, and are opposed to welfare interests that do contribute to one’s well-being. The idea is that when a person invests in some project or commitment—that is, expends effort in service of furthering something one values—that generates an interest, on the part of that person, in that thing. This implies prima facie moral obligations on the part of others to support, or at least not to hinder, the realization or maintenance of the aim of that project or commitment. Investment interests thus represent a kind of moral value that falls under neither the principle of respect for autonomy nor the duty to promote well-being. One way of understanding the moral importance of honoring investment interests is to view doing so as a way of showing respect for a person (even if it is not by respecting that person’s autonomy). Simply, we respect a person by honoring her valued projects and commitments. And if, as Davis suggests, investment interests can survive the loss of the values that generated them, then such interests stand to explain the way in which we might wrong people by failing to honor their former values even if we neither harm them nor violate their autonomy. This sort of picture seems reasonable, but questions will arise about cases in which honoring a person’s investment interests conflicts with other morally important considerations that apply to that person. In such cases, we will want to know just how much moral weight investment interests carry. For instance, we should wonder whether the moral importance of honoring a person’s former commitments—for example, the commitment that one not be remembered as someone who lacked fully intact cognitive faculties—outweighs the importance of promoting that person’s well-being—for example, one’s welfare interest in staying alive in a happy yet demented state. Indeed, it would be highly controversial to claim that the level of harm caused by such a person’s death would be outweighed by the respect we would show that person by honoring her former desire to be remembered as an intelligent person, given that respecting a person in such a way is not a way of respecting her autonomy. Therefore, relying on investment interests, it seems, will not be a particularly convincing way of arguing that honoring someone’s former values, in the kind of case at issue, is morally required.

Conclusion The widely accepted, primary ethical basis for treating medical patients, incompetent patients included, is respect for the patient’s autonomy. As we have seen, there are various ways of understanding that view’s theoretical

124

Eric Vogelstein

foundation—that is, the fundamental explanation of how we respect the autonomy of incompetent patients. I have suggested that counterfactual accounts are not among the most promising theories. I have also proposed that the current values view ought to be viewed as a plausible alternative to the prior values view, and while the implications of those two views converge in most cases, they offer differing instructions in certain cases of severe dementia and severe neurologic damage (e.g., vegetative state). And while some philosophers seek to ground the welfare of patients in their prior values, it is important to recognize that such views rest on highly controversial theoretical assumptions. Therefore, it behooves us to focus on how to respect the autonomy of incompetent patients—in that regard, the debate between the current and prior values views is crucial.

Related Topics Chapter 6, “Autonomy, Competence, and End of Life” by James Stacey Taylor Chapter 8, “Change of Mind: An Issue for Advance Directives” by Paul T. Menzel Chapter 12, “Grief and End-of-Life Surrogate Decision-Making” by Michael Cholbi Chapter 13, “Solidarity Near the End of Life: The Promise of Relational Decision-Making in the Care of the Dying” by Bruce Jennings

Note 1. According to the standard interpretation of counterfactual claims, a counterfactual statement “if it were the case that p, it would be the case that q” is true just in case q is true in the possible world(s) in which p is true which are closest to the actual world (Lewis 1986).

References Barnbaum, D. (1999) “Interpreting Surrogate Consent using Counterfactuals,” Journal of Applied Philosophy 16: 167–172. Beauchamp, T.L. and J.F. Childress (2013) Principles of Biomedical Ethics (7th Edition), New York: Oxford University Press. Buchanan, A. (1988) “Advance Directives and the Personal Identity Problem,” Philosophy and Public Affairs 17: 277–302. Buchanan, A.E. and D.W. Brock (1989) Deciding for Others: The Ethics of Surrogate Decision-Making, Cambridge: Cambridge University Press. Cruzan v. Director, Missouri Department of Health, 497 U.S. 261 (1990). Davis, J.K. (2002) “The Concept of Precedent Autonomy,” Bioethics 16: 114–133. ——— (2006) “Surviving Interests and Living Wills,” Public Affairs Quarterly 20: 17–30.

Deciding for the Incompetent 125 ——— (2007) “Precedent Autonomy, Advance Directives, and End-of-Life Care,” In B. Steinbock (ed.) The Oxford Handbook of Bioethics, New York: Oxford University Press, 349–374. Dworkin, R. (1993) Life’s Dominion: An Argument about Abortion, Euthanasia, and Individual Freedom, New York: Knopf. Feinberg, J. (1984) Harm to Others: The Moral Limits of the Criminal Law, New York: Oxford University Press. In re Conroy, 98 N.J. 321 (1985). In re Eichner, N.Y. Ct. App., 52 N.Y.2d 363, 420 N.E.2d 64 (1981). In re Quinlan, 70 N.J. 10 (1976). In re Westchester County Medical Center, 72 N.Y.2d 517 (1988). Johansson, M. and L. Broström (2011) “Counterfactual Reasoning in Surrogate Decision Making—Another Look,” Bioethics 25: 244–249. Jonsen, A., M. Siegler, and W. Winslade (2010) Clinical Ethics: A Practical Approach to Ethical Decisions in Clinical Medicine (7th Edition), New York: McGraw-Hill. Lewis, D. (1986) On the Plurality of Worlds, Oxford: Blackwell. Locke, J. (1690/1979) An Essay concerning Human Understanding, New York: Oxford University Press. Nagasawa, Y. (2008) “Proxy Consent and Counterfactuals,” Bioethics 22: 16–24. Olson, E. (1997) The Human Animal: Personal Identity without Psychology, New York: Oxford University Press. Parfit, D. (1971) “Personal Identity,” Philosophical Review 80: 3–27. Pearlman, R.A. (2008) “Substitute Decision-Making,” in P.A. Singer and A.M. Viens (eds.) The Cambridge Textbook of Bioethics, Cambridge: Cambridge University Press, 58–64. Phillips, J. and D. Wendler (2015) “Clarifying Substituted Judgment: The Endorsed Life Approach,” Journal of Medical Ethics 41: 723–730. Superintendent of Belchertown State School v. Saikewicz, 373 M.A. 728 (1977). Vogelstein, E. (2016) “Autonomy and the Moral Authority of Advance Directives,” Journal of Medicine and Philosophy. doi:10.1093/jmp/jhw019. Wierenga, E. (1983) “Proxy Consent and Counterfactual Wishes,” Journal of Medicine and Philosophy 8: 405–416. Wrigley, A. (2011) “The Problem of Counterfactuals in Substituted Judgment DecisionMaking,” Journal of Applied Philosophy 28: 269–187. ——— (2015) “Moral Authority and Proxy Decision-Making,” Ethical Theory and Moral Practice 18: 631–647.

8

Change of Mind An Issue for Advance Directives Paul T. Menzel

Summary Advance directives for health care for times of future incompetence face numerous challenges to their moral authority. One of the most direct occurs when a person has a change of mind, but sometimes it is difficult to determine whether that has occurred. After a person becomes incompetent, what behavior and expression might still constitute a change of mind sufficient to alter a directive’s authority? Refining our understanding of relevant change of mind will be pursued through three cases. They point to this conclusion: when the judgments and desires involved in the reasons that people had for their directive have changed, then even though they have lost the capacity to understand and revise their directive, the change is sufficient to call its moral authority into question.

Introduction Individuals make advance directives (ADs) about both medical care and other measures for times when they will not be able to make their own decisions (Brodoff 2010). The person who competently created the directive, often referred to in the literature as the “then-self,” may request, for example, that no treatment that would prolong life in a significantly diminished or painful condition be provided because she does not want to live if she has to “live like that.” Or she may want life not to be prolonged if she is close to death anyhow or treatment is unlikely to lead to significant recovery; then she would be ready to accept death. Yet precisely the situation that motivates writing an AD and brings it into operation—incompetence—may also affect a person’s values, attitudes, and desires. The incompetent individual (the “now-self ”) may not regard “life like that” as unbearable, for example. In her current form she is not autonomous, does not care about autonomy, is not aware of the directive she wrote, and may no longer value many of the things about life that were the basis for writing her directive. So why should it now be followed?

Change of Mind

127

Some—most notably Rebecca Dresser and John Robertson (1989)—argue that the directive is largely irrelevant. We should treat the now incompetent patient solely on the basis of her current best interest, not her previous wishes, even if she clearly expressed those wishes in her directive. To respond that this would fail to respect the patient’s autonomy is hardly persuasive, for it is the relevance to the current patient of that very autonomy that Dresser and Robertson dispute. At this point in the argument, defenders of ADs may choose either to stay with the “precedent autonomy” framework and claim that the autonomy of previous choices simply outweighs the patient’s current best interests, or to move out of the precedent autonomy framework to develop an expanded notion of the patient’s current interests, some of which can be determined by earlier expressed values. The latter option is represented by Ronald Dworkin, among others (1993: 201–232; discussed in this volume in Chapter 10, Menzel, “Refusing Lifesaving Medical Treatment and Food and Water by Mouth”). Whether it is a version of precedent autonomy or expanded best interest (or neither), perhaps the strongest defense of ADs in response to Dresser and Robertson was made by Nancy Rhoden (Rhoden 1990): their view amounts to treating an incompetent person who was previously competent no differently than we would treat an incompetent person who had never been competent. But if it is not acceptable to treat the previously competent the same as the never competent—and it is not—then ADs must have some authority (Menzel and Steinbock 2013: 496). This may be a persuasive response to the then-self/now-self challenge to ADs, but in any case they will sometimes face a simpler, more direct challenge: a person may have had a “change of mind” about her directive. A basic assumption we make about directives is that people may change them, and that later expression removes the previous one. Barring some flaw in voluntariness or competence, a retraction or alteration of a directive supplants it. Even changes of mind communicated orally must be taken seriously. Though change of mind is widely accepted as removing a directive’s normative force, the change-of-mind paradigm still harbors interesting ambiguities and difficulties. How far does the paradigm extend? What kinds of later expressions, short of direct and explicit revocation, constitute relevant change of mind? Different values that the person seems to have formed? Merely different desires that stand in tension with the directive? In dementia, for example, it will be important to discern whether different desires and values constitute a relevant change of mind. In this chapter I will refine what constitutes a change of mind in the morally relevant sense that alters a directive’s moral authority. After clarifying several preliminary matters, I will proceed to three difficult cases. All three involve one level or another of dementia, though dementia itself is not the only circumstance in which discerning a relevant change of mind is difficult.

128

Paul T. Menzel

Preliminary Considerations By speaking of an AD’s “authority,” I do not mean to accord them anything like final control. I mean only that they should be respected as having normative force and a substantial role in decision-making. By “authority” I also always mean directives’ moral, not necessarily legal, authority. A not uncommon kind of situation with directives can be easily confused with change of mind. Suppose someone has said that if, even with the best treatment, she were not likely to get out of the hospital to return home, she wants not to receive “any life-extending treatments, including antibiotics for common infections such as pneumonia.” Now the person is hospitalized, and in precisely this situation, with an incurable illness from which she is not likely to recover sufficiently to return home and, most recently, with pneumonia. Were antibiotics used, she would likely still die from her other complications but several weeks later than if they are used. Suppose, though, that she will likely be more comfortable in whatever time she does have left if antibiotics are used. Her proxy decision-maker asks what the patient would decide if she were now competent and concludes that she would consent to the antibiotic. Using “substituted judgment” reasoning, he claims she would change her mind. Substituted judgment reasoning here may be sound, but it is not an instance of what I mean by change of mind. It is only an observation that a person would change her directive if she were sufficiently aware to speak to the matter. A hypothetical change of mind, no matter how justifiably claimed, is not a real change of mind. My focus is on actual change of mind. Some may think that is not possible once the patient is incompetent. I will argue that it is. Our discussion will benefit from noting the somewhat limited extent to which we are concerned about change of mind. On the one hand, it is important to be on the lookout for it, so that when a directive is implemented, we are justifiably confident that it really does still represent the will of the patient. This implies that a directive’s authority is stronger the more recently it has been reiterated. On the other hand, we do not carry to an extreme the concern that people might have changed their mind without bothering to express it. If we did, we would dispense with ADs altogether, along the following lines: It is always possible that someone changed her mind. Now, precisely because she is incompetent and unable to express herself, we do not know that she has not changed her mind. Therefore, without knowing that a person has not changed her mind, we can never regard a directive as having authority once the time of incompetence to which it was meant to apply has arrived.1 The very acceptance of ADs in the last half century implies that we reject this line of argument. Hidden change of mind is always possible, but mere possibility is no reason to discount directives. One consideration is “balance”: yes,

Change of Mind

129

according ADs any authority runs some risk that the person has changed her mind, but on the other hand, according directives no authority runs a greater risk that patient choice will not be respected. Another consideration is that the outcome of the argument—according directives no role in decision-making— amounts to treating persons who have been competent as if they had never been competent (Rhoden 1990). Yet who among us would accept the proposition that the whole narrative of our competent years, and the wishes we express from them in a directive, should make no difference in how we are treated at the end of our lives? Thus, the general authority of ADs is not destroyed by the ever-present possibility that the person who wrote the directive, unbeknownst to us, has changed her mind. For change of mind to disrupt a directive, there must be evidence that it has occurred.

Three Cases The first of these, Richard, is a fictional adjustment from a real case with which the author is familiar. The second, Margo, is a published real case, modified with the addition of an AD and an incident of pneumonia. The third, Ms. Bentley, is also a real case, the subject of a Canadian court decision. Richard, 94, has been gradually declining, physically and mentally, for ten years. Based on brain scans, his dementia is diagnosed not as Alzheimer’s but as more generic—“severe diffuse cerebral atrophy,” said the radiologist’s report. He needs help getting dressed, going to bed, and with most toileting. He walks with a walker or by firmly holding someone’s arm, but only for short distances; most of his ambulation is now by wheelchair. He can usually feed himself when food is appropriately prepared and placed. His appetite is erratic, and nutritional smoothies are important in keeping up his strength and weight. He has lived in the same excellent comprehensive senior facility for over a decade. Despite some intervening years of frustration with his decline and boredom, he is now consistently good-natured. Sometimes he will grimace with pain when moving to or from chair or bed, but he frequently smiles when someone touches his arm or talks directly to him. He can barely carry on any conversation and has seen his social relationships, on which he truly thrived, dissipate. He occasionally utters a brief partial sentence, usually in response to a story about himself told directly to him. In his directive Richard requested that he receive all life-extending care that does not prolong significant pain and suffering, which he is not currently experiencing. Several times recently, however, he has made comments such as “I’m ready to go,” usually in the context of his daughter or son having noted to him how gratifying some experience in the past must have been. Once his daughter probed him about this, affirming explicitly how good his life has been, then

130

Paul T. Menzel

asking gently whether he was afraid of death. He looked at her for a bit, and then faintly shook his head. Now, for the second time, a cold has morphed into pneumonia. Antibiotics would likely treat it effectively. Should they be provided, as his AD clearly asks, or should his “ready to go” comments be interpreted as a change of mind and pneumonia be allowed to take its course? Margo, 55, is not quite yet in the severe stage of her early-onset Alzheimer’s (Firlik 1991). She often behaves as if she knows people but she never uses their names, and it is doubtful that she relates to them as individuals. She is physically frail and hunched, and ambles unsteadily—but she does amble, persistently, to the point of roaming Central Park in her nightgown once when her facility’s safety locks were faulty. Andrew, the medical student who has visited her regularly for months, notes that while she claims to read mysteries, “her place in the book jumps randomly. . . . She feels good just sitting and humming to herself . . ., nodding off liberally, occasionally turning to a fresh page” (Firlik 1991). Once or twice a week she attends an art class for dementia patients, where for years she has painted virtually the same picture each time. She takes abundant pleasure in simple acts like eating peanut butter and jelly sandwiches. She is not frustrated, depressed, or angry. Andrew writes that “despite her illness, or maybe somehow because of it, Margo is undeniably one of the happiest people I have ever known” (Firlik 1991). We do not know much about Margo’s past life and concerns, but they were undoubtedly different than what she has now. Suppose that before her dementia had progressed very far she wrote a directive, reiterated clearly after its onset, requesting that once she can neither recognize people, converse with them as specific individuals, nor read more than a paragraph continuously, all lifesaving treatments are to be refused, including antibiotics. Now, like Richard, she has contracted pneumonia, from which she will undoubtedly die if not treated. Should antibiotics be withheld, in line with her directive, or do her new enjoyments that arguably constitute a desire to live amount to a change of mind? Mrs. Bentley, an 84-year-old former nurse, was diagnosed with Alzheimer’s 15 years ago (British Columbia Supreme Court 2014; Court of Appeal for British Columbia 2015). For the last nine she has been in a residential care facility, and for the last two in very severe dementia, stage seven of Alzheimer’s. She does not recognize or respond to any of her friends or family, does not speak, and can make only very limited physical movements. She spends her days “motionless in bed or slumped in a wheelchair” with her eyes closed (from the court record) and requires spoon-feeding. She will usually open her mouth when food is held to it, though this varies. Sometimes caregivers have to offer it several times. She accepts different types and amounts of food on different days, seeming to prefer certain flavors, and she stops opening her mouth when apparently she feels full.

Change of Mind

131

Mrs. Bentley first executed a directive 23 years ago, stating, “If . . . there is no reasonable expectation of my recovery from extreme physical or mental disability, I direct that I be allowed to die and not be kept alive by artificial means or heroic measures.” She listed specific instructions, including “no nourishment or liquids,” and designated her husband and daughter as proxy decision-makers (Court of Appeal for British Columbia 2015: 5). Fifteen years ago, after diagnosis of her Alzheimer’s, and consistent with her earlier directive, she “repeatedly told her family that she wished to be allowed to die when she reached a stage of advanced dementia” (Pope 2015: 1). She reached that stage two years ago. Does Mrs. Bentley’s apparent willingness to accept food constitute a change of mind about what she requested in her directive?2

What Change of Mind Must Be About The paradigm of a change of mind by a still competent person is a new directive or addendum that speaks to the same issues but alters what was previously requested. It cannot be merely a supplement to the earlier directive. If the revision is only spoken, that, too, will constitute a relevant change of mind if there is no reason to doubt voluntariness or competence. Once the patient becomes incompetent, things get more complicated. Generally, competence is relative to what is being decided. Advance directives are usually reflective and fairly complex matters. A person in even moderate dementia, such as Margo, may not have much understanding of what a directive is about. We may be tempted to say that relevant change of mind is then no longer possible, for the mental outlook of which the person is capable can no longer be about the directive. I will argue, however, that this construes “change of mind” too narrowly. While the person may not have the capacity to write a new directive, she may still be capable of making new and different judgments about the very matters that were the reasons she had for writing her directive. Let’s apply this view to the three cases. Richard. In his directive he requests all life-extending care unless it extends life in significant pain and suffering. Suppose an important reason that he included was that, barring such pain and suffering, he thought he would never be “tired of life” or “ready to go.” He thought life was too precious no matter how diminished his physical and mental capacities might be or how long his life would go on. Now he seems to have changed his attitude. If he has changed his judgment on the very matter that is a primary reason behind his directive, then has he not changed his mind in a way that invalidates the directive? Why does he have to have the capacity to change his mind more comprehensively and understand what his directive is all about? That he has had a change of mind about what was the basis of his directive calls into question whether what the directive expressed is still his will.

132

Paul T. Menzel

It is revealing to contrast the situation with Richard as I have been describing him (“Original Richard”) with “Revised Richard.” Revised Richard had a different reason for the same substantive requests in his directive—not his anticipation that he would never be tired of life or “ready to go,” but his belief that God meant for life always to be lived as long as it could be, barring significant pain and suffering. God, he believed, was merciful—therefore the pain and suffering exception. But God ordained life as precious, and not only so long as it took its “natural” course but also when it could be preserved by medical advances made possible by God-given human talents. Suppose that Revised Richard now expresses the same “ready to go” attitudes of Original Richard. In writing his directive, Revised Richard did not rely on any judgment that he would never be “ready to go,” but on beliefs about life as ordained by God. Therefore, when he now has “ready to go” feelings, they do not constitute a change of mind about the directive. Those feelings are perhaps not what we would expect from a person with the religious beliefs Revised Richard holds, but that is beside the point. Regardless of whether expected, they do not constitute a change in what he relied on to make his directive. By now, in fact, Revised Richard is no longer capable of a change of mind about the beliefs that were the basis of his directive. He is past the point where he can have a relevant change of mind. If this analysis is correct, then to discern whether a mental change that people undergo after losing general competence is a relevant change of mind, we need to know something about the reasons people had for their directive. If a directive itself conveys little about those reasons, and if the persons to whom the patient is close and are now responsible for speaking for him know little about them, it will not be possible to discern relevant change of mind. To be sure, that is not the end of the matter, as if we then just follow his directive and treat the pneumonia. The kind of situation Revised Richard represents properly leaves us hesitant. We cannot justify departing from his directive by citing a change of mind, but the “ready to go” utterances seem to constitute an acceptance of death, perhaps even a desire for it. If he is truly ready for death, then despite his overall lack of competence and capacity to have a change of mind that alters the directive, should we not just do what his current mental readiness would seem to justify and allow him to die? Our hesitation about sticking to his directive likely reflects the “then-self vs. now-self ” problem that will always be a moral challenge to ADs even when there has been no relevant change of mind (see Chapter 10, Menzel, “Refusing Lifesaving Medical Treatment and Food and Water by Mouth”). Margo certainly has new interests and enjoys activities that would not have engaged her before her dementia progressed to her current moderate-to-severe stage. Agnieska Jaworska (1999) argues that in moderate dementia people can still form new values, not just desires, and that their self-concept can change;

Change of Mind

133

they care about those interests and have a different conception of themselves. Suppose Jaworska is correct about what constitutes valuing and self-concept, and about what is actually happening in moderate dementia. Exactly how to assess Margo, though, is less clear. She does have a new appreciation of repetitive “art exercise,” of eating peanut butter and jelly sandwiches nearly every day, of conversing with people who are familiar but whose names and particular identities she does not recall, and of holding a book and turning its pages, though she barely reads even a paragraph. In a directly experienced sense, Margo has had a change of mind about what makes living worthwhile, though she no longer has sufficient memory or anticipation to consider a new directive. But has she changed her mind about her existing one? In it she requested that once she cannot recognize people, converse with them as specific individuals, or read more than a paragraph continuously, all lifesaving treatments, including antibiotics, should be withheld. She now has an apparently positive attitude about these very states that she did not want to continue to be kept alive in; Firlik (1991) calls her “one of the happiest people” he has ever known. If her reason for saying in her directive that she wants not to be kept alive in these states was that she thought she could not enjoy them, then her current enjoyment of life is a change of mind about the foundation of her directive and the directive has lost its normative force for whether to treat her pneumonia. Suppose, however, that Margo’s reasons for wanting not to be kept alive in advancing dementia were different and more complex. Let me refer to this Margo with the different directive as Complex Margo: (a) she did not want her life as a whole to end that way because its level of engagement is not what she has stood for and supported people to aspire to all her life; (b) she very much preferred that her financial resources go to other things she cared about more strongly than continuing to live in such condition; and (c) she wanted her loved ones not to have to care for her in such a state year in and year out, with the likely result that they would have less memory of her as the creative, personally attentive, intellectually engaged person she was. If these considerations, not thoughts that she couldn’t enjoy life in a condition like her current one, were her reasons, then Complex Margo’s positive accommodation to her dementia is not a change of mind about her directive. This would be especially clear if right in her directive she insisted that her reasons for it should continue to hold even if she turned out to be one of the happy demented. In progressing into dementia, people undergo many mental changes. Of course they have changed their minds about some things, but what have they changed their minds about? As dementia deepens there will come a point when they pass the point where they have enough of a mind to change what they need to change to have had a change of mind about their directive. As Ron Berghmans poignantly puts it, with dementia, “at the time you would most

134

Paul T. Menzel

likely ‘change your mind’ [about your directive, because you have adapted to your situation and are no longer distressed by your dementia], you do not have enough mind left to change” (2000: 107, emphasis added). Revised Richard no longer has enough of a mind to change the beliefs that were the reason for his directive. Neither does Complex Margo. This should not by itself lead to the conclusion we should carry out Margo’s directive and let her die from pneumonia. With her directive and in her currently largely happy state, Margo constitutes a classic example of the then-self/ now-self challenge that still needs addressing. The current Margo does not care about her directive and is not autonomous. This Margo clashes with the back-then Margo, who had strong convictions about the larger arc of her life and how it should best end. Insofar as the competent Margo—the one with a narrative about her whole life—is the only Margo who has defined Margo with this narrative, it is this Margo who creates critical interests about how and when her life should end. It is this Margo, too, who continues to constitute Margo first and foremost in the minds of her survivors; were it not for that Margo, they would not see themselves as her survivors. Margo, therefore, still has current interests that are not simply erased by her present appreciation of a diminished life. Dworkin (1993: 201–232) calls these “critical interests.” They live on at the current moment even though they are not experienced by the person anymore and are therefore not “experiential interests.” Since Margo’s currently experienced interest in continuing to live—she does enjoy her life—is still considerable, the most defensible decision may be to say “not yet” about implementing her directive. It would be treacherous to say, already now when she is happy and still engaged, that it is better for her sake that her directive be followed and she be allowed to die. Nonetheless, there will certainly come a day in the progression of her dementia when the experiential, subjective value to her of her own survival is so diminished that we should see it as outweighed by her previously formed critical interests (Menzel and Steinbock 2013: 492–493). Mrs. Bentley, especially with her experience as a nurse, has presumably written a knowledgeable directive. She has now been in advanced dementia for several years, the condition she explicitly states in her directive when she is not to be given life-extending measures, including nourishment and liquids. The conflicting factor now, though, is that she selectively opens her mouth to accept food. She appears to discriminate between foods she likes and those she does not, and between eating when hungry and not when she feels full. Does her discriminating acceptance of food constitute a relevant change of mind about her directive? This question is not settled if we classify Mrs. Bentley’s acceptance of food as “consenting behavior,” more than a non-voluntary reflex. She is far past being capable of considering her directive, so for her eating behavior

Change of Mind

135

to constitute a relevant change of mind it will have to convey a new attitude toward something that was a reason for what she requested in her directive. It is possible that the reason she stipulated no nourishment and liquids was that she thought eating or drinking would be uncomfortable, but it is hardly plausible. If it was her reason, why would she also have stipulated that all the other interventions she mentions not be provided? Her plausible reasons more had something to do with her life as a whole not concluding with years of advanced dementia, not merely any anticipation that she would not be willing to eat. Mrs. Bentley is no longer capable of any change of mind about these reasons. Her acceptance of food therefore does not constitute a relevant change of mind. Here, again, as with Richard and Margo, it is easier to discern relevant change of mind if we know some of the people’s reasons for their directive’s substance. An optimal directive will mention those reasons and briefly explain them. Conversations between the person and their later decision-makers can also add to understanding the reasons for what their directive conveys. To summarize the cases: •

•

•

•

•

Richard’s new “ready to go” attitudes constitute a change in what was a major reason for his directive requesting all life-extending care that does not prolong significant pain and suffering. His directive should not be followed. His pneumonia should not be treated. Revised Richard’s reasons are too sophisticated for him to any longer have the capacity to change. His directive’s authority still stands and favors treating his pneumonia despite his apparent disposition now to accept death. Margo’s new appreciation of her diminished life constitutes an alteration in the basis for her AD. She has had a relevant change of mind. We should treat her pneumonia. Complex Margo no longer has the capacity to have a change of mind about the basis for her directive not to provide life-extending care when she gets to her current stage of dementia. Her directive retains its authority. Her current positive appreciation of life, however, may outweigh the critical interests conveyed in her directive. In that case we should say “not yet” to its implementation and treat the pneumonia. Mrs. Bentley no longer has the capacity to have a change of mind about the reasons for her directive not to receive food and water once in severe dementia. In face of the then-self/now-self problem, this alone is not sufficient reason to follow her directive, but given that her own currently experienced survival is of very low value to her, we should follow her directive and not provide food and fluids except as strictly necessary to keep her comfortable.

136

Paul T. Menzel

Conclusion Relevant changes of mind limit the moral authority of an advance directive. Not just any change in a person’s attitudes or evaluative outlook, however, constitutes a relevant change of mind. Consideration of three challenging cases leads to the following conclusion: when the judgments and desires involved in the reasons people had for a directive have changed, then even if they have lost the capacity to understand and revise their directive, they have had a change of mind sufficient to call their directive’s authority into question. This points to the importance of articulating within a directive the most significant reasons for its content and for discussing these reasons with the person’s future decision-makers. On the other hand, a quite different situation might obtain. If the reasons people have for their directives are relatively complex and involve convictions about the shape of their whole lives, people with significantly diminished cognitive capacity will have passed the point where they can any longer have a relevant change of mind. This latter circumstance, however, does not necessarily imply that we should carry out the directive. It means only that if we should not carry it out, change of mind is not the appropriate reason. Even when its authority is undisrupted by any relevant change of mind, a directive will still have to meet the then-self/ now-self challenge. Sometimes incompetent persons’ experiential interests in surviving will outweigh the critical interests in not surviving that they conveyed in their directive; in that case we should say “not yet.” In other cases the low value to them of survival will not outweigh the critical interests against surviving. There is no formula by which the balance of critical interest in not continuing to live and experiential interest in living can be clearly discerned, but sensible judgments can be made by attending to the details of the person’s real current life, the directive itself, and the testimony of proxies and loved ones about why the person saw living long into the conditions that currently obtain as a contradiction of what she lived for.

Related Topics Chapter 10, “Refusing Lifesaving Medical Treatment and Food and Water by Mouth” by Paul T. Menzel Chapter 11, “Suicide, Physician-Assisted Suicide, the Doing-Allowing Distinction, and Double Effect” by Thomas S. Huddle

Notes 1. I owe my recognition of this line of thinking to Lawrence Nelson (Santa Clara University). It may explain why the legal authority for advance directives was originally doubted.

Change of Mind

137

2. This is not the question the British Columbia courts addressed. They addressed whether Mrs. Bentley had the capacity to consent to eat, and whether her behavior of sometimes accepting food constituted consent. It answered both in the affirmative and then determined that her previous wishes could not control (British Columbia Supreme Court 2014; Court of Appeal for British Columbia 2015; Pope 2015: 2). The latter court also found that providing food and water by mouth was “basic personal care,” not medical care, and as such not within the purview of any AD. Other nuances not noted in my discussion were also important to the decision of the court.

References Berghmans, R.L.P. (2000) “Advance Directives and Dementia,” Annals of the New York Academy of Sciences 913: 105–110. British Columbia Supreme Court (2014) Bentley v. Maplewood Seniors Care Society, BCSC 165. Brodoff, L. (2010) “Planning for Alzheimer’s Disease with Mental Health Advance Directives,” The Elder Law Journal 17: 239–308. Court of Appeal for British Columbia (2015) Bentley v. Maplewood Seniors Care Society, BCCA 91. Dresser, R. and J.S. Robertson (1989) “Quality of Life and Non-Treatment Decisions for Incompetent Patients,” Law, Medicine and Health Care 17(3): 234–244. Dworkin, R. (1993) Life’s Dominion: An Argument about Abortion, Euthanasia and Individual Freedom, New York: Alfred A. Knopf. Firlik, A.D. (1991) “Margo’s Logo,” Journal of the American Medical Association 265(2): 201. The case was brought to prominent attention in philosophical circles in Dworkin 1993: 220–232. Jaworska, A. (1999) “Respecting the Margins of Agency: Alzheimer’s Patients and the Capacity to Value,” Philosophy and Public Affairs 28(2): 105–138. Menzel, P. and B. Steinbock (2013) “Advance Directives, Dementia, and PhysicianAssisted Death,” Journal of Law, Medicine and Ethics 41(2): 484–500. Pope, T.M. (2015) “Prospective Autonomy and Ulysses Contracts for VSED,” one of two sections of Thaddeus M. Pope and Bernadette J. Richards, “Decision-Making: At the End of Life and the Provision of Pretreatment Advice,” Journal of Bioethical Inquiry 12(3): 389–394. doi:10.1007/s11673-015-9652-6. Rhoden, N. (1990) “The Limits of Legal Objectivity,” North Carolina Law Review 68: 845–865.

Further Reading Buchanan, A.E. and D.W. Brock (1990) Deciding for Others: The Ethics of Surrogate Decision-Making, Cambridge, UK: Cambridge University Press. Cantor, N. (1990) Advance Directives and the Pursuit of Death with Dignity, Bloomington: Indiana University Press. Cebuhar, J.K. (2015) The Practical Guide to Health Care Advance Directives, West Des Moines, IA: Murphy. Fagerlin, A. and C.E. Schneider (2004) “Enough: The Failure of the Living Will,” Hastings Center Report 34(2): 30–42. Menzel, P. and M.C. Chandler-Cramer (2014) “Advance Directives, Dementia, and Withholding Food and Water by Mouth,” Hastings Center Report 44(3): 23–37.

9

Medical Futility and Respect for Patient Autonomy Nancy S. Jecker

Summary Withholding and withdrawing medical interventions on the basis of medical futility is well-established in US law and ethics. However, when this entails overriding the wishes of patients or family members, the concern has been raised that this violates the ethical requirement to respect patient autonomy. This chapter clarifies the ethical principle of autonomy, and shows that properly understood, respect for patient autonomy is compatible with refusing patient or family requests for medically futile interventions.

Introduction Withholding and withdrawing medical treatment on the basis of medical futility is a well-established practice, reflected in law, professional standards, institutional policies, and medical practice. Yet it remains controversial in situations where a patient or surrogate requests that “everything possible” be done and a physician must override the patient’s wishes when withholding or withdrawing futile interventions. In such situations, it has been argued that not offering treatment, or stopping treatment if it is already underway, is objectionable because it violates the ethical principle of respect for patient autonomy and cedes too much authority to the physician. This chapter critically evaluates this objection and argues that it does not withstand careful scrutiny. The first two sections introduce the topic of medical futility, providing a definition and exploring futility’s ethical implications. The next section sets forth an objection to medical futility. The objection claims that physicians who withhold or withdraw futile treatment against the wishes of the patient or family exercise an unjustified, unilateral authority over the patient. I consider two versions of this objection. The first version claims overriding patients’ wishes in this way violates patient autonomy; the second version claims that it is exploitive and takes unfair advantage of patients’

Respect for Patient Autonomy 139 vulnerability and physicians’ relative power. The final sections provide a response to these objections.

The Meaning of Medical Futility Medical futility emerged as a focus of scholarly attention during the 1990s, when empirical studies first reported on a large scale that health professionals were utilizing interventions, such as cardiopulmonary resuscitation (CPR), in situations where they had little chance of helping patients. In 1993, Solomon and colleagues studied 687 physicians and 759 nurses in five U.S. hospitals looking at attitudes toward care of hospital patients near the end of life (1993). They reported that almost half (47%) of those surveyed reported having acted against their conscience in providing care. Of those, 55% reported feeling that sometimes the treatments offered were overly burdensome, compared with only 12% overall concerned about undertreatment. More recent findings show that overtreatment continues to be a major concern among health care professionals. In 2011, Sirovich and colleagues surveyed 627 U.S. primary care physicians and found that nearly half (42%) believed that patients within their own practice were receiving too much medical care, while just 6% believed that their patients were receiving too little care. More than one quarter (28%) said they themselves were practicing too aggressively. Not only was overtreatment found to be potentially burdensome for patients, but also it was consistently associated with moral distress and “burnout” among clinicians (Solomon et al. 1993; Sirovich et al. 2011). When providers act contrary to their sense of what is right, overtreatment creates “moral distress,” understood as the suffering that occurs when the right course of action is known but not carried out. Providers may feel unable to do the right thing due to patient and family demands, fear of legal liability, or the broader culture of medicine. Moral distress is not just a “bad feeling”; it has been demonstrated to be seriously detrimental to family relationships and professional performance. It affects physical well-being, self-image, and job satisfaction, often culminating in a loss of focus, reduced patience, and burnout (Silen et al. 2008). To address the problem of overtreatment, it is helpful to introduce the concept of medical futility. The word “futility” comes from the Latin word futilis, meaning leaky. According to the Oxford English Dictionary, a futile action is “leaky, hence untrustworthy, vain, failing of the desired end through intrinsic defect.” In health care, the futility of a particular treatment may be evident in two distinct ways (Schneiderman et al. 1990). First, an intervention may be futile in the quantitative sense. This means that there is no significant chance of achieving the goal of helping the patient. For example, Schneiderman, Jecker, and Jonsen maintain that if the odds that a patient will survive following in-hospital cardiopulmonary resuscitation (CPR) are only one in one hundred, CPR does not

140

Nancy S. Jecker

confer a significant benefit on the patient and should be considered futile in the first, quantitatively sense. Setting the threshold for quantitative futility at one in one hundred is not arbitrary. Instead it matches the widely accepted justification used in the statistical analysis of controlled clinical trials. In the statistical evaluation of clinical trials, when an intervention produces an effect in less than one in one hundred cases, this is not considered significant, because it is just as likely that the effect is due to chance as to the intervention (Schneiderman et al. 1996a). Of course, there may be disagreement about where the specific threshold for quantitative futility lies. One person may regard a 5% chance of success as futile rather than a 1% chance. However, the more general idea to keep in mind is that at some threshold the likelihood of success is so slim it qualifies as futile. A second dimension of futility is qualitative futility. Qualitative futility refers to cases where an intervention does not achieve the goal of helping the patient because the quality of outcome associated with it falls well below a threshold considered minimal. One example is continued ventilator support of a patient in a persistent vegetative state. Although this intervention produces an effect on the patient’s body—namely, expanding and contracting the patient’s lungs—medicine’s goal has never been to merely produce physiologic effects. Instead, medicine aspires to do good—that is, to help the patient. Ventilator support for a patient who will never regain consciousness fails to confer a benefit that the patient can ever appreciate. Thus, we distinguish between “effects,” on the one hand, and “benefits,” on the other. Interventions produce all kinds of effects on organ systems and body parts. But medicine’s goal has never been to produce effects on bodies. Instead, medicine is beneficence-based and aspires to help the sick. Other examples of qualitative futility include conditions requiring permanent and constant monitoring, ventilator support, and intensive nursing care; and conditions associated with overwhelming suffering for a predictably brief period of time. Admittedly, these kinds of cases fall along a continuum, and there are well-known examples of the most remarkable achievements of life goals despite the most burdensome handicaps. However, if survival requires the patient’s entire preoccupation with intensive medical treatment, to the extent that he or she cannot achieve any other life goals, then treatment does not provide a significant benefit to the patient. Again, we may debate what the specific paradigm cases of qualitative futility include; however, we may nonetheless reach consensus regarding the more general point that at some threshold level, the quality of outcome is so poor it is futile.

The Ethical Duty to Withhold or Withdraw Futile Interventions If the health care team agrees that an intervention is futile in the quantitative or qualitative sense, what does the responsible health professional do? There are

Respect for Patient Autonomy 141 three positions one might take (Schneiderman and Jecker 2011). First, it could be argued that individual physicians should exercise discretion about using or continuing with futile interventions. According to this approach, physicians are ethically permitted to either offer or not offer futile interventions. They are ethically free to either continue with or withdraw futile interventions that are already underway. A second position holds that physicians should be discouraged from using futile interventions. In other words, it is generally better if they avoided interventions that do not help the patient. A third view maintains that physicians should not use futile interventions, and should withhold or withdraw such interventions. Ethical support for these views comes from ethical principles and professional standards. The ethical principle of beneficence requires that physicians do good and act in a way that benefits patients. Thus, if an intervention will not produce a significant benefit, there is not a duty to use or continue it. This principle supports the first stance, which states that physicians are ethically allowed to either use or not use futile treatments. The ethical principle of non-maleficence requires the physician to omit actions that harm the patient. Futile interventions are often invasive and add to the patient’s pain or discomfort at the end of life. Thus, a positive duty to refrain from using futile interventions is supported by this principle. The ethical principle of autonomy holds that physicians should not stand in the way of what a competent patient wishes. Expressed differently, physicians should not interfere with what a competent adult patient chooses. It might, at first glance, seem that the principle of autonomy conflicts with stopping medically futile interventions in situations where a patient requests them. Yet a proper understanding of the ethical requirement to respect patient autonomy distinguishes between weak and strong interpretations of autonomy (Beauchamp and Childress 2012). The concept of autonomy associated with the ethical principle of autonomy in the medical setting is weak autonomy. Weak autonomy refers to noninterference, and the associated ethical requirement is not to intervene to stop a person from doing what he or she wants to do. On this reading, autonomy is about omitting actions that interfere with the patient’s preferences. So understood, the ethical principle of autonomy implies that a patient can ethically refuse any treatment and health care providers should not stand in the patient’s way. Strong autonomy, by contrast, requires performing actions and doing what a patient requests in order to help a patient carry out his or her goals. Contrary to what some have claimed, respect for autonomy does not entail that a patient can demand that a physician perform whatever positive actions a patients requests (Veatch and Spicer 1992: 17). Instead physicians have a duty to provide competent care. Medicine is not practiced on demand, but instead is a profession that aspires to do good—that is, to help the patient. Thus the obligation of the physician is to offer a range of medically suitable treatment options, and not to interfere with what the patient

142

Nancy S. Jecker

chooses to do. Consistent with negative autonomy, physicians are expected to educate and recommend viable treatment options, but the final decision belongs to the patient. A final ethical principle, justice, requires fairness in the distribution of scarce health care resources. Medical care is a scarce resource, not only because the funds to pay for it are limited but also because there is a limit in the raw materials required to make a health service available. For example, society does not have an unlimited pool of physicians and nurses, unlimited hospital and intensive care unit (ICU) beds, or an unlimited supply of every prescription drug. While futile interventions are always wasteful, they may be more than this if they result in disadvantaging other patients. When futile interventions are scarce or expensive, using a futile intervention in one patient may result in another patient not receiving that treatment. For example, the choice to continue to treat a patient with respiratory failure in the ICU may result in another patient not being admitted to the ICU if beds are limited. Generally speaking the use of futile interventions is at odds with the physician’s responsibility to promote the public’s health and to use society’s resources in a responsible manner. Support for refraining from using futile interventions also comes from ethical standards of medicine. Professional medical associations have long endorsed the view that physicians have a duty to refrain from offering or continuing futile interventions. The American Medical Association, for example, held in 1999 that “When further intervention to prolong the life of a patient becomes futile, physicians have an obligation to shift the intent of care toward comfort and closure” (American Medical Association 1997). More recently, the American Thoracic Society, together with the American Association of Critical Care Nurses, the American College of Chest Physicians, the European Society for Intensive Care Medicine, and the Society of Critical Care Medicine, put forth a 2015 multi-society policy statement which held that “Clinicians should explain and advocate for the treatment plan they believe is appropriate . . . [and] should not provide futile treatments” (Bosslet et al. 2015: 1). These professional standards reflect ethical values internal to medicine and other healing professions. Medicine, nursing, and the other healing professions are not practiced merely on demand, but instead aspire to moral goals, such as helping the sick. According to Brody and Miller there is an “internal morality of medicine,” meaning that physicians aspire not just to meet patients’ requests but also to accomplish moral ends, such as: 1. 2. 3. 4. 5.

Reassuring the worried well who have no disease or injury Diagnosing disease or injury Helping the patient to understand the disease Preventing disease or injury Curing the disease or repairing the injury

Respect for Patient Autonomy 143 6. 7. 8.

Lessening the pain or disability caused by disease/injury Helping the patient to live with pain or disability Helping the patient die with dignity and peace (Brody and Miller 1998: 386–387)

The problem with futile interventions is that they often frustrate these ends, thereby undercutting the integrity of medicine—for example, futile interventions often conflict with (3), by frustrating efforts to help the patient understand the disease; with (6), by increasing, rather than lessening, the pain caused by the disease; and with (8), by standing in the way of helping the patient die with dignity and peace. The foregoing analysis shows that in situations where patients (or surrogates) insist on “doing everything,” the responsible physician upholds professional standards of care and offers patients competent care (Jecker and Schneiderman 1995b). Competent care always includes palliative and comfort care (Schneiderman, Faber-Langdeon, Jecker 1994), and may include a range of other interventions that help the patient and fall within the scope of professional standards (Jecker and Schneiderman 1995b). It is important to emphasize that when a treatment is futile, health professionals remain obligated to pursue other treatments that maximize comfort and dignity for the patient. From a practical standpoint, this means that rather than saying to the patient, “There is nothing I can do for you,” health professionals should instead affirm that everything possible will be done to ensure the patient’s comfort and dignity. Too often, patients and families insist on futile efforts because such efforts symbolize caring. Yet providers should identify and practice better ways of caring for patients.

An Objection to Withholding and Withdrawing Futile Interventions Those who object to invoking futility as a basis for withholding or withdrawing treatment have often done so on the ground that patients have an ethical and legal right to make medical decisions affecting them. Opponents maintain that overriding patient (or surrogate) requests is tantamount to granting physicians “unilateral authority” to make treatment decisions. Critics thus characterize appeals to futility as an “abuse of power” (Jecker and Pearlman 1992; Jecker 1995; Jecker and Schneiderman 1995a; Schneiderman et al. 1996b). While it clearly resonates with highly regarded values, such as respect for patient autonomy, this section argues that the autonomy-based objection frames medical decision-making in an overly simplistic manner: either patients exercise autonomy or physicians unilaterally override patients’ autonomy.

144

Nancy S. Jecker

The Externality Thesis A more nuanced account draws on what we might call “the externality thesis.” The externality thesis holds that individual and group decisions depend in part on concepts and values that are not properly seen as internal to the individual or group. Applied to medicine, the externality thesis suggests that medical morality depends, at least in part, on concepts and values that are not properly seen as internal to medicine. The externality thesis reminds us that medicine is practiced in social and cultural contexts that shape and support the morality we regard as “internal” to medicine. In this sense, the morality associated with medicine is porous—that is, it is able to influence and be influenced by “external” morality. As Miller and Brody suggest, the distinction between internal and external is perhaps best understood metaphorically: The internal/external distinction is a matter of perspective. One can start from outside medicine and ask how the various principles and rules of the common morality, or of other general sources of morality, apply to medicine. Or one can begin from within medicine, and ask how understanding medicine as an evolving professional practice helps us to discern what is right or wrong, good or bad, for physicians to do. (Miller and Brody 2001: 590) Situating the patient-physician encounter within the broader social and cultural contexts in which medicine is practiced makes evident that neither physicians nor patients make treatment decisions in isolation. Instead, such decisions occur against a background understanding between members of society and members of the medical profession, a point we shall explore in more detail momentarily. Applied to medicine, the externality thesis underscores that sources outside medicine influence physicians’ decisions and the broader morality of medicine. Yet what about the decisions of patients? When patients with decisional capacity make autonomous treatment choices, to what extent do “external” influences enter into their decisions? To address this question we might consider what it means to be “autonomous.” Many standard accounts assume that autonomous decisions arise entirely from within the individual agent and are authenticated by the patient’s higher-order desires (Dworkin 1970; Frankfurt 1988) or by a psychological core (Noggle 2005) that represents the true or authentic self. According to such analyses, the externality thesis is not particularly relevant to decision-making by autonomous patients. Yet an alternative understanding of autonomy is suggested by relational views of autonomy. Relational views often begin with by regarding “outside influences” not as a problem to be solved but as integral to the very nature of who we are as

Respect for Patient Autonomy 145 persons. According to such accounts, the self is partly constituted by the social roles and relationships in which it stands. A person is self-governing if and only if the sources of the desires that move that person to act are that robust, relational self. The broad idea here is that we are motivated by interests and reasons that can be fully understood only with reference to other people and to the roles and relationships in which people stand. This is because who we are is essentially connected to other people, groups, institutions, social roles, and histories. Relational views endorse the externality thesis, implying that patients’ values depend, at least in part, on concepts and value judgments that are not properly seen as internal to the patient. According to relational accounts, what disrupts autonomy is not outside influences but instead situations where relationships or social conditions involve domination, cruelty, or oppression that interferes with healthy self-government (Brison 2003).

Reinterpreting Autonomy-Based Objections to Medical Futility Where does a relational account leave autonomy-based objections to medical futility? The quintessential test of whether overriding patient preferences is ethically justified becomes the kind of patient-physician relationship that is fostered and supported by these actions. Are physicians who refuse to continue with futile treatment being exploitive to the patient? Does withdrawing futile treatment disrupt healthy self-government by patients? It is to these questions that we now turn. To be “exploitive” means to take unfair advantage of someone (Wertheimer 1999). This is often understood as implying a defect in the process by which a decision or action comes about. For example, unilateral appeals to medical futility would qualify as exploitive if they harm or coerce patients. On this interpretation, to defeat the objection requires showing that appeals to medical futility are both mutually advantageous and consensual, rather than harmful or coercive (Wertheimer 2012). It might at first glance seem that unilateral decisions to withhold or withdraw a futile intervention over the objections of a patient could not possibly meet the requirement of being consensual. After all, the patient’s and physician’s preferences are at odds. Yet the reality is not so simple. First, physicians’ authority to render medical judgments is not granted solely by the individual patient before them. Instead, as a member of a medical profession, the physician has already been recognized as having authority to make medical judgments and to establish professional standards of care in concert with other physicians. This authority is granted by the society in which physicians practice. We grant such authority to physicians on the basis of the specialized training and knowledge physicians possess. In return, we expect members of the medical profession to utilize their knowledge and

146

Nancy S. Jecker

ability in a manner that benefits society as a whole—that is, helps the sick (Freidson 1970). In the United States, we can trace the roots of this agreement between physicians and society to the nineteenth century, when medicine first professionalized. It was during this time that medicine’s orientation became increasingly scientific, placing emphasis on empirical investigation (Jecker 2005). Society granted physicians sole authority to deliver medical services, and did so for the benefit of its members. In this way, citizens benefitted from receiving competent care and from the protection this arrangement offered from charlatans or others who lacked the requisite skills to practice. The authority of the medical professional in establishing and implementing standards of care is illustrated in numerous ways. Thus, we grant physicians exclusive authority to perform surgeries, prescribe drugs, and diagnose medical conditions. As a case in point consider the physician’s special role in declaring a patient’s death. Since the original Uniform Determination of Death Act (UDDA), “death” has been defined as irreversible cessation of all functions of the entire brain, together with irreversible cessation of respiratory and circulatory functions. The American Academy of Neurology (AAN) is the medical group tasked with identifying standards for implementing the UDDA definition in medical practice. The most recent, 2010, guidelines require findings of coma with a known cause, absence of brain stem reflexes, and apnea. Unlike the prior, 1995, guidelines the 2010 guidelines require the absence of respiratory drive after a CO challenge and a determination that brain cessation is “irreversible,” by excluding mimicking medical conditions and observing the patient for a period of time to exclude the possibility of recovery (Widjicks et al. 2010). In this example, science is brought to bear in an ongoing way, and physicians, rather than religious leaders, laypersons, or family members, are granted special authority to pronounce a patient’s death. Another case in point is the diagnosis of persistent vegetative state. Society recognizes physicians as having exclusive authority to establish diagnostic criteria for this medical condition. After reviewing the best available scientific evidence, a multi-society task force of medical professionals from various groups, including AAN, Child Neurology Society, American Neurological Association, American Association of Neurological Surgeons, and the American Academy of Pediatrics, developed the following definition of the medical aspects of a persistent vegetative state (PVS): [A] clinical condition of complete unawareness of the self and the environment, accompanied by sleep-wake cycles with either complete or partial preservation of hypothalamic and brain-stem autonomic functions. The condition may be transient, marking a stage in the recovery from severe

Respect for Patient Autonomy 147 acute or chronic brain damage or permanent, as a consequence of the failure to recover from such injuries. (Multi-Society Task Force 1994: 1500) The task force goes on to distinguish PVS from related terms and conditions, such as unconsciousness, coma, and vegetative state, and to identify clinical signs and laboratory tests that assist with diagnosis. In many other areas of medicine, the medical profession establishes and implements standards of competent care, defines criteria for diagnosing disease, and identifies evidencebased treatment options. The determination of what falls outside the scope of appropriate medical care is likewise a medical determination, and society gives the medical profession exclusive authority to make this determination. The foregoing analysis suggests that both of the requirements for avoiding exploitation are met. There is mutual agreement among patients and physicians that physicians may render medical decisions and deliver medical services. There is mutual benefit, because patients benefit from having those with relevant knowledge and skills deliver competent care. Properly understood, the practice of medicine and the rendering of medical judgments are authorized in an ongoing way by the whole society. As Misak et al. note, no medical decisions are taken unilaterally or in isolation from a dense network of mostly shared values. Rather than asking when the values of physicians can trump the values of patients, we should be asking how to balance the multitude of values within the context of current societal norms. (Misak et al. 2014: 1670) By contextualizing medical decisions, we expose the false assumptions that sometimes underlie autonomy-based objections to medical futility. Rather than considering conflict cases as a choice between granting patient or physician unilateral authority, we should instead recognize that “Medical decisions are made in the context of an implicit and evolving social contract among patient, physicians, and societies at large” (Misak et al. 2014: 1668). Society assigns physicians the authority and responsibility to determine what constitutes competent medical care, and what represents medically appropriate treatment options. Patients with decisional capacity are ethically and legally entitled to choose from among such options, or refuse all options. However, patients are neither ethically nor legally entitled to demand treatments that fall outside the scope of competent care. In contested cases, where it is clear that a treatment is futile yet patients or surrogates insist on receiving it, the social contract clearly gives physicians not only the authority but also the responsibility to refuse. Physicians are obligated to offer only those interventions that are medically

148

Nancy S. Jecker

indicated—that is, have a reasonable prospect of helping the patient. If a patient or family member disagrees with a physician’s judgment, the ethical principle of autonomy does not authorize acting outside the social contract, or requiring the physician to do whatever the patient or family member demands.

A Response from Opponents of Medical Futility In response to the foregoing analysis it could be objected that in individual cases, patients and families are allowed to challenge professional guidelines and to seek an exception in a particular case. Thus, in the determination of death, cases in California and Nevada have led to exceptions to the general standard of neurological criteria when patients or families hold philosophical or religious views that are at odds with such criteria. In California, Jahi McMath was declared dead following complications from surgery, yet the patient’s mother refused to accept either the determination of death by medical experts or the coroner’s death certificate. She subsequently took her child’s body to New Jersey, a state that allows an exception to the neurological criteria set out in UDDA. New Jersey allows substituting cardiovascular criteria for neurological criteria when a patient’s family has a religious or philosophical basis for requesting an exception. Likewise in the Nebraska case of Aden Hailu, a twenty-year-old patient was declared dead following a heart attack that led to low blood pressure and lack of oxygen to the brain. Refusing to accept the declaration of death, the family insisted on continuation of the ventilator and artificial nutrition and hydration, stating that their daughter was not dead and that the hospital wanted to get rid of her due to money. Similarly, in the Terri Schiavo case in Florida, parents rejected the medical claim that their adult daughter met the definition of “persistent vegetative state,” insisting that the hospital continue artificial nutrition and hydration against the wishes of the patient’s husband and legal guardian. The parents insisted that their daughter could track objects with her eyes and listen to commands, although an autopsy later found that the patient lacked an optic nerve (meaning that she could not see) and 50% of her brain had been reabsorbed (meaning that she was incapable of following directions) (Klugman 2015). In all three of these cases, medical determinations properly rest with medical practitioners, rather than with judges, elected officials, or lawyers. Physicians have the medical training and expertise to determine death and diagnose medical conditions; judges, elected officials, and lawyers do not (Klugman 2015: 2). Involvement of a hospital ethics committee can be helpful in such cases by facilitating conversations among family members and the health care team, or by providing a fair and public process for resolution of difficult cases. The death of a loved one is inevitably difficult to understand and accept, and providers should respond with compassion, offering grief counseling and support

Respect for Patient Autonomy 149 to family members. Yet it should be clear that a physician who continues medical interventions for a deceased patient at the request of family members is not helping the family or benefiting the patient. Instead, the physician’s actions make it more difficult for the family to come to terms with the loss of a loved one. In these cases, physicians can and should find other, better, ways of caring for patients and families.

Conclusion The analysis sketched in the preceding pages shows that respect for patient autonomy is consistent with allowing physicians to withhold and withdraw futile medical interventions. Properly understood, autonomy requires that competent adults be allowed to choose from among medically appropriate options, or reject all options. Physician authority to render medical judgments is ethically supported by the social agreement that exists among the medical profession and members of society. This agreement is mutually beneficial to physicians and society. Rather than falling outside societal norms, the principle of respect for patient autonomy is properly understood as a principle situated within social contexts. Autonomy is not a trump card that patients can play to obtain interventions that fall outside the range of competent care. Nor is physician’s power over patients absolute or unilateral. Instead, medical decisions reflect physicians’ role-related responsibilities to benefit patients and the society.

Acknowledgments A version of this chapter was presented in 2015 at Tsinghua University Department of Philosophy, Beijing; I am grateful to Professor Wei Xiao and graduate students in the Department of Philosophy for helpful suggestions that advanced my thinking on this topic.

Related Topics Chapter 1, “Is It Possible to Be Better Off Dead?” by Geoffrey Scarre Chapter 5, “Defining Death” by John K. Davis Chapter 7, “Deciding for the Incompetent” by Eric Vogelstein

References American Medical Association (1997) “Opinion 2.037—Medical Futility in End-ofLife Care,” URL = . Beauchamp, T. and T. Childress (2012) Principles of Biomedical Ethics (7th Edition), Oxford: Oxford University Press.

150

Nancy S. Jecker

Bosslet, G.T., T.M. Pope, G.D. Rubenfeld, B. Lo, R.D. Truog, C.H. Rushton, J.R. Curtis, D.W. Ford, M. Osborne, C. Misak, D.H. Au, E. Azoulay, B. Brody, B.G. Fahy, J.B. Hall, J. Kesecioglu, A.A. Kon, K.O. Lindell and D.B. White, on behalf of the American Thoracic Society ad hoc Committee on Futile and Potentially Inappropriate Care (2015) “An Official ATS/AACN/ESICM/SCCM Policy Statement: Responding to Requests for Potentially Inappropriate Treatments in Intensive Care Units,” American Journal of Respiratory and Critical Care Medicine 191(11): 1318–1330. Brison, S. (2003) Aftermath: Violence and the Remaking of the Self, Princeton, NJ: Princeton University Press. Brody, H. and F.G. Miller (1998) “The Internal Morality of Medicine,” Journal of Medicine and Philosophy 23(4): 384–410. Dworkin, G. (1970) “Acting Freely,” Nous 4: 367–383. Frankfurt, H. (1988) “Freedom of the Will and the Concept of a Person,” in G. Frankfurt (ed.) The Importance of What We Care About, Cambridge: Cambridge University Press. Freidson, E. (1970) Professional Dominance, New York: Atherton Press, 11–25. Jecker, N.S. (1995) “Is Refusal of Futile Treatment Unjustified Paternalism?” Journal of Clinical Ethics 6(2): 133–137. Jecker, N.S. (2005) “Health Care Reform: What History Doesn’t Teach,” Theoretical Medicine and Bioethics 26: 277–305. Jecker, N.S. and R.A. Pearlman (1992) “Medical Futility: Who Decides?” Archives of Internal Medicine 152: 1140–1144. Jecker, N.S. and L.J. Schneiderman (1995a) “Judging Medical Futility: An Ethical Analysis of Medical Power and Responsibility,” Cambridge Quarterly of Healthcare Ethics 4(1): 23–35. Jecker, N.S. and L.J. Schneiderman (1995b) “When Families Request that ‘Everything Possible’ Be Done,” Journal of Medicine and Philosophy 20: 145–163. Klugman, C. (2015) “The Bell Tolls for Death by Neurologic Criteria,” URL = . Miller, F.G. and H. Brody (2001) “The Internal Morality of Medicine: An Evolutionary Perspective,” Journal of Medicine and Philosophy 26(6): 581–599. Misak, C.J., D.B. White and R.D. Truog (2014) “Medical Futility: A New Look at an Old Problem,” Chest 146(6): 1667–1672. Multi-Society Task Force (1994) “Medical Aspects of the Persistent Vegetative State” (First of Two Parts), New England Journal of Medicine 330: 1499–1508. Noggle, R. (2005) “Autonomy and the Paradox of Self-Creation,” in J.S. Taylor (ed.) Personal Autonomy: New Essays on Personal Autonomy and Its Role in Contemporary Moral Philosophy, Cambridge: Cambridge University Press, 87–108. Schneiderman, L.J., K. Faber-Langendoen and N.S. Jecker (1994) “Beyond Futility to an Ethic of Care,” American Journal of Medicine 96: 110–114. Schneiderman, L.J. and N.J. Jecker (2011) Wrong Medicine: Doctors, Patients and Futile Treatment (2nd Edition), Baltimore: Johns Hopkins University Press. Schneiderman, L.J., N.S. Jecker and A.R. Jonsen (1990) “Medical Futility: Its Meaning and Ethical Implications,” Annals of Internal Medicine 112(12): 949–954. Schneiderman, L.J., N.S. Jecker and A.R. Jonsen (1996a) “Medical Futility: Response to Critiques,” Annals of Internal Medicine 125(8): 669–674. Silen, et al. (2008) “Workplace Distress and Ethical Dilemmas in Neuroscience Nursing,” Journal of Neuroscience Nursing 40(4): 222–231.

Respect for Patient Autonomy 151 Sirovich, B.E., S. Woloshin and L.M. Schwartz (2011) “Too Little? Too Much? Primary Care Physicians’ Views on US Health Care,” Archives of Internal Medicine 171: 1582–1585. Solomon, M.Z., L. O’Donnell, B. Jennings, V. Guilfoy, S.M. Wolf, K. Nolan, R. Jackson, D. Koch-Weser and S. Donnelley (1993) “Decisions Near the End of Life: Professional Views on Life-Sustaining Treatments,” American Journal of Public Health 83(1): 14–23. Veatch, R.M. and C.M. Spicer (1992) “Medically Futile Care: The Role of the Physician in Setting Limits,” American Journal of Law and Medicine 18(1–2): 17. Wertheimer, A. (1999) Exploitation, Princeton: Princeton University Press. Wertheimer, A. (2012) “Exploitation,” Stanford Encyclopedia of Philosophy, URL = . Widjicks, E.F.M., P.N. Varelas, G.S. Gronseth and D.M. Greer (2010) “Evidence-based Guidelines Update: Determining Brain Death in Adults,” Neurology 74: 1911–1918.

Part III

How to End Life

10 Refusing Lifesaving Medical Treatment and Food and Water by Mouth Paul T. Menzel

Summary In all but the most unusual circumstances, a competent person has the moral and legal right to refuse lifesaving medical treatment (LSMT). Voluntary refusal of care by an informed and competent patient is thought to have very strong moral protection—others would be committing a grave moral wrong if they imposed treatment. What other sorts of refusal by which people can hasten their deaths also gain moral protection? One is refusing food and water— voluntarily stopping eating and drinking (VSED). Others are the extension of refusing LSMT to circumstances of incompetence by advance directive (AD), and a similar extension of VSED by AD. As one moves beyond the refusal of lifesaving medical treatment, and beyond refusal by currently competent patients, how much of the moral justification of the original basic right of a competent patient to refuse LSMT is retained—by VSED, by withholding or withdrawing LSMT according to an AD, and by withholding oral feeding by AD, respectively? Comparative assessment reveals notable moral differences. A comparative assessment also reveals that the moral case for these extensions is relatively strong, allowing people to control the end of their lives not only by refusing LSMT but also by VSED and by advance directives.

Introduction The right to refuse lifesaving medical treatment is not only a legal right but also a moral right. The obligations that others have to respect it are moral, not merely legal. The scope of the care that patients may refuse within this right includes both acute lifesaving care and the continual care referred to as medical “life support.” It encompasses withdrawing care midstream, as well as withholding it initially, and the right holds regardless of whether the patient’s illness is “terminal.” The additional means of hastening death on which I will focus require extensive justification. Paced and managed well with palliative care, VSED

156

Paul T. Menzel

is usually a peaceful and comfortable process of dying (Byock 1995; Menzel and Chandler-Cramer 2014: 36 and ftns 14–22). When paired with the right to refuse the medically delivered food and water that family or caregivers may want to provide in response to a refusal to eat and drink, refusing to eat and drink ensures death. It also unquestionably causes death—it does not merely “allow death to come” or “let nature take its course.” These are notable differences between VSED and refusing LSMT. With refusing LSMT, even when one correctly judges that a medical treatment is truly lifesaving—that without it, the person will die—one does not know this with absolute certainty. It remains possible that the patient will live despite refusing the treatment that’s seen as life-sustaining. No one, however, can live for long without food and water. These differences of certainty and causal role are taken by some to claim that any alleged right to VSED as morally much more problematic than the right to refuse treatment. VSED’s death-ensuring and death-causing qualities, for example, arguably make it “suicide”; refusing lifesaving treatment only “lets nature take its course.” Similarly, the step from refusing LSMT when people are competent to refusing it for them after they wrote an AD may require additional justification. Patient autonomy is thought to justify the first; with incompetence that is lost. The patient doesn’t remotely understand autonomy and does not care about it. So what gives the choices of the previously competent patient (the “then-self ”) authority over the “best interest” of the currently incompetent patient (the “now-self ”), especially if the competent patient is never going to reemerge? We may initially think that this has an easy answer—it’s still her life, so why should not her choices, expressed in her AD, control what happens to her? Ultimately that point may carry the day, but not without qualifying considerations. VSED by AD has the same initially compelling logic as refusing LSMT by AD: if people have a right voluntarily to refuse oral feeding, and if no one, merely by becoming incompetent, loses her basic rights—they just have to be exercised by others—then it would seem that anyone has as much right to stop eating and drinking (SED) by clear and well-informed AD as they have a right to VSED when competent. Here, though, not only does the then-self/now-self quandary emerge, as it can with any AD, but also an additional question confronts us: should caregivers be bound by a previous directive when the care at issue is basic personal care like feeding, not medical treatment? An array of considerations about human caregiving come into play with SED by AD that are not present when medical treatment is refused per AD. On the other hand, SED by AD cannot just be brushed aside as too problematic from the start to warrant moral and legal recognition. It is the only way short of preemptive action—suicide or VSED done when one is still decisive but before the point when one thinks it best not to live—by which people can assuredly avoid

Refusing LSMT and Food and Water by Mouth 157 living into long years of severe dementia. Living into such years is a prospect many want to avoid more than suffering for a few months with terminal illness. Without venturing further, into the even more controversial arena of physicianassisted suicide and euthanasia, a moral assessment of these extensions of the right to refuse LSMT is rich and interesting.

Refusing Lifesaving Medical Treatment Administering medical treatment to competent persons without their informed consent is, in legal terms, battery—the unauthorized harmful or offensive invasion of the body or contact with it. One may wonder why, in cases where we have every reason to think a medical treatment is not harmful (in fact greatly helpful), it is still offensive to administer it without patient consent. The most plausible explanation, perhaps, is that to touch or invade the bodies of persons capable of consent without their consent is ipso facto morally offensive. Those with capacity should not be subjected to even helpful treatment without being asked, for two reasons (Cantor 2006: 106). First, bodily integrity: treatment is typically an invasion of the body, and if anything should be within persons’ province of control, their own bodies should be. Second, patient autonomy: they can be asked, they are capable of responding if they are asked, and therefore they should be asked. A stubborn, basic moral judgment is involved here: beings with the capacity to consent should be asked, and their autonomous choice should be respected. If we do not ask, we do not treat them as persons with agency, which they are. And clearly, asking is not genuine if responses do not matter. The right of informed consent to medical treatment means nothing if not the right to refuse. There are other reasons for the right to refuse treatment beyond these that are not as immediate and involve more contingent, empirical claims: (a) treatments may often be more effective if the patient is involved as a decisionmaker; (b) treatments often involve risks, even risk of death, and those should be evaluated by the person who is taking the risk; and (c) given the risks involved and the inevitability that treatment will sometimes harm, the trust of patients in their physicians and caregivers will erode if the responsibility for making treatment decisions is not shared. For an array of strong reasons, therefore, the right to refuse treatment has a solid ethical foundation. That said, many questions still arise about how far the right should extend. It has its strongest justification in situations of terminal illness and continual pain. Who better to discern the value of the remaining life with suffering than the person who is experiencing it? But how much of this point changes when the person who would refuse lifesaving care is not terminally ill? On the one hand, very little. If anything, in non-terminal illness suffering may be longer and all the worse to endure. Moreover, while an illness may not

158

Paul T. Menzel

be officially “terminal” in the sense of death being likely within six months, life may still be on a steady downhill progression in which the patient is losing the things she most associates with being “alive” (e.g., in progressive dementia). On the other hand, what if the refusal occurs in a situation that is not terminal at all—an anorexic 25-year-old refuses tube feeding, for example? May we add a tasteless liquid to her drink to revive her appetite? At least two further factors in such non-terminal situations are relevant to how strongly we stick to a strong right of refusal: the degree of bodily invasion and the potential for “retrospective consent.” Degree of invasion matters: surgery and tube feeding are different than a tasteless pill or food additive. With the former it is especially objectionable to treat against the patient’s refusal. The prospect of retrospective consent may also affect our judgment. If providers have very good reason to think that a burn patient, for example, will end up saying, after recovery, that it was right to give him lifesaving treatment though now he refuses it, we may think treating him without his current consent is justified.1 Such justification is not straight hard paternalism, claiming the treatment has so much benefit for the patient that we simply override his autonomous choice, but an appeal to the patient’s own eventual consent. Overriding his refusal for this reason arguably still respects his autonomy. Even if low degree of bodily invasion and likelihood of retrospective consent may occasionally justify treating despite a patient’s refusal, however, the default rule remains the right to refuse. The burden of proof falls not on the patient to show that consent is required, but on any provider or family wanting to treat to show that limitations on the right of refusal are justified in the instant case because degree of bodily invasion is minor or likelihood of retrospective consent is very high. Neither consideration is reason to dispute the basic right to refuse treatment. They are only reason to limit the right, not reason to say people do not have it. Moreover, probably the single most frequent good reason for treating patients despite their refusal is doubts about their capacity to consent. This exception is not at all a compromise of the right to refuse LSMT but part of its very nature. We must not simply turn our disagreements with patients into claims that they lack adequate capacity; mistaken perceptions and difficult to defend value judgments are not by themselves sufficient reasons to claim lack of capacity. But as long as we assess capacity on grounds other than our own normative judgment about the treatment, we respect the right to refuse treatment when we do not follow the choice of a patient who lacks capacity.

Voluntary Stopping Eating and Drinking (VSED) In respect to rights, how different, morally, is VSED from refusing lifesaving medical treatment? If people have a moral right to refuse treatment, do they also have a right to stop eating and drinking?

Refusing LSMT and Food and Water by Mouth 159 Legally, in the United States at least, in effect they do, though the case law is less clear (Cantor 2006: 109–112; Pope and Anderson 2011). By one explanation, the right to VSED even derives from the right to refuse lifesaving treatment. VSED, of course, is not per se the refusal of lifesaving treatment. The right to refuse treatment reflects not only the right of informed consent but also a focus on bodily intrusion. Providing food and water by mouth, unlike LSMT, is not an intrusion. Yet a moment’s reflection on what happens if we do not respect the decision of a person to embark on VSED reveals how tightly connected the right to it is to the right to refuse medical treatment. The only way to keep alive someone who steadfastly refuses to eat and drink is to insert some sort of feeding tube, using whatever force is necessary—coercive medical treatment. In practice the right to VSED is thus a right not to be force-fed, an instance of the right to refuse lifesaving treatment. The prospect of physically restraining competent patients to insert feeding tubes against their will is so demeaning and repugnant that we quickly acknowledge a right to VSED (Cantor 2006: 112). Another explanation of the right is more direct. U.S. Supreme Court chief justice Rehnquist surmised that it is a direct manifestation of bodily integrity: “It seems odd that your bodily integrity . . . [would be] violated by sticking a needle in your arm but not by sticking a spoon in your mouth.”2 Without the consent of the patient, neither is a warranted intrusion on the body. An entirely different view of VSED is taken by the U.S. Roman Catholic bishops. It begins with the long-standing Catholic position that intentionally ending life—by suicide or euthanasia, for example—is impermissible. Acting to intentionally end life—killing—includes not only active killing but also some omissions. Often the purpose of an omission like refusing treatment is “to relieve the patient of a particular procedure that was of limited usefulness . . . or unreasonably burdensome for the patient and the patient’s family or caregivers.” Even if it is nutrition and hydration that are omitted, the omission is then not a decision to kill and should not be equated with suicide. But at other times the harsh reality is that . . . withdrawal of nutrition and hydration . . . directly intend[s] to bring about a patient’s death. . . . Whether orally administered or medically assisted . . . [food and fluids] are sometimes withdrawn not because a patient is dying, but precisely because a patient is not dying (or not dying quickly) and someone believes it would be better if he or she did. . . . We must be sure that it is not our intent to cause the patient’s death—either for its own sake or as a means to achieving some other goal. (U.S. Bishops’ Pro-Life Committee 1992: 37) The bishops often couch their view in terms of what can be “withheld” or “withdrawn,” language that fits medically assisted food and water better than

160

Paul T. Menzel

self-initiated refusal of food and water by mouth. The view is nonetheless clear: in intentionally and certainly causing one’s death, VSED is impermissible suicide. When family and caregivers cooperate with it and facilitate it with good palliative care, therefore, they assist a suicide. Neither the act itself nor the assistance is morally permissible. Outside a framework such as Roman Catholic moral theology where intentional killing is inherently wrong, however, the matter can look very different. Why should hastening death intentionally, and with certainty ensuring it, and being the primary cause of it, make the act morally wrong if what it brings about, death sooner rather than later, is a good thing compared to living with great suffering or a chronic condition like severe dementia? These characteristics of VSED’s relationship to death—intention, certainty, causation—may make VSED “suicide,” not mere “aid-in-dying” or “letting death come,” but do they make it wrong? Intentionally ensuring and causing a good result would, if anything, be better, one would think, than bringing about the result less directly with less certainty. As long as SED is voluntary and not foisted on patients by others, why is it not within their moral prerogative as much as refusing LSMT? To be sure, either can be done foolishly or wrongly, but there is no more reason to think that questionable or mistaken exercise of a right occurs more often with VSED than it does with refusing lifesaving treatment. In any case, these moral rights are presumably “strong rights”—protections from the interference of others even when a person does not do the right thing in choosing to exercise the right (Dworkin 1977: 188–192). Those who believe that VSED can be suicide and that suicide is morally wrong still need to provide arguments that people should not be at liberty to voluntarily stop eating and drinking. Nonetheless, perhaps the strongest argument against the legal and moral acceptability of VSED is connected with its character as suicide. Though in U.S. law suicide is not a crime (the person who accomplishes or unsuccessfully attempts it is not punished), but it is not a right, either: others may stop it, and in most jurisdictions others are not allowed to assist it.3 The resulting complication for VSED is clear. For VSED to be a comfortable path to death, it should be medically managed and accompanied by palliative care. If descriptively VSED is “suicide,” not merely letting death come, comfortable and humane VSED will then often involve assisted suicide. This poses little moral difficulty if there is nothing wrong per se with assisting suicide, but it complicates the legal and moral situation. In practice, however, doubts about the right to VSED because it involves assisted suicide are likely overcome by our negative reaction to the realistic alternatives: by our “revulsion at the prospect of physically overcoming and restraining people” to feed them against their will (Cantor 2006: 112), or by the hypocrisy of letting people embark on hastening death by VSED but then pulling them back when they become incapable of resistance, or by the heartlessness of not providing them with palliative care

Refusing LSMT and Food and Water by Mouth 161 when they are carrying it out anyhow. Medically assisting VSED is thus likely to be regarded as permissible even if VSED is suicide and assisting suicide is not generally permissible. Such permissibility is crucial if VSED is going to play its important role in assuring a hastened death in the many jurisdictions where physician-assisted death by direct lethal means is illegal. Adequate justification for a right to VSED is not provided by simply reiterating the reasons that ground the right to refuse LSMT. Sufficient additional reasons are available, however, in various moral frameworks: if suicide is not inherently wrong; if VSED is not suicide; or if, though it is, assisting it is still permissible.

Refusing Medical Treatment by Advance Directive If a patient lacks the mental capacity to make medical decisions, the right to refuse LSMT does not apply. This limitation is a constitutive part of the right itself. When we do not follow the wishes of patients with insufficient capacity, we have not disrespected their right to refuse treatment. If, though, someone now incapacitated previously had such capacity, and if when she did, she expressed distinct and informed preferences about treatments for conditions she could one day be in, we tend to look past her current incapacity and regard her previous directive as tantamount to her decision for her current treatment. The life that will be affected, after all, is still her life, regardless of whether she is competent now. To treat her as a person we need to look to the choices she made when she could make them. I shall refer to this as the precedent autonomy view of ADs. To see that precedent autonomy is respected, people write directives for a range of medical care, including LSMT. ADs can be substantive about what care, when, should be refused, and what care, when, should not be—often referred to as “living wills.” Or they can be procedural, appointing someone to be one’s proxy decision-maker—appointments with “durable power of attorney.” The moral force of any substantive directive must track the same basic considerations that give force to a competent patient’s choice: (a) how informed the person is about the medical condition to which she understands her treatment would apply, about her prognosis with and without the treatment, about its risks, and so forth; (b) how adequate her mental capacity is to absorb such information and make decisions based on it, referred to as her “capacity”; and (c) how voluntary and free from pressure and coercion her choices are. These same three dimensions affect the normative force of choices expressed in an AD. When by directive a person refuses a kind of treatment, to what conditions, with what range of prognoses, does she understand this refusal to apply? Was she knowledgeable about those when she wrote the directive? Did she

162

Paul T. Menzel

have the capacity to comprehend them? Was her directive made voluntarily and without coercive pressure? The clearer, more knowledgeable, and more unquestionably voluntary a directive is, the greater its moral force for later caregivers and decision-makers. We cannot justify ADs on the basis of precedent autonomy if they do not meet the same standards that contemporaneous choices have to meet to command respect for patient autonomy. We can summarize these considerations: a directive has normative force in a given situation if it is “valid”—made voluntarily and with sufficient understanding of the things to which it speaks—and it is now “applicable”—the conditions to which the person intended it to apply now in fact obtain. A good directive must be clear and substantive enough to later be applicable. If the directive conveys little understanding of and direction about the conditions for which this treatment later might be refused, how can it be an appropriate substitute for the informed choice the person would make if she were still competent? Few directives, probably, contain enough clear content to make them truly applicable. If one is serious about one’s substantive directive, it is thus virtually always best to supplement it with another directive appointing a proxy, someone whom the patient trusts to best understand her values and wishes for the later situations that she herself is not likely to understand at the time she writes the directive. These difficulties in meeting a reasonably high standard of validity and applicability are ongoing, everyday challenges to the authority of ADs in a precedent autonomy framework. Perhaps we tend not to be very demanding of directives on these scores because we stubbornly hold two moral beliefs: that people do not lose their rights when they become incompetent, and that the life they live in incompetence is still their life. Thus, despite the difficulties, we reach back to their earlier will, adding to this the most appropriate substitute decision-maker to fill in the inevitable substantive gaps in any actual directive. Even if we deal generously with ADs in this way, they still face another fundamental difficulty, the “then-self ” vs. “now-self ” problem. The directive’s competent writer (often referred to as the “then-self ”) may request, for example, that no treatment that would prolong life in a significantly diminished or painful condition be provided because she does not want to live if she has to “live like that.” Or she may not want life to be prolonged if she were close to death or treatment were unlikely to lead to significant recovery, in which case she would be “ready to accept death.” Yet precisely the situation that motivates writing an AD and brings it into operation—incompetence—may also affect a person’s values, attitudes, and desires. The now incompetent individual (the “now-self ”) may not regard “life like that” as unbearable, or she may not be “ready to accept death” if she no longer has the cognitive capacity required for such a mental state. In the form of herself as she is now, she is not autonomous, does not care about autonomy, is not aware of the directive she wrote, and may

Refusing LSMT and Food and Water by Mouth 163 no longer value many of the things about life that were the basis for writing her directive. So why should her directive now be followed? Some critics—most notably Rebecca Dresser and John Robertson (1989)— conclude that the directive is irrelevant. We should treat the now incompetent patient solely on the basis of her current best interest, not her previous wishes, even if her directive clearly and accurately expressed those wishes. “Precedent autonomy” is a misnomer: the moral force of autonomy is lost when autonomy is no longer present. To respond that this would fail to respect the patient’s autonomy manifested in the directive is hardly persuasive, for what Dresser and Robertson are disputing is the relevance to the current patient of that autonomy. At this point in the argument, ADs’ defenders may choose to stop citing autonomy as what still outweighs best interests. They can make a very different claim, expanding the notion of what constitutes the patient’s interests— namely, that those interests are even now partly defined by what the person said earlier (Dworkin 1993: 201–232). If this is the way debate proceeds, the focus may still touch on the tussle between autonomy and best interest, but it will also turn on whether current best interest plausibly includes elements from an earlier directive. Much of the debate since Dresser and Robertson posed their famous challenge has been focused on not only the moral relationship between autonomy and best interest but also how best interest itself should be conceived. A strong response to the Dresser/Robertson view is Nancy Rhoden’s (1990), which I will refer to as “the Rhoden response.” The Dresser/Robertson position, Rhoden notes, amounts to treating previously competent patients no differently than patients who were never competent: both are to be treated only on the basis of best interest. But should we ever treat a person who has had a long life as a competent, full, and vital person as if she had never been competent and deserves no respect for having been an autonomous person? We should not. Rhoden (1990: 860) elaborates: Something is wrong . . . when we treat formerly competent patients as if they were never competent. Someone who makes a prior directive sees herself as the unified subject of a human life. She sees her concern for her body, her goals, or her family as transcending her incapacity. . . . One . . . component of treating persons with respect [is] that we view them as they view themselves. If we are to do this, we must not ignore their prior choices and values. If this is correct, ADs must have some authority.4 Dresser and Robertson are right in noting that the patients incapacitated at the time we would act on their ADs do not understand autonomy, exemplify it, or care about it, but that simply

164

Paul T. Menzel

does not mean we may ignore previous wishes. It is not autonomy per se that we are dealing with here, but whole persons with narrative identity in their lives. Norman Cantor states the view this way: “A person, by nurturing and developing a body, character, and relationships, has earned a prerogative to shape a life narrative—including the medical fate of a succeeding incompetent persona” (Cantor 2015: 5). Ronald Dworkin attempts to capture essentially this claim within the framework of best interest, not precedent autonomy, through the notion of “critical interests” (Dworkin 1993: 201–232). Yes, a person’s “experiential” interests can change significantly with chronic incompetence, but a person’s overall best interest is still connected with previous choices. Competent, autonomous persons have beliefs, desires, and values that form something more than experiential interests: “critical” interests based on the “second-order” desires that emerge in reflecting on events that just happen and first-order desires one just has. They are based on evaluations that are more than mere preferences—convictions, for example, about “what helps to make . . . [one’s] life good on the whole” (Dworkin 1993: 201). For some, that involves how one’s life should end. Unlike experiential interests, critical interests can exist even after one no longer experiences them. If someone cares about what happens to his property, his family, or his reputation after he dies, for example, he has critical interests in these things—in these things after he has died, when he will have no experience of whether his interest is thwarted or satisfied. This is also true for any belief he might have that his life as a whole would be better if it did not extend into years of severe dementia, for example. Violating his directive would harm him just as unfairly tarnishing his reputation after death would, though he will never be aware of either. For incompetent persons experiential and critical interests can conflict. Imagine, for example, James at the age of 90. He has for years been gradually declining, physically and mentally, in unsurprising ways. After several minor strokes in his late 80s, a severe stroke six months ago paralyzed him on one side and left him barely capable of verbal or facial response. In his recovery, he has now apparently plateaued. He has regained enough physical strength and control to walk short distances with a walker and slowly feed himself, though he usually ambulates by wheelchair and requires spoon-feeding assistance. With his stroke-induced dementia he is able to utter no more than a few words. He occasionally seems to recognize his most frequent caregivers, though it is not clear he recognizes them as particular persons. He does recognize his wife or son as his wife or son. He needs help getting dressed, going to bed, and with most toileting. Despite all this he is consistently good-natured, smiling when someone looks at him directly and speaks kindly, and making soft appreciative sounds when someone clasps his hand or arm. He can carry on no conversation and has seen his social relationships, on which he previously thrived, dissipate.

Refusing LSMT and Food and Water by Mouth 165 James has an AD, reiterated in writing a few months before his most recent stroke: that once he cannot recognize his closest friends and loved ones, converse with them as specific individuals, or occupy himself with activities such as TV or Word Find, he should receive no lifesaving treatment of any sort, including antibiotics. In the directive he briefly explains his reasons. Now he has contracted pneumonia, readily curable with standard antibiotics, but from which he will undoubtedly die if not aggressively treated. Applied to the current situation, his directive clearly tells us not to provide the antibiotic. Yet he is not suffering or in pain, and he experiences some real though very modest enjoyment of life—that is, he has some experiential interest in surviving, though if we depart from his directive, he will not notice that we have. In regard to experiential interests, “best interest” would seem to call for treating his pneumonia. Does he not still, however, have a critical interest in ending his life in this condition, an interest he expressed in his AD? Unless we are willing to end up considering and treating him no differently than we would if he had never been competent, we must say he does. On the assumption, then, that he still has a critical interest in his life not continuing long into a condition like his current state, that interest may constitute enough of his overall interest to have us conclude that it is still in his current best interest to follow his directive. Call this the “balancing” or “sliding scale” model of handling the tension between thenself and now-self: weigh up together the person’s respective experiential and critical interests (Menzel and Steinbock 2013: 495–496; Menzel and ChandlerCramer 2014: 28–29). The interest of the now-self includes both. This is only the beginning of an analysis of how to resolve the challenge of the then-self vs. now-self tension inherent in ADs. A stronger view that valid and applicable directives have dominant authority even within a bestinterest framework is defended by Ronald Dworkin (1993: 220–232), claiming that critical interests always trump experiential ones. An equally strong version of the authority of ADs is defended by Norman Cantor (2015: 4–5) on the basis of precedent autonomy, not subtle best interest. And the DresserRobertson view, even if it is ultimately implausible given Nancy Rhoden’s response, remains appealing to caregivers whose professional role is caring for the patient immediately in front of them. ADs to refuse LSMT are both compelling and problematic. They morally command our attention and support, but their discerning application is far from simple. Stopping Eating and Drinking by Advance Directive (SED by AD) The moral prerogative of people to refuse LSMT by AD is ethically more complex than refusing treatment as a contemporaneously competent patient. VSED

166

Paul T. Menzel

by AD is more complex yet. Still, a directive to stop eating and drinking may have normative force if the circumstances are right. All the characteristics required to justify following a directive to refuse LSMT will apply to a directive for SED. First, it needs to convey clearly the condition(s) in which feeding is to be stopped. That could be, for example, a defined stage on dementia’s Functional Assessment Staging Test scale (Reisberg 1988; Menzel and Chandler-Cramer 2014: 33). Second, the person writing the directive must be reasonably knowledgeable about what SED involves, about the likely course of decline it causes, and about the medical situations that in the person’s directive he or she says it should apply to. Third, the directive must have been made voluntarily, without pressure or coercion. The last two characteristics are required for the directive to be valid; the first is necessary for it to be applicable. Assuming that the directive is valid and applicable, implementation will still be vulnerable to the then-self/now-self problem that always potentially confronts ADs. We must have a morally cogent strategy for handling this problem, whether we are implementing a directive for SED or of any other sort. For example, if we are operating within a framework of current patient best interest, we may see the patient as having both a modest experiential interest in surviving and a strong critical interest in not living longer, and we would weigh up those competing interests. Even if the basic then-self/now-self problem about survival is thus resolved, however, SED raises further problems. I will focus on two: a potential desire to eat that is more explicit than the typical experiential interest in survival, and the fact that food and water are basic personal care that caregivers may have a more absolute obligation to provide than medical treatment. (1) Suppose that James’s directive, discussed earlier in regard to refusing LSMT, also asks that his eating and drinking be stopped when he has reached a certain stage of dementia. But also suppose that when that time comes and we do not feed him, he gets exercised. When not brought to his eating table, he grimaces, looks puzzled, holds out his hand, and utters apparently pleading sounds. His behavior is not likely a change of mind about his directive; he is not cognitively capable of that (see Chapter 8, Menzel, “Change of Mind: An Issue for Advance Directives”). There is no reason to think he has any greater experiential interest in surviving than the James discussed earlier in regard to refusing lifesaving medical care; since he does not anticipate being alive tomorrow, and tomorrow he will not remember have survived from today, survival has little subjective value for him. With a clear and knowledgeable directive, moreover, he still has the strong critical interest in not living further into dementia as the previous James had. Yet here he is, evidently wanting to eat. Even if he does not have much experiential interest in survival, he has a direct immediate interest in being fed. It is difficult to know how to weigh this interest with his others; seeing our way to a conclusion about his overall best interest becomes more difficult.

Refusing LSMT and Food and Water by Mouth 167 If James did not address this kind of complication in his directive, we cannot reasonably infer that he would still want us to follow it and not be fed. We might then make a “not yet” decision: don’t stop food and drink yet, but still pay close attention to his directive, knowing that as his dementia continues to deepen he will undoubtedly reach a point fairly soon where he will not have much if any desire for food (Menzel and Chandler-Cramer 2014: 34). Suppose, on the other hand, that James did address this kind of situation in his AD, requesting that we administer distress-relieving drugs or appetite suppressants and proceed with withholding food and water even if he then still gives some signs of wanting food. He made clear both his wish that he not live long into severe dementia and his wish to accomplish that by SED if necessary, even if he expresses a desire to eat. Although the then-self/now-self problem will still need to be addressed, we can address it with the same consideration of critical interests and narrative identity that contribute to the authority of any AD. If the directive itself has clarified what to do in the circumstance where he still desires to eat, what difference should it make that it is SED, not just refusal of LSMT, that is requested? (2) There is still another complication with VSED by AD, however: should a person have the prerogative of directing later caregivers not to provide basic personal care, not just lifesaving medical care? Providing food and water by mouth and manually assisting with it are not medical care. If the same considerations of bodily intrusion that make coercive imposition of medical care so objectionable do not so clearly apply to feeding, does refusal of food and water by mouth fall within the proper scope of an AD? At least one major court has said that it does not (Court of Appeal for British Columbia 2015; Pope 2015). In relation both to personal autonomy and to Rhoden’s objection to treating a person who has previously been competent no differently than a never-competent person, however, it is hard to see why personal care should be any less within a person’s control than medical care. It’s care for that person all the same, and it’s that person’s life. ADs can address all sorts of considerations about mental deficiencies and mental health care, for example, as clearly and elaborately as ADs for regular medical care (Brodoff 2010). ADs for SED raise additional moral complications not already inherent in ADs to refuse lifesaving treatment, but they still have considerable moral pull. People do not generally lose their rights by becoming incompetent, and a person’s moral right to VSED is virtually as firm as her right to refuse LSMT. So why should that right, too, not apply when she becomes incompetent, especially if she left specific instructions about what to do if she expresses a desire to eat when the directive becomes applicable? Perhaps the personal/medical care distinction makes a moral difference because of stronger caregiver obligations, but if that is the case, the difference needs to be justified, not merely asserted.

168

Paul T. Menzel

Conclusion The right of competent patients to refuse lifesaving medical treatment when competent is the anchoring right of patients in end-of-life decisions. Extending this right to VSED requires us to address challenging questions about suicide and assisting suicide that refusing LSMT seldom needs to confront. Extending it to refusing LSMT by AD requires confronting the significant additional problem of conflicts between then-self and now-self. Extending the reach further yet, to SED by AD, can raise even more poignant and difficult then-self/ now-self conflicts, as well as the claim that personal care like being fed is not within the scope of advance directives in the same way as medical care is. Each extension’s ethical challenges can be effectively met by careful distinction and sound argument, but with each extension the challenges must be taken seriously, not brushed aside.

Related Topics Chapter 8, “Change of Mind: An Issue for Advance Directives” by Paul T. Menzel Chapter 11, “Suicide, Physician-Assisted Suicide, the Doing-Allowing Distinction, and Double Effect” by Thomas S. Huddle

Notes 1. The retrospective consent, “he will be glad later” justification infuses much of the reasoning of doctors who treated burn victim Dax Cowart terribly painfully against his wishes in 1973–74. (See Burton 1989; Kliever 1989; Cowart and Burt 1998.) 2. In oral argument in Vacco v. Quill, 521 U.S. 793 (1997) (No. 95–1858) at 13: 39 to 13: 46, available at http://www.oyez.org/cases/1990–1999/1996/1996_95_1858/ argument. As oral argument, Rehnquist’s view here is only dictum, not case law precedent. See Pope and Johnson (2011: 414). 3. Assisted suicide (assisted by physicians) is legal in only five U.S. states (Oregon, Washington, Vermont, Montana, and California) and Switzerland, Belgium, and The Netherlands, and there it is typically only in restricted circumstances, such as terminal illness or unbearable suffering. The exception is Switzerland, where neither terminal illness nor a minimum level of suffering is required; the only restriction is that the person assisting must be acting unselfishly. See Lewis (2015). 4. In speaking of ADs’ “authority” I do not intend to accord them anything like final control. It means only that they should be respected as having normative force and therefore a substantial role in decision-making.

References Brodoff, L. (2010) “Planning for Alzheimer’s Disease with Mental Health Advance Directives,” The Elder Law Journal 17: 239–308. Burton, K. (1989) “Dax’s Case as It Happened,” in L.D. Kliever (ed.) Essays in Medical Ethics and Human Meaning. Dallas, TX: Southern Methodist University Press, 1–12.

Refusing LSMT and Food and Water by Mouth 169 Byock, I. (1995) “Patient Refusal of Nutrition and Hydration: Walking the Ever-Finer Line,” American Journal of Hospice and Palliative Care 8: 8–13. Cantor, N. (2006) “On Hastening Death without Violating Legal and Moral Prohibitions,” Loyola University Chicago Law Journal 37: 101–125. Cantor, N. (2015) “Is It Immoral for Me to Dictate an Accelerated Death for My Future Demented Self ?” posted December 2, 2015, URL = . Court of Appeal for British Columbia (2015) Bentley v. Maplewood Seniors Care Society, 2015 BCSC 91, Appealing British Columbia Supreme Court Decision Bentley v. Maplewood Seniors Care Society, 2014 BCSC 165. Cowart, D. and R. Burt (1998) “Confronting Death: Who Chooses, Who Controls?” Hastings Center Report 28(1): 14–24. Dresser, R. and J.S. Robertson (1989) “Quality of Life and Non-Treatment Decisions for Incompetent Patients,” Law, Medicine and Health Care 17(3): 234–244. Dworkin, R. (1977) “Taking Rights Seriously,” in R. Dworkin (ed.) Taking Rights Seriously, Cambridge, MA: Harvard University Press, 184–205. Dworkin, R. (1993) Life’s Dominion: An Argument about Abortion, Euthanasia and Individual Freedom, New York: Alfred A. Knopf. Kliever, L.D. (ed.) (1989) Dax’s Case: Essays in Medical Ethics and Human Meaning, Dallas: Southern Methodist University Press. Lewis, P. (2015) “Assisted Dying: What Does the Law in Different Countries Say?” BBC World Report, URL = . Menzel, P. and M.C. Chandler-Cramer (2014) “Advance Directives, Dementia, and Withholding Food and Water by Mouth,” Hastings Center Report 44(3): 23–37. Menzel, P. and B. Steinbock (2013) “Advance Directives, Dementia, and PhysicianAssisted Death,” Journal of Law, Medicine and Ethics 41(2): 484–500. Pope, T.M. (2015) “Prospective Autonomy and Ulysses Contracts for VSED,” one of two sections of T.M. Pope and B.J. Richards, “Decision-Making: At the End of Life and the Provision of Pretreatment Advice,” Journal of Bioethical Inquiry, doi:10.1007/s11673-015-9652-6. Pope, T.M. and L.E. Anderson (2011) “Voluntarily Stopping Eating and Drinking: A Legal Treatment Option at the End of Life,” Widener Law Review 17(2): 363–427. Reisberg, B. (1988) “Functional Assessment Staging Test (FAST),” Psychopharmacology Bulletin 24: 653–659. Rhoden, N. (1990) “The Limits of Legal Objectivity,” North Carolina Law Review 68: 845–865. U.S. Bishops’ Pro-Life Committee (1992) “Nutrition and Hydration: Moral and Pastoral Reflections,” The Linacre Quarterly 59(4): 33–49.

Further Reading Buchanan, A.E. and D.W. Brock (1990) Deciding for Others: The Ethics of Surrogate Decision-Making, Cambridge, UK: Cambridge University Press. Cantor, N. (1993) Advance Directives and the Pursuit of Death with Dignity, Bloomington: Indiana University Press. Therein, especially “Testing the Limits of Prospective Autonomy,” Reprinted in B. Steinbock, A. London, and J. Arras (eds.) 2013, Ethical Issues in Modern Medicine (8th Edition), New York: McGraw-Hill, 366–367.

170

Paul T. Menzel

Cebuhar, J.K. (2015) The Practical Guide to Health Care Advance Directives, West Des Moines, IA: Murphy. Fagerlin, A. and C.E. Schneider (2004) “Enough: The Failure of the Living Will,” Hastings Center Report 34(2): 30–42. Sumner, L.W. (2011) Assisted Death: A Study in Ethics and Law, Oxford: Oxford University Press.

11 Suicide, Physician-Assisted Suicide, the Doing-Allowing Distinction, and Double Effect Thomas S. Huddle

Summary It is likely that most philosophers view suicide as a legitimate exercise of autonomy, while most physicians would decline to participate in physicianassisted suicide (PAS). This chapter surveys recent discussion supporting and opposing PAS. Highlighted are two argumentative strategies essential to PAS opponents: the doctrine of doing and allowing (the DDA) and the doctrine of double effect (the DDE). Opponents of PAS claim that the DDA and the DDE can specify a moral difference between PAS and common medical practices. Once PAS is thus set apart, the case against it rests upon its incompatibility with the patient’s moral status or with the physician’s role. Some supporters of PAS deny that the DDA and the DDE can morally distinguish licit and illicit actor contributions to negative outcomes. Without the aid of the DDA and the DDE, PAS must be morally assimilated to withdrawal of support and terminal sedation and opponents of PAS who accept these common medical practices must be deemed to be inconsistent. Other advocates of PAS accept the DDA, the DDE, or both, but deny their bearing on PAS, holding that death for the terminally ill patient requesting it is not a negative outcome. The author concludes that the DDA and the DDE have so far withstood the objections of their critics but that opponents of PAS are now on the defensive insofar as their opposition relies upon asserting the inviolability of human life or an incompatibility of PAS with medical practice.

Introduction While suicide has been of perennial interest to philosophers, the contemporary debate over physician-assisted suicide (PAS) and active euthanasia (AE) dates from the exchange between Glanville Williams and Yale Kamisar in 1958 (Kamisar 1958; Williams 1958). Williams argued for permitting PAS on the grounds of respecting autonomy and relieving suffering. Kamisar argued that legalization would lead to abuses. While their debate was conducted as a

172

Thomas S. Huddle

policy dispute, the issue was taken up by philosophers in the ensuing decades. Suicide continues to be a disputed topic in the philosophical literature, but PAS has become somewhat less controversial there, having served as a kind of limiting case of the broader topic. Whatever philosophers may say about suicide more generally, most view PAS in the right circumstances as permissible and, perhaps, obligatory for physicians. Strikingly, the medical profession in the United States remains mostly at odds with this growing consensus among philosophers (with, of course, notable exceptions to the prevailing group opinion in both groups). In what follows I will briefly discuss the ethics of suicide and then turn to physician-assisted suicide and active euthanasia. My focus will be on two argumentative strategies used by opponents of PAS and AE to morally distinguish those acts from two common medical practices: the withdrawal of life support and terminal sedation. The two strategies respectively invoke the doctrine of doing and allowing (the DDA) and the doctrine of double effect (the DDE). Much of the philosophical dispute over PAS and AE has turned on whether they can be meaningfully distinguished from pervasively accepted medical practices using the DDA and the DDE. If not, PAS opponents would be in the uncomfortable position of denying the permissibility of physician actions that do not differ in principle from other actions that physicians routinely perform.

Suicide Whether it is permissible to take one’s own life has, of course, been debated since ancient times. Much turns on how one chooses to set the boundaries of “suicide.” The more capacious one’s concept of suicide, the more plausible it is to suppose that some suicides, at least, are permissible. Durkheim notoriously believed that suicide could be identified in acts absent an actor’s intention to die. Suicide for Durkheim was any volitional act known by an actor to be likely to result in death, so that the martyr, the soldier braving mortal danger to save a comrade, or the mother sacrificing herself for a child are every bit as much suicides as the despondent man’s successfully intended life-ending overdose (Durkheim 2010: 43). Most philosophers have sought a narrower definition that demands a suicidal actor’s intention of dying, with a broader or narrower sense of intention. For instance, Michael Cholbi suggests that we should consider as suicide intentional self-killing, with a person’s self-killing being intentional “just in case her death has her rational endorsement in the circumstances in which she acts so as to bring about her death” (Cholbi 2011: 28). Rational endorsement of one’s death need not imply the usual sense of intention. That is, it need not imply that one acts having one’s death as a goal. Such a definition would thus include acts resulting in the death of the actor that many would deny to be suicide, such as the soldier covering the grenade with his

Suicide 173 body to save his comrades. This move draws the normative sting of “suicide” and opens the way to the suggestion that suicide may sometimes be permissible. Opponents of suicide’s permissibility are more likely to prefer a narrower definition, according to which admirable acts such as the soldier’s sacrifice of himself for his comrades would be excluded, reserving the suicide designation for acts more narrowly intended as self-destructive. On the latter view, suicide would properly retain its traditional sense as a concept normatively charged in a negative way, analogous to murder. The semantic dispute thus bears on the substantive issue of suicide’s permissibility. Recent arguments for the permissibility of suicide give the most weight to the importance of respecting rational autonomy, given that choosing to end one’s life may sometimes be a rational choice. Criteria for judging a suicidal choice to be rational vary from relatively demanding accounts, according to which death must be not only “what the patient really wants, at the deepest, most reflective level” (Battin 2003: 697) but also in her interest as judged independently of her wants, to accounts that would allow a choice of death to be permissible even if it were against the chooser’s interest so long as the choice was competently made (Cholbi 2012: 288–292). Opposition to the permissibility of suicide often appeals to the sanctity of life, the position that human life is intrinsically valuable and, properly, inviolable apart from some carefully managed exceptions (often including selfdefense, capital punishment, and just war). The sanctity of life position is often characterized by advocates of suicide’s permissibility as a religious doctrine (Battin 2005: 94). While it is certainly associated with religion (Ramsey 1968), opponents point out that the conviction that human life is valuable in itself and ought not to be violated is a widespread and powerful moral sentiment that transcends religion (Shils 1968). I suspect that R.F. Holland was right to identify the fault line in discussions of suicide not at the divide between religion and irreligion nor even at the appeal (or lack thereof ) of the notion that human life is somehow sacred or special. Holland cites the example of Captain Oates, the member of Scotts’s expedition in the Antarctic who was endangering his fellow explorers because of his increasing weakness. On the day before his death Oates had said that he could not go on and had proposed that the rest of the party should leave him in his sleeping bag. “That we could not do” says Scott, whose account of the upshot is as follows: “He slept through the night before last, hoping not to wake; but he woke in the morning-yesterday. It was blowing a blizzard. He said, ‘I am just going outside and may be some time.’ He went out into the blizzard and we have not seen him since. . . . We know that poor Oates

174

Thomas S. Huddle was walking to his death, but although we tried to dissuade him, we knew it was the act of a brave man and an English gentleman” (Scott’s Last Expedition, vol. i, p. 462). (Holland 1980: 150–152)

Ought one to call Oates’s act a suicide? Many who argue for suicide’s permissibility likely would; many opposed would wish to call it something else—an act of self-sacrifice the best description of which is determined by Oates’s primary intention in walking into the blizzard: that of relieving his fellow explorers of the burden of his presence. He took himself off so that his companions could go on unencumbered by him. His death in the blizzard was a side effect of that action. Holland suggests that the divide in attitudes to suicide is between those for whom a construal of Oates’s action as both noble and non-suicidal is important and those for whom it is of little moment whether Oates is held to be no suicide or a suicide but an honorable one. On one side of this fault line is the sense that our human life is prior to the stream of choices made and directions taken that it makes possible; our life’s continuation should therefore remain outside the realm in which we assert our wills. Self-sacrificial acts resulting in death do not do violence to this conviction as the act of suicide is in part constituted by an actor’s primary aim of selfdestruction—which was not Oates’s primary aim. On the other side of the line is the determination that my life is my own—it may be placed in my scale of value and retained or relinquished at my pleasure. My choice to end my life is simply one more choice in the series of my choices in how to live my life. That choice deserves respect just as do my other self-regarding choices that harm no one else. Oates’s decision is well described as a choice to end his life and we ought to view it none the worse for that. Holland labels those on these opposing sides as respectively bearers of a religious ethics (regardless of whether accompanied by religion) or as humanists. Moving the ground from suicide considered generally to PAS and the plight of the terminally ill moves us to the context in which proponents of suicide’s permissibility can make their strongest case. For whom if not for a suffering, terminally ill patient might a decision to end one’s life be warranted? Accusations of disloyalty to life have little bite for the suffering patient who is near death. Philosophers hesitant to endorse suicide in general often find an option for PAS in the terminally ill far easier to favor. The role of physicians introduces the complication of another party’s participation in the suicidal act. In addition to justifying PAS in general, those of its proponents who see physician participation as (at least sometimes) obligatory have the additional task of arguing for PAS as an essential implication of the physician’s role. The rest of this chapter will focus on suicide and euthanasia in the medical context—that is, on PAS and physician participation in

Suicide 175 active euthanasia. Philosophical positions range from those holding all PAS to be intrinsically wrong through those holding at least some PAS to be permissible to those holding some PAS to be obligatory. Much of the argument has revolved around the cogency of moral distinctions asserted by opponents of at least some PAS between actor contributions to intended vs. merely foreseen negative outcomes (differentiated by the doctrine of double effect, the DDE) and between active and passive contributions to negative outcomes (differentiated by the doctrine of doing and allowing, the DDA). These distinctions are necessary for those seeking to prohibit certain kinds of actions as bad in themselves (e.g., suicide) while acknowledging the inevitability of actor causal involvement in negative outcomes. Both are implicated in Holland’s suggestion that Captain Oates can be exonerated of the charge of suicide. Can Oates’s act be best described as relieving his comrades of a burden while avoiding its description as suicide? That is, was Oates’s death a side effect of his walking into the blizzard to remove himself from his comrades? Or must we say that Oates must have intended his death and was thus a suicide? In Oates’s case the DDA and the DDE run together, as (at least some) opponents of suicide would wish to characterize his death-producing act as both producing death passively (through allowing rather than doing) and producing it unintentionally (as a side effect). It is important to note that this parallelism does not always hold. The DDA and the DDE are complementary; an active relation to a negative outcome may be foreseen or intended, as may be a passive relation. The DDA maintains that sometimes allowing a negative outcome is permissible when doing it is not; the DDE maintains that sometimes foreseeing a negative outcome from what one does or allows is permissible when acting or omitting while intending that outcome would not be. In the debate over PAS, the negative outcomes (according to PAS’s opponents) in question are the deaths of patients. Physician involvement in patient death may be active or passive (doing and allowing in this case corresponding to killing and allowing-to-die; the K/ATD distinction) and death may be intended or merely foreseen. Opponents of PAS maintain that PAS (or at least some PAS) may be categorically prohibited while passive and unintended causal involvement in patient death may sometimes be permitted. Advocates of PAS have often proceeded by attacking these distinctions and claiming that PAS opponents are inconsistent in permitting some physician actions resulting in patient death and forbidding others that, PAS advocates hold, are morally indistinguishable.

Physician-Assisted Suicide, the Doctrine of Doing and Allowing, and the Killing/Allowing-to-Die Distinction The DDA asserts that one may sometimes morally distinguish acts/omissions resulting in similar negative outcomes according to the kind of actor agency

176

Thomas S. Huddle

involved (active or passive). That it can matter whether a harm comes about through one’s doing it or allowing it seems, at first glance, highly plausible. As Philippa Foot remarked, there appears to be a moral difference between buying a winter coat and thereby omitting to send the cost of the coat to Africa (such that a child dies whom the money might have helped) and murdering someone to get the money to buy a winter coat (Foot 2002). Specifying the distinction between doing and allowing in a way both true to our usual ways of drawing it and also properly exclusive of counterexamples has, however, proven extraordinarily difficult (see Woolard 2012). Much of the difficulty may be due to two presumptions commonly held by both attackers and defenders of the distinction: (1) that the target of the distinction is causal relationships among natural facts, and (2) that one may consider the moral importance (or not) of the distinction separately from the way the distinction is drawn on the natural fact/causal level. In what follows I will presume that a cogent distinction between doing and allowing (and between killing and allowing-to-die) can be drawn and consider both attacks on and defenses of the moral importance of the distinction. The K/ATD Distinction Is Never of Moral Importance Per Se The most thoroughgoing attacks on the moral importance of the distinction between killing and allowing-to-die have been mounted since the late 1970s by James Rachels, Michael Tooley, Dan Brock, and Peter Singer, among others (Rachels 1979; Brock 1985; Singer 1993; Tooley 1994). The strategy has been to suggest that PAS is not different in any morally important way from the medical strategy of allowing terminally ill patients to die. If that is true, PAS opponents should either accept PAS or disown a pervasively common medical approach to terminal illness. Moral equivalence between the two ways of acting is demonstrated through comparing cases in which actor intentions, patient wishes, and an outcome of death are held equal and the only factor differing is whether the act leading to death is a doing or an allowing. Our intuitions lead us (claim proponents of PAS) to the conclusion that there is no intrinsic moral difference between killing and allowing-to-die. That is, there is no important moral difference between otherwise similar cases in which a difference resides solely in the kind of causal relevance (doing vs. allowing) an actor has to a patient’s death. A recent example of such cases is L. W. Sumner’s Douglas, who is dying of pancreatic cancer and may choose either to refuse treatment or be euthanized: Chemotherapy holds out the possibility of slowing the dying process somewhat but the side-effects of the treatment are severe; without it Douglas will soon die from his cancer. Let us say that Douglas has two means

Suicide 177 of managing his dying process: refusing treatment (letting death happen) and requesting euthanasia (making death happen). So as to keep other factors equal, as much as possible, we will assume that euthanasia is legal in Douglas’s jurisdiction, that he would satisfy the criteria for it, and that his physician is willing to administer it. Furthermore, we will stipulate that Douglas will experience no more suffering by refusing treatment than by requesting euthanasia. Douglas is indifferent between the two options: he will give informed refusal to chemotherapy and informed consent to euthanasia. (Sumner 2011: 95) In case 1, we are invited to imagine Douglas refusing treatment and then dying; in case 2, Douglas requests euthanasia, receives it, and dies. The two cases are similar except for the “bare difference” of Douglas dying by killing in case 1 and by allowing-to-die in case 2. The conclusion Sumner suggests we should draw is that there are no grounds for asserting an ethical significance to the causal “bare difference” in this case. If there is no such ethical significance, then relevantly similar extensions from licit allowing-to-die to licit killing may be made in terminally ill patients. How do pairs of cases such as Sumner’s show that there is no moral importance in whether a death comes about through doing or allowing? The cases are constructed so that morally important aspects of the transition from life to death are equalized. Once all of these have been identified and stipulated to be the same in each case, then one is forced to conclude (so advocates of this strategy contend) that the bare fact of death coming about through doing vs. allowing cannot be morally important. For instance, James Rachels suggests that wrong-making aspects of killing and allowing-to-die might include the patient’s death itself, the actor’s intentions in being causally related to said death, and effects of the patient’s death on other people. Rachels notes that all of these aspects of a terminally ill patient’s death may be the same whether the death comes about by killing or allowing-to-die. Ergo, these aspects are sufficient to determine the moral valence of either killing or allowing-to-die, and that valence must be independent of the bare fact of death occurring through doing or allowing (Rachels 1979: 165). Morally important differences between cases of actively and passively causing death reside in agent and patient wishes and intentions and in outcomes, not in activity or passivity per se. From the suggestion that there is no difference between activity and passivity in producing death in the two Douglas cases, it is a further step to conclude that activity or passivity in causing death is invariably of no moral importance, whereas intentions and outcomes in such cases can always be morally determining independently of activity or passivity. The latter conclusion follows from presuming a fixed relation between moral and natural properties of cases.

178

Thomas S. Huddle

Dan Brock suggests that moral properties must supervene on descriptive properties (e.g., whether an act is a doing or an allowing) in an invariable way. Thus the absence of moral importance in the differing causal structure of otherwise similar cases implies that moral significance could never reside in the causal structure of a case (as regards an agent doing vs. allowing) per se (Brock 1985: 861). If once a descriptive property is not morally important, then it should never (of itself ) be important. While this position has a certain intuitive plausibility, it has been shown to fit badly with our intuitions about important cases. For instance, the moral difference between doing and allowing seems profoundly important in Foot’s pair of cases mentioned earlier. Yet in a pair of cases in which harm is done vs. allowed by an agent defending himself against an aggressor, the moral difference between harm coming to the aggressor through doing versus allowing may be trivial or nonexistent (Kagan 1988: 18). It would seem that the same kind of causal difference between agents and agent-related harms in contrasting pairs of cases can be of variable moral importance across situations. The position that moral and natural properties are related in a fixed way cannot accommodate that result. The K/ATD Distinction May Sometimes Be Morally Important The contention that an agent’s active or passive causal relation to a patient’s death is invariably of no moral importance in itself has been attacked by nonconsequentialist philosophers, beginning with Philippa Foot. Foot attacked the equivalence of a pair of cases offered by Rachels. In Rachels’s cases, Smith drowns his nephew in the bathtub whereas Jones observes his nephew drowning but fails to rescue him. Both intend their nephew’s death (Rachels 1975). Foot acknowledged that Smith and Jones were blameworthy in each case but denied that the cases were equivalent: drowning the nephew was an offence against justice, whereas the failure to rescue an offence against charity. Our rights not to be drowned extend further than our rights to be rescued. While Smith’s act and Jones’s omission might have the same moral valence, the moral importance of their respective involvement in their nephew’s death differed (Foot 1977). In this pair of cases, the moral difference between Smith and Jones does not affect the moral valence of their involvement in their nephews’ death but moral differences between acts and omissions might be important in other cases—as the moral significance of killing or allowing-to-die is not determined by the causal character of doing or allowing in any possible case. It is determined instead by how activity or passivity figures morally in the situation in which an actor is causally related to an outcome. And the moral significance of activity or passivity will vary in differing situations. The same point is made by Frances Kamm, who considers a doctor transplanting organs to save needy patients while knowingly using a chemical that

Suicide 179 will seep into the next room and kill a patient there. Such an act would be impermissible. On the other hand, if faced with one patient needing care and other patients needing transplants, the doctor might permissibly transplant the needy patients while leaving the one patient unattended to. Presuming that both the one patient and the patients needing transplants will die without treatment, it is permissible to let the one patient die in this latter case to save more than one. She takes these cases to show that when death is merely foreseen but not intended as in both cases, killing (in the case of using the chemical) is of a different moral valence than letting die (the one patient unattended to) (Kamm 2007: 147). Even if the K/ATD distinction is not morally important in all cases, it may be important in some. The distinction may be morally important to physicians facing requests from terminally ill patients in the following way: requests to be allowed to die issue from a patient’s right to non-interference or bodily integrity, which demands observance by physicians. Requests for assistance in dying appeal to no such negative right but to a positive right (if right it is) to a particular service. As negative rights are more pressing than positive rights, a patient’s demand in the case of withdrawing therapy is exigent in a way that an otherwise similar demand for PAS cannot be (Kamm 1999: 590–591). Proponents of a physician obligation to provide PAS suggest in response to this way of pressing the K/ ATD distinction that a patient’s right of bodily integrity or non-interference is derivative from a right to self-determination, which is primary. Rights to withdrawal of treatment and to assistance in suicide ought to be viewed as equally pressing implications of the patient’s right to autonomy (Brock 1999: 525–527; Dworkin 2009: 379). Foot, Kamm, Sumner, and others who deny that an invariable moral significance always attaches to causal relations of doing or allowing (and that that significance is none) generally view PAS as permissible (or obligatory) in at least some circumstances. Patients that are terminally ill who wish assistance in dying do not have differing rights violated by physicians who may either kill them or allow them to die. Willing patients may waive rights to life and rights of non-interference do not apply. While whether an act is a killing or an allowing-to-die may sometimes be morally important, it is not important, these ethicists would maintain, in the usual kinds of cases in which assistance in suicide might be requested by terminally ill patients. That being so, PAS is presumptively at least permissible if patients wish for it and if the outcome of patient death is beneficial rather than harmful. This position on PAS is probably the most common in bioethical discussion and is held by many who would oppose the legalization of PAS on policy grounds. If the permissibility of PAS is not foreclosed by an agent’s activity in achieving a patient’s death, then foreclosure, if PAS is foreclosed, must issue in some manner from the patient’s status or the doctor’s role. That is, PAS may

180

Thomas S. Huddle

be forbidden on the grounds of the sanctity of life, on grounds of respect for persons, or for reasons specific to the doctor-patient relationship. Opponents of PAS focusing on physician participation in suicide generally suggest that PAS is incompatible with the physician’s role. The K/ATD Distinction Is Always Important (or Always Important in the Medical Context) Those who would prohibit most or all PAS share with the liberal nonconsequentialists who would permit it an acknowledgement of the possible importance of the K/ATD distinction and the DDE. Where they part company is in an unwillingness to consider one’s own death as simply one more harm or benefit to be weighed against others and factored into one’s autonomous welfare-promoting decisions. That unwillingness follows from the special status of human life according to those who take these views: Kantians and proponents of natural law. A. Natural Law The natural law tradition takes there to be valid ethical norms and standards that are prior to human choices. These take the form of “basic goods” that are known to be such either through reflection or through human inclinations that are the occasion for reason’s recognition of given goods as basic (Murphy 2001: ch. 1; Finnis 2004). Unlike the goods of consequentialists, the basic goods in natural law are not to be weighed against one another. They are fundamental and incommensurable. Human life is one of the basic goods. Natural law theorists generally assess the good of life differently than their consequentialist and non-consequentialist opponents. These latter tend to see the value of life in the higher rational capacities made possible by biological life in human beings. In the absence of these capacities, or even in their presence given an autonomous choice to dispense with life if, say, its benefits are judged to be exceeded by its burdens, a decision to die through either active or passive means may be reasonable. The natural law theorist replies that human life’s value is intrinsic rather than instrumental; and that if its intrinsic value is especially evident in human life’s higher capacities, those capacities cannot be severed from the biological substrate in which they appear for purposes of assigning value to life (e.g., Murphy 2001: 101–105). To maintain otherwise is to suppose that bodily life is something persons possess but that remains distinct from what persons are; to commit oneself, that is, to a dualistic theory of personhood that, the natural law theorist contends, flies in the face of human experience (Finnis 1993: 567–569). Human life, then, is of fundamental value and may be taken only in strictly limited contexts (e.g., war, self-defense, capital punishment) outside of

Suicide 181 which killing is forbidden. Allowing-to-die, on the other hand, is sometimes permissible when the intention of withholding or withdrawing treatment is avoidance of burden rather than death. B. The Kantian Tradition Contemporary Kantians tend to be more permissive than natural law theorists as to suicide, but an argumentative strategy common among consequentialists and non-Kantian liberal non-consequentialists is closed off to them, as it is to natural law theorists: that is, assessing one’s future life (or death) in terms of harms and benefits and possibly opting for death on grounds that, given the overall balance, one would be better off dead. The Kantian denies that the person is simply a nexus of interests the balance of which may rightly determine life or death decisions. Value inheres in persons on account of our rational nature, a value expressed in the specifically Kantian notion of dignity. Personal interests deserve to be furthered for the sake of the person whose interests they are. But the value inhering in the person is independent of her interests and she is not permitted to extinguish her own person on behalf of her interests. To do so is incoherent—it is to act against an end (the person) on behalf of that which has value only in relation to that end (Velleman 1999: 611–613, 624–625). This stance appears to close off the possibility of prudential suicide. Although Kant himself indeed viewed suicide as prohibited, contemporary Kantians have generally avoided this implication. The most plausible route to permissible suicide for the Kantian has been the suggestion that if a person’s rational nature is sufficiently degraded, then a point might be reached at which its Kantian dignity might be better honored by destruction rather than by preservation of its degraded remnant.1 While so far more permissive than natural law theory as to suicide, Kantian views may be more demanding in regard to withdrawal of lifesaving therapy. While the Kantian is not obligated to preserve life at all costs, she has a defeasible duty to preserve her rational agency. That duty may give way to important moral projects or to the necessity of avoiding immorality. Kantians (and non-Kantians) plausibly suggest, however, that on the Kantian view, the duty of self-preservation cannot be defeated by the prospect of future suffering. So a refusal of lifesaving treatment, possibly justifiable if aimed at reducing the burdens of others, cannot be grounded in the prospect of one’s own future life made burdensome under the treatment (McMahan 2002: 482; Gunderson 2004). C. The Medical Context The most common medical approach to physician involvement in patient death is to forbid active participation categorically but to hold allowing-to-die to

182

Thomas S. Huddle

be not only licit in the terminally ill but also obligatory in cases of treatment refusal. The medical community offers a contrast with the bioethics community in that the medical community in the United States is still substantially opposed to PAS.2 As with Kantians and natural law theorists, those physicians who hold to the traditional proscription on PAS vindicate the importance of the K/ATD distinction and the DDE in regard to their relevance to the patient outcome of death on the ground of the special importance of human life. Much medical argument against PAS presumes that patient death is a negative outcome if it is a result of active physician agency—and focuses on vindicating the DDA or the DDE as means of distinguishing PAS from withdrawal of support or terminal sedation (e.g., Sulmasy and Pellegrino 1999). And medical argument favoring PAS generally echoes the bioethics case in its favor (e.g., Quill et al. 1997). The case against PAS in the medical context turns on considerations about human life as it does elsewhere, but it invokes considerations specific to illness and the relations of the ill to the physicians who care for them. The most important such arguments have been offered by Edmund Pellegrino and Leon Kass. Both suggest that scrutiny of the doctor-patient relationship leads to dispositive conclusions about the norms of medical practice antithetical to PAS. The suggestion is that there are facts about illness such as the weakness, frailty, and vulnerability of patients who are ill; and facts about the doctor’s healing role—her aims of relieving suffering and restoring health and her position of power in relation to the patient’s weakness. The facts of illness and of the physician’s role imply norms that should govern medical practice (Pellegrino 2001). These tell against the bioethicist ideal of a patient’s rational autonomy as an end toward which physicians should aim (Pellegrino and Thomasma 1987). What patients need from physicians are, primarily, care and guidance aimed at help, healing, and the relief of suffering. PAS is incompatible with what patients need from physicians and with what physicians offer patients in proper doctor-patient relationships. Why, it might be asked, should the physician’s aim of relieving suffering not sanction PAS in circumstances in which only death might end suffering? Kass echoes the natural law suggestion that one can benefit a person by killing him only if a person’s body is a possession of that person rather than the person himself—that is, only if some form of dualism is true. His objection to dualism, however, is founded not upon abstract philosophical considerations but upon the character of the doctorpatient relationship: The patient presents himself to the physician, tacitly to be sure, as a psychophysical unity, as a one, not just a body, but also not just as a separate disembodied person who simply has or owns a body. The person and the body are self-identical. True, sickness may be experienced largely as

Suicide 183 belonging to the body as something other, but the healing one wants is the wholeness of one’s entire embodied being . . . This human wholeness is what medicine is finally all about. (Kass 2002: 33) So, the argument goes, some form of body-self monism rather than dualism is presumed by the doctor-patient relationship and, granted such monism, a decision to commit suicide to benefit oneself is incoherent. Needless to say, arguments against PAS based upon the character of the doctor-patient relationship have not impressed PAS advocates. The most important difficulties are first, that the doctor-patient relationship is not a timeless ideal but instead a product of the varying cultures in which it occurs. Why ought we to view the traditional Western ideal as dispositive on the matter of PAS? Second, the doctor-patient relationship, however idealized, underdetermines the ethics of medicine at the end of life. It is simply not clear that the imperatives to heal and relieve suffering such as physicians in most (or all?) times and places might acknowledge necessarily proscribe PAS if one does not build such a proscription into those imperatives at the outset (Arras 2001). To their opponents, Kass and Pellegrino appear either to simply stipulate the norms of medicine or else to illicitly derive them from particular extant medical practices.

The DDE The DDE is essential for physicians wishing to maintain a prohibition on PAS or active euthanasia while conducting medical practice in generally accepted ways, including withholding and withdrawing life-preserving treatment in the terminally ill and relieving pain and suffering when such relief might contribute to patient death. The DDE asserts that acts with similar negative outcomes may sometimes be morally distinguished by whether the negative outcomes are intended or merely foreseen. Formally stated, the DDE holds that acts/ omissions with both good and bad outcomes may sometimes be permissible: 1. 2. 3. 4.

if the act/omission is itself licit. if the good outcome is the intended outcome. if the bad outcome is neither intended as an end nor as a means to the actor’s end. if the (at least permissible) intended end in acting is sufficiently valuable that the bad outcome (a side effect of the act) is tolerable in light of that end.

As with the sanctity of life, the DDE is often considered to issue from religious doctrine, and it is indeed closely associated with Catholic theological

184

Thomas S. Huddle

reflection. Its philosophical champions point out, however, that it is an expression of the demands not merely of Catholic religion but of many traditions of serious moral thinking and of ordinary morality (Wiggins 2006: 253–254; Wedgwood 2011: 391). In medicine at the end of life, the DDE permits physicians to withhold or withdraw life-preserving treatment while avoiding a description of those acts as the active killing of patients, in spite of the fact that death is among the outcomes of treatment withdrawal. In withholding or withdrawing treatment the physician invoking the DDE intends not the patient’s death but relief of a burden imposed by the treatment or avoidance of imposing that burden while tolerating the foreseen side effect of the patient’s death. Thus according to the DDE, withholding and withdrawing life-preserving treatments, if intended as burden relief, may be properly described as acts of burden relief that have the side effect of allowing the patient’s disease to end her life. Consequentialist and Non-consequentialist Assessment of the DDE Thoroughgoing consequentialists, naturally, will have none of the DDE.3 As what matters about acts are their outcomes, the distinction between intended effects and side effects is simply contrived. Actors are similarly responsible for all of the effects of their actions, and there are no distinctions to be drawn between their morally important effects for purposes of assigning agency. The DDE is simply a smokescreen for obscuring consequentialist judgments. In the case of physician care for patients at the end of life, physicians often judge a patient’s death to be preferable to continued life under treatment. They invoke the DDE to disguise from themselves that bringing about the patient’s death is the best description of what they are doing when they administer life-ending pain relief or withdraw life-preserving treatment (Rachels 1986: 93–96; Singer 1993: 209–210; Harris 1995: 36–39). Of course if one grants the consequentialist premise that only outcomes matter in act description and assessment, this analysis readily follows. Most bioethicists have not been able to accept the counterintuitive results of the consequentialist assimilation of intended and unintended but foreseen effects of action as that account offers no means of distinguishing agency in pairs of cases that strike most as morally very different. Such pairs of cases typically contrast scenarios in which the outcomes are similar but in which agent intention figures differently in producing those outcomes. The pairs of cases might include the following: Terror vs. Strategic Bomber. The strategic bomber intends to destroy the munitions factory but foresees that deaths of innocent civilians will also result from his bombing. The terror bomber intends the deaths of innocent civilians as a means to hasten the end of the war.

Suicide 185 Trolley vs. Bridge. In trolley, a trolley is on a track headed toward five people confined to the track who will die if the trolley continues unimpeded; the actor may switch the trolley to another track where one person confined will die. In bridge, the trolley is headed down a track toward five but on top of a bridge above the track is a fat man whose body would stop the trolley. The actor has the option of pushing the fat man off the bridge onto the track in the path of the oncoming trolley to save the five. Craniotomy vs. Hysterectomy. In craniotomy a fetus in the birth canal threatens the life of the mother; the doctor, aiming at saving the mother’s life, crushes the fetus’s skull to remove the fetus from the birth canal. In hysterectomy, the gravid uterus is cancerous and the doctor must remove it to save the mother’s life. In pairs of cases like these, although the outcomes are similar, most judge the respective acts in strategic bomber, trolley, and hysterectomy to be permissible while the analogous acts in terror bomber, bridge, and craniotomy are not (or are less so). That is, in the first case of each of the three pairs one may act to produce a good effect while foreseeing harm. In the second case of each pair it is not (or less) permissible to act intending harm as a means to one’s legitimate end. The DDE would appear to be well suited to distinguish the cases in these pairs. In each pair, the plausibly permissible case is one of seeking a legitimate end while causing an unintended harmful effect. The good end is not produced directly by means of the harmful effect and the end is important enough to permit the harmful effect. In the contrasting case in each pair, the harmful effect is a direct means to the good end. Although non-consequentialist philosophers generally view the cases in pairs like these to be morally distinguishable, many deny that the DDE can do the work required to distinguish them. Attacks have focused on whether intending harms as means to benefits is actually impermissible and, perhaps more fundamentally, on whether intended and merely foreseen harms can be consistently and reliably distinguished if a sufficiently broad account of intention is in play. The latter point has been pressed by James Rachels, Judith Thomson, Alison McIntyre, and, more recently, Thomas Scanlon. All suggest that absurd consequences follow from holding the moral permissibility of actions to turn on the intentions of individual actors. On the other hand, if intention is construed sufficiently broadly, then the distinction between intended and foreseen in the important cases disappears (McIntyre 2001: 242–247). Scanlon argues the inadequacy of actor intentions for underwriting act permissibility using the case “prime minister,” in which the prime minister of a country at war is asked to decide whether to bomb a munitions factory when civilian casualties

186

Thomas S. Huddle

in the area surrounding the factory may be anticipated (Scanlon 2009: 19–20). According to the DDE as per Scanlon, the permissibility of the bombing will turn on whether the pilot intends the deaths of the civilians—which seems highly implausible. If intentions have a decisive role in determining the permissibility of actions with good and bad effects, it would also seem to follow that actors could manipulate the permissibility of such actions by simply aiming their intentions at the good effect and away from the bad effect. Thomson posits a dying patient in pain so severe that the dose of morphine required to relieve the pain will also kill the patient. Is administering that dose permissible? According to the DDE, the answer is yes if the doctor intends only relief of pain and no if the doctor intends the patient’s death (Thomson 1999: 514–515). But Thomson suggests that the permissibility of this or any act ought to follow from facts about the act, not about the state of mind of the actor. Intentions have a bearing on whether an act may be virtuous or vicious for an actor; they have none on whether a given act is morally permissible. Thomson and Scanlon hold that differences between the pairs of cases such as those earlier, which the DDE has been wrongly taken to illuminate, are actually accounted for by other principles. Others are reluctant to jettison the importance of intentions to the moral assessment of acts and defend variants of the DDE as attacked by Scanlon and Thomson. The suggestion is that the importance of intentions to the moral assessment of acts is built into our morality—it is simply not plausible that intentions may be dispensed with in the moral assessment of acts (Wedgwood 2011: 391–392). Much of criminal law presumes that intention can either partly constitute offences (e.g., murder) or aggravate them. And in ordinary life we are concerned not merely with what is done but also with the spirit and intention of an actor’s doing. The DDE simply voices what ordinary morality demands of actors faced with situations in which acting will produce both positive and negative outcomes: when it is permissible that innocents be the victims of acts performed for good ends, it cannot be that they become so through the actor’s use of them to attain the good end. Thus while the corvette commander may depth-charge the U-boat to save the convoy, incidentally killing survivors in the water, he may not save the convoy by an agreement with the U-boat commander that the convoy will be spared if the U-boat commander is permitted to machine-gun the survivors in the water (Wiggins 2006: 250). Ralph Wedgwood, Jeff MacMahan, and David Wiggins suggest that the costs of giving up the DDE are too high, in spite of the challenges presented to it by DDE skeptics, such as Thomson and Scanlon. An alternative response to Thomson and Scanlon is offered by FitzPatrick, who suggests that they have misconstrued the DDE at the outset. The DDE does not speak to the intentions

Suicide 187 of particular agents; it is directed instead at specifying what intentions would be acceptable in any agent performing a given action in a good cause. In the case of the pilot flying the mission against the munitions plant, the function of the DDE is not to affirm or condemn the actions of pilots flying the mission with differing intentions. It is merely to deem the mission permissible because a pilot could fly the mission seeking victory in the just war without being involved in bad intentions (as a tactical, but not as a terror bomber). The criticisms of Thomson and Scanlon are evaded by construing the DDE not as a constraint on act justification operating through scrutiny of individual actor intention; the DDE instead warrants acts through the specification of intentions with which any actor might permissibly perform an action in a good cause (FitzPatrick 2013).4 The Medical Context Most liberal non-consequentialist philosophers who defend the DDE would nevertheless deny that it prohibits PAS or active euthanasia in terminally ill patients in the right circumstances—that is, if such patients’ lives had become burdensome to them and they wished for aid in dying. The suggestion would be that the DDE does not bear on such cases; death for such terminally ill patients is not a negative outcome and there is therefore no difficulty with intending a patient’s death while acting in such a way as to bring it about. Or, if the DDE does bear, it might actually be held to prefer acting while intending death to acting while merely foreseeing it—as the moral significance of intending rather than foreseeing harm would reverse if the outcome were beneficial rather than harmful. Jeff McMahan suggests an analogous result for the DDA, as does Robert Young (Young 2007: 96). That is, one should prefer bringing about death actively to doing so passively in terminally ill patients when death would benefit them. If the DDA is valid and death is beneficial rather than harmful, active euthanasia should be more imperative than passive euthanasia (McMahan 2002: 461).5 As in the case of the DDA, physicians who invoke the DDE to distinguish acceptable withdrawal of support or terminal sedation from unacceptable active euthanasia do so from a determination that their role forbids their participation in killing (Sulmasy and Pellegrino 1999). As we have seen, that conclusion follows from considerations about human life or about human life as exhibited in the patient role in relation to the physician role. Even if one grants physician opponents of PAS the premise that patient death at physician hands is a negative outcome if a result of active physician agency, however, PAS opponents must contend with a more elaborated version of the attack on the DDE through the posited manipulability of intentions as per Thomson mentioned earlier: the so-called problem of closeness.

188

Thomas S. Huddle

The DDE and the Problem of Closeness According to some opponents of the DDE, if the distinction between intending and foreseeing can be morally meaningful in cases such as strategic bomber, it can be similarly meaningful in cases such as terror bomber and, thus, permit acts such as terror bombing that the DDE is meant to block. The thought is that a sophisticated terror bomber can make use of the DDE if he simply aims his intention away from the deaths of civilians as means to achieving the end of the war; what is necessary for terrorizing the enemy population is not civilian deaths; it is merely the widespread conviction on the enemy side that civilians have died. This will come to pass if motionless civilian bodies are observed in the rubble of the bombing; that is all that is necessary to achieve the terror bomber’s end and that is all that the sophisticated terror bomber intends. That civilians will be killed in the course of producing dead-appearing bodies in the rubble is, for this terror bomber, a regrettable necessity that he foresees but does not intend. The sophisticated terror bomber has fixed his intentions solely on the good aspects of his action’s outcome, which may then be deemed intended; the bad aspects of the outcome can be relegated to side effects (Bennett 1995: 210–213). If the DDE blesses the action of sophisticated terror bomber, principled distinctions between outcomes intended and those merely foreseen would appear to be impossible. A physician opposed to PAS who disconnects a terminally ill patient’s ventilator may say to herself that she intends only relief of the burden of the ventilator and not the killing of her patient. But then the greedy nephew disconnecting his rich uncle’s ventilator may intend only his own receipt of his uncle’s bequest; he may view his uncle’s death as a mere side effect of his action aimed at getting the bequest.6 If we convict the nephew of intending his uncle’s death, because his action is “too close” to killing his uncle for that act to be deemed unintended, must we not also convict the physician of intending her patient’s death, whatever she may say? Either intentions are sufficient to determine morally relevant act descriptions, in which case actors can justify acts the DDE would usually be held to condemn; or natural facts determine proper act description, in which case actors are condemned for acts we generally seek to excuse through the DDE.

Possible Ways Forward from the Problem of Closeness It is unclear that there is, at present, a way out of the DDE’s problem of closeness. The difficulty arises through the question of how one picks the appropriate description of an act for moral assessment by the DDE. DDE opponents suggest two alternatives, both of which present difficulties. One is to accept an actor’s view of her act as regards what is intended and what foreseen. This

Suicide 189 leads to implausible results. The sophisticated terror bomber in bombing civilians intends only motionless bodies in the rubble, not dead civilians. His act is therefore bombing-while-intending-motionless-bodies-but-not-civilian-deaths. Thus his “act itself ” is licit, meeting the first condition of the DDE. Even if one goes on to disqualify the sophisticated terror bomber’s act by the DDE’s proportionality condition, this conclusion as to the apposite moral description of the sophisticated terror bomber’s act seems wrong. Yet the alternative approach to apposite act description is no less problematic for the DDE. If one holds that certain alternative act descriptions are “too close” to what the actor is doing to be deemed unintended no matter that said descriptions are said to be unintended by the actor, there is no clear way to exempt analogous act descriptions from being deemed intended in the case of those actors whose actions we wish the DDE to deem licit—strategic bomber, bystander in trolley, and hysterectomy doctor. In each case, the action just is, respectively, killing civilians, killing the one, or killing the fetus in spite of actor intentions otherwise. Attempts by the DDE’s defenders to cope with the problem of closeness divide over their view of the apposite description of acts for moral assessment by the DDE. A minority are willing to bite the bullet of an actor-determined view of apposite act description. More often, following Anscombe (Anscombe 2005), defenders seek a principled way to deem some action descriptions “too close” to what an actor ostensibly intends to be deemed unintended for purposes of moral act assessment. Finally, some defenders of the DDE similarly insist on a role for natural facts in determining apposite act description but find the problem of closeness to be intractable. These discussants seek to replace the traditional DDE with other principles that would do similar work without depending upon a distinction between outcomes intended and those merely foreseen. Actors Determine Morally Apposite Act Descriptions The problem of closeness does not arise for John Finnis because actors doing evil while directing their intentions away from said evil fall foul not of the DDE but of constraints on producing evil as a side effect of one’s intended act. Finnis argues for an actor-determined view of apt description of intentional actions. So the sophisticated terror bomber may be said to really intend only motionless bodies in the rubble and not the deaths of the civilians he bombs to hasten the end of the war (Finnis 1991). The DDE, then, does not bear on a comparison between sophisticated terror bomber and strategic bomber, as the difference between them is not one of a negative outcome intended vs. foreseen. The sophisticated terror bomber is condemned independently of the DDE: because it is prohibited, in the circumstances, to produce the evil of dead civilians, even as a side effect, if the only end to be gained by doing so is terror.

190

Thomas S. Huddle

Natural Facts Determine Morally Apposite Act Descriptions Most adherents of the DDE seek to block an exclusively actor-determined construal of what is intended by an act. In the case of the sophisticated terror bomber, their suggestion would often be that bombing-while-intending motionless bodies in rubble is “too close” to bombing-while-intending civilian deaths for there to be a meaningful moral distinction between the two putative intentional acts. For purposes of construing intention they are the same act, because bombing-while-intending motionless bodies in rubble just is bombingwhile-intending civilian deaths. The sophisticated terror bomber is trying to decompose an intention and attach himself to a subpart of it while repudiating the remainder; but the intention in question, the DDE adherent maintains, resists such decomposition as the candidate subparts are “too close” to be pulled apart. The strategic bomber has no such difficulty because bombingwhile-intending the destruction of the munitions factory is far enough apart from bombing-while-intending civilian deaths that the one act need not be equated to the other, so that civilian deaths resulting from strategic bombing can be classed as side effects rather than intended effects. While this approach to the problem of closeness has a good deal of intuitive appeal, the difficulty is in specifying the distinction between acts “too close” to be separated in intention by an actor and those sufficiently far apart to permit such separation. Thomas Cavanaugh argues that this distinction follows straightforwardly from a sufficiently rigorous theory of intention. He considers cases such as craniotomy doctor and sophisticated terror bomber and contends that an adequate theory of intention, such as Michael Bratman’s, implies that each of these actors intends harm whereas hysterectomy doctor and strategic bomber do not. Bratman suggests that intention involves volitional commitment to a plan encompassing means and ends. In craniotomy and terror bomber, the deaths of the fetus and of civilians figure as means in the respective actors’ plans. In hysterectomy and strategic bomber they do not (Cavanaugh 2006: 112–117). The problem of closeness, then, does not arise. The difficulty for Cavanaugh’s contention, as Dana Nelkin and Samuel Rickless point out, is that intentions to harm in craniotomy doctor and sophisticated terror bomber do not follow from Bratman’s theory without a stipulation the theory does not make: the equation of destroying the fetus’s head and bombing civilians until they are motionless with harm (Nelkin and Rickless 2015). That is, it must be held that each of these actions just is harming. Yet if that is so, the actions of hysterectomy doctor and strategic bomber must be similarly harmful vis-à-vis the fetus and the civilians—as in each case their action also just is killing the fetus and the civilians. The problem of closeness is back. William FitzPatrick suggests that the closeness problem can be surmounted by distinguishing constitutive and causal relations of acts to outcomes (FitzPatrick 2006). If a given act merely causes an outcome, the outcome may be

Suicide 191 unintended. If an act constitutes an outcome, then the outcome is too close to the act to be unintended. So being rendered motionless in the rubble of bombing constitutes being killed; if the first is the intended aim, then so is the second, whereas bombing while aiming at the destruction of a munitions factory may cause civilian deaths but said bombing doesn’t constitute those deaths. It is therefore intelligible to consider civilian deaths in strategic bomber as side effects rather than intended effects. Natural Facts Determine Apposite Act Descriptions and the Resultant Problem of Closeness Is Intractable Warren Quinn argued that there was no clear way to distinguish outcomes of acts intended from those foreseen and that a cogent DDE could not be founded upon the intended vs. foreseen distinction (Quinn 1989). His suggestion was to modify the DDE, removing its dependence on distinguishing intended from foreseen outcomes. Quinn’s DDE focuses instead on the relation between actor intention and the harms conveyed to victims of acts having both good and bad outcomes. The DDE should be taken to prefer, ceteris paribus, harmful indirect agency (harm conveyed by an act to victims incidental to the actor’s project) to harmful direct agency (harm conveyed to victims through victims’ deliberate involvement in the actor’s project). For Quinn, sophisticated terror bomber is morally suspect compared to strategic bomber because his civilian victims are part of his terror-producing project, whereas strategic bomber’s victims, which similarly count as intended, are completely incidental to the mission of bombing the munitions factory. A variant of Quinn’s solution to the closeness problem has recently been defended by Dana Nelkin and Samuel Rickless (Nelkin and Rickless 2014). The difficulty with Quinn’s tack is that it solves the closeness problem by redirecting the DDE away from its original target, the wrongness of intending harm or evil as such. Many of the DDE’s defenders would like to say that the special wrongness of sophisticated terror bomber is not intending harm as part of a project but simply intending (rather than merely foreseeing) harm. Quinn’s strategy shares the weakness of other patient rather than agentcentered accounts of deontological constraints: of failing to acknowledge the force of a constraint not merely against evil happening but also against one’s own participation in it.7 It is indeed worse to intend evil deliberately than incidentally, but the moral force of the DDE follows from its condemnation of intending evil, simpliciter. Other Possibilities for Addressing the Problem of Closeness Recently some have suggested the possibility of a middle position between the actor-determined and natural facts-determined view of morally apposite act

192

Thomas S. Huddle

description for purposes of act assessment by the DDE. The suggestion is that neither actor intentions nor natural facts suffice to specify what an act “just is” in a given context. Instead, act descriptions for purposes of scrutiny by the DDE are generated by normative standards. One such suggestion is that of Ralph Wedgwood. Wedgwood (2011) shares Finnis’s view that actor intentions determine morally apt descriptions of intentional acts. In contrast to Finnis, however, Wedgwood suggests that the DDE compares not acts producing intended vs. foreseen evil (as construed by the actor) but instead intended vs. foreseen bad states of affairs (considered agent neutrally). So while the sophisticated terror bomber intends only motionless bodies in the rubble, this state of affairs in our world is also civilian deaths, and is, hence, bad. The strategic bomber intends the munitions factory’s destruction and produces the same number of civilian deaths as sophisticated terror bomber. Both bombers produce similarly bad states of affairs as regards civilian deaths, but one is intended and the other only foreseen. So the DDE holds strategic bomber’s act to be so far preferred to sophisticated terror bomber’s. The problem of closeness appears to be averted by moving from act descriptions as determined by actor construals to descriptions as mandated by agentneutral standards of goodness and badness. The other suggestion here is that the normative standards determining apt act descriptions for scrutiny by the DDE are not those of an impersonal scheme of agent-neutral values but instead the standards of relevant human practices or conventions.8 It is in the context of these practices that acts take on their moral significance and, hence, their proper descriptions. Actors intend as they do when they act, but what they are doing (the apposite act description for purposes of the DDE) is not determined by reference either to their mental states or to what they do purely as a matter of natural facts. Terror bombing, in the context of just war theory and practice, is an illicit act. Seen in that light, no matter what the terror bomber’s intentions it fails the first condition of the DDE, that an act be licit in itself. Married to FitzPatrick’s gloss on the DDE, that it is intended to pass judgment not on individual actors’ act tokens but on act types, this approach yields a particular understanding of the DDE’s first provision: the act itself—that is, the act as characterized by the norms that bear on the actor in the human practice(s) in which the act is situated—must be licit. It is human practices or conventions, on this view, that determine not only the morally apposite description of an act but also how lines may be drawn between intended and foreseen outcomes of given acts for purposes of the DDE. In medicine, the physician need not intend death when disconnecting the terminally ill patient’s ventilator. The greedy nephew who disconnects his rich uncle’s ventilator cannot be held not to intend his uncle’s death. The difference follows from the norms that bear on the physician and on the nephew. Applied to the DDA, the approach would hold that lines between doing and allowing

Suicide 193 (or between killing and allowing-to-die) are similarly dependent upon practice norms. The approach is broadly consonant with an increasingly large body of research suggesting that our ascriptions of both causation and intention are value-laden.9 It holds, its champions suggest, the promise of intelligible lines between intended and foreseen that can avoid the problem of closeness and so far vindicate the DDE.

Conclusion The past forty years has seen an increase both in philosophical acceptance of suicide and in skepticism of argumentative strategies often used to maintain a prohibition on suicide while permitting some acts that have one’s own death as an outcome: the DDA and the DDE. Defenders of the DDE and DDA are likely now a minority among moral philosophers (McMahan 2009: 345). Both “doctrines” continue, however, to be vigorously defended in current discussion. Physicians, most of whom would deny PAS to be a necessary aspect of their role, have not (yet?) been deprived of these particular avenues for defending common medical practices while proscribing PAS and active euthanasia. That said, physicians opposed to PAS are on the defensive as to their reluctance to countenance physician participation in PAS. Positions upholding human life’s inviolability (“sanctity” is falling out of currency) have lost ground in the past forty years of philosophical discussion even if far less so in U.S. medical practice. While the prospects of PAS in the United States policy arena remain murky, it is generally favored in philosophical if not in medical discussion. Lines of philosophical resistance may emerge in discussions of philosophical anthropology, in particular in bioethical debates over the nature and implications of personhood. The character of personhood as exhibited in the patient role may prove a fruitful starting point for developing a better account of the implications of assisted suicide for physicians than we now possess.

Related Topic Chapter 10, “Refusing Lifesaving Medical Treatment and Food and Water by Mouth” by Paul T. Menzel

Notes 1. The state at which this point is reached varies in the writings of Kantian theorists. The mere prospect of loss of rational agency is enough for Thomas Hill, Jr. David Velleman suggests that a patient tyrannized by pain or who finds pain unbearable might be at the requisite point. Michael Cholbi, who argues that neither pain nor the prospect of dementia robs patients of Kantian dignity, suggests that that dignity is sufficiently compromised for suicide to be an option in the patient who is so

194

2.

3.

4.

5.

6. 7. 8. 9.

Thomas S. Huddle

depressed as to have lost all interest in her own well-being and happiness. (See Hill 1983; Velleman 1999; Cholbi 2010.) Of course the generalization obscures various complexities. Polling suggests that U.S. physician support of PAS availability in recent years has been up to the 50% range, although far fewer physicians express a willingness to participate. U.S. medical organizations are and have been unanimously opposed to PAS with the recent exceptions of the Oregon and California Medical Associations. The Oregon Medical Association took a position of neutrality on the 1994 ballot measure to legalize assisted suicide in Oregon but reversed itself three years later. The California Medical Association announced a policy of neutrality on PAS legalization in California in May 2015. In spite of vigorous challenges to medical prohibitions on PAS, there is still a consensus at the level of recommendations by influential medical organizations that PAS is impermissible. That consensus relies on the use of the DDA and the DDE in distinguishing PAS and active euthanasia from withdrawal of support and terminal sedation. (See Truog et al. 2008.) That is, consequentialists committed exclusively to maximizing the good as a moral measure of human action. Those preferring “satisficing” to maximizing or who avoid maximizing by introducing into consequentialism a distinction between positive and negative duties would demur from the view attributed in this paragraph (e.g., Slote and Petit 1984). A similar critique of Scanlon is offered by Stuchlik (2012). Stuchlik suggests a distinction between permissibility (of acts independently of actors) and liceity (what a given actor can permissibly do); the DDE is about the former rather than the latter. Frances Kamm, who accepts the DDA but rejects the DDE, argues for a physician duty to provide PAS from more general physician duties to seek greater medical goods for patients in spite of producing at the same time lesser medical evils. In the case of terminally ill patients for whom death would be a benefit, acting while intending the patient’s death is not morally different from similarly acting while merely foreseeing it (Kamm 2007: 150, 153–154). This case is Dan Brock’s. (See Brock 1992: 10–12.) For a discussion of agent and patient-centered accounts of deontological constraints, see Alexander and Moore (2015). The general approach to the DDE suggested in the paragraphs ahead is taken by Chappell (2013). A similar approach to the DDA in the context of medicine is suggested by Huddle (forthcoming). The seminal experimental work suggesting that intention ascriptions are often valueladen was done by Joshua Knobe (2003). For a recent summary of the development of this line of research and attempts to explain its results, see Alfano and Loeb (2014: 2.3). Judgments that given acts are doings or allowings are similarly affected by moral appraisals. (See Cushman et al. 2008.)

References Alexander, L. and M. Moore (2015) “Deontological Ethics,” in E.N. Zalta (ed.) The Stanford Encyclopedia of Philosophy, URL = . Alfano, M. and D. Loeb (2014) “Experimental Moral Philosophy,” in E.N. Zalta (ed.) The Stanford Encyclopedia of Philosophy, URL = . Anscombe, G.E.M. (2005) “Action, Intention, and Double Effect,” in M. Geach and L. Gormally (eds.) Human Life, Action and Ethics: Essays by G.E.M. Anscombe, Upton Pyne: Imprint Academic, 207–226.

Suicide 195 Arras, J. (2001) “A Method in Search of a Purpose: The Internal Morality of Medicine,” Journal of Medicine and Philosophy 26(6): 643–662. Battin, M. (2003) “Euthanasia and Physician-Assisted Suicide,” in H. LaFollette (ed.) The Oxford Handbook of Practical Ethics, Oxford: Oxford University Press, 673–703. ——— (2005) “Is a Physician Ever Obligated to Help a Patient Die?” in M. Battin (ed.) Ending Life: Ethics and the Way We Die, Oxford: Oxford University Press, 88–107. Bennett, J. (1995) The Act Itself, Oxford: Clarendon Press. Brock, D. (1985) “Taking Human Life,” Ethics 95(4): 851–865. ——— (1992) “Voluntary Active Euthanasia,” The Hastings Center Report 22(2): 10–22. ——— (1999) “A Critique of Three Objections to Physician-Assisted Suicide,” Ethics 109(3): 519–547. Cavanaugh, T.E. (2006) Double-Effect Reasoning: Doing Good and Avoiding Evil, Oxford: Clarendon Press. Chappell, T. (2013) “What Have I Done?” Diametros 38: 86–112. Cholbi, M. (2010) “A Kantian Defense of Prudential Suicide,” Journal of Moral Philosophy 7: 489–515. ——— (2011) Suicide: The Philosophical Dimensions, Calgary: Broadview Press. ——— (2012) “What Is Wrong with ʻWhat Is Wrong with Rational Suicide,’ ” Philosophia 40: 285–293. Cushman, F., J. Knobe, and W. Sinnott-Armstrong (2008) “Moral Appraisals Affect Doing/Allowing Judgments,” Cognition 108: 281–289. Durkheim, E. (2010) Suicide: A Study in Sociology, G. Simpson (ed.), New York: Simon and Schuster. Dworkin, R. et al. (1997) “Assisted Suicide: The Philosopher’s Brief,” New York Review of Books 44(5): 41–47. Finnis, J. (1991) “Intention and Side-Effects,” in R.G. Frey and C.W. Morris (eds.) Liability and Responsibility: Essays in Law and Morals, Cambridge: Cambridge University Press, 32–64. ——— (1993) “The ‘Value of Human Life’ and ‘The Right to Death’: Some Reflections on Cruzan and Ronald Dworkin,” Southern Illinois University Law Journal 17: 559–571. ——— (2004) “Natural Law: The Classical Tradition,” in J. Coleman and S. Shapiro (eds.) The Oxford Handbook of Jurisprudence and Philosophy of Law, Oxford: Oxford University Press, 1–39. FitzPatrick, W. (2006) “The Intend/Foresee Distinction and the Problem of ‘Closeness,’ ” Philosophical Studies 128: 585–617. ——— (2012) “The Doctrine of Double Effect: Intention and Permissibility,” Philosophy Compass 7(3): 183–196. ——— (2013) “Intention, Permissibility and Double-Effect,” Oxford Studies in Normative Ethics 2: 97–126. Foot, P. (1977) “Euthanasia,” Philosophy and Public Affairs 6: 85–112. ——— (2002) “Killing and Letting Die,” Moral Dilemmas and Other Topics in Moral Philosophy, Oxford: Clarendon Press, 79–87. Gunderson, M. (2004) “A Kantian View of Suicide and End-of-Life Treatment,” Journal of Social Philosophy 35(2): 277–287. Harris, J. (1995) “The Philosophical Case against the Philosophical Case against Euthanasia,” in J. Keown (ed.) Euthanasia Examined: Ethical, Clinical and Legal Perspectives, Cambridge: Cambridge University Press, 36–45.

196

Thomas S. Huddle

Hill, T.E. (1983) “Self-Regarding Suicide: A Modified Kantian View,” Suicide and Life-Threatening Behavior 13(4): 254–275. Holland, R.F. (1980) “Suicide,” in Against Empiricism: On Education, Epistemology and Value, Totowa, NJ: Barnes and Noble Books, 143–157. Huddle, T.S. (forthcoming) “What’s a CIED Got to Do with It? The Implications of Deactivating Life-Sustaining CIEDs for the K/ATD Distinction,” Cambridge Quarterly of Health Care Ethics. Kagan, S. (1988) “The Additive Fallacy,” Ethics 99: 5–31. Kamisar, Y. (1958) “Some Non-Religious Views against Proposed ‘Mercy-Killing’ Legislation,” Minnesota Law Review 42(6): 969–1042. Kamm, F. (1999) “Physician-Assisted Suicide, the Doctrine of Double Effect, and the Ground of Value,” Ethics 109(3): 586–605. ——— (2007) “Ending Life,” in L.P. Francis, A. Silvers, and R. Rhodes (eds.) The Blackwell Guide to Medical Ethics, Oxford: Wiley-Blackwell, 142–161. Kass, L. (2002) “ ‘I Will Give No Deadly Drug’: Why Doctors Must Not Kill,” in K. Foley and H. Hendin (eds.) The Case against Assisted Suicide; for the Right to Endof-Life Care, Baltimore: The Johns Hopkins University Press, 17–40. Knobe, J. (2003) “Intentional Action and Side Effects in Ordinary Language,” Analysis 63(3): 190–194. McIntyre, A. (2001) “Doing Away with Double Effect,” Ethics 111: 219–255. McMahan, J. (2002) The Ethics of Killing: Problems at the Margins of Life, Oxford: Oxford University Press. ——— (2009) “Intention, Permissibility, Terrorism, and War,” Philosophical Perspectives 23: 345–372. Murphy, M. (2001) Natural Law and Practical Rationality, Cambridge: Cambridge University Press. Nelkin, D. and S. Rickless (2014) “Three Cheers for Double Effect,” Philosophy and Phenomenological Research 89(1): 125–158. ——— (2015) “So Close, Yet So Far: Why Solutions to the Closeness Problem for the Doctrine of Double Effect Fall Short,” Nous 49: 376–409. Pellegrino, E. (2001) “The Goals and Ends of Medicine: How Are They to Be Defined?” in R.J. Bulger, J.P. McGovern, and D.P. Sulmasy (eds.) Physician and Philosopher: The Philosophical Foundation of Medicine: Essays by Dr. Edmund Pellegrino, Charlottesville, VA: Carden Jennings, 56–65. Pellegrino, E. and D. Thomasma (1987) “The Conflict between Autonomy and Beneficence in Medical Ethics: Proposal for a Resolution,” Journal of Contemporary Health Law and Policy 3: 23–46. Quill, T., R. Dresser, and D. Brock (1997) “The Rule of Double Effect—A Critique of Its Role in End-of-Life Decision-Making,” New England Journal of Medicine 337: 1768–1771. Quinn, W. (1989) “Actions, Intentions, and Consequences: The Doctrine of Double Effect,” Philosophy and Public Affairs 18(4): 334–351. Rachels, J. (1975) “Active and Passive Euthanasia,” The New England Journal of Medicine 292: 78–80. ——— (1979) “Killing and Starving to Death,” Philosophy 54(208): 159–171. ——— (1986) The End of Life: Euthanasia and Morality, Oxford: Oxford University Press. Ramsey, P. (1968) “The Morality of Abortion,” in D.H. Labby (ed.) Life or Death: Ethics and Options, Seattle: University of Washington Press, 60–93.

Suicide 197 Scanlon, T. (2009) Moral Dimensions, Cambridge: Harvard University Press. Shils, E. (1968) “The Sanctity of Life,” in D.H. Labby (ed.) Life or Death: Ethics and Options, Seattle: University of Washington Press, 2–38. Singer, P. (1993) “Taking Life: Humans,” in P. Singer (ed.) Practical Ethics, Cambridge: Cambridge University Press, 175–217. Slote, M. and P. Petit (1984) “Satisficing Consequentialism,” Proceedings of the Aristotelian Society, Supplementary Volumes 58: 139–163+165–176. Stuchlik, J. (2012) “A Critique of Scanlon on Double Effect,” Journal of Moral Philosophy 9: 178–199. Sulmasy, D. and E. Pellegrino (1999) “The Rule of Double Effect: Clearing Up the Double Talk,” Archives of Internal Medicine 159: 545–550. Sumner, L.W. (2011) Assisted Death: A Study in Ethics and Law, Oxford: Oxford University Press. Thomson, J. (1999) “Physician-Assisted Suicide: Two Moral Arguments,” Ethics 109(3): 497–518. Tooley, M. (1994) “An Irrelevant Consideration: Killing versus Letting Die,” in B. Steinbock and A. Norcross (eds.) Killing and Letting Die (2nd Edition), New York: Fordham University Press, 103–111. Truog, R.D., M.L. Campbell, J.R. Curtis, C.E. Haas, J.M. Luce, G.D. Rubenfeld, C.H. Rushton, and D.C. Kaufman (2008) “Recommendations for End-of-Life Care in the Intensive Care Unit: A Consensus Statement by the American Academy of Critical Care Medicine,” Critical Care Medicine 36(3): 953–963. Velleman, D. (1999) “A Right of Self-Termination?” Ethics 109(3): 606–628. Wedgwood, R. (2011) “Defending Double Effect,” Ratio 24: 384–401. Wiggins, D. (2006) Ethics: Twelve Lectures on the Philosophy of Morality, Cambridge: Harvard University Press. Williams, G. (1958) The Sanctity of Life and the Criminal Law, London: Faber and Faber. Woolard, F. (2012) “The Doctrine of Doing and Allowing I: Analysis of the Doing/ Allowing Distinction,” Philosophy Compass 7(7): 448–458. Young, R. (2007) Medically Assisted Death, Cambridge: Cambridge University Press.

Part IV

Other Parties and the End of Life

12 Grief and End-of-Life Surrogate Decision-Making Michael Cholbi

Summary Because an increasing number of patients have medical conditions that render them incompetent at making their own medical choices, more and more medical choices are now made by surrogates, often patient family members. However, many studies indicate that surrogates often do not discharge their responsibilities adequately, and in particular, do not choose in accordance with what those patients would have chosen for themselves, especially when it comes to end-of-life medical choices. This chapter argues that a significant part of the explanation of such surrogate failure is that family surrogates are likely to undergo anticipatory grief when making end-of-life decisions. After clarifying both the emotional structure and object of grief, I propose that the pending death of a loved one induces an emotional conflict in surrogates between the care demanded by their responsibilities as surrogates and the attachment surrogates feel toward their dying loved one, a conflict surrogates “resolve” in the direction of attachment rather than care. This hypothesis helps to explain both surrogates’ general inability to exercise “substituted judgment” on behalf of their loved ones and a wide swath of the particular data regarding this inability (e.g., that surrogates more often err by choosing overtreatment). I conclude by considering possible clinical and philosophical responses to this hypothesis.

Introduction One unmistakable feature of modern medicine is its increasing reliance on surrogate decision-makers, individuals charged with making medical care decisions on behalf of individuals whose illnesses make them incapable of making such decisions. Because of increasing longevity and the greater prevalence of chronic illnesses associated with aging, more and more of the responsibility for medical decision-making falls on these surrogates. The purpose of this chapter is to explore a novel set of practical concerns regarding surrogate decision-making at the end of life—namely, the role

202

Michael Cholbi

played by the emotions in such decision-making. The clinical settings in which end-of-life care is typically provided are hothouses of emotion. For medical professionals, providing care to end-of-life patients is itself an intensive act of emotional labor, putting great strain on their capacities for empathy (Bagdasarov and Connelly 2013). For end-of-life patients, emotions such as fear, shock, anger, sadness, and regret can compound the physical toil of a body in its last throes (Kübler-Ross 1969). For family members of the critically ill, anxiety, anger, and emotional fatigue are common (Pochard et al. 2001, Davidson et al. 2007). Add volatile interpersonal dynamics to these emotional stresses, and the result is an environment where what moral philosophers call the “moral emotions”—guilt, shame, resentment, and so forth on the “negative” side, pride, gratitude, hope, and so forth on the “positive” side—are never too far below the surface (Dillon 2010). Unfortunately, ethicists have said little to illuminate how moral emotions shape end-of-life care and decision-making. A general account of this relationship is beyond the scope of this chapter. My purpose is narrower: to investigate the role that one particular moral emotion, grief, has in end-of-life medical decisions. I shall argue that that even if grief itself is often a rational response to one’s loss and the relationship transformed by that loss, we have good reason to think that grief nevertheless compromises the specific sort of rationality needed to act effectively as a loved one’s medical surrogate. More specifically, I propose that the anticipatory grief experienced by surrogate decision-makers helps to explain the well-corroborated finding that surrogates often fail to satisfy the “substituted judgment” standard widely thought to provide surrogate decision-making its ethical foundation. Irrespective of whether I am correct in these conclusions, I hope to inspire other ethicists to examine end-of-life decisions, and medical decision-making more generally, through the lens of the moral emotions. The first section provides an overview of the current clinical environment surrounding surrogate decision-making and describes the currently accepted rationale and standards for such decision-making. In the second section, I adumbrate the evidence indicating that surrogates routinely fail to satisfy such standards. I defend a crucial philosophical claim regarding grief in the third section—namely, that grief’s object—what it is that individuals grieve for across the course of their bereavement—is the transformation or loss of the intimate relationship the grieving individual had with the deceased. In the fourth section, I apply this claim to surrogate decision-making to argue that such decision-making often occurs in the midst of anticipatory (i.e., premortem) grief, and that decision-making in that context involves a powerful emotional conflict between the decision-maker’s attachment to the ill patient and the demand that the decision-maker exercise care on that patient’s behalf. The difficulties surrogates face in exercising the kind of “substituted

Grief and Surrogate Decision-Making 203 judgment” called for by their roles as surrogates can be partially explained by their resolving this conflict (understandably, mind you) in the direction of attachment rather than care. The fifth section considers some possible responses to this phenomenon.

The Rise and Rationale of Surrogate Medical Decision-Making In modern industrialized societies, dying has changed rapidly in the space of just a few generations. Where our grandparents were more likely to die of accidents or communicable diseases, we are now much more likely to die of chronic or degenerative diseases. The leading causes of death at the turn of the twentieth century included flu, pneumonia, and tuberculosis, whereas the leading causes of death today include heart disease, cancer, and lung disease (Jones et al. 2012). As a result, far more individuals now reach old age than in the past (Arias 2014). That life spans have increased so dramatically has its downside, though. For one, the conditions that cause the majority of deaths these days tend to lengthen the process of dying and lead to sustained stretches wherein patients have depleted or intermittent consciousness. Moreover, far more individuals live long enough to become demented for some portion of their lives. A principal result of these changes in how we die is that individuals are now more likely to undergo periods in which their ability to make medical decisions for themselves is sufficiently compromised to render them incompetent. Their medical conditions make it impossible for them to understand their diagnoses, rationally engage with alternative treatment options, communicate their decisions, and so forth. About 40 percent of hospital patients are sufficiently cognitively impaired that they lack the capacity to make competent medical decision for themselves (Fritsch et al. 2013: 126). Prior to the 1960s, incompetent patients would likely have had their medical decisions made by their medical caregivers (Rothman 1992). However, as patient autonomy has become the cornerstone of contemporary clinical ethics, the right to make medical decisions is increasingly placed in the hands of a decision-making surrogate or proxy. Surrogates now make 75 percent of the medical decisions for patients hospitalized with serious illnesses (Hiltunen et al. 1999) and around half of the medical decisions for patients in nursing homes (Kim et al. 2002). Who serves as a patient’s surrogate is determined by an established legal and professional protocol. In some cases, the patient will have designated a surrogate prior to becoming incompetent. Usually this surrogate is a family member or loved one. In cases where a patient has not assigned the power of surrogacy to anyone, the patient’s next of kin is designated as her surrogate (and if this is not possible, an administrative surrogate—a social worker, attorney, etc.— may be appointed).

204

Michael Cholbi

The underlying rationale for using surrogates in such circumstances is that this is the best way for patients who are not competent to nevertheless exercise autonomy regarding their own medical care. Since they themselves are incompetent and so cannot exercise autonomy on their own behalf, a surrogate, ideally one knowledgeable about the patient’s preferences and values, is designated to make medical choices regarding the patient. The patient’s “autonomy” is thus exercised at one remove: when a patient cannot choose which interventions will or will not be administered to her, she will nevertheless receive treatments that reflect her preferences and values. Note, then, that the surrogate is not supposed to be engaging in “hard paternalism,” choosing for the patient what the surrogate believes to be best according to some supposedly objective or all-things-considered standard (Feinberg 1986: 12). Rather, surrogates are charged with choosing that course of care that answers to the patient’s preferences and values (Buchanan and Brock 1990). This autonomy-based rationale for surrogate decision-making thus implies a counterfactual standard for responsible surrogate decision-making: a responsible surrogate chooses those treatment options that the patient would have chosen for herself were she competent and able to do so. The surrogate is thus to exercise substituted judgment on the patient’s behalf. Those who have longstanding relationships with a patient—the patient’s spouse, children, parents, siblings, and so forth—appear best suited to exercise this sort of substituted judgment. After all, who is in a better position to know what someone wants than those who know her best and longest?

The Unreliability of Surrogate Decisions But do surrogates in fact discharge the decision-making responsibilities assigned to them? How often and to what extent do patient-designated or next of kin surrogates exercise substituted judgment properly—that is, choose those courses of treatment that patients would have chosen for themselves? Such questions are not easy to investigate, especially because establishing what patients would prefer in contexts where they would be incompetent is not straightforward. Nevertheless, empirical investigations of surrogate decisionmaking are discouraging. Surrogates are not particularly reliable with respect to the counterfactual standard of exercising substituted judgment. Shalowitz et al. (2006) is the most comprehensive metastudy of the relevant research. Among its key findings: •

•

Patient-designated and next of kin surrogates accurately predict patient preferences about two-thirds of the time, whereas attending physicians predict patient preferences just over half the time. Surrogates are slightly more accurate regarding choices about current patient health (e.g., whether to administer more pain medication) than

Grief and Surrogate Decision-Making 205

• • • • •

they are about projected future medical conditions, such as dementia or late-stage cancer. Surrogate accuracy does not significantly vary based on type of medical intervention proposed (e.g., surgery versus pharmaceuticals). Surrogates tend to err in the direction of overtreatment, advocating treatments that patients would refuse. Surrogates’ specific relationship to patient (sibling vs. spouse vs. child, etc.) is not correlated with surrogate predictive accuracy. Patient-designated surrogates are not more accurate than legally assigned administrative surrogates. Surrogates having prior discussion of patient treatment preferences slightly reduced accuracy of subsequent surrogate predictions.

Needless to say, the hope that surrogates will “stand in” for the patient and thereby enable the patient to receive those treatments that reflect her preferences does not seem well founded. Shalowitz et al. dismally conclude that patient-designated or next of kin surrogates satisfy the counterfactual condition implied by the substituted judgment standard at a rate “only slightly better than chance” (2006: 495). That surrogates are not as reliable in their decision-making as we might hope can be explained based on what surrogates report about their own methods of decision-making. Indeed, if surrogates’ own reports of how they go about making medical decisions for incompetent or incapacitated patients can be trusted, they often make use of methods that probably do not correlate with patient preferences (Vig et al. 2006). Some of the methods surrogates report using at least purport to track patient preferences. For example, just over twothirds of surrogates report relying on conversations with patients as a guide to the decisions they make as surrogates, and about half of surrogates infer patient preferences from their background knowledge of the patient. But among methods that at least purport to track patient preferences, some warrant skepticism. For instance, about one in six surrogates reports that she used the values or life experiences she presumed to have in common with the patient to make decisions for the patient. These surrogates often report that they have an “inner sense” of what the patient would want or that they will “just know” what the patient would want when the time comes to exercise medical judgment on the patient’s behalf. (Only about 10 percent of surrogates report having relied on written statements made by patients.) Whatever their reliability, at least these methods are facially focused on the patient and her preference. Worse still, many surrogates concede that they use methods of decision that bear no apparent relationship to patient preferences at all. About one quarter of surrogates report simply using their own medical care preferences to make decisions for the patient. About 15 percent make these decisions based on their own interests and 12 percent make them based on their own religious convictions.

206

Michael Cholbi

The methods surrogates draw upon to make their medical decisions are thus a hodgepodge of suspect reliability and clear irrelevance. That surrogate choice often fails to reflect patient preferences is unsurprising (Fritsch et al. 2013).

The Nature of Grief Exactly why surrogates are so ineffective in exercising substituted judgment has not been investigated in much depth. Perhaps surrogates’ unreliability has a cognitive origin: surrogates lack sufficient information regarding patient preferences or do not understand what the substituted judgment standard asks of them. Alternatively, their unreliability may have a moral explanation. Surrogates may simply be indifferent to the moral demands implied by the substituted judgment standard. While these considerations can no doubt play an explanatory role, I would like to explore a partially affective explanation— namely, that surrogates face an emotional conflict between their own grief and the emotional detachment needed to choose as others would choose for themselves. Artists are fond of depicting grief as a flirtation with madness and unreason. Hamlet’s paramour Ophelia jabbers in riddles after the death of her father, Polonius. In his memoir A Grief Observed, C. S. Lewis speaks of the “mad words, the bitter resentment, the fluttering in the stomach, the nightmare unreality” that sent him “round and round” after the death of his wife Joy (2009: 69). Anne Roiphe’s grief leads her into an insomniac haze (2008). Indeed, that grief can seem antithetical to rational self-control is part of the reason why ancient philosophers typically saw grief as a threat to moral virtue and to a well-lived life (LeBarge 2012). To contemporary eyes, such worries appear hyperbolic. Nevertheless, the powerful affective manifestations of grief raise questions about the rational capacities of the grieving. To get a handle on how grief influences end-of-life decisions, we must first get a handle on the emotional complexities of grief itself. Categorizing grief as an emotion risks oversimplifying it. While philosophers disagree about whether emotions are best thought of as cognitive states such as beliefs, conative states such as desires, or states of bodily awareness, they standardly think of emotions as responses to particular facts that elicit them. Anger, for example, is generally analyzed as the emotion elicited by the sense that one has been wronged; fear is generally analyzed as the emotion elicited by the belief that one’s well-being is under threat; and so on. The state that elicits the emotion is the emotion’s object. In the case of grief, one complication here is that grief does not seem to be a single response. Rather, grief unfolds as an emotional process. Since the work of Kübler-Ross (1969), the notion that grief fits into a five-stage

Grief and Surrogate Decision-Making 207 pattern (denial-anger-bargaining-depression-acceptance) has been ascendant. While this model is certainly incorrect at the level of detail—few bereaved persons undergo precisely those five stages in that order (Maciejewski et al. 2007; Konigsberg 2011)—that grieving involves multiple affective states is widely recognized. Grief is more than simply sadness or suffering at the death of another. Rather, grief often involves anxiety, joy, fear, and other affective states along the way. As Wittgenstein noted, grief differs from other emotions in that it is less a momentary sensation than a prolonged emotional engagement with a change in one’s condition (Wittgenstein 1958: 174–175). In this respect, grief seems to resemble a kind of “narrative” response, a complex emotional arc with a beginning, a climax, a resolution, and so forth (Goldie 2011; Higgins 2013). But the composite emotional nature of grief only seems to make its object more elusive: imagine an individual whose loved one dies and she undergoes an episode of grief that conforms to Kübler-Ross’s denial-anger-bargainingdepression-acceptance pattern. In what sense are these particular emotional stages parts of the same grieving episode? If grief is a story, what is this story about? What object unites these stages as elements of a larger emotional episode? That we “grieve for” a person might be taken to imply that the object of grief is the bereaved individual’s belief that the deceased has been harmed by death, perhaps by being deprived of further opportunities to enjoy the goods of life. Admittedly, some of what we experience as bereavement is otherregarding in this way: we grieve for what the deceased have lost by dying. But this is implausible inasmuch as we do not grieve for everyone, despite the fact that many people “lose out” on the goods of life by dying. Thousands die each day, but we individually do not grieve for each of them, or for even a tiny fraction of them. If grief’s object were what the deceased have lost by dying, we would grieve for far more deaths than we in fact do. Genuine grief instead requires some sense of familiarity or intimacy with the deceased, a fact that implies that grief’s object is more self-regarding. In other words, the object of grief is not our awareness of what the deceased lose by dying but what we the bereaved lose by their dying. But loss of what exactly? We might think that what is lost are the various goods that the deceased individual provided to the bereaved: companionship, comfort, and so on. However, this claim is at odds with the fact that we can grieve those who disappointed us by failing to provide these goods (as when an adult grieves the death of a parent who abandoned her at a young age). We instead grieve for those in whom we invest our hopes regarding our wellbeing, even if they ultimately do not live up to those hopes (Nussbaum 2001: 81–82). Furthermore, to think that the loss that is grief’s object is the loss of the goods provided by the deceased overlooks the sense of irreplaceability that

208

Michael Cholbi

the bereaved attach to the deceased. To suppose that the right response to having lost, say, one’s spouse is to immediately find another spouse with the same desirable attributes is to overlook that the value of our relationships with one another, especially our loving relationships, cannot be reduced to the value of the goods that others provide us. We relate to others as persons first and foremost, not as baskets of goods. The object of grief, I propose, resides in the fact that death necessitates a transformation in the relationship that a surviving individual has with the deceased. Our relationships with the deceased do not necessarily end with their deaths, but they cannot but be changed by death. Obviously, the relationship becomes much more one-sided, given the impossibility of the deceased communicating with the bereaved. Lacking agency, the deceased can no longer be directly addressed with our pleas. Certain characteristic attitudes a person may have had toward the deceased no longer make apparent sense. Certain ethically salient acts (e.g., forgiveness or the keeping of promises) can no longer be performed in a straightforward way. The activities that constitute the relationship—conversations, dining, enjoying nature, and so on—cannot continue. With the death of one of its members, a relationship crosses a critical Rubicon. As social animals, human beings’ sense of themselves and their orientation in the world is significantly constituted by their relationships with others. Many bereaved persons report feeling a loss of a part of themselves. This is not surprising, since the death of the partner in a crucial relationship can lead to a sense of identity crisis in the bereaved as they struggle to grasp their own situation in light of a reconfigured relationship with the deceased. Note that the relationship need not be lost in its entirety. It is instead transformed by the other’s death. Grief, I suggest, is the emotionally fraught process by which we reground our relationship with a deceased individual on new terms. My concern here is with how grief in turn shapes end-of-life decisionmaking by patient-appointed or next of kin surrogates. Of course, the patients in question are not yet deceased. Hence, the surrogates are undergoing anticipatory grief, grief predicated on the awareness that the patient is dying (or is likely to die). Yet anticipatory grief has the same source as “ordinary” grief— namely, the awareness that one’s relationship with a particular person is (or is about to be) transformed by that person’s death. It may seem odd that we can undergo an emotion or emotional process prior to its object being present. But this is commonplace. We sometimes experience anticipatory anger in the expectation that we are about to be wronged. We sometimes experience anticipatory joy in the expectation we are about to have our sense of wellbeing augmented. As an emotion, fear thrives on the anticipation of a threat. In this respect, that grief can precede the object that, properly speaking, triggers it is to be expected.

Grief and Surrogate Decision-Making 209

Surrogate Competence and the Care-Attachment Conflict Many surrogates, I claim, are asked to make end-of-life medical decisions in the midst of anticipatory grief. Anticipatory grief is the first stage in a process whose psychological object is the crisis that an intimate’s decline and impending death bring about in the surrogate’s relationship with that intimate. This crisis, I now argue, compromises the surrogate’s capacity to exercise the substituted judgment that, according to the standard model of surrogate decision-making, a surrogate is obliged to exercise on the intimate’s behalf. More specifically, surrogates undergoing anticipatory grief experience a conflict between two central emotional facets of this relationship—namely, attachment and care. As we saw in the third section, we experience grief at the loss of relationships with those who matter to us. But others can matter to us in multiple ways. On the one hand, we typically care for those we grieve. To care for another is to regard her as important in a specific way—to wit, to see her well-being or concerns as an especially powerful sources of reasons for acting. In caring about another, we render ourselves emotionally vulnerable to their states. To care about another is to feel joy and satisfaction when the object of one’s care is flourishing and frustration over its misfortunes; anger at agents who heedlessly cause such misfortunes; pride in the successes of the object of care and disappointment over its failures; the desire to help ensure those successes and to help avoid the failures; fear when the object of care is in danger and relief when it escapes unharmed. (Jaworska 2007: 560–561) The care stance is thus one in which the other is a source of second-personal concern—that is, their weal or woe is given greater weight in one’s choices and actions than that of strangers or of those with whom one does not have a caring relationship. On the other hand, we generally grieve for those with whom we stand in a relationship of attachment. To experience attachment to a person is for that individual to serve as a source of security. Monique Wonderley (2016) analyzes attachment as having four features: 1. 2. 3.

The attached individual seeks to be proximate to, and to interact, with the person to whom she is attached. The attached individual experiences distress upon separation from the person to whom she is attached (or at the prospect of such separation). The attached person derives a sense of security from the presence of the person to whom she is attached.

210 4.

Michael Cholbi The person to whom she is attached is a “non-substitutable particular” for the attached individual—that is, only that person instantiates features 1–3 in precisely the way she does.

To the extent that security is a human need, we thus come to need those to whom we are attached. Attachment thus has a decidedly first-personal quality. Our attachments to others give us reasons to act that are inextricably tied to our own sense of identity and well-being. Not every important relationship involves both care and attachment: Adult siblings, for instance, may care for another but not have the sense of felt neediness or security that characterizes attachment. A person may be romantically obsessed with someone to whom he is attached but about whom he does not care. However, many relationships that matter to us will include both attachment and care. Long-term romantic partners, for example, generally are attached to one another, but also care for one another. And the balance between care and attachment in relationships may shift over time. A relationship with one’s parents may shift from one dominated by attachment to one dominated by care once one’s parents become aged, debilitated, or ill. Similarly, as one’s children mature, care may begin to recede in importance as attachment grows in importance. Critically for our purposes, attachment and care generate two competing sets of emotional prerogatives. The desires rooted in attachment to another lead us to hold fast to her as a source of security, to fear (and hence avoid) the distress associated with loss of that individual, and to seek to sustain the relationship with that individual on familiar and welcoming terms. The prerogatives of care, in contrast, lead us to be concerned for another’s welfare for her own sake, occasionally even at cost to ourselves. My hypothesis is that in the midst of anticipatory grief and the relationship crisis wrought by it, surrogates undergo a clash between these two sets of prerogatives, a clash in which (unsurprisingly) attachment’s prerogatives often prove stronger than those of care. Grieving surrogates must in effect undertake two cognitively and emotionally taxing tasks simultaneously: reorienting or reconfiguring a relationship with a person who is rapidly withdrawing from that relationship while also exercising impartial judgment on that person’s behalf. Grief can thus cloud surrogate decision-making thanks to the difficulties of shifting between the first-personal attachment-based evaluative perspective characteristic of grief and the secondpersonal care-based evaluative perspective demanded of competent surrogates. This hypothesis helps explain, in both general and specific terms, the empirical findings laid out in the second section. Surrogates report significant stress surrounding the accuracy of the medical choices they make (Wendler and Rid 2011). This stress, I suggest, can partially be traced to the conflict between care and attachment that surrogate grief evokes. In a moment when attachment

Grief and Surrogate Decision-Making 211 is threatened, surrogates are asked to undertake an unusually demanding care task: deciding what is best for an incompetent other by that other’s lights. This conflict can also account for specific findings about how surrogates tend to err. First, that surrogates undergoing a care-attachment conflict will often err in favor of overtreatment relative to patients’ own wishes is readily explained by our hypothesis. Surrogates wish to maintain their attachment with the patient on familiar terms and so may opt for aggressive treatments in the hope that such treatments will enable that attachment to continue. Similar considerations indicate how my hypothesis explains why surrogates are particularly unreliable regarding end-of-life decision-making. With other medical decisions, the life of the intimate is not so directly at stake and so the surrogates’ attachment is less threatened in those decision-making contexts. My hypothesis also helps account for surrogates’ common and almost magical belief in the congruence of their values and preferences with those of the patient (the putative “inner sense” of the patient’s values and preferences). Surrogates in the midst of a relationship crisis will understandably want to maintain attachment with the dying patient. The belief in this congruence likely reflects a yearning for that attachment. After all, by identifying her preferences as the patient’s preferences, the surrogate sustains a sense of unity or presence with the patient. Finally, prior communication between surrogates and patients regarding patient values and preferences often fails to improve subsequent surrogate decision-making. Here my hypothesis implies that this failure may be due to the fact that such communication will often occur near the onset of the surrogate’s anticipatory grief that takes her rapidly changing relationship with the dying patient as its object. The stresses of anticipatory grief may make it difficult for surrogates to process or fully attend to statements made by the patient. Obviously the hypothesis that surrogates undergo anticipatory grief and an ensuing attachment-care conflict that is often resolved in favor of the former does not explain every instance in which surrogates fail to exercise substituted judgment or every facet of those failures. However, given the absence of extant explanations of surrogate unreliability in either the philosophical or psychological literatures, and the hypothesis’s fruitfulness in explaining the available evidence concerning surrogate unreliability, I contend that this hypothesis deserves serious consideration. From a broader philosophical perspective, this hypothesis implies that in defending surrogacy and the standard of substituted judgment, many bioethicists have unwittingly adopted too harmonious or unified a picture of human rationality. Note that I have not said that anticipatory grief, or the attachmentcare conflict it tends to invite, is an irrational reaction on the part of surrogates charged with end-of-life decision-making responsibilities. Quite the contrary. Such reactions will often be perfectly rational responses to the changes that

212

Michael Cholbi

the impending death of a loved one brings about or anticipates (Cholbi 2016). However, human rationality is multifaceted and in some respects fragmented. One and the same action or reaction may be rational along one dimension but irrational along another dimension. Such is the case with respect to anticipatory grief and surrogate medical decision-making, I propose. A bereaved surrogate can be torn between a legal (and moral duty) to be a reliable decision-maker on a patient’s behalf and a prudential imperative to grieve so as to sustain whatever attachment she may have with that patient. The former duty and the latter imperative can fail to coincide, despite the fact they owe their existence to a common source—namely, the surrogate’s presumed intimacy with the patient. In such circumstances, it may prove effectively impossible to divorce the rational, attachment-based responses evinced by anticipatory grief from the irrationality surrogates often exhibit when they attempt to exercise substituted judgment on behalf of the very individual for whom they are grieving. At the very least, faulting surrogates for resolving this conflict by prioritizing their own attachment to the dying individual over their care-based duty to exercise substituted judgment for that dying individual seems unsympathetic, even hard-hearted.

Possible Responses The claims I have defended regarding how anticipatory grief undermines the sort of rationality necessary for surrogates to exercise substituted judgment on behalf of patients undergoing end-of-life care are troubling. After all, a going assumption of the existing regime of medical ethics is that surrogates are well situated to make rationally grounded judgments regarding another’s care precisely because most surrogates stand in an intimate relationship with the other individual (Lindemann and Nelson 2014: 163). If sound, my arguments demonstrate that exactly the opposite is the case: intimate surrogates undergo a conflict between their duties of care and the attachment threatened by the patient’s death, a conflict that makes surrogates substantially less rational qua proxies—that is, far less able to make those choices that the patient herself would have made while competent. Of course, my hypothesis remains somewhat speculative. To be empirically vindicated, investigations would have to be made targeting two relevant populations: surrogates facing end-of-life decisions undergoing anticipatory grief and surrogates facing end-of-life decisions not undergoing anticipatory grief. My hypothesis would predict greater difficulties in exercising substituted judgment in the former population than in the latter. At the very least, my hypothesis suggests an agenda for research bridging clinical psychology and clinical ethics. In the meantime, however, how ought the bioethics and clinical communities best respond to this conclusion? The most radical response would be to declare the regime of surrogacy a failure and jettison it altogether—that is, to

Grief and Surrogate Decision-Making 213 have medical decisions for incompetent patients not be made by designated or legally mandated familial surrogates but by institutional professionals. This option may seem attractive inasmuch as such professionals, because they lack the intimate ties to particular end-of-life patients, would not undergo the emotional conflict we have articulated and so would be better positioned to exercise substituted judgment on a patient’s behalf. I am reluctant to embrace this option, however. While I have made clear that there are strong reasons to doubt that intimate surrogates can circumvent their anticipatory grief and exercise the proper form of surrogate rationality, the specter of such decision-making being automatically delegated to physicians, social workers, attorneys, and so forth is likely to be highly traumatic for families. To deny family members or other intimates the right to make medical decisions for incompetent patients may increase its epistemic reliability, but at the cost of alienating the former from the latter just when the desire for a sense of connection or intimacy is at its greatest. And the great majority of patients prefer family members or loved ones to make their medical decisions for them when they are incompetent, in large measure due to the belief that such surrogates will in fact make the medical decisions that the patient would (Kelly et al. 2012). A second possibility would be to change the standard of judgment applied to surrogates. Under the substituted judgment standard, a surrogate’s duty is to decide in accordance with what the patient would have wanted while competent. However, alternative standards are possible. One would be to require surrogates to decide only on the basis of a patient’s past express statements of her wishes—that is, not on the basis of potentially iffy conjectures about the patient’s preferences or values but solely based on relatively unambiguous patient declarations. Another possibility would be to hold surrogate choice to standards that are not indexed to patient preferences, whether explicit or inferred, at all. For instance, surrogates could be asked to decide solely on what is in the patient’s best interests, objectively understood. The difficulty with proposals to apply these other standards to surrogate decision-making is that it is not clear they evade the challenges presented by anticipatory grief. A surrogate asked to decide solely on the patient’s past explicit statements can nevertheless, because of the care-attachment conflict I have identified, disregard or unwittingly misinterpret those statements. Indeed, the earlier evidence that communication between surrogates and patients does not contribute much to the reliability of surrogate decisions supports this contention. Similarly, surrogates asked to decide based on the patient’s objective best interests can nevertheless have their decisions clouded by the very same mechanisms of anticipatory bereavement that I have argued cloud the decisions of surrogates asked to exercise substituted judgment. In all likelihood, surrogates in anticipatory bereavement will be prone to represent judgments of their own attachment-based interests as judgments of a patient’s best interests.

214

Michael Cholbi

But so long as surrogates are employed and the presumption is that these surrogates will typically be intimately familiar with the patient, what can be done about the apparent gap between surrogate choice and patient preferences, a gap that (I have argued) can be partially explained by surrogates undergoing anticipatory grief that impedes their capacity to exercise substituted judgment? One clinical response to the way in which anticipatory grief undermines the rationality needed for surrogates to exercise substituted judgment would be to attempt to counteract the effects of anticipatory grief on surrogate decisionmaking. Such a strategy could involve promoting awareness of anticipatory grief among clinicians, patients, and their families. At the very least, my conclusion implies that grief counseling, a common feature of the end-of-life care environment nowadays, should begin not when the patient dies but when it becomes evident that the patient’s death is imminent. Such a policy would require more candor than is sometimes in evidence in end-of-life settings: traditionally, the medical profession has believed in the therapeutic value of hope and has thus been reluctant to tell families outright that their loved ones have a negligibly small chance of recovery and will almost certainly die (McMillan et al. 2014). Still, providing “pre-mortem” grief counseling may lead surrogates to be more aware of the effects of their own grief on their decision-making and thus enable them to compartmentalize that grief so as to focus on their duty to exercise substituted judgment for the patient. Beyond this, my conclusion suggests that the ethical rationale for surrogacy may need to be reconfigured so as to displace surrogate reliability as the sole factor in judging surrogate performance. In other words, it may be possible to retain surrogacy but ground it, at least partially, on a different ethical basis. From the patient perspective, surrogacy is ostensibly valuable because it enables the medical decisions made for the patient to reflect the patient’s values or preferences. However, attitudes toward surrogacy are complex. First, while patients do anticipate that their surrogates will make a decision reflecting their (the patients’) values and preferences, patients recognize that surrogates have a difficult responsibility and are therefore reluctant to blame their surrogates for decisions that deviate from their preferences and values (Fins et al. 2005). Indeed, patients generally understand that surrogate decision-making is burdensome and that surrogates must live with the medical choices they make. These considerations suggest that while patients do in fact value their autonomy and so desire a surrogate to exercise it on their behalf, they are fairly forgiving about how surrogates exercise their autonomy so long as it is done conscientiously. Second, there may be reasons besides the value of autonomy, especially reasons rooted in patients’ relationships with their surrogates, for valuing surrogacy arrangements. We might attempt to identify these reasons by considering whether, if we knew beforehand that whomever we designate as our medical surrogate would nevertheless make medical decisions at odds with those we

Grief and Surrogate Decision-Making 215 would make for ourselves, we would nevertheless proceed to designate our chosen individual as our surrogate. If the answer is “yes,” that implies that something besides the purported reliability of the surrogate may explain the desire to have surrogates make our medical choices for our incompetent selves. One possibility is that designating a surrogate may be seen as an act within a relationship. In selecting a surrogate, we may not be so much identifying the individual who we believe will best exercise the quasi-fiduciary responsibilities of a surrogate. Rather, we may be signaling to an individual our love or esteem for her by selecting her as our surrogate. If so, then the message sent by patients to their medical surrogates may be less egocentric and more other-regarding, less “you know what I want (or will want)” and more “you are the one that I want to choose for me.” In that case, surrogacy is more akin to an honor bestowed upon the surrogate than a moral responsibility transferred to him, for the patient’s concern is less with “what is decided” than with “who does the deciding” (Lindemann and Nelson 2014: 166). On this model, the surrogate’s authority is not simply a reflection of her presumed knowledge of the patient’s preferences. And although it may still be desirable for surrogates to exercise appropriate substituted judgment, it may be less crucial to the patient that the surrogate make the medical decisions her counterfactually competent self would have made.

Conclusion In tandem, the two measures described in the previous section—greater emphasis on pre-mortem grief counseling and less concern that surrogates replicate the decisions patients themselves would have made while competent— narrow the gap between surrogate decisions and the moral expectations imposed on them, both by improving surrogate capacity to exercise substituted judgment and by de-emphasizing the centrality of reliability in evaluating surrogate performance. All the same, grief is a ubiquitous part of the human condition, and it would be vain to think that its negative effects could be entirely mitigated or ethically cabined. Hence, the common gap between surrogate decisions and patient preferences or values may be one we must tolerate to some degree. Indeed, it may be effectively impossible to mitigate grief’s negative effects on medical decisionmaking without stanching grief altogether, an ethical price that may be too great to bear.

Acknowledgments Katie Stockdale, Monique Wonderley, and audiences at the 2014 Western Michigan Medical Humanities conference and the Hastings Center provided valuable feedback on earlier versions of this chapter.

216

Michael Cholbi

Related Topics Chapter 7, “Deciding for the Incompetent” by Eric Vogelstein Chapter 13, “Solidarity Near the End of Life: The Promise of Relational Decision-Making in the Care of the Dying” by Bruce Jennings

References Arias, E. (2014) United States Life Tables, 2010, National Vital Statistics Report (Centers for Disease Control and Prevention, U.S. Department of Health and Human Services). Bagdasarov, Z. and S. Connelly (2013) “Emotional Labor among Healthcare Professionals: The Effects Are Undeniable,” Narrative Inquiry in Bioethics 3: 125–129. Buchanan, A.E. and D. Brock (1990) Deciding for Others: The Ethics of Surrogate Decision-Making, Cambridge: Cambridge University Press. Cholbi, M.J. (2016) (forthcoming) “Grief’s Rationality, Backward and Forward,” Philosophy and Phenomenological Research 92. Davidson, J.E., K. Powers, K.M. Hedayat, M. Tieszen, A.A. Kon, E. Shepard, V. Spuhler, I.D. Todres, M. Levy, J. Barr, R. Ghandi, G. Hirsch, and D. Armstrong (2007) “Clinical Practice Guidelines for Support of the Family in the Patient-Centered Intensive Care Unit,” Critical Care Medicine 35: 605–622. Dillon, I.L. (2010) “Emotions in the Intensive Care Unit,” in D.W. Crippen (ed.) Endof-Life Communication in the ICU: A Global Perspective, New York: Springer, 107–113. Feinberg, J. (1986) Harm to Self, Oxford: Oxford University Press. Fins, J.J., B.S. Maltby, E. Friedmann, M.G. Greene, K. Norris, R. Adleman, and I. Byock (2005) “Contracts, Covenants, and Advance Care Planning: An Empirical Study of the Moral Obligations of Patient and Proxy,” Journal of Pain Symptom Management 29: 55–68. Fritsch, J., S. Petronio, P.R. Helft, and A.M. Torke (2013) “Making Decisions for Hospitalized Older Adults: Ethical Factors Considered by Family Surrogates,” Journal of Clinical Ethics 24: 125–134. Goldie, P. (2011) “Grief: A Narrative Account,” Ratio 24: 119–137. Higgins, K.M. (2013) “Love and Death,” in J. Deigh (ed.) On Emotions: Philosophical Essays, Oxford: Oxford University Press, 159–178. Hiltunen, E.F., C. Medich, S.E. Chase, L. Peterson, and L. Forrow (1999) “Family Decision-Making and End-of-Life Treatment,” Journal of Clinical Ethics 10: 126–134. Jaworska, A. (2007) “Caring and Internality,” Philosophy and Phenomenological Research 74: 529–568. Jones, D.S., S.H. Podolsky, and J.A. Greene (2012) “The Burden of Disease and the Changing Task of Medicine,” New England Journal of Medicine 366: 2333–2338. Kelly, B.A., A. Rid, and D. Wendler (2012) “Systematic Review: Individuals’ Goals for Surrogate Decision-Making,” Journal of the American Geriatrics Society 60: 884–895. Kim, S.Y.H., J.H.T. Karlawish, and E.D. Caine (2002) “Current State of Research on Decision-Making Competence of Cognitively Impaired Patients,” American Journal of Geriatric Psychiatry 10: 151–165. Konigsberg, R.D. (2011) The Truth about Grief: The Myth of Its Five Stages and the New Science of Loss, New York: Simon and Schuster.

Grief and Surrogate Decision-Making 217 Kübler-Ross, E. (1969) On Death and Dying, New York: Simon and Schuster. LaBarge, S. (2012) “How (and Maybe Why) to Grieve Like an Ancient Philosopher,” in R. Kametkar (ed.) “Virtue and Happiness: Essays in Honour of Julia Annas,” Oxford Studies in Ancient Philosophy, Oxford: Oxford University Press, 321–342. Lewis, C.S. (2009) A Grief Observed, New York: Harper Collins (reprint of London: Faber, 1961). Lindemann, H. and J.L. Nelson (2014) “The Surrogate’s Authority,” Journal of Medicine and Philosophy 29: 161–168. Maciejewski, P.K., B. Zhang, S.D. Block, and H.G. Prigerson (2007) “An Empirical Examination of the Stage Theory of Grief,” Journal of the American Medical Association 297: 716–723. McMillan, J., S. Walker and T. Hope (2014) “Valuing Hope,” Monash Bioethics Review 32: 33–42. Nussbaum, M. (2001) Upheavals of Thought: The Intelligence of Emotions, Cambridge: Cambridge University Press. Pochard, F., E. Azoulay, and S. Chevret (2001) “Symptoms of Anxiety and Depression in Family Members of Intensive Care Unit Patients: Ethical Hypothesis Regarding Decision-Making Capacity,” Critical Care Medicine 29: 1893–2026. Roiphe, A. (2008) Epilogue, New York: Harper Collins. Rothman, D.J. (1992) Strangers at the Bedside: A History of How Law and Bioethics Transformed Medical Decision-Making, New York: Basic Books. Shalowitz, D.I., E. Garrett-Mayer, and D. Wendler (2006) “The Accuracy of Surrogate Decision-Makers: A Systematic Review,” Archives of Internal Medicine 166: 493–497. Vig, E.K., J.S. Taylor, H. Starks, and E.K. Hopley (2006) “Beyond Substituted Judgment: How Surrogates Navigate End-of-Life Decision-Making,” Journal of the American Geriatrics Society 54: 1688–1693. Wendler, D. and A. Rid (2011) “Systematic Review: The Effect of Surrogates on Making Treatment Decisions for Others,” Annals of Internal Medicine 154: 336–46. Wittgenstein, L. (1958) Philosophical Investigations (2nd Edition), Oxford: Blackwell. Wonderly, M.L. (2016) “On Being Attached,” Philosophical Studies 173: 223–242.

Further Reading Bonnano, G. (2010) The Other Side of Sadness: What the New Science of Bereavement Tells Us about Life after Loss, New York: Basic Books. Cholbi, M. (2016) (forthcoming) “Grief’s Rationality, Backward and Forward,” Philosophy and Phenomenological Research 92. Goldie, P. (2011) “Grief: A Narrative Account,” Ratio 24: 119–137. Shalowitz, D.I., E. Garrett-Mayer, and D. Wendler (2006) “The Accuracy of Surrogate Decision-Makers: A Systematic Review,” Archives of Internal Medicine 166: 493–497. Solomon, R.C. (2004) “On Grief and Gratitude,” in In Defense of Sentimentality, Oxford: Oxford University Press, 75–107.

13 Solidarity Near the End of Life The Promise of Relational DecisionMaking in the Care of the Dying Bruce Jennings

Summary Social perspectives on ethical obligations for medical treatment and the care of persons with critical and terminal illness have changed significantly in the past forty years and are in a process of transformation again today. This chapter explores the role of the concept of solidarity and closely related notions in the ethics of end-of-life care. It argues that solidarity is not only relevant to issues of health policy and access to health care but also significant in more personal and intimate settings of clinical and family decision-making. The importance of solidarity has become more apparent as ethical analysis of the goals and practices of end-of-life care have shifted in recent years. This shift has many facets, but in this chapter I characterize it as the transition from an individualistic emphasis on the rights of the dying person to a relational and communicative emphasis on shared decision-making, institutionally shaped transactions (curated caring), and dialogic practices of judgment and decision-making.

Introduction Social perspectives on ethical obligations for medical treatment and the care of persons with critical and terminal illness have changed significantly in the past forty years and are in a process of transformation again today (Sabatino 2010; Wolf et al. 2015). This chapter explores the role of the concept of solidarity and closely related notions in the ethics of end-of-life care. It argues that solidarity is not only relevant to issues of health policy and access to health care but also significant in more personal and intimate settings of clinical and family decision-making. The importance of solidarity has become more apparent as ethical analysis of the goals and practices of end-of-life care has shifted in recent years. This shift has many facets, but in this chapter I characterize it as the transition from an individualistic emphasis on the rights of the dying person to a relational

Solidarity Near the End of Life 219 and communicative emphasis on shared decision-making, institutionally shaped transactions (curated caring), and dialogic practices of judgment and decisionmaking. As a background to considering this shift in the ethics of end-of-life care, it is important to bear in mind that much work in bioethics has grown out of the liberal Enlightenment tradition in which the freedom and flourishing of the human individual are the primary social and ethical values. Social relationships exist to serve the rational interests of those who are parties to them; individuals do not exist to serve institutions and relationships. This intellectual and philosophical tradition helps explain the appeal of individualistic autonomy and rights to control over what is done by others to one’s own body. Such rights have loomed large in end-of-life care, especially once medicine and technology advanced to a point where these matters could be seriously and regularly at issue. Alternative perspectives that question this ethical individualism are often considered collectivist, communitarian, or metaphysically holistic. The concept of solidarity can easily be classified on the collectivist side of this debate, but in this chapter I hope to show that it would be a mistake to do so. Neither in theory nor in practice is the emerging relational reorientation in end-of-life medical decision-making a rejection of patient autonomy and the various ethical standards and practical tools that have been developed to hear and heed the voice of those who are dying. But it is a reinterpretation and a reconceptualization of autonomy and the values of independence and subjective choice that have structured the social dynamics of end-of-life care and interpersonal interactions at the bedside. An emphasis on rights and an emphasis on relationships are conceptually distinct, but substantively they point toward many of the same ends. Both seek to foster a social ecology of good dying—by which I mean a relational system of communication, medical decision-making, and treatment that facilitates both the values and obligations of respect for persons and the obligations of care—human presence, relief of suffering, and preservation of dignity. More specifically, the concept of autonomy is central to both rights-based and relational approaches to end-oflife care, but for the former it is an individualistic interpretation of autonomy focusing on noninterference and control, and for the latter it is a relational interpretation of autonomy focusing on contextualized, dialogic judgment and self-affirmation in and through mutuality. The discussion in this chapter will proceed in three steps. First, I introduce and discuss the concept of solidarity and the ways in which it changes the orientation of ethical analysis relative to the individualistic orientation in bioethics. Next, I explicate in more detail the rise of patient autonomy, patient empowerment, and the right to refuse treatment. Finally, I review and assess the new emphasis on relationality and communication now emerging in end-of-life

220

Bruce Jennings

care. I argue that appeal to the concept of solidarity and autonomy informed by relational judgment can extend and strengthen this new orientation. My notion is that the challenge facing the relational orientation in end-oflife care is twofold: (1) to provide a more practical and feasible template for medical planning and decision-making that more effectively serves the needs and rights of dying persons; and (2) to avoid bluntly rejecting or abandoning ideals of personal autonomy and respect for persons. The move from rights to relationships can accomplish both of these objectives; it need not purchase the good of practicality or greater patient well-being at the expense of respect for patient agency and personhood. The concept of solidarity can play an important role in meeting both of these objectives. This is due, in part, to the fact that solidarity is constituted by a broader conceptual framework consisting of what I shall call “right recognition” of moral standing and membership, and “right relationality” of mutual care and concern for others. It is also due to the ways in which solidaristic practices can inform judgment and reflective communication in end-of-life care settings.

Solidarity: Moral Standing and Standing Up Solidarity is a multifaceted concept that is grounded on right recognition and right relationship (Prainsack and Buyx 2011, 2012; Prainsack 2014; Jennings 2015; Jennings and Dawson 2015). Right recognition is a condition of moral membership and equality—namely, the recognition of what has been called the “moral considerability”—the equal and impartial moral standing of each person and the corresponding respect due to the person (Goodpaster 1978). Right relationship is a condition of mutuality—the mutuality of interdependence, care for (and of ) others, and concern for their human flourishing, understood in developmental and interdependent terms (Gergen 2009; Nussbaum 2011). It is from these socially and culturally constituted conditions of membership and mutuality that solidarity arises. And it is to them that solidarity makes its basic contribution to ethical theory generally and to the ethics of end-of-life care. The idea of solidarity carries with it a number of conceptual orientations and commitments. When it interacts with other concepts, such as justice, autonomy, liberty, and dignity, it opens up a relational interpretation of them. For this reason, it is best not to view solidarity as one ethical principle among others (Jennings and Dawson 2015: 34–35). It is not a rule prescribing certain types of conduct or the mental motivation or cause of certain kinds of consequential actions. Nor is it moral virtue in the sense of being a disposition or characterological motivation in the mind of an individual. Solidarity is a social practice. To grasp the role that solidarity can play in the ethics of end-of-life care, it is better to think of ethics discursively rather than architectonically (Fraser 1986).

Solidarity Near the End of Life 221 Here is what I mean by that distinction. Discursive ethics is a way of articulating and interpreting ethical arguments, vocabularies, standpoints, and narratives; it is a mode of critical and interpretive “discursive practice” that takes place in a historical and social context made up of a number of other contrasting and competing discourses. By a “practice” I mean a normatively structured form of activity (including linguistic and symbolic activity) through which agents pursue a variety of capabilities and activities (or “functionings”) that conduce to their human self-realization and constitute a life well lived (MacIntyre 2007: 187–203). Architectonic ethics, by contrast, does not see itself as one discourse among many but instead as a more detached, foundational, and contemplative enterprise. It is concerned with the construction of first principles designed to form the structure and scaffolding for many other, more specific rules. These first principles and secondary rules direct and govern the functioning of institutions and the modes of action and interaction in a society. John Rawls’s great work A Theory of Justice (1971) is an example of architectonic ethics in that he derives two complex principles of justice (on equal liberty and on equal opportunity and distribution of important resources) to oversee what he calls the “basic structure” of a society. Rawls argues that these principles of justice would emerge from an ahistorical, generic situation of impartial, rational choice and, for that reason, should morally be embraced by members of actually existing societies. (In later work [1996], he argues that these principles reflect and clarify the consensual core of values that contemporary liberal societies have already embraced.) The discourse and practice of solidarity, on the other hand, do not construct an idealized situation of original choice. It begins with the particularities of society, culture, and psychology—its starting point is the latent possibilities of a given place at a given time and with ongoing forms of conviction and contentment, we might say. It builds on senses of historical memory and tradition, and it feeds on the gratitude felt when one recognizes the service and contributions that others have made to one’s way of life. Solidarity begins with the recognition of reciprocal and symbiotic interdependence and then intervenes in—interrupts—an ongoing form of life in order to redress the injustice it contains and win through to a better moral community. Solidarity tends to move our ethical attention away from the generic, abstract, and universal and toward the practical and concrete particularity of a social, historical, or personal situation. The logic of its conceptual structure is inseparable from what might be called the phenomenological understanding of its practice and the experience of standing in solidarity with others. I shall describe aspects of that phenomenology in a moment. Solidarity inherently leads us to view our own lives and agency as bound together with the rights, well-being, health, and dignity of others here and now. That does not mean that discursive ethics is therefore ethically relativistic, situational, opportunistic,

222

Bruce Jennings

or ad hoc. It means that discursive ethics hazards the wager that the capacity for critical reasoning and moral imagination can grow developmentally out of an engagement with the particularity of the world. However, when you think about it, this is no bolder a wager than the one made by architectonic theorists in their belief that objective moral knowledge can achieve an Archimedean point of eternity and prescind from concrete lived experience in a cultural place or a historical time. We can begin to get a better sense of what solidarity means by examining what is missing from social and moral life when the idea and the practice of solidarity are absent. Imagine a society in which the well-being of each individual (including one’s health status and the circumstances of one’s own dying) is seen as sui generis and fundamentally distinct from the well-being of others. Sometimes, to be sure, one person’s success may depend in some way on the success and cooperation of others, but then the success of others is valuable to the individual only incidentally. It is as if each person says to herself, if I can do without your cooperation, I shall: all the more to be secure, selfreliant, and independent. There is no moral encumbrance here, no normative push of commitment or obligation, no motivational pull of mutual recognition and resemblance. But why then should healthy citizens support an expensive health care system for those near the end of life who have progressive, degenerative, and incurable conditions and will never return to economically productive lives (Houtepen and ter Meulen 2000a, 2000b; ter Meulen et al. 2001)? At the very least, solidarity and related strong notions of ethical recognition of membership in a human moral community and the mutuality of care, dignity, and respect give us important moral ideas with which to answer that question. I turn now to one way to think about what solidarity means and what practicing it involves. We can distinguish among different moments or stages in the practice of solidarity as it engages with moral recognition and mutuality. I suggest that the fundamental gesture and stance of solidarity are standing up beside another, thereby signaling publicly one’s recognition of that person’s (or group’s) moral standing or considerability. In addition, there are two distinct modes of mutual relationality that follow from this recognition, and that are particularly germane to end-of-life care. These may be called standing up for and standing up with. Standing up for. Solidarity as standing up for is advocacy. It involves assisting, defending, and pleading the cause of the other. In general discussions of solidarity, the other is often taken to be a stranger, but that is not necessary; it makes perfect sense to talk about being in solidarity with one’s friends, although in this case it may well be that we build the concept of solidarity into the concept of friendship without noticing what we are doing or commenting on it. Moreover, the other need not be a human individual: one can stand up

Solidarity Near the End of Life 223 for other species, an ecosystem, or a cultural way of life. What is crucial is that there is some kind of power or knowledge differential between self and other in a relationship of solidarity. Otherwise the doing-for aspect of the relationship would be less salient. However, there is a serious shortcoming inherent in standing-up-for solidarity. It can assume and perpetuate a continued differentiation, even condescension, in the relationship between those individuals or groups standing up and those on whose behalf the standing is done. This kind of solidarity can advocate for improved treatment or benefits for an oppressed or vulnerable group, but does not necessarily challenge the underlying basis for their subordinate social status. Solidarity as standing and doing for others can become self-perpetuating. Inasmuch as it does not undermine structural inequalities, it serves to perpetuate the need for its own continuing existence. Standing up with. Solidarity as standing up with takes another step in the direction of mutuality and recognition of shared moral standing. Moving from a mode of mutuality for to mutuality with requires deeper engagement with the experience and lifeworld of the other person. This closer understanding can tend to humanize and personalize the relationship further, and the potential benefit becomes more reciprocal. Those who stand with others may find their own initial prejudgments concerning others transformed by the encounter. Solidarity can be stopped in its tracks and thwarted by fear of such self-transformation. Those who practice solidarity must open themselves to the opportunity—and risk—it poses. Relating to other people or groups in the specificity of their values and vocabularies of self-interpretation can simultaneously develop respect for the specific standpoints of others (Dean 1995, 1996; Bartky 2002). It can also enable a greater capacity for intercultural and transpersonal interpretive understanding (Forst 1992). Solidarity contains the possibility of being common readers of the diverse and distinct lives we each author. In the end-of-life care context, an example of this transformation often arises in hospice programs that make room for dialogues of memory and reconciliation between a dying person and members of their family and friends. Family members often learn things during the hospice caregiving process that they did not previously know or understand about a loved one. Having the opportunity to discover this can be of lasting benefit after the person’s death. If relational autonomy requires respect for the personhood of each and the moral membership of all, and if the practice of solidarity in the modes of standing up for and with involves a deep understanding and interpretation of a lived experience perceived through a distinct cultural and personal lens, then autonomy and solidarity find their point of convergence and reciprocal support. Solidarity involves finding a kind of covering connection that does not negate the autonomy of either oneself or the other with whom one stands in solidarity. Instead, it establishes the grounds of respect for and the social perpetuation of autonomy. Reconceiving autonomy in relational terms and bringing it together

224

Bruce Jennings

with solidarity allows us to say that we stand here both as human beings made capable through the functioning interdependence of a complex social community and as human persons with rights, needs, and vulnerabilities common to our kind, our humanity. One of these vulnerabilities is mortality; another is suffering. As Dame Cicely Saunders, describing St. Christopher’s Hospice in London, aptly put it, solidarity and autonomy are found in “the common giving of people who share the cost of being vulnerable” (Saunders 1971: 43).

Standing Alone before Death: Empowering the Autonomous Choices of the Dying Beginning in the early 1970s, innovations in ethics, law, and medicine began a movement toward empowering dying patients and their surrogates to have a greater voice in medical treatment decision-making, particularly in the face of progressive, chronic, and ultimately fatal disease (Webb 1997). The landmark New Jersey Supreme Court decision (In re Quinlan 1976) that permitted the removal of a life-supporting mechanical ventilator from Karen Ann Quinlan— a young woman who had suffered catastrophic anoxic brain injury some years before and was being kept alive in a permanent unconscious state—gave this movement tremendous momentum. Over time, the authoritarian, albeit benevolent, paternalistic decision-making style of many physicians was called into ethical question and came to be supplemented—if not entirely displaced— by the rights of patients to determine the nature of their medical treatment in accordance with their own preferences and values. In a growing number of court decisions, a new body of law established the legal right of patients to informed consent and to informed refusal of even life-sustaining medical treatment, using what was often called a “right to privacy” or a fundamental “liberty interest” protected by the constitution. State legislatures followed suit by enacting statutes reinforcing these rights and providing means, such as advance directives, for realizing them in practice (Meisel, 1993; National Center for State Courts 1993; Meisel and Cerminara 2004; Berlinger et al. 2013). In ethics writings and in standards promulgated by various professional bodies in medicine and nursing (Cassel and Foley 1999), this kind of recognition and empowerment of the voice of the patient was expressed as the ethical principle of respect for autonomy or self-determination. What was meant, at bottom, by “autonomy” in bioethics at that time was an individualistic idea of liberty as noninterference and respect for self-sovereignty (Jennings 2007). Being respected by others as autonomous was taken to be an intrinsic value in its own right, but the patient empowerment strategy was also oriented toward goals of beneficence and improvement in the quality of care and the quality of life experienced by the patient and others during the dying process. Paradoxically, changes taking place at about this same time in medical science

Solidarity Near the End of Life 225 and technology were making it possible to sustain vital function and to prolong biological life much more effectively than before. However, there was a widespread perception that due in part to this very technology, the dying individual was being objectified in the medical setting, and that physicians, now preoccupied with the technical challenge of intensive, invasive procedures, were losing sight of the whole person. Medical treatment was becoming divorced from medical care (DelVecchio Good et al. 2004). Under these circumstances, relying on more or less unilateral physician expertise and paternalism was no longer ethically trustworthy or practically adequate to ensure a good dying marked by dignity, respect, relief of suffering, and human presence and witness (Field and Cassel 1997; Meier et al. 2010; IOM 2015). Can life-extending medical technology be deliberately withheld or withdrawn when the patient refuses to consent to those treatments or when they offer no benefit but serve only to prolong the patient’s dying? There was a time when a physician was on solid ethical ground in doing “everything possible” to stave off death. The person confronted with a life-limiting illness was seen as the victim of an attack who enlists the knowledge, skills, and technology of physicians’ medicine as the defender and rescuer. This image of the physician’s role in the face of impending death can certainly be seen as the solidarity of standing up for—the solidarity of rescue, so to speak. Solidarity also provides one potential way to understand the role and responsibility of a health care surrogate or proxy in end-of-life decision-making. Moreover, the situation of end-of-life care poses an ethical challenge for the concept and practice of solidarity in an especially poignant way. In many instances, medical decision-making near the end of life presents a tension between the duty to respect individual autonomy and the duty to protect and promote the best interests of the dying person. At stake is the timing of a person’s death and the manner of his or her dying. If a patient’s stated wishes or advance directive calls for treatment that is judged to be futile and unduly burdensome, a dilemma is created for surrogates, family, and physicians. The same problem arises if the patient’s wishes indicate forgoing treatment that others believe would be beneficial. I think that the concept of solidarity per se does not resolve this ethical tension in end-of-life care or elsewhere. It does perhaps cast the tension in a new light and a new sensibility, however. The call of solidarity to stand up for a vulnerable other seems naturally to suggest primary attention to the best interests of the other, and hence there is a tendency for this mode of solidarity to embrace a paternalistic stance (if by paternalism we mean overriding subjective patient will and choice in favor of the patient’s objective best interests). There is no logical or necessary reason why this should be the case, for there is nothing in the meaning of the concept of solidarity as I understand it that would preclude being in solidarity with the autonomy of the other. Nonetheless, it has been my observation in clinical

226

Bruce Jennings

settings that solidarity with the dying often means protecting them from something, and that something is most often neglect or nontreatment, rather than the use of interventions and treatments, which, in truth, are often actually abusive in their effects. In any case, the distinction between solidarity as standing up for and solidarity as standing up with can be helpful here. The closer degree of imaginative connection with the other as a unique person that comes during standing up with acts as a counter to the advocacy mode of standing up for and protecting. It leads one toward viewing the situation through the eyes of the other—the internal properties of the situation, so to speak—rather than focusing on the objective external properties of the other’s situation. Perhaps the challenge of the tension between the duty to respect individual autonomy and the duty to care for the individual’s best interests is not so much to resolve it as to keep in view simultaneously the external consequences and the internal meaning in a person’s life; in other words, to adopt both a third-person and a secondperson standpoint in ethical judgment. That is the zone in which solidarity does its conceptual work. And all practices of solidarity structure experience and agency in particular ways. Undergoing treatment for a serious illness is very much like entering an alien world, a kind of despotic, totalitarian world in which one must suspend one’s normal rights, freedoms, expectations, and routines and then submit to physical and emotional handling that under nonmedical circumstances would surely be condemned as abuse. Acute medical treatment is also akin to social dislocation—leaving one’s normal social space and entering an alien world— because it involves the abandonment of one’s personal identity and everyday roles and the assumption of a very different persona and role. This experience can precipitate a feeling of loss of self-identity. It can cause an alienation from one’s past—and perhaps even from what one fears will be the future—when it is precisely that past that one is desperate to regain and the avoidance of that future on which the entire treatment enterprise is predicated. It can involve a detachment from one’s family and closest loved ones, who, still healthy and whole, reside on the other side of the looking glass. Paradoxically, it was medicine’s success rather than its failure that called such a passive patient encounter with medical expertise and intervention into serious question. Beginning in the 1960s, powerful new drugs and medical technologies were prolonging life-functions in critically and terminally ill patients—mechanical ventilators, cardiac interventions, kidney dialysis, and the fine-tuning of vital functions made possible by intensive care units. It soon came to pass that, for perhaps the majority of patients at the end of life, the exact timing and circumstances of one’s death were not beyond anyone’s control but were in fact the result of specific decisions and choices made by those caring for (physicians) and speaking for (usually family) the patient. Gradually,

Solidarity Near the End of Life 227 thoughtful leaders in medicine, and many family members who had observed the new end-of-life care firsthand, became aware that the experience of dying (for the individual, for the family, and also at times for health care providers) was often a horror. Few deaths were easy or “natural”; most occurred only after a serious of invasive and burdensome interventions, well-intentioned no doubt, but nonetheless very deleterious to the dying person’s quality of life and disruptive to connection and communication between the patient and family during the final weeks or days of life. In sum, while life was being prolonged, many were living through an avoidably bad dying—a dying with inadequate relief of pain and suffering, inadequate compassion, inadequate human presence and witness; a dying marked by fear, anxiety, loneliness, and isolation; a dying that effaced dignity and denied individual self-control and choice. How did this happen? Powerful new technology was not being used wisely and judiciously but had succumbed to a kind of technological imperative. Aggressive interventions were becoming the default in hospitals and sometimes pushing the patient’s own wishes and values to the margins of decision-making (SUPPORT 1995). Reductionistic medical perspectives, which were the flip side of the new technology and the medical training that went along with it, were displacing more holistic perspectives, and this made it more difficult to see “the patient as person,” as one leading theologian-bioethicist aptly put it (Ramsey 1970). By about 1990, this situation was changing, although the forces of technology and patterns of inadequate palliative care and poor advance care planning persist today. In the 1990s, the right to refuse unwanted and burdensome lifesustaining medical treatment was solidly established and the ethical obligation to respect patient autonomy was widely recognized in mainstream medicine. Nonetheless, the patient empowerment strategy of the past forty years has not been entirely successful (Jennings et al. 2005). The shortcomings of the patient empowerment strategy take several forms, and they have been summarized and extensively documented in two successive studies by the Institute of Medicine (Field and Cassel 1997; IOM 2015). Technologically driven orientations remain powerful in medicine and physicians do not always accept or cooperate with patient and family wishes; or else they influence and shape those wishes unduly. Family dynamics and emotions complicate the identification of a patient’s autonomous preferences and values, and hence make it unclear what precisely others are obliged to respect when they respect autonomy. And cultural or religious traditions and ways of understanding right conduct in the face of illness and dying have often not embraced the seemingly individualistic orientation of the concept of autonomy and the underlying concern for power and control implicit in the patient empowerment reform strategy in end-of-life care (Koenig and Gates-Williams 1995).

228

Bruce Jennings

Standing Together before Death: Enabling the Relational Judgment in the Lifeworld of Dying Prompted by problems such as these, another new reorientation—more focused on relationality, communication, and shared decision-making—is getting underway (Sabatino 1999, 2010; Mackenzie and Stoljar 2000; Jennings et al. 2005; Code 2006; Baylis et al. 2008; Nedelsky 2011). There are several facets to this relational approach. It emphasizes shared, informed medical decision-making and consensus building among providers, surrogates, and family members, and the dying person, if capacity has not been lost. Attempts are also being made to improve continuity of communication and care planning among rotating providers and different health care settings. Staff training resources are being developed to improve communication styles and messaging so that they are culturally sensitive and competent. Changing processes and practices such as these may be characterized in the phrase “Reconstrue rights, work through relationships.” Ethical and conceptual clarity will be needed to guide this shift if it is to realize its full potential, and also to protect against the undue erosion of important patient-centered rights and values. It is important to emphasize that while the rights of dying patients should be reconstrued, they should not be rejected. And it is especially important to insist on this distinction in the United States today, for this is a time of concerted, often less than careful or discerning financial costcontainment policy measures across the board that can have deleterious effects on end-of-life care (Jennings and Morrissey 2011). The discourse of solidarity can be helpful in our social discussion of endof-life care at two levels. One is on the system and policy level, where solidarity, just recognition, and just mutuality can appeal to rights and autonomy in a relational way and provide justification and legitimacy to public expenditures, universal insurance, and social investment in a humane and dignified dying process for all persons (ter Meulen et al. 2001). The other is on the more intimate level of care planning and decision-making at the bedside, where a new understanding of the roles and obligations of surrogates and other family members, as well as physicians and nurses, can be illuminated by seeing the situation in terms of the trajectory of solidarity from standing for to standing with. Now surrogates tend to be seen legalistically as advocates or formal representatives (Ventura et al. 2013). To develop this line of thinking further, we need to consider the substantive content and purposes of dialogic communication and relational judgment in end-of-life decision-making. I turn first to dialogic communication. Relief of suffering, respecting individual rights, and preserving human dignity and fulfillment at the close of a person’s life are the core values that good communication in the care process serves. Most dying patients—and the vast

Solidarity Near the End of Life 229 majority of grieving family members too—do not actually desire medical treatments that are known to be futile in advance or that most likely will leave the patient worse off, not better off. Moreover, while prognosis concerning life expectancy and quality is still difficult and uncertain, objective scientific information and tools are being developed that enable physicians to make reasonable judgments about the effectiveness and benefit of a given treatment to a particular patient. If patients, or more often, family members for dying patients who have lost decision-making capacity request overly aggressive and burdensome life-sustaining treatment interventions, it is not usually because they have certain principled moral or religious beliefs of a vitalistic kind, but more often because they do not accurately understand the nature and consequences of what they are requesting. Good communication is communication that will level with family members about prognosis and treatment modality outcomes and keep them apprised in a timely way of changes in the patient’s condition so they have time to assimilate disturbing new information emotionally before suddenly being called upon to make a wrenching nontreatment decision as a surrogate. Good communication is also based on an understanding of the social and emotional needs of family members and their cultural worlds. In these worlds, kinship roles, spiritual orientations, and the responsibilities of caregiving in the face of sickness and death all have a powerful meaning. We have created a health care system in which medical choices are taken as the medium of fulfilling these social and emotional meanings and goals. Fortunately, however, cultural and personal meaning can be supported and respected without having to make unreasonable decisions to request unduly burdensome life-sustaining medical treatments. Instead of solely focusing on how to empower the decisions of individual patients one at a time, we should ask how we can design a relational ecology of communication, recognition, respect, and good judgment in using technology and in meeting the needs of the dying patient as a whole person. I believe that the cognitive biases of all parties to end-of-life decision-making—providers, patients, and families alike—can be remediated by reflective judgment and participation in dialogic communities offering parity of voice and engagement to all members (Nikulin 2006). Autonomy, human rights, and interdependent relationships of membership and mutuality are best understood as interwoven in the fabric of solidarity through which individuals acting together can each attain their own mode of flourishing and self-realization in ways that no one of them could attain by acting alone. That is the case in social life generally, and it is the case as well at the end of life. How would viewing end-of-life decision-making through the prism of solidarity—that is, within an ecosystem of dialogic discourse and social practices that affirm moral recognition and mutuality—help preserve the ideals of personal autonomy and of respect for persons and their dignity? Social

230

Bruce Jennings

practices already permeate everyday life so it is not an architectonic question of building such forms of relationship and activity from the ground up. It is a more situated, practical question of mending or healing broken practices or sending out extensions of existing practices into new domains. In this way, I believe, we can arrive at a notion of autonomous medical decision-making as the exercise of relational judgment. By relational judgment, I mean the capacity to discriminate among available courses of action on the basis of an interpretive understanding of shared values embedded in an ongoing institutional practice and in a broader form of communal life (Garsten 2006; Thiele 2006). Judgment is context-sensitive. It does not seek to make global and ultimate determinations but rather provisional, local ones—decisions for the here and now. In saying that it involves discrimination among available options, I mean to underscore the fact that the options are immediate, tangible, doable. Since the options are real, not hypothetical, and since the social world normally offers a rather narrow range of effective options at any time, the options are not usually grossly different from one another. They present instead subtle, fine-grained differences of meaning and implication. This is manifestly the case in end-of-life care. Catching the nuance, “discriminating” aptly, is highly sensitive to present attitudes, understandings, and probable reactions by others. End-of-life care has seen professional paternalism, which is indeed imperious and controlling, albeit benevolent. It has seen individualistic autonomy, which has been very difficult for patients and families to sustain (Schneider 1998; Kittay 1999; O’Neill 2002; Conly 2013). Relational judgment is an alternative to both of these. It is an enmeshed practice of nurture, protection, provision, care, and support for another person. Relational judgment does not override autonomy, but contextualizes it by giving the individual receiving care a new insight about his or her own agency, a new interpretation of the intentional and expressive possibilities that are actually present in a given situation, even the situation of curated caring during the time of dying. Judgment is fundamentally dialogic. It requires a space for open dialogue and room for maneuver. Particular interpretations of the values embedded in a common form of life can always be challenged as misleading or lacking in depth or insight. One person’s judgment and reasons in action are not totally incomprehensible or inaccessible to others, nor need the judgment of one person be a foreign or wholly extrinsic imposition on others (Gadamer 1990; Hutto 2008). Judgment, in the sense I intend here, is not imposed but rather imbibed. It is absorbed and shared in a dialogue of question and answer among persons who are autonomous in the sense that they are indeed the authors of their own acts and lives, but they are symbiotic, not solipsistic authors. This agency of relational autonomy informed by judgment is a matter of mutual adjustments, mediations, trial and error innovation. It is an autonomy that savors the shared

Solidarity Near the End of Life 231 engagement of reflection and learning for its own sake as much as the attainment of its ends or the granting of its claims. When autonomy is respected and practiced within a framework of reflective judgment in this sense, it will be open to reconciliation with caring counsel or intervention, just as caring practice will be open to respect for autonomy. The obligation to respect the self-determination of a relational agent will embrace the obligation to provide care and to promote the person’s well-being.

Conclusion There are some fundamental aspects of the human condition and human experience that naturally lead us up a path of increasing moral awareness. In the social world, solidarity traces one such path; we grow in moral awareness as we come to recognize the moral standing of others to whom that standing is being—or could easily be—denied, and then as we move from understanding ourselves as in relationship for others to being in relationship with them. That means that we not only are willing to support and defend others but also recognize their vulnerability as our own. We grow to understand that linkages of interdependence and symbiosis are not exceptional but pervasive and essential in human life (Benhabib 1987; Code 2006; Gergen 2009; Prainsack 2014). A serious reflection on death and dying—and even more so having gone through the lived experience of caring for a dying person and working to support continued human flourishing to the end—is an occasion, some would say the foundational occasion, for this sort of moral growth and learning. Those who are dying are among the least valuable assets in a society limited to instrumental cooperation and individual self-interest. When the conversation centers, as it so often does these days, on economic value and finding the most cost-effective ways to provide care to persons in the last year or few months of their lives, the equally important question of the moral value of such care to individuals, families, and society as a whole gets swept under the rug. Many would like to maintain that the ethical justification of high-quality end-of-life care is based on its goodness (the facilitation of human flourishing) and rightness (equitable respect for the dignity and worth of all persons). The economics of that care, real and valid though it is, is but a prudent means to those ends. Yet it is difficult—perhaps impossible politically if not logically— to sustain this argument without recourse to the concept of solidarity.

Related Topics Chapter 7, “Deciding for the Incompetent” by Eric Vogelstein Chapter 12, “Grief and End-of-Life Surrogate Decision-Making” by Michael Cholbi

232

Bruce Jennings

References Bartky, S.L. (2002) Sympathy and Solidarity, London: Roman and Littlefield. Baylis, F., N.P. Kenny and S. Sherwin (2008) “A Relational Account of Public Health Ethics,” Public Health Ethics 1(3): 196–209. Benhabib, S. (1987) “The Generalized and the Concrete Other,” in S. Benhabib and D. Cornell (eds.) Feminism as Critique, Minneapolis: University of Minnesota Press, 77–95. Berlinger, N., B. Jennings, and S.M. Wolf (2013) The Hastings Center Guidelines for Decisions on Life-Sustaining Treatment and Care near the End of Life (Revised and Expanded 2nd Edition), New York: Oxford University Press. Cassel, C.K. and K. Foley (1999) Principles for Care of Patients at the End of Life: An Emerging Consensus among the Specialties of Medicine, New York: Milbank Memorial Fund. Code, L. (2006) Ecological Thinking: The Politics of Epistemic Location, New York: Oxford University Press. Conly, S. (2013) Against Autonomy: Justifying Coercive Paternalism, Cambridge, UK: Cambridge University Press. Dean, J. (1995) “Reflective Solidarity,” Constellations 2(1): 114–140. Dean, J. (1996) Solidarity of Strangers, Berkeley: University of California Press. DelVecchio Good, M.J., N.M. Gadmer, P. Ruopp, M. Lakoma, A.M. Sullivan, E. Redinbaugh, R.M. Arnold, and S.D. Block (2004) “Narrative Nuances on Good and Bad Deaths: Internists’ Tales from High-Technology Work Places,” Social Science and Medicine 58(5): 939–953. Field, M.J. and C.K. Cassel (eds.) (1997) Approaching Death: Improving Care at the End of Life, Washington, DC: National Academy Press. Forst, R. (1992) “How (Not) to Speak about Identity,” Philosophy and Social Criticism 18(3/4): 293–312. Fraser, N. (1986) “Toward a Discourse Ethic of Solidarity,” Praxis International 5(4) (January): 425–429. Gadamer, H.-G. (1990) Truth and Method (2nd Revised Edition), J. Weinsheimer and D.G. Marshall (trans.), New York: Crossroad. Garsten, B. (2006) Saving Persuasion: A Defense of Rhetoric and Judgment, Cambridge, MA: Harvard University Press. Gergen, K.J. (2009) Relational Being: Beyond Self and Community, New York: Oxford University Press. Goodpaster, K. (1978) “On Being Morally Considerable,” The Journal of Philosophy 75(6) (June): 308–325. Houtepen, R. and R. ter Meulen (2000a) “The Expectation(s) of Solidarity: Matters of Justice, Responsibility and Identity in the Reconstruction of the Health Care System,” Health Care Analysis 8(4): 355–376. Houtepen, R. and R. ter Meulen (2000b) “New Types of Solidarity in the European Welfare State,” Health Care Analysis 8(4): 329–340. Hutto, D.D. (2008) Folk Psychological Narratives: The Sociocultural Basis of Understanding Reasons, Cambridge, MA: MIT Press. In re Quinlan, 70 N.J. 10, 355 A.2d 647, cert. denied, 429 U.S. 922 (1976). IOM (Institute of Medicine) (2015) Dying in America: Improving Quality and Honoring Individual Preferences near the End of Life, Washington, DC: National Academies Press.

Solidarity Near the End of Life 233 Jennings, B. (2007) “Autonomy,” in B. Steinbock (ed.) The Oxford Handbook of Bioethics, New York: Oxford University Press, 72–89. Jennings, B. (2015) “Relational Liberty Revisited: Membership, Solidarity, and a Public Health Ethics of Place,” Public Health Ethics 8(1) (April): 7–17. Jennings, B. and A. Dawson (2015) “Solidarity in the Moral Imagination of Bioethics,” Hastings Center Report 45(5) (September-October): 31–38. Jennings, B., G. Kaebnick, and T.H. Murray (eds.) (2005) “Improving End of Life Care: Why Has It Been So Difficult?” Hastings Center Report, Special Supplement 356: S52–S57. Jennings, B. and M.B. Morrissey (2011) “Health Care Costs in End-of-Life and Palliative Care: The Quest for Ethical Reform,” Journal of Social Work in End of Life and Palliative Care 7(4) (December): 300–317. Kittay, E. (1999) Love’s Labor: Essays on Women, Equality, and Dependency, New York: Routledge. Koenig, B.A. and J. Gates-Williams (1995) “Understanding Cultural Difference in Caring for Dying Patients,” Western Journal of Medicine 1633: 244–249. MacIntyre, A. (2007) After Virtue (3rd Edition), Notre Dame, IN: University of Notre Dame Press. Mackenzie, C. and N. Stoljar (eds.) (2000) Relational Autonomy: Feminist Perspectives on Autonomy, Agency, and the Social Self, New York: Oxford University Press. Meier, D.E., S.L. Isaacs, and R.G. Hughes (eds.) (2010) Palliative Care: Transforming the Care of Serious Illness, San Francisco: Jossey Bass. Meisel, A. (1993) “The Legal Consensus about Forgoing Life-Sustaining Treatment: Its Status and Its Prospects,” Kennedy Institute of Ethics Journal 2(4): 309–345. Meisel, A. and K.L. Cerminara (2004) The Right to Die: The Law of End-of-Life DecisionMaking (3rd Edition), New York: Aspen Law and Business. National Center for State Courts, Coordinating Council on Life-Sustaining Medical Treatment Decision-Making by the Courts (1993) Guidelines for State Court Decision-Making in Life-Sustaining Medical Treatment Cases (Revised 2nd Edition), St. Paul: West. Nedelsky, J. (2011) Law’s Relations: A Relational Theory of Self, Autonomy, and Law, New York: Oxford University Press. Nikulin, D. (2006) On Dialogue, Lanham, MD: Lexington Books. Nussbaum, M.C. (2011) Creating Capabilities: The Human Development Approach, Cambridge: Harvard University Press. O’Neill, O. (2002) Autonomy and Trust in Bioethics, Cambridge, UK: Cambridge University Press. Prainsack, B. (2014) “Personhood and Solidarity: What Kind of Personalized Medicine Do We Want?” Personalized Medicine 11(7): 651–657. Prainsack, B. and A. Buyx (2011) Solidarity: Reflections on an Emerging Concept in Bioethics, London: The Nuffield Council. Prainsack, B. and A. Buyx (2012) “Solidarity in Contemporary Bioethics—Towards a New Approach,” Bioethics 26(7): 343–350. Ramsey, P. (1970) The Patient as Person, New Haven: Yale University Press. Rawls, J. (1971) A Theory of Justice, Cambridge, MA: Harvard University Press. Rawls, J. (1996) Political Liberalism (Expanded Edition), New York: Columbia University Press. Sabatino, C.P. (1999) “The Legal and Functional Status of the Medical Proxy: Suggestions for Statutory Reform,” Journal of Law, Medicine and Ethics 27(1): 52–68.

234

Bruce Jennings

Sabatino, C.P. (2010) “The Evolution of Health Care Advance Planning Law and Policy,” Milbank Quarterly 88(2): 211–239. Saunders, C. (1971) “The Patient’s Response to Treatment: A Photographic Presentation,” in Catastrophic Illness in the Seventies: Critical Issues and Complex Decisions, New York: Cancer Care, 33–46. Schneider, C. (1998) The Practice of Autonomy: Patients, Doctors, and Medical Decisions, New York: Oxford University Press. SUPPORT Principal Investigators (1995) “A Controlled Trial to Improve Care for Seriously Ill Hospitalized Patients: The Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT),” Journal of the American Medical Association 274: 1591–1598. Ter Meulen, R., W. Arts, and R. Muffels (eds.) (2001) Solidarity in Health and Social Care in Europe, Dordrecht: Kluwer Academic. Thiele, L.P. (2006) The Heart of Judgment: Practical Wisdom, Neuroscience, and Narrative, New York: Cambridge University Press. Ventura, C.A.A., A. Gallagher, R. Jago, and I.A.C. Mendes (2013) “Rethinking Ethical and Legal Issues at the End of Life in the UK and Brazil: A Role for Solidarity?” Medicine and Law 32: 481–496. Webb, M. (1997) The Good Death: The New American Search to Reshape the End of Life, New York: Bantam. Wolf, S.M., N. Berlinger, and B. Jennings (2015) “Forty Years of Work on End of Life Care—From Patients’ Rights to Systemic Reform,” New England Journal of Medicine 372(7) (February 2012): 678–682.

14 Justice and the Aging of the Human Species Colin Farrelly

Summary Does justice require we extend the human lifespan by slowing the aging process? Or could such actions actually be considered unjust if they exacerbate existing health inequities or increase problems of population density and climate change? If there are conflicting or competing demands of justice in this context, how can we be begin to disentangle them so that we respond rationally and fairly to addressing the diverse stakes that arise with respect to the aspiration to extend (healthy) life? This chapter addresses these important questions by examining how the aging of the human species alters the moral landscape in novel and important ways.

Introduction Philosophers invoke or develop an account of justice to help bring focus and attention to pressing (yet often neglected) societal problems or ills that are in need of addressing. These ills include problems like patriarchy, institutional racism, systemic socioeconomic inequalities, health disparities, and so forth. By subjecting the institutions of the family, criminal justice, the political economy, and health care to critical scrutiny, a theory of justice seeks to help shed light on what would constitute a more fair and humane state of affairs than the one provided by the (admittedly unjust) status quo. Our “injustice radar” is frequently activated when: 1. 2.

there is significant human suffering or disadvantage; and the cause of (1) is something within our collective control. That is, our institutions or culture cause/permit (1) to occur/persist.

If we altered or reformed our institutions and societal practices in some manner we could go a significant way toward reducing or eliminating (1), and that fact is at the core of the moral judgment that mitigating (1) is a requirement of justice.

236

Colin Farrelly

While not necessarily an exhaustive definition of all injustice, I think it is the case that most things we describe as an “injustice” are cases of unnecessary suffering or disadvantage. Most pressing claims of justice thus have a HARM element and a CAUSATION element. Patriarchal practices, for example, harm women (e.g., treating them as the property of males, imposing on women an unfair division of unpaid labor within the home) and the cause of such harms come from formal legal (e.g., property or divorce law) and social (e.g., the family) institutions. The harmful effects of climate change pose significant problems for developing countries and humanity’s reliance on fossil fuels is (at least partly) responsible for this vulnerability. Where do the stakes involved in the debates surrounding the ethics of developing novel life-extending technologies fit into the picture of justice? More specifically, does justice require we extend the human lifespan by slowing the aging process? Or could such actions actually be considered unjust if they exacerbate existing health inequities or increase problems of population density and climate change? If there are conflicting or competing demands of justice in this context, how can we be begin to disentangle them so that we respond rationally and fairly to addressing the diverse stakes that arise with respect to the aspiration to extend (healthy) life? This chapter hopes to shed some light on these important questions. There are no existing theories of justice (at least that I know of!)—liberal, egalitarian, libertarian, or feminist—that were developed to take seriously the phenomenon of global aging, let alone the prospect of manipulating our biology to extend the average lifespan. I think it is an interesting question to ask why this is the case. However, that question will not be our focus here. Instead, I hope to help lay out some of the theoretical foundations necessary for us to see the force of the claim (which I offer as a provisional moral thesis) that there is indeed a justice-based case for aspiring to slow the aging process, thereby extending the period of time humans could enjoy healthy life. To make the case for this claim attention must be given to the HARM and CAUSATION dimensions of aging. Before turning explicitly to those two topics, it is important that we begin from an accurate understanding of what senescence is, why we age, and how slowing the rate of molecular and cellular decline would likely affect our health prospects in late life.

“Framing” Aging 101 Addressing the topic of the ethics of life extension by asking a general question like “Should we aspire to live to age 120+?” is, in my opinion, wrought with problems. I think it is the wrong question to ask, let alone try to answer. The information the question attempts to get us to consider is ambiguous and vague. For example, does living to age 120+ mean we enjoy more years of

Justice and the Aging of the Human Species 237 health and vitality in late life (so there is an increase in the human healthspan)? Or would we reach age 120+ by simply keeping people alive longer in a frail and diseased state (thereby increasing the frailspan)? Furthermore, are we being asked to consider what the pros and cons would be of just me, as an individual, living to age 120+? Or for my society? Or for the global population as a whole and all future generations? Unless attention is given to the different questions, and different stakes, invoked with these issues one is likely to encounter some pretty strong opposition to the suggestion that we are morally obligated to extend the human lifespan. Richard Miller (2002: 170), a scientist who works on aging in mice, describes the public relations challenge the field of biogerontology must overcome as “gerontologiphobia”—“the irrational fear that ageing research is a public menace bound to produce a world filled with nonproductive, chronically disabled, unhappy senior citizens consuming more resources than they produce.” Which questions might be better questions to ask if one is interested in building a sound moral analysis of the ethics of longevity science? Here are three foundational questions I believe it is prudent to begin with: 1. 2. 3.

What is aging? Why do we age? What impact would slowing aging have on our health prospects in late life?

Aging is “the progressive loss of function accompanied by decreasing fertility and increasing mortality with advancing age” (Kirkwood and Austad 2000: 233). A helpful distinction to note is that between primary and secondary aging. The definition just noted, by two of the world’s most eminent experts on the biology of aging, is that of what can be called “primary aging.” We all experience, regardless of lifestyle and environmental factors, the loss of function, decreased fertility, and increased mortality in late life. Primary aging is inevitable. Eating vegetables and exercising, for example, do not alter primary aging. But these lifestyle modifications can of course improve one’s health in late life, because they do have an effect on secondary aging. Secondary aging is the “deterioration in tissue structure and biological function that is secondary to disease processes and harmful environmental factors” (Holloszy and Fontana 2007: 1). Regular exercise, for example, can increase the number of years a person can expect to live, even though it does not alter primary aging or the lifespan. Moderate and high physical activity levels have been demonstrated to lead to 1.3 and 3.7 years more in total life expectancy and 1.1 and 3.2 more years lived without cardiovascular disease, respectively, for men aged

238

Colin Farrelly 50 years or older compared with those who maintained a low physical activity level. For women the differences were 1.5 and 3.5 years in total life expectancy and 1.3 and 3.3 more years lived free of cardiovascular disease, respectively. (Franco et al. 2005: 2355)

It is important to begin by acknowledging that even ideal lifestyle does not alter primary aging, and if a population like the United States complied with the requirements of an ideal lifestyle, life expectancy would not likely exceed age 85 (Olshansky et al. 1991). The average lifespan of humans is approximately 85 years, “with a distribution which includes 99 percent of individuals between the ages of 70 and 100” (Fries 2005: 808). The focus of biogerontology (more recently the field has been coined “geroscience”)—the scientific field that studies the biology of aging—is on ways to modulate the rate of primary aging. Instead of studying what causes cancer, heart disease, or stroke, biogerontologists study the biology of longlived species—like the bowhead whale, naked mole rat, and centenarians and supercentenarians—and short-lived species (like fruit flies, yeast, and mice). “Anti-aging” interventions like caloric restriction, genetic modification, and castration are thus some of the central foci of the field. The potential payoff of an applied gerontological intervention is not to deliver a cure to any one specific disease of aging (like cancer), but rather to develop an intervention that simultaneously delays all the afflictions of senescence (and possibly compresses the period of morbidity in late life). Our second question to consider is, why do humans age? Not all living species appear to age. Cnidarian hydra, for example, can reproduce asexually and regrow lost body parts and escape senescence. But for a sexually reproducing species like humans, aging is a fact of life. But why do we age? The answer lies with our distinct evolutionary and life history, a history very different from that of the hydra. The disposable soma theory (Kirkwood 1977; Kirkwood and Holliday 1979) maintains that aging occurs because natural selection favors a strategy in which reproduction is made a higher biological priority (in terms of the utilization of resources) than the somatic maintenance needed for indefinite survival. Understanding why we suffer disease, frailty, and death in late life thus begins with the story of the world’s extrinsic risks, a story succinctly summarized by Bruce Carnes (2007), which goes like this. The world is a dangerous place. Death is, for all living things on this planet, inevitable. In order for any species’ existence to persist over time a solution to death must be found. And that solution, for us and for other sexually reproducing species, is reproduction. There is thus a real race between reproduction and death, and all the species alive today are, at least for the moment, winning this race. The winning strategy for humans has been to prioritize reproduction over indefinite maintenance.

Justice and the Aging of the Human Species 239 While this winning biological strategy of prioritizing reproduction over indefinite maintenance has permitted our species to survive in a hostile world, the success human innovation has had with reducing the extrinsic threats of the world has resulted in one of the most significant (but often neglected) developments in our species’ history—population aging. “Prehistoric human remains have never revealed individuals older than about 50 years of age, and humans had a life expectancy at birth of 30 years or less for more than 99.9% of the time that we have inhabited this planet” (Hayflick 2000: 269). But with innovations like the sanitation revolution, immunizations (especially the eradication of small pox), improvements in material prosperity, changes in behavior, and so forth, life expectancy at birth for the global population now surpasses 70 years (67.5 for males, 73.3 for females) (Wang et al. 2012) and is expected to rise to age 81 by the end of this century (UN 2011: xviii). “Globally, the number of persons aged 60 or over is expected to more than triple by 2100, increasing from 784 million in 2011 to 2 billion in 2050 and 2.8 billion in 2100” (UN 2011: xvi). The story of the aging of our species is a true success story; we have dramatically reduced early and mid-life mortality. But a consequence of our success in saving the young from disease and death is that humans now live long enough that they experience the suffering, disease, and frailty of the inborn aging process. This was something most humans born into the hostile world did not have to grapple with. But now millions of people, especially those in populous middle-income countries, like China and India, suffer and die from chronic disease in late life each year. The size of the global population, coupled with the fact that populations are aging, means that an unprecedented numbers of humans will suffer the afflictions of multi-morbidity in late life—cancer, heart disease, stroke, Alzheimer’s disease (AD), arthritis, and so forth. Caring for older persons also imposes stringent duties on caregivers, typically women, to care for their older family members. And scarce healthcare resources will come under greater strain as the number of people requiring assistance to manage multi-morbidity in late life continues to grow. If aging was immutable then this state of affairs concerning the impact aging has on the health prospects of those living in this century would simply be an unfortunate, but unavoidable, development. It certainly could not be described as an “injustice.” Recall I stipulated that an injustice arises when the human suffering is unnecessary, something that could be prevented or ameliorated in some fashion. Outside of the patriarchal dimension of existing caring responsibilities I mentioned earlier, it might appear that the other deleterious impacts of global aging, such as a rise in the number of people suffering chronic disease, are unavoidable. Saving people from disease and death in early and midlife simply means delaying those things until later in life, but they cannot be eliminated.

240

Colin Farrelly

The CAUSATION element of claims of injustice requires us to give attention to at least two things. First, the extent to which individuals, groups, or societies are responsible for creating the suffering or disadvantage in question. And second, the extent to which individuals, groups, or societies could successfully ameliorate or redress the suffering or disadvantage in question. I turn now to the first issue.

Aging Is an Artifact of Human Civilization Scientists who study the biology of aging have long been aware of the fact that the aging of any species is an artifact of human intervention. In the wild, species of mice, dogs, and fish do not age because the extrinsic risks of the world are so high they typically die long before they reach the end of their lifespan. Our pets can live long enough to develop cancer because we feed and protect them, so they are shielded from threats like starvation, infection, and predation. The same is true for laboratory mice being studied for aging. Because they live in a protected environment, they can live long enough to experience the effects of senescence, something that a mouse in the wild would not typically experience. In 1953 Peter Medawar described senescence as something “revealed and made manifest by the most unnatural experiment of prolonging human life by sheltering it from the hazards of its natural existence” (Medawar 1952: 13). “Humans, and the animals we choose to protect, are the only species in which large numbers experience ageing” (Hayflick 2000: 269). Without the benefits of the knowledge of hygiene and biomedicine, only a small percentage of people would live long enough to experience what occurs when our “biological warranty” (Carnes et al. 2003) period expires. That warranty is estimated to be approximately seven decades. So humanity is responsible for the aging of our species (indeed the aging of any species). But could we actually do anything about this situation, short of forfeiting the health innovations that permitted us to reduce early and mid-life mortality in the first place? If the only choice is between a population having high morbidity and mortality in early versus late life, then it would appear the aging of the human species does not raise any novel moral challenges, as the CASUATION requirement for claiming there is an injustice would not be satisfied. Those at risk of chronic disease in late life would have no basis for claiming that their high risks of disease and frailty are in any way “unjust” as the alternative was to have a hostile environment where most persons would not live long enough to suffer the chronic diseases of late life. But insights into the biology of aging reveal that the rate of aging is malleable and influenced by a trade-off between reproduction and longevity. The disposal soma theory predicts that a greater investment in longevity should come at a cost

Justice and the Aging of the Human Species 241 to reproductive fitness. And a series of studies support that conjecture. For example, one study compared the fertility rates of men and women with exceptional longevity (Tabatabaie et al. 2011). These individuals were young adults in the 1920s, before reliable methods of birth control were widely available. The study found that the exceptionally long lived (both males and females) had an average of 2.01 children versus 2.53 children for the control group. These differences in fertility were not related to gender or to education level. But there were developmental differences among the individuals with exceptional longevity. They tended to reach menarche a year later than average, have their first child three years later, and their last child 2.5 years later than average. Other studies have examined the impact castration has on the longevity of men, and they effectively illustrate the longevity/reproduction trade-off. Castrated men residing in a mental hospital lived 14 years longer than intact men in the same hospital (Hamilton and Mestler 1969). And historical Korean eunuchs had an incidence rate of centenarians at least 130 times higher than that of present-day developed countries (Min et al. 2012). Such findings support what the disposable theory predicts—that longevity comes at a cost to reproductive fitness. What if we could safely alter this longevity/reproduction trade-off ? The process of evolution through natural selection is a blind, unintentional process. If humans possessed the knowledge to be able to purposefully alter the tradeoff, to promote more health in late life, would this be morally permissible? Indeed, could such an aspiration be morally obligatory?

Modulating the Aging Process May Soon Be a Reality A vocal contingent of longevity scientists argues that slowing human aging ought to be considered a top global priority of public health (Butler et al. 2008; Kaeberlein et al. 2015). Unlike medical research that aspires to prevent, treat, or cure a specific disease, retarding aging could have a much more significant health impact precisely because it targets the major risk factor for chronic disease—aging itself. Some have estimated that, if the effects of the magnitude of the benefits of caloric restriction on mice were replicated in humans, this would mean it would take individuals 112 years to have the abilities and risks of today’s 78-year-old (Miller 2002: 165). It is important to note that this means extending the healthspan rather than simply adding more years to the end of the lifespan, when people experience frailty and multi-morbidity. The longest lived humans (centenarians, age ≥100, and supercentenarians, age ≥110) perhaps provide us with the best insight into what the health prospects of slowing aging might offer for the population. Individuals who live such exceptionally long lives experience a compression of morbidity in late life (Perls 1997; Andersen et al. 2012). Centenarians markedly delay or even

242

Colin Farrelly

escape age-associated diseases (Perls et al. 2002: 360). Many of the longest lived humans can enjoy 100 years of life free of the prevalent diseases that typically afflict humans in the seventh and eighth decades of life (like cancer, heart disease, and stroke). Slowing the rate of aging for a population would mean fewer resources need to be allocated to treating the specific diseases of aging, fewer people would be in nursing homes (and those in such homes would spend less time there), less human suffering would be experienced, and less caregiving for older persons would be required. The social and economic benefits of keeping older populations more healthy would be enormous. And individuals in late life would have more time to pursue their passions, contributing to their communities, spending time with loved ones, and so forth. How could the rate of aging be safely modulated? There are a number of potential routes to altering the rate of primary aging. Perhaps this could involve the development of a pill that mimics the effects of caloric restriction or that mimics the natural biology of centenarians. The intervention might target telomeres (the tips of the chromosomes that get shorter each time a cell divides) or mitochondrial DNA damage. The fact that we do not yet know exactly which intervention will prove successful in permitting us to safely and effectively slow aging does not negate the conjecture that aspiring to slow aging is now a requirement of justice. The support we provide to the field of longevity science will determine if, and when, retarding aging becomes a reality. To remain indifferent to the suffering of multi-morbidity in late life is to treat older persons as less than moral equals. Treating a person as a moral equal means respecting (rather than dismissing) the importance of their interests in remaining healthy and vigorous. If a potential anti-aging intervention is forfeited because of public apathy or ageism then this can constitute an “injustice” because significant harms occurred that might have been preventable.

Objections vs. Concerns with Increased Lifespan Having spent some time explaining what biological aging is, why we age, and how we might slow aging to delay the onset of disease and frailty in late life, the case for claiming that global aging raises new concerns of justice is hopefully more evident than it would be if we simply starting by asking, “Does justice require extending the human lifespan?” But even if the rapid rise of the chronic diseases of global aging now appears on our potential “injustice radar,” to establish the case that justice requires we aspire to slow aging we must address many more concerns than those simply raised by gerontologiphobia. Not all potential opposition to increasing the human lifespan is predicated upon ignorance of the aging process or ageist attitudes. There are significant moral concerns that arise when considering the ethics of a technology that could increase lifespan. The most common worries raised concern the fair

Justice and the Aging of the Human Species 243 diffusion of such a technology, overpopulation, and the potential impact on climate change. To help us gain some perspective concerning how to address these issues I believe it is useful to start by noting a distinction Allen Buchanan (2011) makes with respect to the ethics of enhancing humans. Buchanan notes that there is a difference between voicing a concern about enhancement and raising an objection to enhancement. The former is merely a “con,” a reason against it. But an objection to an enhancement is a much stronger claim. An objection is an “all-things-considered” judgment that the cons outweigh any potential pros. I am more than willing to concede that the extending the lifespan of humans raises significant concerns. If we pursued an aging intervention that increased the human lifespan but did nothing to ensure it was fairly accessible to the world’s aging populations then it could hardly be described as a just course of action. Furthermore, if we increased lifespan but did nothing to tackle problems like population density and climate change we could be heading for disaster. Population density and climate change are significant problems, and slowing aging might exacerbate those problems. But it would be too hasty to jump from the concession that slowing aging raises concerns about population size and climate change to the “all-things-considered” stance that this constitutes an objection to pursuing age retardation. To come to the latter conclusion one would have to weigh all the pros and cons of slowing aging with those of population growth and climate change under a scenario of decelerated aging vs. normal aging. No one, including myself, has taken on this Herculean task. But it may not be necessary, especially if we recognize that the choice need not be a simple dichotomy between slowing aging or tackling population size or climate change. We can, and should, be tackling all these pressing issues. It might be instructive to consider how the concern vs. objection distinction applies to a mundane decision like balancing safety and efficiency when deciding if one should drive one’s car to work versus walking. There are reasons to prefer driving one’s car to work over walking to work. If walking to work would take one hour each way, and driving 15 minutes, then you can save 90 minutes a day on your commute by driving rather than walking. And if you have children to collect from school, family meals to cook, and so forth, you probably don’t have the luxury of an extra 90 minutes each day, especially during busy periods of the day when work and parenting duties are demanding. However, there are safety concerns to acknowledge because cars can be dangerous. It is more likely you will die in a traffic fatality than walking to work, assuming you don’t live in a very dangerous area. So you have some reason to prefer walking over driving (in addition to the health and environmental benefits of walking over driving). But does this safety concern constitute an objection to driving? Does it constitute an “all-things-considered” judgment that it is better to walk to work instead of driving? For most people

244

Colin Farrelly

who drive to work, the answer is mostly likely “no.” While they would concede safety is of course a concern, they can take precautions to minimize those risks—like wearing a seat belt, driving at a safe speed, ensuring they are well rested and not distracted with things like texting while they drive, and so forth. While the risks of a car fatality cannot be completely eliminated, they can be managed. I believe we should think about age retardation in a similar manner. There are certainly potential cons with slowing aging. Densely populated regions of the world might feel the strain of resource depletion if their populations are living longer. Longer lives for humans might mean higher levels of carbon emissions, thus adding to the problems of climate change. These are certainly worries that need to be addressed. Indeed, they are significant problems even if we do not alter the aging of the human species. But do these worries constitute an objection to age retardation? Do they represent a sound “all-things-considered” judgment that it is better to forfeit a potential late-life health intervention? One way of keeping our moral sensibilities in perspective about these issues is to gauge our attitude toward other health innovations. Imagine the following three important health measures/medical innovations were, miraculously, realized overnight. 1. 2. 3.

100% compliance with smoking cessation 100% compliance with ideal lifestyle (e.g., regular exercise and healthy diet) A cure for all 200+ types of cancer

What is our initial reaction to this prospect? Is it “Hooray!” or “Boo!”? I suspect for most of us it is “Hooray!”. And once the cheers and celebrations subside, I suspect some concerns would arise. If people are living longer because they are not dying from lung cancer or obesity or any form of cancer, this means there will be an impact on population size, climate change, social security, and so forth. But, and this is the crucial point for my purpose, I suspect no one would raise these concerns as an objection to our having pursued and accomplished any of (1)–(3). No one would say, “It is best to combat climate change by ensuring humans die sooner from obesity and/or cancer.” If I am correct about this, then the objections raised against retarding aging begin to look less persuasive as slowing aging might confer benefits even more significant than what (1)–(3) combined offer. We do not tend to adopt the attitude that forfeiting health innovations is the way to tackle overpopulation or climate change when we think of public health measures that prevent disease. And we should be consistent with those sensibilities when considering a possible aging intervention. I suspect the tendency among some to be inconsistent stems from confusion over what the

Justice and the Aging of the Human Species 245 potential benefits of age retardation actually are (e.g., extending the “frailspan” vs. “healthspan”), as well as ageism. It is interesting to consider why there aren’t widespread sentiments of “exercisephobia” or “cancer-freephobia” like there are of “gerontologiphobia.” Those concerned with population density and climate change do not typically propose people die sooner from obesity and cancer as a way of addressing those concerns. Such a proposal would strike us as being predicated upon a faulty “all-thingsconsidered” judgment because (a) human lives should not be treated in such a harsh, instrumental fashion, and (b) there are more humane, and effective, ways of tackling these problems. For the concern about overpopulation, the solution is to try to reduce fertility rates in those regions of the world that are poor and have high fertility rates. Combating poverty and patriarchy, for example, are humane and effective solutions to tackle the problem of overpopulation. Eschewing exercise or cancer research is a morally and empirically suspect stance to take. And yet people often take this very stance with respect to slowing aging. What would the benefits of slowing aging actually be? If aging is slowed by even just a few years it would mean individuals would have lower risks of disease, frailty, and death in late life. It would mean a simultaneous delay in the onset of cancer, heart disease, stroke, AD, and so forth, and possibly a compression of morbidity at the end of life. This would mean fewer health care resources have to be allocated helping older persons manage multi-morbidity and the demands of caring for those in late life could also be diminished. Slowing aging is an initiative that overlaps with not only the goals of public health but also those of global justice and feminism more generally. That is why I believe it is important to distinguish between concerns about slowing aging and objections to increasing lifespan. The former are often collapsed into the latter. But objecting to longevity science is, in my view, analogous to objecting to smoking cessation or cancer research because the benefits at stake are the same—more health and vitality in late life.

Conclusion There are many competing claims of justice in the world. And it can be difficult to discern between those claims that are most pressing and substantial, and those that are less so. In this chapter I have outlined the case, albeit only provisionally, for thinking that the aspiration to retard the rate of aging is a requirement of justice and our indifference to directly modulating our biology should now, given where the science of aging is, be considered an injustice. It is an injustice because (a) billions of persons throughout the world will suffer the chronic diseases of late life, and (b) this suffering could have been delayed and possibly reduced had we been more aggressive in our commitment to the science that examines how aging could be modulated.

246

Colin Farrelly

If I have persuaded the reader that the aging of the human species adds new considerations to our “injustice radar,” what do I propose individuals or societies do to remedy this problem? At the most general of levels I believe a three-pronged strategy is required. The first is to prioritize medical research on aging itself rather than the current approach of prioritizing only research on the diseases of aging. This is not to suggest that we abandon the latter, but at the moment most medical research in the developed world is invested into the study of specific diseases of aging rather than aging itself. Calling on the government to take on a more proactive role in modulating our biology by supporting basic scientific research on the biology of aging would be a necessary first step. Second, the aspiration to retard aging must be coupled with domestic and global policies to tackle related problems, like population size and climate change. And thirdly, should an applied gerontological intervention be developed, we should ensure the fair diffusion of any aging intervention so all of the world’s aging populations can benefit from extra years of health and vitality (and not just those living in the richest countries). The aging of the human species alters the moral landscape in novel and important ways. In this chapter I have tried to make the case that humanity does have a pressing collective moral responsibility to aspire to promote health in late life. This was not a pressing responsibility for most of our species’ evolutionary history given the fact that the planet’s hostile environment meant high rates of early and mid-life morbidity and mortality. But our success in reducing those risks means that our populations now age. Furthermore, recent studies on the aging process reveal that the rate of aging is not fixed, but is malleable. I hope the argument developed in this chapter will help philosophers take more seriously the complex moral stakes that arise in the context of an aging world.

References Andersen, S., P. Sebastiani, D.A. Dworkis, L. Feldman, and T.T. Perls (2012) “Health Span Approximates Life Span among Many Supercentenarians: Compression of Morbidity at the Approximate Limit of Life Span,” Journal of Gerontology: Medical Sciences 67A(4): 395–405. Buchanan, A. (2011) Beyond Humanity, Oxford: Oxford University Press. Butler, R., R.A. Miller, D. Perry, B.A. Carnes, T.F. Williams, C. Cassel, J. Brody, M.A. Bernard, L. Partridge, T. Kirkwood, G.M. Martin, and S.J. Olshansky (2008) “New Model of Health Promotion and Disease Prevention for the 21st Century,” British Medical Journal 337: 149–150. Carnes, B. (2007) “Senescence as Viewed through the Lens of Comparative Biology,” Annals of the New York Academy of Sciences 1114: 14–22. Carnes, B., S.J. Olshansky and D. Grahn (2003) “Biological Evidence for Limits to the Duration of Life,” Biogerontology 4: 31–45.

Justice and the Aging of the Human Species 247 Franco, O., C. de Laet, A. Peeters, J. Jonker, J. Mackenbach, and W. Nusselder (2005) “Effects of Physical Activity on Life Expectancy with Cardiovascular Disease,” Archives of Internal Medicine 165: 2355–2360. Fries, J. (2005) “The Compression of Morbidity,” The Milbank Quarterly 83(4): 801–823. Hamilton, J.B. and G.E. Mestler (1969) “Mortality and Survival: Comparison of Eunuchs with Intact Men and Women in a Mentally Retarded Population,” Journal of Gerontology 24: 395–411. Hayflick, L. (2000) “The Future of Ageing,” Nature 408: 267–269. Holloszy, J.O. and L. Fontana (2007) “Caloric Restriction in Humans,” Experimental Gerontology 42(8): 709–712. Kaeberlein, M., P. Rabinovitch, and G. Martin (2015) “Healthy Aging: The Ultimate Preventative Medicine,” Science 350(6265): 1191–1193. Kirkwood, T. (1977) “Evolution of Aging,” Nature 270: 301–304. Kirkwood, T. and S. Austad (2000) “Why Do We Age?” Nature 408(9): 233–238. Kirkwood, T. and R. Holliday (1979) “The Evolution of Ageing and Longevity,” Proceedings of the Royal Society of London: Biology 205: 531–546. Medawar, P. (1952) An Unsolved Problem of Biology, London: Lewis. Miller, R. (2002) “Extending Life: Scientific Prospects and Political Obstacles,” The Milbank Quarterly 80(1): 155–174. Min, K.J., C.K. Lee, and H.N. Park (2012) “The Lifespan of Korean Eunuchs,” Current Biology 22(18): R792-R793. Olshansky, S.J., M.A. Rudberg, B.A. Carnes, C.K. Cassel, and J.A. Brody (1991) “Trading Off Longer Life for Worsening Health,” Journal of Aging and Health 3: 194–216. Perls, T. (1997) “Centenarians Prove the Compression of Morbidity Hypothesis, but What about the Rest of Us Who Are Genetically Less Fortunate?” Medical Hypotheses 49: 405–407. Perls, T., L. Kunkel, and A. Puca (2002) “The Genetics of Exceptional Longevity,” Journal of American Geriatrics Society 50(2): 359–368. Tabatabaie, V., G. Atzmon, S.N. Rajpathak, R. Freeman, N. Barzilai, and J. Crandall (2011) “Exceptional Longevity Is Associated with Decreased Reproduction,” Aging 3(12): 1202–1205. United Nations, Department of Economic and Social Affairs, Population Division (2011) World Population Prospects: The 2010 Revision, Highlights and Advance Tables (Working Paper No. ESA/P/WP.220). Wang, H., L. Dwyer-Lindgren, K.T. Lofgren, J.K. Rajaratnam, J.R. Marcus, A. LevinRector, C.E. Levitz, A.D. Lopez, and C.J.L. Murray (2012) “Age-Specific and SexSpecific Mortality in 187 Countries, 1970–2010: A Systematic Analysis for the Global Burden of Disease Study 2010,” The Lancet 380: 2071–2094.

Index

active euthanasia (AE) 171–2 Ad Hoc Committee 76 advance directives (ADs): change of mind and 128–9, 131–5; degrees of dementia case studies 129–31; experiential vs. critical interests and 164–5; introduction to 126–7; lifesaving medical treatment and 161–5; moral authority 128; stop eating/drinking by 156, 165–7; summary of 126; then-self/now-self challenge to 126–7, 165; voluntarily stopping eating/drinking and 165–7; see also change of mind concept afterlife: annihilation assumption of 47–8, 51; conceptions of 50–2; conclusions 58–9; death and 29–30; deprivation theory and 53–4; existence requirement of 53; experience requirement of 52–3; harm of death and 48–54; immortality and, desirability of 54–8; summary of 47 age, severity of death and 62–3 agent 109 aging 236–40; as artifact of human intervention 240–1; castration and 241; defined 237; fertility rates and 241; ideal lifestyle and 238; modulating, process 241–2; moral analysis of ethics of 237; multi-morbidity and 239; physical activity and 237–8; population 239; primary 237–8; reasons for 238–9; reproduction and 238–9; secondary 237 American Academy of Neurology (AAN) 146 American Academy of Pediatrics 146 American Association of Critical Care Nurses 142

American Association of Neurological Surgeons 146 American College of Chest Physicians 142 American Medical Association 142 American Neurological Association 146 American Thoracic Society 142 anatomy, defining death and 75 annihilation assumption 47–8, 49, 51 Anscombe, G. E. M. 189 anti-aging interventions 238 anticipatory grief, care-attachment conflict and 209–12 apposite act description: actors determine 189; natural facts determine 190–1 architectonic ethics 221 assisted suicide, moral propriety of 18 atemporalism 37 attachment, features of 209–10 auto-integration definition of death 80–4; criterion matching 84 autonomy: analyzing 101–3; Christman’s analysis of 100–1; concept of 97–9; defined 3; end-of-life decision-making and 104–6; ethical principle of 141; Frankfurt’s analysis of 99–100; Kant’s view of 98–9; personal 98, 99, 102; precedent 105, 115–16; respect for 103–4, 109–10; summary of 93 autonomy, respect for 109–10; desires and values, current vs. past 120–2; precedent autonomy and 115–18 autonomy-based objections to medical futility 145–8 badness of death 30; afterlife and 48–54; deprivation theory and 50–1; Epicurus’s view of 48–9; existence requirement 49; experience requirement 49–50, 52–3; Lucretius

250

Index

view of 49; for person who dies 62; wrongness of killing and 62 basic personal care 156, 166, 167 battery 157 Beauchamp, T. L. 94 Belliotti, R. A. 16 Belshaw, C. 19 Benatar, D. 16 Berghmans, R. 133–4 Bernat, J. L. 80, 81, 83, 84 betrayal case 33, 49–50 better off dead (BOD): conclusions 26; introduction to 11–13; life worth living and 17–20; non-existence and 13–17; summary of 11; see also lives BIID see bodily identity integrity disorder (BIID) biogerontology 238 BOD see better off dead (BOD) bodily identity integrity disorder (BIID) 96–7 Borges, J. L. 57–8 Boswell, J. 12 Bradley, B. 66 brain, defining death and 75 brain-dead 78 Bratman, M. 190 Brock, D. 176, 178 Brody, H. 142–3, 144 Brongersma, E. 23 Brueckner, A. 42, 43–5, 51, 53 Buchanan, A. 243 burnout, clinicians 139 Cantor, N. 164, 165 capacity 161 cardiopulmonary criterion of death 75–6 care-attachment conflict, surrogate competence and 209–12 Carnes, B. 238 categorical desires 19–20, 55, 56 causation element of justice 236 Cavanaugh, T. 190 ceremonial suicide 23 change of mind concept 131–5; advance directives and 127, 128–9; conclusions 136; degrees of dementia and 129–31; hypothetical vs. real 128; Margo case of 130, 132–4; Mrs. Bentley case of 130–1, 134–5; Richard case of 129–30, 131–2 Child Neurology Society 146 Childress, J. F. 94 chivalry 23–4 Cholbi, M. 172

Christman, J. 100–1 circulatory criterion of death 75 climate change 243 competence: concept of 94–7; decisionmaking 110; descriptive 94; end-of-life decision-making and 96, 104–6; normative 94–5; summary of 93; see also autonomy concurrentism 37–8 conditional desires 55, 56 consequentialist assessment of doctrine of double effect (DDE) 184–5 counterfactual account for surrogate decision-makers 111–15; adequateness of 112; changing of mind and 113–14; court cases 111–12; described 111; objections to 112–15 Crime and Punishment (Dostoevsky) 22 critical interests 121–2, 134, 164 Culver, C. M. 81 cure case, the 67 current values view 117, 124 Davis, J. K. 122–3 DDA see doctrine of doing and allowing (DDA) DDE see doctrine of double effect (DDE) death: afterlife and 29–30; Bernat criteria for 83; brain-dead and 78; concept of 77; criterion of 77–8; definition of (see death, defining); event of 62; intrinsic/ extrinsic value of 62, 66–8; legal definition of 78; Lucretius on 40; past and future goods and 44; severity of (see severity of death); special case of 16; in terms of loss 17 death, defining 77, 146; anatomy and 75; auto-integration definition 80–4; cardiopulmonary criterion for 75–6; concepts used in 77–8; conclusions 89; criterion of 75–8; higher brain criterion for 77; integrated function definition 79–80; introduction to 74–5; legal definition and 78; personhood definition 84–6; sentience definition 86–9; summary of 74; whole brain criterion for 76–7 “Death” (Nagel) 29 decision-making competence 110 degree of bodily invasion 158 Deprivation Approach: challenges to 64–5, 68–9; the cure case 67; described 63–4; intrinsic/extrinsic value and 62, 66–8; in preemption cases 69–72; psychological unity and

Index 65–9; to severity of death 63–5; wellbeing and 64 deprivation theory: afterlife and 53–4; conclusions 45; Epicureanism and 31; explained 30; harm of death and 50–1; introduction to 29–31; No-Harm-Done Argument 31–3; No-Subject Argument 34–7; summary of 29; Symmetry Argument 39–45; Timing Argument 37–9 De Rerum Natura (Lucretius) 39–40 descriptive competence 94 desirability of immortality see immortality, desirability of desires and values, patient’s: current vs. past 120–2; evidence of 120; wellbeing and 120–2 desire-satisfaction: critical interests and 121; theories of well-being 18, 19 determination of death cases 148 dialectical stalemate 36–7 discursive ethics 221–2 discursive practice 221 disposable soma theory 238 doctrine of doing and allowing (DDA) 171 doctrine of double effect (DDE) 171, 183–93; consequentialist assessment of 184–5; medical context of 187; non-consequentialist assessment of 185–7; permissible acts 183; problem of closeness and 188–93; religious doctrine and 183–4 doctrine of informed consent 3 Dresser, R. 127, 163, 165 durable power of attorney 161 Durkheim, E. 172 Dworkin, R. 14, 127, 134, 164, 165 dying 30; process of 62 Elena Makropulos (EM) case 55–6, 57–8 emotion, grief as 206–7 end-of-life ethics: development of 1; introduction to 1–7 enhancing humans, ethics of 243 Epicureanism 31 Epicurus 29, 37, 50, 51; annihilation assumption 48–9; Letter to Menoeceus 31, 34; see also deprivation theory essentiality of origins thesis 41–2 eternalism 37 European Society for Intensive Care Medicine 142 Euthyphro question 50 existence condition 16

251

existence requirement of harm 49 experience requirement of harm 49–50 experiential blank, death and 29–30 experiential interests 121, 134, 164–5 exploitive, defined 145 externality thesis, medical futility and 144–5 Feit, N. 71–2 Feldman, F. 15, 40–1, 71 Finnis, J. 189, 192 Firlik, A. D. 133 first-order desires 100 Fischer, J. M. 20–1, 33, 36–7, 42, 43–5, 51, 53 FitzPatrick, W. 186–7, 190, 192 Foot, P. 176, 178, 179 frailspan 237 Frankfurt, H. 99–100 futile interventions: ethical duty to withhold/withdraw 140–3; see also medical futility futility, defined 139 Generalized Deprivation Approach 70–1 gerontologiphobia 237 Gert, B. 81 Gilgamesh 48 grief 206–8; as emotion 206–7; loss and 207–8; pattern of 207 Grief Observed, A (Lewis) 206 Hailu, A. 148 hara-kiri 23, 25 harm of death 30; afterlife and 48–54; deprivation theory and 50–1; Epicurus’s view of 48–9; existence requirement 49; experience requirement 49–50, 52–3; Lucretius view of 49; for person who dies 62; wrongness of killing and 62 harm element of justice 236 Harris, J. 84–5 hedonism, critical interests and 121 hedonistic theories of well-being 18–19 higher brain criterion for defining death 77, 88–9 holistic view of life 21–2 Holland, R. F. 173–4 human healthspan 237, 241 Hume, D. 11, 12, 13, 15 hypothetical vs. real change of mind 128 “Immortal, The” (Borges) 57–8 immortality, desirability of: afterlife and 54–8; desire types and 55, 56–7; EM

252

Index

case of 55–6, 57–8; immortality types and 57–8 indefinitism 38 injustice, defined 235, 236 Institute of Medicine 227 integrated function definition of death 79–80 internal morality of medicine 142–3 investment interests, incompetent patient’s 122–3 Jaworska, A. 132–3 Jecker, N. S. 139–40 Johansson, J. 37 Johnson, S. 12 Jonsen, A. R. 139–40 justice: elements of 236; mitigating as requirement of 235 Kagan, S. 85 Kamisar, Y. 171 Kamm, F. 178–9 Kant, I. 98–9 Kantian theory, K/ATD distinction and 181 Kass, L. 182–3 K/ATD see killing and allowing-to-die (K/ATD) distinction Kaufman, F. 42–3 killing and allowing-to-die (K/ATD) distinction 175; Kantians and 181; medical context and 181–3; moral importance of 176–8; as morally important 178–80; natural law and 180–1 Kübler-Ross, E. 206–7 legal definition of death 78 Letter to Menoeceus (Epicurus) 31, 34 Lewis, C. S. 206 life expectancy, severity of death and 63, 66–8 life extension: aging explained 236–40; causation dimension of 240; conclusions 245–6; harm dimension of 239; introduction to 235–6; objections vs. concerns with 242–5; summary of 235; see also aging life satisfaction, critical interests and 121 life worth living 17–20 lives: contents/shapes of 20–2; holistic view of 21–2; investments in 14; narrative view of 20–1; objective/ subjective judgments of 19–20; sanctity and 14; suffering and 16, 20; value of, judging 23–5; worth living 17–20

lifesaving medical treatment (LSMT): advance directives and 161–5; conclusions 168; introduction to 155–7; refusing 157–8; summary of 155 living wills 3, 4, 105, 109, 161 locked-in syndrome 11 loss, death and 17 Lucretius 31, 34, 44, 48, 50–1; De Rerum Natura 39–40; symmetry argument 40, 49, 53; see also deprivation theory Luper, S. 37–8 MacMahan, J. 186, 187 McIntyre, A. 185 McMahan, J. 14–15, 16, 20, 66–7, 69 McMath, J. 148 Medawar, P. 240 medical futility: autonomy-based objections to 145–8; conclusions 149; ethical duty to withhold/withdraw 140–3; externality thesis and 144–5; internal morality of medicine and 142–3; introduction to 138–9; meaning of 139–40; moral distress and 139; objection to withholding/withdrawing 143; opponents of 148–9; patient autonomy and 138; qualitative dimension of 140; quantitative dimension of 139–40; summary of 138 medical immortality 57 medical life support 155 medical power of attorney 109 Mill, J. S. 18 Miller, F. G. 142–3, 144 Miller, R. 237 Milton, J. 12 Misak, C. J. 147 momentarily competent 112 moral considerability 220 moral distress, clinicians 139 moral emotions 202 moral standing, solidarity and 220–4 Mortal Questions (Nagel) 42 multi-morbidity 239 Myth of Er 51, 52 Nagel, T. 16, 29–44, 49–50, 58; see also deprivation theory narrative view of life 20–1 natural law theory, K/ATD distinction and 180–1 necessary boredom thesis 54, 55, 56 Nelkin, D. 190, 191 next of kin, as surrogate 109 No-Harm-Done Argument 31–3 No-Subject Argument 34–7

Index non-autonomous decisions 110 non-consequentialist assessment of doctrine of double effect (DDE) 185–7 non-existence 13–17 non-maleficence, ethical principle of 141 nonperson 119 normative competence 94–5 now-self 126 Nussbaum, M. 34, 35–6 objective-list theories of well-being 18, 19, 24–5 other-focused desires 56–7 Paradise Lost (Milton) 12 Parfit, D. 19, 42, 43–4 patient autonomy, end-of-life decisionmaking and 224–7 Pellegrino, E. 182 persistent vegetative state (PVS) 146–7, 148 person: defined 85; Kaufman and 42–3 personal autonomy: political approach to 98, 99, 102; value-creating approach to 98, 99, 102 personal identity, surrogate decisionmakers and 118–20 personhood definition of death 84–6; argument for 85–6 physician-assisted suicide (PAS): DDA and 175–6; DDE and 183–93; introduction to 171–2; K/ATD distinction and 176–83; philosophical positions on 174–5; summary of 171 Plato 51–2 population aging 239 population density 243 posthumous harm, critical interests and 122 posthumous nonexistence 40 power of attorney for health care 109 precedent autonomy 105; cases against 116–17; current values view and 117–18; defined 115–16; prior values view and 115–18 preemption cases: Deprivation Approach and 69–72; severity of death in 69; Young Pedestrian case 69–71 premise, Symmetry Argument and 40–1 prenatal nonexistence, Symmetry Argument and 40–1 President’s Commission 76, 79, 80–1, 82, 84 President’s Commission for the Study of Ethical Problems in Medicine and Biomedical and Behavioral Research see President’s Commission

253

pre-vital nonexistence 40 prima facie 99, 103, 104–5; best interests and 123; case against precedent autonomy 116; prior values view as 115; rights 109–10 primary aging 237–8 prior values view for surrogate decisionmakers 115–18, 124; defined 115; precedent autonomy and 115–18 priorism 38 problem of closeness, DDE and 188–93; actors determine apposite act description 189; described 188; middle position possibility for 191–3; natural facts determine apposite act description 190–1; ways forward from 188–9 proxy 109 psychological theory of personal identity 119–20 psychological unity, Deprivation Approach and 65–9 qualitative futility 140 quantitative futility 139–40 Quinlan, K. 111, 224 Quinn, W. 191 Rachels, J. 176, 177, 178, 185 rational desire 19 Rawls, J. 221 relational judgement: communication and 229; described 230; end-of-life decision-making and 228–31 reproduction 238–9 Republic (Plato) 51–2 respect for autonomy see autonomy, respect for retrospective consent 158 Rhoden, N. 127, 163, 165 Rickless, S. 190, 191 right recognition: as condition of moral membership and equality 220; solidarity and 220 right relationship as condition of mutuality 220 Robertson, J. 127, 163, 165 Roiphe, A. 206 sanctity of life, suicide and 173 sati 25 Saunders, C. 224 Scanlon, T. 185–6 Schiavo, T. 148 Schneiderman, L. J. 139–40 secondary aging 237

254

Index

self-determination 109; see also autonomy, respect for self-focused desires 56 self-governing person 145 self-sacrificial acts, suicide and 173–4 Sen, A. 24 senescence 238, 240 sentience definition of death 86–9; argument for 87–8; criterion matching 88–9 severity of death: age and 62–3; clarifying points 61–2; conclusions 72; Deprivation Approach to 63–5; introduction to 61; life expectancy and; in preemption cases 69–72; psychological unity and 65–9; summary of 61 Shalowitz, D. I. 204–5 Singer, P. 176 Sirovich, B. E. 139 Society of Critical Care Medicine 142 Socrates 51–2 solidarity, end-of-life decision-making and: conclusions 231; ethics of 220–1; importance of 218–19; introduction to 218–20; moral standing and 220–4; patient autonomy and 224–7; patient empowerment and 224–7; relational judgement and 228–31; right recognition/relationship and 220; summary of 218 Solomon, M. Z. 139 spatially distant 38 standing-up-for solidarity 222–3, 226 standing-up-with solidarity 223, 226 Strawson, G. 22 subsequentism 37 substituted judgment principle 109, 128, 204–5 sufferings 44 suicide 172–5; Durkheim and 172; Hume rationale for 13; intentional selfkilling as 172–3; permissibility of 173; sanctity of life and 173; self-sacrificial acts and 173–4 Sumner, L. W. 176–7, 179 surrogate decision-makers, moral framework for: conclusions 123–4; counterfactual view of 111–15; desires and values, current vs. past 120–2; introduction to 108–10; investment interests and 122–3; personal identity and 118–20; prior values view 115–18; summary of 108; well-being and 120–3 surrogate decision-making, end-of-life: autonomy-based rationale for 204;

care-attachment conflict and 209–12; clinical/philosophical responses to 212–15; conclusions 215; grief and 206–8; introduction to 201–3; rise and rationale of 203–4; substituted judgment and 204–5; summary of 201; unreliability of 204–6 surrogates: decision making methods of 205–6; hard paternalism and 204; people serving as 203; rationale for using 204; standard of judgment applied to 213 Symmetry Argument 39–45, 49 temporal distance 38–9 termination thesis 15 then-self 126 Theory of Justice, A (Rawls) 221 thick person 42–3 thin person 42–3 Thomson, J. 185, 186 Timing Argument 37–9 Tooley, M. 176 true immortality 57–8 Uniform Determination of Death Act (UDDA) 146 ventilators, defining death and 76 voluntarily stopping eating and drinking (VSED): advance directives and 165–7; basic personal care and 156, 166, 167; characteristics of 160; conclusions 168; introduction to 155–7; right to 158–61; as suicide 160–1; summary of 155 voluntary euthanasia: Hume rationale for 13; moral propriety of 18 weak autonomy 141 Wedgwood, R. 186, 192 well-being 120–2; critical vs. experiential interests and 121–2; levels of 64; theories of 18–19 whole brain criterion for defining death 76–7, 84 Wiggins, D. 186 Williams, B. 16, 19–20, 54–8; see also immortality, desirability of Williams, G. 171 Wittgenstein, L. 207 Wonderley, M. 209–10 wronging a person 122–3 Young, R. 187 Young Pedestrian case 69–71