Moral Reasoning in a Pluralistic World 9780773597563

Citation preview

Acknowledgments

moral reasoning in a pluralistic world

i

ii

preface

This page intentionally left blank

preface

Moral Reasoning in a Pluralistic World patricia marino

McGill-Queen’s University Press Montreal & Kingston • London • Chicago

iii

iv

preface

© McGill-Queen’s University Press 2015 isbn 978-0-7735-4614-1 (cloth) isbn 978-0-7735-4615-8 (paper) isbn 978-0-7735-9756-3 (epdf) isbn 978-0-7735-9757-0 (epub) Legal deposit third quarter 2015 Bibliothèque nationale du Québec Printed in Canada on acid-free paper that is 100% ancient forest free (100% post-consumer recycled), processed chlorine free. This book was published with the help of a grant from the Canadian Federation for the Humanities and Social Sciences, through the Awards to Scholarly Publications Program, using funds provided by the Social Sciences and Humanities Research Council of Canada. McGill-Queen’s University Press acknowledges the support of the Canada Council for the Arts for our publishing program. We also acknowledge the financial support of the Government of Canada through the Canada Book Fund for our publishing activities.

Library and Archives Canada Cataloguing in Publication Marino, Patricia, 1966–, author Moral reasoning in a pluralistic world / Patricia Marino. Includes bibliographical references and index. Issued in print and electronic formats. isbn 978-0-7735-4614-1 (bound). – isbn 978-0-7735-4615-8 (pbk.) isbn 978-0-7735-9756-3 (epdf). – isbn 978-0-7735-9757-0 (epub) 1. Ethics. 2. Reasoning. 3. Pluralism. I. Title. bj1025.m37 2015

170

c2015-902833-7 c2015-902834-5

This book was typeset by True to Type in 10.5/13 Sabon

List of Tables and Figures

For Jonathan

v

ii

preface

This page intentionally left blank

preface

vii

Contents

Acknowledgments

ix

Introduction 3 1 Basic Concepts 12 2 Value Pluralism, the Unity of Value, and the “Delimiting Argument” 34 3 Epistemological Arguments and Rich Coherence 60 4 Case Consistency 84 5 Pluralist Coherence 108 6 Conclusion 142 Notes 161 Bibliography 183 Index 193

ii

preface

This page intentionally left blank

List of Tables and Figures

ix

Acknowledgments

I have been thinking about the ideas in this book for a number of years, and I am grateful to the many people who contributed to their development. In graduate school I was fortunate to take a class on moral dilemmas with the late Ruth Marcus, and it was there I began to think about morality and conflict. Though I was working on different topics at that time, my PhD supervisor Penelope Maddy has taught me so much about philosophical thinking and writing over the years that all of my work bears the imprint of her teaching. Lorraine Besser-Jones, Luke Brunning, Dave DeVidi, Emma Dewald, Mathieu Doucet, Doreen Fraser, Allan Gibbard, Alan Goldman, Patricia Greenspan, Nicole Hassoun, Tom Hurka, Chris Kaposy, and Tim Kenyon all contributed insightful comments on earlier versions of this work. Jonathan Dewald, Penelope Maddy, and Paul Thagard not only helped at many points along the way but also read the whole manuscript; their thoughts and suggestions greatly improved the final manuscript. I have presented parts of this work at the Western Canadian Philosophical Association Meeting in 2006, the Canadian Philosophical Annual Meetings in 2007 and 2010, the American Philosophical Association Central Division Meetings in 2007 and 2008, the North American Society for Social Philosophy conference in 2009, the American Philosophical Association Eastern Division Meeting in 2009, the American Philosophical Association Pacific Division Meeting in 2010, the Departments of Philosophy at Wilfrid Laurier University, Ryerson University, University of Toronto, University of Guelph, Southern Illinois University, and University of Waterloo, and the Joint Centre for Bioethics in Toronto. Discussion with audiences there improved this

x

Acknowledgments

work substantially. I’m especially grateful to Nathan Brett, Eugene Heath, Tristram McPherson, Sergio Tenenbaum, and Mark van Roojen for their probing questions. Elements of this work appeared previously in “Moral Dilemmas, Collective Responsibility, and Moral Progress,” Philosophical Studies 104 (2001): 203–25, “Moral Rationalism and the Normative Status of Desiderative Coherence,” Journal of Moral Philosophy 7 (2010): 227–52, “Ambivalence, Valuational Inconsistency, and the Divided Self,” Philosophy and Phenomenological Research 83 (2011): 41–71, “Moral Coherence and Value Pluralism,” Canadian Journal of Philosophy 43 (2013): 117–35, “Moral Coherence and Principle Pluralism,” Journal of Moral Philosophy 11/6 (2014): 727–49. The critical and constructive comments that I received from reviewers for these journals have been immensely helpful. This research was supported by the Social Sciences and Humanities Research Council of Canada, in part through a generous Standard Research Grant. My research assistants Rosalind Abdool and Kirsten MacDonald, funded through this grant, not only helped with details but also provided insightful feedback. This book has been published with the help of a grant from the Federation for the Humanities and Social Sciences, through the Awards to Scholarly Publications Program, using funds provided by the Social Sciences and Humanities Research Council of Canada. The University of Waterloo also supported this work through sabbatical leaves. Two anonymous referees for McGill-Queen’s University Press provided reports that were the perfect mix of encouragement and criticism. Mark Abley, my editor at MQUP, made the publication process smooth, constructive, and pleasant, and Kathryn Simpson’s perceptive questions during the copy-editing phase led to substantial improvements in the text. For a long time now my partner Jonathan Dewald has been a source of intellectual inspiration and my most penetrating philosophical interlocutor. With love and gratitude, this book is dedicated to him.

Introduction

moral reasoning in a pluralistic world

1

ii

preface

This page intentionally left blank

0 Introduction

This book concerns moral reasoning: the process we engage in – or ought to engage in – when we try to figure out what is right and wrong and why. Moral reasoning is puzzling. On the one hand, moral disagreement and the diversity of moral thinking around the world might seem to show that moral justification is simply impossible. Moral judgments vary greatly from place to place, time to time, and even from person to person. If morality is based on reason, we should expect there to be a single morally correct point of view; it would follow that beliefs that disagree with this single truth are simply mistaken. This strikes many people as both implausible, since it entails that so many people would have to get so many things wrong over a long period of time, and dangerous, since it seems close to imperialism to say that we have got it right and everyone else has got it wrong. On the other hand, almost everyone engages in moral reasoning, if only to find fault with those they disagree with. We know that if a person were to say, “I believe abortion is morally wrong, but I believe it’s okay for me to have one,” then something would have gone wrong. Arguments against the permissibility of abortion are often criticized for ignoring the rights of women to control what happens in and to their bodies. Arguments for the permissibility of abortion are often criticized for failing to apply the same standard to some forms of human life that are applied to others. These are examples of using reason to figure out what to believe – and of trying to convince others what to believe – in the moral realm. Moral arguments often make a difference to us. It used to be thought that a man could not rape his wife: marriage implied unity of a man

4

Moral Reasoning in a Pluralistic World

and his wife, and forced sex could only be morally wrong if it violated someone’s chastity – something impossible in such a context. Over time, this belief came to seem contradictory to other beliefs about individualism, equality of persons, and rights to one’s body. Feminists fought to have the contradiction resolved through a legal recognition that marital rape is rape.1 One important element in the development of attitudes against slavery in antebellum US was the noted inconsistency of enslavement in “the land of the free.”2 Recently Mark Grisanti, a Republican member of the New York State Senate who had pledged to vote against legalizing same-sex marriage, voted in favour of legalizing it, on the grounds that after doing months of research and thinking, he could not “legally come up with an argument against [it].”3 Most broadly, the question of this book is how moral reasoning is best understood in a world with varying moral outlooks, a diversity of moral views, and substantial moral disagreement. Though moral diversity has sometimes been associated with moral skepticism, recent empirical research suggests that such diversity is best explained by appeal to the fact that the way people value is “pluralistic” in two senses. First, we have a number of common but distinct moral cares, such as benevolence, fairness, honesty, and loyalty. Second, different people direct those cares at different objects and prioritize among them in different ways. These pluralisms help explain why there is diversity in moral judgments: in deciding whether it is appropriate to lie to protect a friend, a person who values honesty most may say no, while one who values loyalty most may say yes. This connection between diversity and conflict explains why so much moral disagreement seems to arise in the contexts of dilemmas, in which there is more than one value at stake and we feel pulled in different directions. I focus here on reasoning through methods of moral coherence. We employ methods of moral coherence when we work back and forth among our considered judgments – such as “you ought to keep this particular promise” – and the general principles we endorse – such as “promise keeping is generally morally required” – seeking an acceptable fit among these. Such a method is typical in ordinary moral reasoning, as for instance when we reason about abortion by asking whether abortion is more like other forms of killing or more like other forms of exercising personal autonomy. Coherence asks us to universalize our judgments and avoid arbitrary distinctions, and thus

Introduction

5

explains what has gone wrong in the abortion reasoning above: it makes no sense to make an exception for oneself for no good reason. But traditional conceptions of moral coherence, I argue, are inappropriate for use in pluralist contexts like our own. These “rich” forms of coherence require that we pursue overarching values in terms of which others are understood, and to pursue moral belief systems with as few, simple principles as possible. An example of a richly coherent theory is utilitarianism, according to which the right action is the one that maximizes happiness or pleasure overall. Utilitarianism offers a unified view of value, a single principle to follow, and a single metric for making decisions about what to do. David Brink expresses a commonly held view of coherence when he writes that “other things being equal, the more interconnections and mutual support among beliefs, the better justified those beliefs are ... Coherentism, therefore, favours unified over nonunified or fragmented moral theories.”4 Understood this way, coherence methods are dramatically at odds with our practice of valuing pluralistically. They deny that values are really multiple in the relevant sense. Adopting them would thus require us to reject our pluralistic way of valuing. How should we respond to this problem? Does the correctness of such methods show that we ought, contrary to everyday practice, to unify our cares and systematize our beliefs? I argue here that the answer is no. These interpretations of coherence rest on epistemological assumptions that cannot be justified in pluralist contexts. While it may seem appealing to resolve difficult moral matters by finding a single principle to apply to all cases, I show that these approaches to coherence are misguided: we have no reason to prefer unified and systematized theories. And while it may seem intuitive to say we ought to systematize as much as possible, I argue this isn’t so: we have no reason to think the beliefs in more unified theories support and explain one another better than those in more pluralistic ones. I propose an alternative way of thinking about coherence, moral reasoning, and to some extent moral progress, that makes sense in our pluralistic world. Central to this proposal is the importance of prioritizing among distinct and conflicting values, and using this prioritization in addressing moral dilemmas – situations in which our principles conflict and our obligations are incompatible. Prioritizing requires making sensible compromises and trade-offs among values such as benev-

6

Moral Reasoning in a Pluralistic World

olence, fairness, honesty, and fidelity. The aim is to find a principled way of weighing values against one another. Finding such a principled way, I argue, does not require finding a single overarching principle or value in terms of which others can be understood. It requires instead consistency from one case to another: when we prioritize, we aim for a set of judgments about relative importance that are consistent from case to case. This does not require a ranking of values or principles, where one always takes precedence over others. It instead entails judging similarly cases that are similar in morally relevant ways, and being willing to stand behind morally significant differences in cases we judge differently. For example, most of us are highly susceptible to framing effects: we judge differently cases that are identical in all the facts just because of the way the cases are presented and described. When the same tax policy, for instance, is described in terms of exemptions, it is judged to be unfair, but when described in terms of benefits, it is judged to be fair.5 We make our moral beliefs principled by ridding them of such anomalies. Because of the effects of emotions, self-interest, and contextual influences, most of us are far from principled in this way. The point of moral theorizing and reflection is to become more so. This form of coherence supports theories that have multiple basic principles – principles such as “Don’t lie,” “Don’t harm others,” and “Keep your promises” – grounded in various values and leading to conflicting obligations. The best-known example of a moral theory that is pluralist in this sense is that of the early twentieth-century ethicist W.D. Ross.6 Ross argues for principles emerging from seven or so kinds of duties: there are duties of fidelity, such as promise keeping, duties of reparation, as when one has done something wrong, duties of gratitude, of justice, of beneficence, of self-improvement, and of not inflicting harm. When these principles engender conflicts – e.g. a person may be required to keep her appointment and save a life in conditions in which she cannot do both; a person who has promised to keep a secret may be under obligation from principles of honesty to tell the truth, and under obligation from principles of promise keeping to lie – Ross says we must use our judgment to prioritize, that is, to decide which obligation is most pressing. Meeting this becomes our overall duty. Ross’s theory has been charged with failing to be sufficiently unified and systematic – with giving us with an “unconnected heap of

Introduction

7

duties.”7 A common contemporary response to this is to show how Ross’s view, or one like it, can be made unified and systematic – by showing, for example, how the various principles and duties follow from something more basic. In Brad Hooker’s view, our duty is simply to follow rules that, if internalized by most people, would produce the most well-being overall.”8 Robert Audi proposes unifying considerations inspired by Kant – we should consider whether our action falls under a maxim we can universalize and whether our actions treat persons as mere means.9 I argue here, however, that to bring coherence to a set of Rossian beliefs requires not unifying and systematizing but rather judging in ways that are case consistent. Deploying case consistency means that we cannot prioritize on a case-by-case basis, as Ross’s view about “judgment” might seem to suggest; instead we must consider how our judgments of priority compare to other judgments of priority in relevantly similar cases. If we take coherence as involving principled compromises and consistency from one case to another, it follows that when we roughly share values and priorities we can use moral reasoning to show that one or another interlocutor is mistaken in failing to be coherent, and this is an important role. However, because we prioritize among and direct our values differently, competing moral belief sets can be internally coherent and still disagree with one another. The person who values honesty most and the person who values fidelity most may use coherence to root out inconsistencies in their own views, since they may fail to apply these priorities consistently or may make special and unjustified exceptions. But they may continue to disagree about what should be done in the case of a conflict where a person must choose whether to lie or betray a friend. In these situations reason cannot tell us who is right, and moral persuasion requires other kinds of social and cultural changes. So reason plays an important but also limited role in justification and persuasion. It follows from this that two kinds of moral progress are compatible with moral diversity. First, we may work – often collectively – to reduce the likelihood of conflicts and dilemmas, and in this way honour more of our shared values. Those who value honesty and fidelity can agree that it is best to avoid situations in which these will conflict – even when they disagree about how to adjudicate those conflicts. Economic inequalities among people can create the conditions for conflicts between the demands of benevolence and those of respect

8

Moral Reasoning in a Pluralistic World

for autonomy, since offers may be coercive if they cannot be refused. Those who share these values but prioritize differently may disagree about what to do in the immediate circumstance, but can agree on the importance of avoiding similar conflicts in the future, perhaps by reducing inequalities. Second, an expansion of our circle of concern is compatible both with multiplicity of values and with diversity and disagreement. Peter Singer writes of the improvement of our moral views that results when we expand our circle of concern beyond our family and friends, our nation, and our species.10 Singer is a utilitarian who endorses maximizing the satisfaction of interests overall, which means that the moral judgments he puts forward are richly coherent in my sense. But one needn’t endorse one value at the expense of others to expand one’s circle of concern: values such as fairness, honesty, and loyalty can be expanded together. The result is a form of moral progress that is compatible with the existence of a range of distinct and conflicting values. Such expansion is also compatible with differing prioritizations and thus with moral diversity. Three implications for our moral practices follow. The first has to do with philosophical methodology. The ideas of compromise and consistency from one case to another are not new, and work in applied ethics frequently makes use of methods like the one I am proposing. But from a theoretical point of view in philosophical thinking, such methods are denigrated: they are treated as mistaken, subpar, or regrettable concessions to the difficulties of finding systematic theories we can all agree on. I’ve already mentioned the way defenders of Rossian theories try to show how those theories can be unified and systematized. Tom Beauchamp and James Childress, the authors of a standard contemporary bioethics text, defend a theory with multiple basic principles but go on to ask: “can the common morality itself be made coherent? If one argues, as we do, that a heap of obligations and values unconnected by a first principle comprises the common morality, is there any hope of coherence, short of so radically reconstructing norms that they only vaguely resemble the norms in the common morality?”11 This means that the methods that are considered most appropriate from a theoretical point of view are not the methods that are typically used in applied contexts. Here I argue, however, that the methods used in applied contexts – of arguing for morally significant differ-

Introduction

9

ences and similarities and drawing out implications – are entirely appropriate from a theoretical point of view as well. The difficulty of using unified theories in application is that different unified theories give different answers to complex problems; from this it is unclear how to proceed. This difficulty manifests itself in a practical way in university courses on moral issues. There it is common to begin with a range of different unified, systematized theories, and then to consider how well those theories work for analyzing particular problems. If I am right about pluralist coherence, this approach is mistaken: instead of considering which systematic or unified theory is most appropriate for a given problem, we ought to consider how to incorporate multiple values into an outlook that can be applied to all cases consistently. Multiplicity, pluralism, conflict, and compromise are not defects in a moral point of view. Two other implications of my approach concern the role of reason in popular moral discourse. It is often observed that spirited moral debate can produce more heat than light. When we disagree morally, arguments and the use of reason do not appear to contribute much toward producing consensus. People seem more likely to appeal to emotion than reason in making such judgments, and more likely to believe what their friends believe morally than what they themselves have reasoned out. We respond more to emotional appeals – such as pictures of suffering animals – than to logical arguments. Given this state of things, pluralism and diversity are sometimes thought to undermine the very possibility of moral reasoning – and, by extension, moral progress. In his 2002 book The Problematics of Moral and Legal Theory Richard Posner argues that appeal to reason in morality requires an assumption that disagreements can be resolved by settling on a few uncontroversial premises and seeing what follows from those. This assumption, he says, is proven false by moral pluralism – that is, by the fact that people’s judgments not only vary but correlate with their social communities. But in my proposal, a role for moral reasoning is compatible with pluralism: the fact that people sometimes value and prioritize differently is compatible with the fact that they ought to be consistent in their views and ought to stand behind the morally significant differences they claim to endorse. Finally, it has been said that the bitterness and futility of moral debate in public life is best explained by the fact that most people’s use of moral reasoning is extremely limited. Jonathan Haidt, for exam-

10

Moral Reasoning in a Pluralistic World

ple, argues that the vast majority of people not only do not use moral reasoning when making judgments, they also fail to change their minds when presented with contrary evidence. When they cannot defend their own moral beliefs, they are stubbornly and irrationally “morally dumbfounded”: unable to articulate a reason or justification, but also unwilling to change. Haidt says that our mistaken faith in our reasoning powers leads us to expect – incorrectly – that if others are reasoning, they will come to agree with us. And thus the bitterness and futility of most moral discourse.12 The point of view being developed here, however, shows how this explanation might be mistaken: the bitterness and futility of much ethical debate are consistent with the use of reason, because those who value differently may be responding to reasons and nonetheless disagreeing. What seems like stubborn moral dumbfounding is really just the apt expression of commitment to certain basic moral beliefs, and is appropriate. That we sometimes justify moral beliefs in terms of certain values without being able to explain why those values matter to us is what we ought to expect, and does not show that we are deficient in our use of moral reasoning. In addressing the question of moral reasoning in a pluralistic world, I adopt a methodology that takes our considered moral convictions as good starting places for moral thinking. This book does not, therefore, attempt any new way of justifying morality or of answering the moral skeptic – the person who doubts whether any moral beliefs are true at all. It also does not contain direct discussion of the debate over objectivity and relativism, and I do not have anything to say about how, politically, we ought to structure our society in response to pluralism. The book does not address the normative question of what the right answers are to moral questions, and though it defends the appropriateness of appealing to morally significant factors in justification, there isn’t much information here about how those factors are, or should be, embedded in one’s values. Finally, it does not present a point of view about the metaphysical or epistemological view of “moral pluralism.”13 It addresses, instead, a question of method: the question of how to proceed with moral reasoning as the people we are in the modern world, people with multiple values, who confront frequent conflicts and dilemmas, and who sometimes must engage with people who have very different moral beliefs from our own. I do not claim to have shown that unified views are untenable

Introduction

11

in themselves. My argument proceeds by examining the pluralistic situation I believe we find ourselves in, showing why the most common strategies in favour of unification are unsuccessful, articulating an alternative approach that fits our pluralist context, and drawing out the implications – many of them quite appealing – of adopting that approach. In chapter 1, I explain the concepts central to my discussion – those of moral disagreement, value pluralism, moral coherence, and so on – and I argue for three claims: that the pursuit of moral coherence does not require the elimination of moral conflicts, that there are reasons to retain conflicts in moral reasoning, and that the need to prioritize in cases of conflict does not support systematizing. In chapter 2, I focus on value pluralism. It might seem that to be properly integrated as a person one must value in a unified way; if this entails interpreting values so that they do not really conflict, it might seem that the way we value is only nominally pluralistic. I argue, however, that rationality and autonomy require prioritization – not unification – of our values. Chapter 3 addresses epistemological approaches to systematicity; there I show that given the pluralistic way we value we have no reason to think that the beliefs in more systematized sets support and explain one another any better than beliefs in more pluralistic ones. Rich coherence is inappropriate for pluralist contexts. In chapter 4 I explain case consistency, saying what it means to judge in accordance with morally relevant similarities and differences and defending the pursuit of such consistency as a useful, novel, and plausible approach to moral reasoning in pluralist contexts. In chapter 5 I say more about pluralist coherence and give some examples, and in the concluding chapter 6 I give a brief overview, address the three implications for moral practice mentioned above, then briefly discuss a couple of further thoughts.

12

Moral Reasoning in a Pluralistic World

1 Basic Concepts

In this chapter I start by explaining some of the concepts central to my discussion: those of disagreement and diversity, value pluralism and principle pluralism, moral coherence, moral conflict, and moral dilemmas. I characterize a norm of systematicity as saying that the general principles of a moral theory should be as few and as simple as possible. I then argue for three conclusions about the status of moral conflict and its relationship to systematicity: first, that coherence does not require eliminating moral conflict; second, that there are reasons to retain such conflicts; and finally, that the need to prioritize among conflicting obligations and duties does not support the systematicity norm. 1

MORAL DISAGREEMENT AND DIVERSITY

I use the terms “moral disagreement” and “moral diversity” interchangeably to refer to those situations in which people make different and incompatible moral judgments. Anyone who has watched the intense recent public debates over issues like abortion, sex work, samesex marriage and assisted suicide has a vivid sense of how entrenched moral disagreement can be, of how committed various parties are to their views, and of the range of people’s opinions. For a particular example of moral disagreement, consider the following. In the late 1990s, controversy erupted over a set of proposed international collaborative medical studies meant to test a new approach for reducing mother-to-child HIV transmission.1 These trials were to be performed in developing countries, and the controversy

Basic Concepts

13

had to do with methodology: in these studies the treatment was compared to a placebo, which meant that some test subjects received no treatment at all. The studies would have been judged unethical in the US, Canada, or Europe because of longstanding guidelines concerning the treatment of subjects in control groups. Well-designed studies must test a treatment against something else: a control is necessary to determine efficacy. These longstanding guidelines specify that the control group ought to receive “the best” current proven treatment.2 Among other things this rule is meant to ensure that there will be no repetition of the horrific Tuskegee experiments, in which poor black sharecroppers were tested in order to watch the natural progression of syphilis – that is, the progression of the disease in the absence of treatment. In this study, the hundreds of people who were found to have syphilis were not told they had it, and were not treated with the standard cure of penicillin between 1947, when that cure became available, and 1972, when the study was called to a halt.3 These subjects were clearly being exploited. In the late 1990s, when the HIV transmission study was done, there was an effective – but expensive – way to limit transmission involving AZT administered intravenously relatively early in pregnancy. The problem was that because of a lack of resources and infrastructure this method was impossible to use widely in certain populations. The new test thus sought to determine the effectiveness of a simpler and cheaper regime: lower doses of AZT, administered later in pregnancy (since the target population would rarely receive early pre-natal care), and administered orally (suiting the medical facilities to which the target population had access). Because the test hoped to determine the effectiveness of a lesser intervention than the known standard one, the question was how much better the proposed treatment was than no treatment at all. The new treatment was therefore tested against a placebo. The international furor that resulted raised several profound ethical questions. Were the people in these studies being exploited because of their poverty? If the studies were unethical in some places, how could they be ethical in others? Could anything justify such a “double standard”? One simple answer to this last question is “no.” Writing in the New England Journal of Medicine, Marcia Angell gave her opinion that “our ethical standards should not depend on where the research is

14

Moral Reasoning in a Pluralistic World

performed” – any other approach, she says, can lead to the exploitation of test subjects.4 If ethical beliefs are justified, they should not vary in application from one place to the next for no good reason. Some argue, however, that this answer is improperly strict and categorical, and that a better answer is “sometimes.” The people in developing countries are desperate for new treatments for diseases like malaria; if using laxer guidelines means that more possible treatments are tested, that makes a difference. The women in the HIV study would not have received any treatment if they were not participating, so they were not being made worse off. If no one is being made worse off by these kinds of tests, and new treatments are developed, isn’t this good? And what if the people in question want the tests to go forward? Is the imposition of our ethical policies where they’re not wanted a kind of ethical imperialism? This is an example of a genuine moral disagreement: a case in which people judge differently about what is right to do, even when they agree about all the ordinary particulars and facts. I am interested here in what is sometimes called “real” as opposed to “apparent” disagreement, and in that which is basic and not simply contextual. Apparent moral disagreements arise not because we disagree about the values and morals in a case, but because we disagree about the non-moral facts. For example, there are people who believe that it is dangerous to give children raw milk, and so one ought not give it to them. Others say this isn’t so, that giving raw milk to children is perfectly safe and even healthy, and thus one ought to give it to them. Though the feelings on each side are strong, and this is surely a disagreement about what is right to do, it is disagreement about the facts, not about what values and moral principles are the right ones. Everyone here agrees on the relevant moral principle: we ought to protect and foster children’s health. The disagreement is over how best to do that. The medical testing disagreement is not like this: even when there is agreement about all the facts, there is disagreement about what to do. Sometimes moral diversity is explained away by saying that what looks like a disagreement isn’t really, it’s just a contextual application of a principle. For example, suppose we allow a general principle in favour of ensuring that everyone has enough to eat. In times of abundance this might be achieved without much redistribution of wealth, but in times of scarcity it might be achieved only through resources being directly given to the less well off. We may agree on the general

Basic Concepts

15

principle while allowing that different conditions and different nonmoral facts require different actions to honour the principle. These kinds of disagreements are “contextual” and not “basic”: there is underlying agreement about what to do in each case; it’s just that the cases considered differ.5 The medical testing controversy is not merely contextual: there is disagreement not because the right action varies by context but because there is disagreement about what to do in particular and specific cases. There are examples of real and basic disagreement in many people’s everyday lives. Imagine the following kind of case, having to do with our duties in family contexts. Marie is a discreet aunt, the sometime confidante of her seventeen-year-old niece. The niece’s father – Marie’s brother – is a very strict parent. The niece approaches Marie for a confidential talk about her romantic life on the condition that Marie promise not to tell the girl’s father anything about the matter. During the talk, she reveals that she is dating a young man Marie is sure the girl’s father would disapprove of. Later, the father asks Marie some uncomfortable questions: “Do you know anything about my daughter’s romantic life? Is she dating anyone? Is there anything you think I would want to know?” Marie judges it appropriate to lie. After all, a promise is a promise, and if she were even to admit having information or say she doesn’t want to discuss it, the father would know something was up. So she lies. When the father learns the truth, he is enraged. He judges that a lie is a lie, and that Marie has wronged him through her dishonesty. This, too, is a case of real and basic moral diversity: Marie believes that she was right to lie; the father believes that Marie ought to have told the truth; both agree on the non-moral facts. This is not merely a case in which context determines that a principle should be applied differently: Marie and her brother disagree about this particular case.

2 diversity and its roots in two forms of pluralism J.L. Mackie famously thought that the best explanation of moral diversity was that moral beliefs just aren’t true or factual at all. Moral judgments, he said, reflect projections of our own attitudes onto the world, and the result of this process is a set of moral beliefs that we mistakenly take to be objective. It is more likely, he said, that a person

16

Moral Reasoning in a Pluralistic World

believes monogamy is good because she lives a monogamous way of life than that she lives a monogamous way of life because monogamy really is good.6 But recent empirical research suggests a more subtle and nuanced explanation based on the fact that the way people value is “pluralistic” in two senses. First, we have a number of common but distinct moral cares, such as benevolence, fairness, honesty, and loyalty. Second, people direct and interpret those cares differently and prioritize among them in different ways. How we direct and prioritize among those values creates moral disagreements.7 The first kind of pluralism is value pluralism. This is the idea that there are various distinct values, and that these are not reducible to a single overarching value or a super-value.8 Psychological, anthropological, and sociological research suggests that for people around the world there are a number of common but distinct moral cares; that is, we generally care about various things, in various ways, to varying degrees.9 We value respect for individual persons and the collective good; we respect fidelity and honesty; we care about liberty, equality, and justice. These not only conflict in particular circumstances but also seem to represent different kinds of goods and resist expression in terms of a single unifying value. The second form of pluralism concerns variation in the way we direct and prioritize among our values. We share many values. But while we share values, we care about them to varying degrees and direct and interpret them in various ways. For example: with respect to prioritization, one group of people may prize justice and fairness above all – with less regard for the collective good. Others may prioritize differently, allowing that in extreme or even moderate cases, the collective good is most important. Americans are famous for valuing and prioritizing autonomy, even when the demands of autonomy seem to conflict with other values such as beneficence. Let me call the way a person values what she does her “value field,” and this second form of pluralism “value-field variation.” In the context of these two forms of pluralism, the explanation for some moral diversity is that we value both pluralistically, in holding various values, and with varying value fields, in instantiating those values in different ways and to different degrees. As Steven Pinker says in discussing some relevant empirical research, how and the extent to

Basic Concepts

17

which people moralize some things and not others creates moral differences around the world.10 For example, it is plausible in Marie’s case that both she and her brother care about both honesty and fidelity. They differ in how they prioritize: Marie judges fidelity in the form of promise keeping most important, and her brother judges honesty most important. This difference in prioritization leads them to divergent judgments about what is best to do, and thus moral diversity arises out of this difference. There are several values concerned in the controversy over international medical testing: a value of respect for individual persons means we ought not coerce, exploit, or use test subjects; a concern for justice means we ought to treat subjects fairly in all times and places; an interest in bringing about the collective good and preventing collective harm means we ought to develop and test as many of the relevant treatments as possible so more people will live healthy lives; an appreciation for liberty means allowing people to do as they please as long as they are not harming anyone. Again, those who disagree over what to do can still plausibly share these values. In some cases, disagreement may be not over what is valuable but in how those values should be prioritized or weighed against one another in particular cases. For example, one question might be whether the good that could come of the studies in question could be great enough to override any injustice of allowing differential treatment. This might depend on factors such as how bad the impact of disease is. Those who value respect for persons or justice above all may say that one may never allow tests with double standards; those who value the collective good, or preventing harm, might judge them to be sometimes appropriate.11 Some disagreements, such as Marie and her brother’s, rest primarily on differences in prioritization. Other more complex cases, though, go beyond differences in prioritization and involve either values that are not shared or values that are shared but are interpreted and “directed” in different ways. For example, justice and fairness can be interpreted in many ways. Do they entail equality, as when we say that equal work deserves equal pay? Or do they entail mere equity, as when we say that richer persons ought to contribute more to taxes than poorer persons? In the medical testing case, taking justice to be equity might support a double standard but taking justice to be equality would not.12

18

Moral Reasoning in a Pluralistic World

Another example of “directing” values differently arises in sexual ethics. One way to think about sexuality is in terms of appropriateness, or as it used to be called, “chastity.” Though we now think of chastity as having to do exclusively with virginity, in its broader sense it means simply sex where and when it is appropriate. In one traditional interpretation, this would be between heterosexually married couples and not on any other occasions. A very different way of thinking about sexuality has to do with autonomy: what determines the morality of a sexual act is whether the participants are consenting to it of their own free will, with no coercion, force, or deception. Obviously, people valuing in these different ways will disagree about the morality of particular sexual acts. Evaluations of sex through the lens of chastity rather than autonomy form one part of the structure that grounded the belief that a man could not rape his wife: since the wrong of rape could only be a wrong of inappropriate sex and chastity violated, and since sex between spouses is generally appropriate, there is no violation in marital rape. Evaluations of sex through the lens of autonomy entail that forced sex is always wrong, in any context.

3 value pluralism, descriptive and normative To say that values are plural is to say that there are various genuinely distinct values that are not reducible to one another or to some supervalue.13 Though the question of appropriate reducibility is complex, and I say more about it in chapter 2, there is evidence that the way people value in practice is pluralistic in the sense of appealing to multiple values.14 In explaining some of this research Jonathan Haidt and Craig Joseph argue that for people around the world there are roughly five common but distinct moral cares that we share to greater and lesser degrees, and which form the primary colours from which more complex moral views are formed. These five are harm/care (helping others and not hurting them); fairness/reciprocity (treating others with justice, not cheating them, honesty); in-group/loyalty (commitment to protecting one’s community); authority/respect (respect for, and obedience to, those in positions of authority); and purity/sanctity (cleanliness, chastity, temperance in desires).15 The descriptive thesis that this is how we value in practice is, of course, distinct from the normative thesis that value pluralism is in some sense correct, that we are right to value pluralistically. Though

Basic Concepts

19

my arguments and conclusions here concern the normative question of how we ought to value in practice, they take as a starting point the descriptive thesis. I take the descriptive thesis to be supported by the simple fact that in practice, most moral reasoners appeal to multiple kinds of moral considerations in thinking about what matters and why, and weigh considerations against one another rather than seeking out a unifying value. For some of the discussion in this book, I will be assuming the normative thesis, to make the exposition clearer and easier to follow; but my conclusions rest ultimately only on the weaker, descriptive claim. Though my conclusions do not ultimately depend on it, note that the normative thesis gets intuitive support from considerations like those canvassed by Thomas Nagel.16 Nagel’s “Fragmentation of Value” argues that value has different irreducible sources, or as he puts it, “not all values represent the pursuit of some single good in a variety of settings.” Different values, he says, give rise to various types of considerations: we have “specific obligations to other people or institutions”; there are “constraints” on the way we behave “deriving from general rights everyone has”; we care about the welfare of others, and want to help and not harm them; we have “perfectionist ends or values,” as when we value a work of art or a scientific discovery for its intrinsic worth; and finally, there are considerations grounded in commitment to our own “projects and undertakings,” which – Nagel explains – “is a value in addition to whatever might have led to [those projects and undertakings] in the first place.” Nagel points out that our intuition that values are plural (or, as he would say, that value is “fragmented”) is supported not only by what is commonly seen as the failure, so far, of attempts to find a single source of value, but also by the nature of the types of considerations themselves. For one thing, he says, these types have important formal differences: considerations of the welfare of others take into account “the number of people whose interests are affected,” while perfectionist values do not.17 Specific obligations to other people arise from some particular interaction we’ve had with them, and are typically felt as binding for personal reasons; when we feel we ought to keep a promise, the feeling and motive are different from those we experience when reflecting that it would be a good thing, impersonally considered, if people generally kept their promises.18 Some of these differences arise because as human beings, we are able to take on a

20

Moral Reasoning in a Pluralistic World

variety of perspectives on our lives. And there is no reason to give up on this variety. As Nagel says, “the capacity to view the world simultaneously from the point of view of one’s relations to others, from the point of view of one’s life extended through time, from the point of view of everyone at once, and finally from the detached viewpoint often described as the view sub specie aeternitatis is one of the marks of humanity. This complex capacity is an obstacle to simplification.”19 I take it as uncontroversial that the way we value is, descriptively, pluralistic in this simple sense. Whether this is the way we ought to value is controversial. In chapter 2 I consider the possibility that values are only superficially plural, and are best understood as derived from a unified source.

4 moral coherence As I’ve mentioned, I focus here on methods that seek to bring coherence to sets of beliefs that contain general principles and particular judgments. As I use the term here, to be sufficiently “general” to count as a “general principle,” a principle must be formulated with no proper names or indexicals such as “I” or “you.”20 In the example of the discreet aunt, Marie must decide whether to keep her promise to her niece (to keep a secret) or tell the truth to her brother, breaking her promise. The belief that Marie ought to keep her promise is thus a particular judgment because it has a proper name, as is the belief you may sometimes have that says, “I ought to call my mother,” since “I” is indexical. The belief that one ought to keep one’s promises to family members is a general principle, as is the belief that one ought to keep one’s promises always.21 At the heart of coherence reasoning is a demand that principles and judgments should fit together in some appropriate way. For example, suppose we start with judgments that some lies are wrong. Then we may seek a general principle that explains why it is so. This leads us to consider the general principle that it is wrong to lie. This, in turn, suggests that other kinds of lies are wrong also, and we might consider the principle “It is wrong to lie.” If we judge that there are some situations in which lying is not wrong after all – perhaps a case like Marie’s, or an even more obvious one in which a small lie could save a life – we try to find a suitable explanation: by finding a more general principle that isn’t contradicted by the case, by changing the principle to encompass

Basic Concepts

21

the case, or by finding a new principle that explains the judgment. I consider coherence methods as they arise within the framework Geoffrey Sayre-McCord calls “conviction ethics.”22 This approach takes our considered moral judgments as the proper starting points for moral reflection. In contrast to some kinds of rationalist, theological, or empirical approaches, conviction ethics understands such judgments as having some prima facie plausibility; the task of moral theorizing is to use whatever tools we have to improve those judgments. The aim, then, is to develop a moral theory that is coherent and that is also “intuitively acceptable” – i.e., fits well with our considered convictions.23 There are many ways of understanding coherence. On pretty much any coherence view, we require that the principles and judgments fit together with one another in the simple sense that each judgment must follow from some principle, and principles must apply to judgments in a way that is appropriately general. Obviously it is incoherent to say that lying is generally wrong but that Marie is right to lie without any further thought about how the judgment and the principle fit. The debate concerns what coherence also requires with respect to how the principles themselves relate to one another or fit together. Should we aim to show how the prohibition against lying and the obligation for promise keeping are somehow related? One minimal kind of constraint on the principles is consistency. I take it that logical consistency is required and I don’t discuss this further here. Beyond logic, however, several kinds of consistency are of interest. What I call “practical consistency” concerns the matter of whether principles issue in conflicting oughts; conflicting oughts arise when principles recommend each of two actions and one cannot do both. For example, as Marie’s case shows the principles enjoining honesty and promise keeping are – on the face of it – practically inconsistent, since they require her to lie and to keep her promise and she cannot do both. For specificity, let me say that a moral theory is “practically consistent” when its principles never issue in conflicting obligations: its principles do not conflict, even in principle, and there is no possible world in which such conflicts would arise. When I say “a possible world,” I do not mean a world the person can necessarily bring about; I mean a way things could be so that the principles would not engender conflict.24 Even if cases like Marie’s happened not to arise, the principles in question would be practically inconsistent, since the relevant kind of conflict is clearly possible.

22

Moral Reasoning in a Pluralistic World

Practical inconsistency is distinct from what I call “essential inconsistency” – which has to do with whether conflicts arise in every possible world.25 Principles that are practically inconsistent but not essentially inconsistent, though they lead to conflict in our world, do not essentially conflict with one another, because there is a possible world in which they do not conflict. To be essentially inconsistent, principles have to conflict with one another in a deeper way. The principles in favour of telling the truth and keeping promises are practically inconsistent, but they are not essentially inconsistent: it may be that circumstances like Marie’s, in which they conflict, just never arise. But the principles that one ought to treat all children equally and that one ought to provide the most care and resources for one’s own children are essentially inconsistent, because there is no possible world in which they can all be obeyed. Principles are practically inconsistent when there is a possible world in which they engender conflict; they are essentially inconsistent when they engender conflict in every possible world.26 Given these definitions, principles that are essentially inconsistent are necessarily also practically inconsistent, and principles that are practically consistent are necessarily also essentially consistent. Typical conceptions of coherence offer various desiderata in addition to consistency, such as unity, simplicity, connectedness, comprehensiveness, and generality.27 Unity is the demand that the principles of a theory be as few in number as possible. For instance, a single principle that tells us only to maximize preference satisfaction is more unified than the set of two principles “Don’t lie” and “Keep your promises.” Simplicity is the demand that the principles be as simple as possible. A principle against lying is simpler than a principle against lying except when one has to keep a promise. Connectedness requires us to prefer systems in which principles are related to one another. The principles “Don’t lie” and “Don’t lie on your taxes” are connected, since the second follows from the first. The principles “Don’t lie” and “Keep promises” are not, on the face of it, connected, though they would be if each followed from a single utilitarian principle in favour of promoting good consequences. Comprehensiveness means preferring principles that, taken together, cover a wide range of cases. Generality means applying a single criterion to a large range of cases. For example, the principle against lying to one’s siblings is less general than the principle against lying to one’s family members, which is in turn less general than the principle against lying.

Basic Concepts

23

I use the term “rich” to describe those conceptions of coherence that add to consistency elements of unity, simplicity, and connectedness. Though comprehensiveness is sometimes mentioned as part of coherence in a moral theory, it is a desideratum of a different kind: a set of disconnected judgments would be comprehensive if there were enough of them, even if they were not coherent in the least. And generality raises more complex issues than other elements, as I discuss in chapter 5. In this book, I consider rich coherence through a norm of systematicity, which I take to be the demand that the general principles of a theory be as few and as simple as possible. Systematicity is an expression of the idea behind rich coherence, because it brings together the idea of unity, simplicity, and connectedness. The more systematic a set of principles is in this sense, the more unified and simple it will be, in the sense of showing how our various moral judgments can be seen as following from few fundamental moral principles. The more systematic a set of principles is in this sense, the more connected it will be, in the sense that sets of principles and judgments that are all inferentially connected to one another will be preferable to those that are patchy – i.e. having parts that are internally connected but not connected to one another. In endorsing a rich form of coherence, David Brink writes that “insofar as a moral theory explains the connections among moral considerations and arranges them in a systematic fashion, as unified theories do, it makes our beliefs more coherent and better justified.”28 In a similar vein, Shelly Kagan writes: “To reduce the many to the few is not yet to support the few. Unless we can offer a coherent explanation of our principles (or show they need no further justification), we cannot consider them justified, and we may have reason to reject them.”29 And Brad Hooker says that, compared to a pluralist theory, a unified theory would contain “information that moral pluralism does not,” “would contain a deep connectedness omitted by moral pluralism,” and “would be able to explain more on the basis of fewer assumptions.”30 As I understand systematicity here, to say that the general principles of a theory should be “as few and as simple as possible” is to say two things. First, it is to say that we should aim to reduce the number of general principles of our theory at the same time that we aim for a good fit with our considered convictions. Let me call such fit “intuitive acceptability.” Then systematicity means sometimes making

24

Moral Reasoning in a Pluralistic World

trade-offs between the goal of reducing that number and the goal of intuitive acceptability.31 Second, it is to say that all else being equal, we should prefer more systematized theories to those that are less so, so that in deciding between two theories that fit equally well with our considered convictions, we ought to prefer the more systematic one. Philosophers are often attracted to richly coherent, systematic theories for their simplicity and seeming explanatory power. Consequentialist theories (such as utilitarianism) have a single principle telling us to maximize pleasure, happiness, or preference satisfaction overall: the right action is simply the one that brings about the best consequences. Applied to Marie’s case, the unified theory of consequentialism would determine whether Marie ought to lie by considering the overall harm caused by lying and by telling the truth. Consequentialism analyzes questions like those of the medical trials by considering the effects – both short-term and long-term – of allowing or not allowing tests like the one in question to go forward. The relevant question would not be whether one person or group is mistreating another in a given interaction, but rather the overall costs and benefits of the proposal as compared with other proposals.32 A different unified view is organized around the single principle of respect for persons, or their autonomous nature. In one Kantian formulation, we look not at the consequences of our action, but at whether we are treating persons always as ends-in-themselves and never as mere means. That is, an action is right if it shows proper respect for individuals, and does not manipulate, coerce, use, or deceive them. In this approach, one question would be whether subjects are being coerced by their circumstances into participating, and another would be whether responding to their difficult life circumstances with offers of drug trials rather than simple offers of assistance is to show immoral indifference to them as people.33 A third unified approach is a libertarian one that favours individual liberties understood through the notion of “full self-ownership.” In some interpretations, such a view leads to the conclusion that we have no duties to help others unless we have contracted to do so voluntarily, or have forfeited our rights through wrongdoing. It follows that we ought to be able to do as we please as long as we are not harming others and thus violating their self-ownership. This, in turn, entails that in the medical testing case, what matters is only whether the subjects have consented; any “double standard” is not morally relevant.

Basic Concepts

25

Note finally that coherence has what Klemens Kappel calls a “dialectical role.”34 This means that, other things being equal, saying that one theory is more coherent than alternatives is a way of recommending it.

5 value pluralism, principle pluralism, conflict, and dilemmas Theories with multiple general principles are called “principle pluralist”; those with a single principle, “monist.” For example, the pluralist theory of Ross has principles of fidelity, reparation, gratitude, justice, beneficence, self-improvement, and not inflicting harm. The question of when principles are appropriately irreducible raises complexities: if we endorse each of several principles because they follow from a single other, this would be pluralistic in name only. So let me say that a “derived” moral principle is one that follows from other more basic or fundamental moral principles together with relevant non-normative facts. For example, the principle that enjoins us to breastfeed infants follows from the moral principle that one ought to nurture one’s children, together with the non-moral fact that breast milk is healthy food for babies. I’ll call non-derived principles “fundamental.” Genuine principle pluralism, then, involves there being multiple fundamental moral principles. I claim that the normative thesis of value pluralism gives us prima facie reason to endorse a set of principles that is pluralistic in the sense of having multiple fundamental principles. If values are plural in the sense described here – if the normative thesis of value pluralism is true – this has implications for how we think about sets of principles and judgments in the following ways: (1) if values are plural, then on the face of it we’ll need multiple principles; (2) such multiplicity is indeed necessary, because the reasons we give for the moral judgments and the principles we endorse will not ultimately be reducible to a single kind of consideration: to say that beneficence and justice are fundamentally different is to say that they resist expression in terms of a single value. If values are plural, they are plural all the way down;35 (3) when various considerations apply in a given case, the principles we endorse will typically conflict; if values are plural, the proper activity in such a case is weighing considerations against one another rather than seeking out a more general point of view from which our principles can be

26

Moral Reasoning in a Pluralistic World

seen to follow. That is, we have to prioritize our cares, rather than unifying them. We’ve seen some examples of conflict and of weighing moral considerations. Marie faces a conflict, and she weighs considerations of honesty against those of fidelity; the participants in the debates over medical testing face conflicts and weigh respect for autonomy, justice, and the value of promoting the collective good. Parents often confront seeming conflicts, since their obligations to their children can conflict with other kinds of obligations. Imagine a father, Anthony, who borrows money from his neighbour, fully justified in his expectation that he will be able to pay it back promptly. Before he can do so his child is diagnosed with a serious form of cancer, for which he must be treated at a cancer centre in a faraway city. Even if the treatment itself will be paid for by insurance, we can imagine that the father owes it to his son to be near him at this time, and that this will require a hotel stay in an expensive city and thus prevent him from paying back what he owes. The father is in a conflict: he ought to be with his son; he ought to pay back what he owes; and he cannot do both. Anthony weighs his obligation to pay back what he’s borrowed against that of being near his son during his son’s illness. If values are plural, then (1) our caring about honesty, fidelity, respect for persons, justice, the collective good, autonomy, and harm prevention result in various principles such as “One must tell the truth,” “Everyone ought to keep their promises,” “Do no harm,” “Always pay back what you’ve borrowed,” and “Everyone should take care of and protect their children”; (2) the reasons supporting the judgments about what to do cannot be understood as expressions of a single value, but rather represent fundamentally different values such as respect for autonomy, justice, honesty, fidelity, etc.; (3) the proper activity in such examples is one of weighing different kinds of considerations against one another, rather than seeking out a more general point of view from which the considerations can be unified. So if values are irreducibly plural, principles are also. How to deal with the possibility of such conflicts is one of the more controversial and difficult questions in moral theorizing. It is often thought that moral conflict ought to be explained away in one of two ways. Ross, for example, says that principles generate only “prima facie” duties, and that in cases of conflict only the strongest becomes one’s actual duty. So, for example, if the principle in favour of truth

Basic Concepts

27

telling generates a duty to tell the truth, and the principle in favour of promise keeping generates a duty to lie, these two conflicting duties are just prima facie duties – by which Ross means not that they are duties “on the face of it,” but rather that they are duties only insofar as they do not conflict with others. If and when duties do conflict, whichever is the strongest is one’s actual duty, and the other simply disappears. A slightly more accurate way of expressing this idea is in terms of “pro tanto” duties and “overall” duties. Principles generate duties that are pro tanto - they are duties insofar as they are not overruled by other duties. The duty that is most pressing, that one ought to act on, is one’s overall duty.36 A second way of explaining away conflict involves articulating the right “exception clauses” for our principles. For example, we may say that true honesty doesn’t require always telling the truth; rather, true honesty requires telling the truth only when doing so does not entail breaking promises, or when we must tell a lie to save a life. For example, instead of “It is wrong to lie” our principle would become “It is wrong to lie unless you must in order to keep a promise, or to save a life, or ... ” Instead of “repay what you have borrowed” we would have a principle such as “repay what you’ve borrowed, unless you need it to care for your children.” This is called “specification,” since it suggests specifying the principles more precisely.37 If the conflict is not explained away by pro tanto or specification, it is a dilemma: a case in which one ought to do each of two things and cannot do both. The question of the reality of dilemmas is controversial for reasons that go beyond considerations of coherence, and I discuss these in chapter 5. For purposes of the discussion here, let me say that principles are (practically, essentially) consistent when they engender no (contingent, necessary) conflicts, either because of the way they are structured or because the conflicts are explained away by one of the strategies of specification or pro tanto. If the conflict is not explained away it is a dilemma: a case in which one ought to do each of two things and cannot do both.

6 coherence, conflict, and dilemmas I end this chapter with arguments to three conclusions. First, we need not include practical or essential consistency in our coherence norms. Second, there are some reasons to retain conflicts in a set of princi-

28

Moral Reasoning in a Pluralistic World

ples. Finally, the need to prioritize among conflicting norms does not support systematicity. It is sometimes suggested – for example by Sayre-McCord – that an important norm in coherence reasoning is associated with avoiding moral conflict, that is, with finding principles that do not issue in conflicting obligations such as those Marie faces. My first claim is that this isn’t so. In his early work on coherence, Sayre-McCord takes practical consistency as I do here: its demands are met “if and only if the theory is structured in such a way that it will not, even in principle, generate conflicting directives.”38 Others, too, think the avoidance of conflict important for consistency. Sidgwick, for example, suggests that the existence of conflict means the principles aren’t properly formulated.39 But why should we think that avoiding conflicting directives is so important for moral coherence? There is one simple answer that can immediately be set aside, which is the idea that just as sentences that generate contradictions cannot all be true, principles that generate conflict cannot all be right.40 This sees the difficulty of conflict as arising directly from plurality. But as has been pointed out, this simple view cannot be correct, since single principles such as “keep promises” can themselves generate conflicting demands. This can happen if, say, a person promises to give ten dollars to one friend and ten dollars to another friend and then has her wallet stolen, leaving her with ten dollars in total.41 A more sophisticated approach presents the difficulty with conflict as a problem about justification. Sayre-McCord focuses on Sidgwick’s idea that principles that generate conflicts will require modification, and might not seem as simple – and thus attractive – in their modified form as they did before. The presence of conflict shows we have mistaken for an “ultimate and independent axiom” one that “is really derivative and subordinate.”42 In this approach, the problem with the promise-keeping principle is that it is not sufficiently general to resolve all conflicts. But there are two difficulties with this line of thought in the present context. The first is that we needn’t modify principles to avoid moral conflict. The suggestion seems to assume that there is only one way to make a theory action-guiding, namely, to reformulate the principles with exception clauses specifying when other obligations take precedence – that is, we need to specify. But this not the only way to avoid moral conflict. We can also, in the style of Ross, say that princi-

Basic Concepts

29

ples are simply pro tanto. This Rossian strategy would allow us to avoid conflicts while keeping our principles general and unspecified. The second, more serious, difficulty is that this justification assumes a rich coherence model, which would be question begging here because the proper interpretation of coherence is what we are investigating. In the rich coherence model, we assume that simpler and more general principles are better; this would show clearly why modifying principles to avoid conflict would make them less attractive. But in the absence of such an assumption, why should we believe that the generation of conflicts makes principles fail to be “ultimate” in their justification? Again, the single principle enjoining promise keeping generates conflicts; in the absence of rich coherence norms there is no reason to think there is anything not fully justified or ultimate in this principle. A third way of thinking about practical consistency ties it to the need for prioritization and action-guidingness. In order for our moral theory to be action guiding, we might say, we need a way of determining – in cases of putative conflict – what our most important or overall obligation is. And this requires us to resolve all the conflicts engendered by our principles. Just so, if by “resolve” we mean merely “determine what to do overall.” But this does not require practical consistency. I discuss prioritization and its connection to justification further in chapters 4 and 5. But here I claim that prioritization is independent of practical consistency. A theory can be fully prioritized without being practically consistent. For suppose we have a set of principles and judgments together with a way of saying, in every case, what one’s overall obligation is. The principles can be made practically consistent in either of two ways already mentioned: by reformulating the principles with exception clauses (specification) or by making the obligations in question pro tanto only. But prioritization does not require that we do either of these things. Having a way to determine one’s overall obligation is consistent with saying that both obligations are real and that both are in force when one acts. In a case in which one ought to do A and ought to do B but cannot do both, it is possible to say that one ought to do A overall, while still saying that both A and B are genuine obligations. When it comes to lying to keep promises, or just generally lying to protect others, a way of prioritizing among honesty, fidelity, and benevolence will issue in a range of judgments about what is best to do overall. A way of prioritizing among our various obligations, and

30

Moral Reasoning in a Pluralistic World

saying what one ought to do overall in every case, does not require us to formulate these principles so that they do not conflict or to say that only our overall obligation is real or actual. Indeed, for any set of moral principles and judgments that avoids conflict, (e.g. through specification or pro tanto), there is a set of principles and judgments that has identical judgments about what to do overall, and differs only in whether or not conflicts are allowed. This is so whether the resolution of conflicts is principled or not. My second claim is that there are reasons to retain conflict in a set of principles. If principles come with exception clauses, or if our only obligation is our most pressing one, then there is no real dilemma in difficult cases – there is just one thing we ought to do overall. But dilemmas play starring roles in our moral phenomenology. People who seldom consciously reflect on morality are most likely to do so when they’ve encountered what they feel is a dilemma, because this is a case in which one must figure out what to do.43 It is common after experiencing a moral dilemma to feel a sense of regret or moral distress, that all is not as it should be. If Marie honours her promise, and judges herself right to do so, she may nonetheless feel bad about having lied to her brother. If she were to lie to her brother, and judge she was right to do so, she may nonetheless feel bad about having broken her promise. If Anthony is a morally sensitive person, he will likely feel bad no matter what he does, either for not paying back what he’s borrowed or for not being with his son. Generally, we tend to think that people who have these feelings of regret in the context of dilemmas are morally sensitive – that this is an admirable way to feel. The moral phenomenon associated with this feeling of regret is called a “remainder.” A remainder is a reflection of the fact that even if one has done one’s best overall there is something one ought to have done that one didn’t do, and this is regrettable. The phenomenon arises even in the case in which one obligation clearly overrides the other and even when it is clear to the person herself what she ought to do overall. It is easy to imagine particulars that would have made Marie’s choice straightforward: if she knew that her brother might be violent or abusive with his daughter, then keeping her promise would seem obviously the right thing to do. Likewise, if Marie knew that keeping the secret would endanger her niece then telling the truth would seem obviously the right thing to do. Still, in these cases it would be natural for Marie to feel that there was some-

Basic Concepts

31

thing morally regrettable in the situation, that things would have been better if she hadn’t had to make the choice in the first place. The remainder explains this feeling. The significance of the remainder and its relation to moral dilemmas raise complex issues which I discuss in chapter 5; there I defend the idea that the remainder plays a role in our moral theorizing and practice. But notice here that, as Ruth Marcus and Bernard Williams both emphasize, the remainder fits the emotional responses we typically have when we experience a moral conflict, as it does in Marie’s and Anthony’s cases.44 The remainder explains why a feeling of regret arises, and why the feeling seems to have a moral component: this reflects the sense that there was something not only sad but also morally regrettable in one’s having acted as one did. If we were to explain away conflicts with exception clauses or pro tanto obligations, we would lose an explanation of the remainder’s existence. In acting for the best overall, Marie and Anthony would have fulfilled all of their obligations; it would then be irrational for them to feel any moral distress.45 But we typically think such emotions not only appropriate but also morally admirable. This means we have some reason to retain moral conflicts: their existence fits our moral experience. My third claim is that the need to resolve conflicts in the sense of prioritizing does not lead to systematicity. Sayre-McCord draws an indirect connection between the two in the following way.46 Suppose a given set of multiple principles generates conflicts. Consistency then requires us to say, in cases of conflict, which obligation is our “actual” or – as I prefer to say – “overall” obligation. This, Sayre-McCord says, must be done either through a justifying priority rule – one that tells us how to prioritize some things over others, and explains why we ought to do so, or by an unjustifying priority rule – one that tells us how to prioritize but is, at bottom, arbitrary. In the first case, he says, we effectively have a “systematized” theory, since the priority rule functions as a single first principle. In the second, the theory fails to be truly coherent because it fails to show how the various prioritized judgments are connected to one another.47 But there are difficulties with applying these conclusions in the present context. First, we might have various methods for determining what to do overall, and these need not take the form of a rule. In the case of a justifying method, if the method does not take the form of a rule, then it is not a way of systematizing. Giving a justification for a

32

Moral Reasoning in a Pluralistic World

prioritization is different from showing how several principles follow from a single one. Consider, for example, John Rawls’s famous justification of his suggested prioritization of his principles of justice. This asks us to imagine what we would choose from behind a veil of ignorance, given certain assumptions, and uses the results to argue for the legitimacy of the prioritization he offers.48 This is intended as a justifying method for determining priority, but it is not a way of rendering the principles more systematic since it does not end in a single principle but rather in a lexical ordering of two.49 Furthermore, in the context we are considering – in which the question is what kind of coherence is appropriate – there is nothing to rule out adopting an unjustifying method. Suppose we flip a coin in any case of conflict and say that determines the “overall” obligation. In doing so, Sayre-McCord says, we violate connectedness and thus do not have a fully coherent theory. Connectedness requires that each principle must be connected by a path of justification to all other principles of the theory; unjustifying priority methods would allow that some principles are unconnected by any path of justification to any other.50 This may well be convincing if we are already in the context of rich coherence. But here this response would be question begging, since the status of rich coherence – and of connectedness as part of it – is what we are questioning. If values are plural, it is possible that our moral theory ought to have multiple principles that are unconnected to one another by any path of justification. It might be thought that the difficulty arises not with “practical” inconsistency but with “essential inconsistency.” Recall that essentially inconsistent principles generate conflict in every world – a set of principles is essentially inconsistent when it cannot help but generate conflicts, no matter how things are arranged. Given this distinction, we might wonder if the difficulty with consistency is not so much with practical but rather essential inconsistency. But everything I have said so far about practical inconsistency applies equally well to essential inconsistency. In particular, one can prioritize among essentially conflicting principles and duties just as well as among those that contingently conflict. An essentially inconsistent set of principles might require that we provide equal resources for all children, and also that we provide more resources for our own children than for others. As odd as such a set of requirements might seem in other ways, there is nothing standing in the way of prioritizing among these concerns, and

Basic Concepts

33

either might be most important. Such prioritization does not require eliminating the conflict. And thus nothing about prioritization per se requires us to prefer principles that are essentially consistent. It may be objected that the distinction I’m drawing between justifying a prioritization and pursuing systematicity shows that my focus on systematicity misrepresents the ongoing dialectic, since the interesting question in this domain is how to prioritize. But as I explained above, it is standard to include elements such as “unity” and “connectedness” in accounts of the nature of (rich) coherence, and systematicity is an apt expression of these. And even committed principle pluralists such as Beauchamp and Childress suggest that the ideal coherence would involve systematicity if only it could be achieved. This shows that systematicity itself – and not just prioritization – is considered an important norm. So the question of whether or not to systematize is a live and meaningful one.

34

Moral Reasoning in a Pluralistic World

2 Value Pluralism, the Unity of Value, and the “Delimiting Argument”

In chapter 1 I characterized value pluralism as the idea that there are a multiplicity of values not reducible to one another or to some single super-value. I then gave some reasons for thinking that the way we value is pluralistic. It is pluralistic in this sense: we endorse various values – harm prevention, fairness, in-group loyalty, etc. – and in deciding what to do we typically weigh these against one another, rather than seeking out a perspective from which they can be seen as manifestations of a single more abstract value. But are the values we endorse genuinely distinct? Or is plurality merely a matter of superficial appearance? Against the belief that our values are genuinely plural, it is sometimes thought that we have reason to believe in or seek out a unified view of value. This is thought to be so either because value monism – the view that there is one ultimate value – is a more plausible theory overall than value pluralism, or because a person’s rationality and well-being require him to endorse values that fit together somehow, thus rendering his values only nominally pluralistic. The “unity of value,” or so this thinking goes, emerges as correct, or desirable, or rationally required. In this chapter, I consider some arguments along these lines. Along the way I draw out some thoughts about what kinds of prioritizations, compromises, and commitments make sense when it comes to valuing generally.

1 value pluralism and value monism The well-known pluralist Isaiah Berlin famously argued that some values, such as liberty and equality, were in conflict, and were even in

Pluralism, Unity of Value, and “Delimiting Argument”

35

some sense incompatible, involving “collisions of value”1 and “perpetual rivalry.”2 In defence of this view he pointed out that some liberties will lead inevitably to inequalities: liberty for the strong is fatal for the weak. But Berlin’s views have always been open to multiple interpretations, and these metaphors leave open the precise sense in which values are thought to fail to fit together. Do values conflict just in the loose sense of sometimes recommending different actions? Are they incompatible in the sense that one cannot honour them both at the same time? Are they opposed, so that honouring one means violating the other? Are they incomparable, or incommensurable, so that there is no way to rank them or prioritize among them?3 In response to Berlin, critics point out that just because equality sometimes requires us to do things we’d prefer not to does not mean that equality and liberty are deeply incompatible, or incomparable, or even conflicting in any deep sense. We might, for example, simply value liberty only insofar as it does not conflict with the demands of equality: what is fatal to the weak is not allowed by the value of liberty, properly understood. Or we might see both as manifestations of some other value, taking liberty and equality to be good only insofar as they lead to what seems best overall. Indeed, given that we often make justified decisions even in the face of seeming conflict, we might simply say that what is best overall is what is “good” or “just.” This seems a way of unifying value that is compatible with our lived experience of conflict. Pluralists appeal to various kinds of evidence to support their claims that values are somehow deeply or genuinely plural in a significant way. Some emphasize the complexity of moral choices: we often don’t know what is the right thing to do, and we honour different values in different ways.4 Others focus on “discontinuities”: cases in which more of one kind of pleasure does not seem to outweigh less of another kind, or on the possibility that values are in some sense “incommensurable” or “incomparable” – impossible to measure with a single scale, or impossible to compare with one another.5 The most common move in favour of pluralism focuses on rational regret. As pluralists such as Bernard Williams and Michael Stocker have pointed out, what is significant about conflicts of value is that it seems possible, after experiencing one, to experience rational regret for the value that it was not possible to honour. This parallels the arguments in chapter 1 for the existence of the remainder in the case of conflicting obligations. The appropriateness of rational regret seems to fit oddly with the claim

36

Moral Reasoning in a Pluralistic World

that value is unified: why regret the loss of one thing, when what you’ve gained is more of the same?6 Defenders of monism respond that we can make sense of these phenomena without endorsing a pluralistic view of value: monism about value is compatible, they point out, with choices being difficult and with the loss of some good things being necessary for the gain of others. In the case of rational regret, Tom Hurka argues that monists can accommodate the relevant phenomena just as well as pluralists can.7 Imagine a person who cares only about pleasure, who faces the choice between giving five units of pleasure to person A and ten units of pleasure to person B. This person, Hurka says, in choosing to give pleasure to B, may nonetheless rationally regret the pleasure not given to A. The simple example of conflicting promises mentioned in chapter 1 (section 6) supports this view: the person who endorses only the value of promise keeping may find himself in the situation of being unable to keep two of his promises, through no fault of his own. This person may naturally and rationally regret the breaking of one of his promises. What this shows is that the mere existence or plausibility of rational regret itself does not immediately entail a pluralist perspective. This suggests that the role of rational regret in understanding pluralism is complex. In what follows I give my own way of understanding that role, and I argue against the idea that values should be interpreted as deeply unified and only superficially plural.

2 the burden of proof and the “delimiting argument” in favour of value monism Descriptively, the values people typically endorse in practice are not only multiple they also conflict, in the sense that they cannot always be mutually fully instantiated or honoured properly. Evidence for this is found in the fact that people experiencing a conflict of values feel “remainders” like those discussed in chapter 1. Recall the example of Marie’s dilemma of whether to break the promise she made to her niece by telling the truth. A person who values honesty and fidelity in such a case experiences a remainder: a reflection of the fact that their values have not been fully honoured. In Marie’s case, the feeling will be different depending on the choice: choosing to keep her promise,

Pluralism, Unity of Value, and “Delimiting Argument”

37

she is sorry to be misleading her brother about matters essential to his well-being; if she had chosen to break her promise, she would be sorry to have betrayed her niece. This shows that people who prioritize among their values may still value in ways that conflict; value conflict is compatible with values being comparable in this sense. On the face of it, if values conflict in this way, it is natural to describe them as genuinely distinct. One reason for this is that, as in the case of Marie, when we experience a conflict of values, what we feel we have failed to honour differs depending on the choice we have made: honouring her promise, she is not fully honest; if she had been honest, she would not have been faithful to her promise or her friend. This, I take it, puts the burden of proof on the defender of monism, to give explanations of why and how we ought to see our values as having underlying unity. A plausible monist response focuses on what I call here the “delimiting argument.” Even if values do conflict, when we make choices from among those values, we must prioritize. That is, we must say what is more and less important in a given situation. But once we see a way of prioritizing among values we might understand those values as delimiting or circumscribing one another, and as fitting together. Where one leaves off the other begins, like the pieces of a jigsaw puzzle. When we curtail liberties to further equality in appropriate ways, we are not failing to honour the value of liberty: instead, true liberty only extends to where it does not do violence to the value of equality. Then if values do fit together in this sense we might say that they are only “plural” in a non-interesting way. Deep down they are unified, in fitting together. A super-value such as “good” or “just” might then unify them. In his recent book Justice for Hedgehogs, Ronald Dworkin makes a kind of delimiting argument.8 He begins by saying that value concepts are “interpretive” – that is, as we use them, we interpret them, define them, create their boundaries, and so on. Since in cases of conflict we must try to determine which value is the most important and which duty is the most pressing, we are engaged in an ongoing task of figuring out how our values ought to be prioritized. Then – and this is the crucial move – once we’ve seen how they ought to be prioritized, we have reason to interpret them as integrated, as delimiting one another. Do we have reasons to interpret our values as integrated in this way? Beyond a prejudice in favour of unity, what would these reasons be?

38

Moral Reasoning in a Pluralistic World

One answer, which I focus on below, has to do with avoiding inconsistency and ambivalence: a person must be “consistent” to act rationally and autonomously, and this requires a delimiting of overlapping and conflicting values. This sees the need for delimiting in terms of what is rationally required for a valuing person. The argument against pluralism would then be that because rationality requires delimiting in the way a person values, the monist would have a response to the burden of proof argument. I’ll argue against this response, on grounds that rationality does not require delimiting. As I explain in section 7 this does not show that monism is inconsistent with the existence of remainders, but it does show reasons against reinterpreting conflicting values to make them fit together.

3 values and “valuations” To address the question of whether and when delimiting is rationally desirable, first we need a more precise account of what it is for a person to value something. The rationality and autonomy arguments for delimiting have to do, broadly, with the valuing self generally – rather than with anything particularly moral. So for purposes of this discussion I’ll be considering all kinds of values, not just moral ones. Here I take valuing to be an affective state of endorsing or standing behind one’s desire for or attitude toward something. When a person values something, I’ll say that she has a “valuation” for it. There are two ways to make the notion of valuation more precise, and here I’ll have something to say about each of them. Both can be expressed using the idea of “endorsement,” though via different interpretations. The first is to say that a person’s valuations are those desires and attitudes that he identifies with, takes on as his own, and regards as properly belonging to himself. On this view, the important questions concern the relationship between the affective state and the self. “I endorse my desire” in this case means “I stand behind it; it is mine.” A person then values whatever he has an endorsed desire or attitude for in this sense. The second interpretation, focusing on the person’s evaluative stance toward life, takes valuing as a matter of approval; in this view the important questions concern the way in which the person experiences the affective state as somehow positive. “I endorse my desire” then means, “I feel what I want is worth wanting.”

Pluralism, Unity of Value, and “Delimiting Argument”

39

To appreciate the distinction, consider these examples. A person who loves expensive clothes may well identify with her desire to spend a lot of money on such things, in the sense of recognizing the desire as reflecting her self, even if she feels some disapproval toward her desire, in the sense that she thinks she might be a better person if she desired differently. This person identifies in the first sense of self-recognition, but not in the second sense of approval. On the other hand, a person who loves expensive clothes but has been practicing being more frugal may surprise herself with a newfound desire to put money into her savings account; she doesn’t quite recognize this desire as her own, but she certainly approves of it. This person identifies in the second sense of approval, but not in the first sense of self-recognition. Neither of these conceptions is without difficulties, and I do not claim to be offering a novel account of valuing, identification, and approval, but they are both useful for our purposes here.9

4 practical and essential valuational inconsistency The delimiting argument concerns conflicts of values, and since there are two ways that values can conflict there are also two ways of interpreting that argument. Recall from chapter 1 the characterizations of two kinds of consistency. Principles are practically consistent when they engender no conflicting demands, even in principle: there is no possible world in which they require mutually incompatible actions. Principles are essentially consistent when there is some possible world in which they can all be obeyed: even if they create conflicts in ours, as long as there is some way that things could be in which no conflicts would arise they are consistent in this sense. Analogously, I say that the way a person values is essentially consistent if there is a way things could be so that those values could each be fully instantiated, and practically inconsistent – I call it “conflicted” – if those values can be fully instantiated in our world. A set of valuations, then, is essentially inconsistent when there is no possible world in which they can all be fulfilled – they conflict essentially. For instance, if I value honesty in the sense of telling the truth, and fidelity in the sense of keeping promises, these may conflict as they do for Marie. But they needn’t do so. The situation may never arise. So

40

Moral Reasoning in a Pluralistic World

this conflict is contingent, and the person who has it is merely conflicted and not essentially inconsistent. But if I value a life of security and contemplation and also a life of adventure and risk, there is no possible world in which I may have both, and I am essentially inconsistent. If I value working and having a career, I may be merely conflicted, but if I value working and being a full-time stay-at-home mom, I am essentially inconsistent, since there is no way things could be that I could do both. The essentially inconsistent person, we might say, has values that “essentially” or “inherently” conflict. Those concerned about the unity of the valuing self often contrast mere conflictedness with ambivalence: that is, valuing A and not-A, caring about something and its opposite. As a simple example of this sort of ambivalence, consider a person who has a friendly rival who is up for some great honour. This person may want his friend to win, insofar as he is his friend, and he may stand behind this desire as reflecting a friendship that is important to him. This person may also want his friend to lose, insofar as he is his rival, and he may stand behind this desire as reflecting a kind of competitiveness and ambition that he recognizes and approves of as a part of himself. Endorsing his desire that his rival win and his desire that he not win, this person is valuationally ambivalent. Ambivalence is an important kind of inconsistency, but it is a special case of what I am calling essential inconsistency, and the latter notion is broader and philosophically richer. Essential conflict is significant because what matters about the essential consistency of valuations isn’t their logical form but rather the more general question of whether there is any way, in the abstract, for the objects of valuations to obtain together. The question of whether my valuations conflict essentially is the crux of the matter with respect to essential inconsistency, rather than the question of whether my valuations are ambivalent. This is because what is strange or troubling about valuing A and notA generally applies as well to valuing A and B where A and B essentially conflict.10 Consider: (1) in each case, one has no coherent valuational outlook, for there is no way the world could be that would instantiate and fulfil all of one’s valuational requirements; (2) in each case, one is inevitably dissatisfied; (3) in each case, to become satisfied, one would have to change not the world, but one’s own valuational stance. When we talk about valuational inconsistency, the intuitive dis-

Pluralism, Unity of Value, and “Delimiting Argument”

41

tinction we are trying to draw is that between the person who is internally troubled – there is something wrong with his valuations, and the person who is externally troubled – there is something wrong with the world. Being “essentially conflicted” signals the former difficulty. Of course, inessential conflictedness is very common. I may value watching the sunset every day even if I cannot because a job I love requires me to be on duty at that hour. I may value giving money to the local soup kitchen even if I also value saving that money for my child’s future. These values conflict contingently, and are not essentially inconsistent: if I had a job with different hours, or if I had more money, I could be satisfied. That most of us play various life roles and make do with various limited resources means that most of us are practically inconsistent in this sense: there are things we value doing that we would do more if conditions allowed. Essential inconsistency, though perhaps less common, is not unheard of. If I value frugality but also generosity, or value giving more to my children than to others but also treating all children equally, I am essentially inconsistent. If I care about sexual monogamy with Jane, and also endorse, as my own, my desire to have sex with Julia, I am essentially inconsistent, since there is no way I can have both. The person who endorses her desire that a friendly rival win an honour, and also endorses her own desire to win the honour for herself, is essentially inconsistent as well. These values conflict essentially: there is no possible world in which they may all be satisfied.11 The person who is merely contingently conflicted may face conflicts in everyday life about what to do, but these conflicts arise because of the way the world is – if the world were different the conflicts in question would not arise; thus contingent everyday conflicts arise for any person who is practically inconsistent whether or not he is essentially inconsistent. It is not hard to see how practically inconsistent valuations may be genuinely reflective of the self: everyone plays various roles in life, and most lives are full of occasions in which one role requires one action and another requires an incompatible one. Consider again the example of the conflicted mother who values being at home with her children and also values having her own career. If her former attitude endorses a desire to be available at the right times, with flexible work hours, but she can’t find the right kind of job, then her conflict is contingent: with the right work arrangements, she could be content. Clearly this “merely conflicted” mother may endorse, in both senses,

42

Moral Reasoning in a Pluralistic World

each of her conflicting desires: there is nothing puzzling about how she may regard both her child-care desires and her career desires as fully her own, reflecting her self, and as desires she approves of. She is practically inconsistent only in the simple sense that she values doing A and values doing B and cannot do both. It may seem more puzzling that essentially inconsistent valuations may each be genuinely reflective of the person’s self. But to see this vividly, consider the second conflicted mother, who has an attitude for being a full-time stay-at-home mom and values working. This woman’s conflict is essential: there is no possible world in which one can be a full-time mom and have another job. Notice, for this person to be essentially inconsistent requires only that whatever mix of family life and work life she pursues, she will be dissatisfied, feeling under any arrangement that something has been lost. One needn’t be extreme in one’s child-care valuations to have a conflict that is essential; one need only feel that the time and energy one cares to spend with one’s children inevitably cuts into the time and energy one cares to spend on one’s work, and vice versa. This renders a person essentially inconsistent. When it comes to valuations and selfhood, the essentially conflicted mother is much like the merely conflicted one, since although there is no possible world in which her valuations may all be satisfied, those valuations may reflect what the person takes to be parts of her self she approves of: she cares about a certain kind of motherhood and a certain kind of work, regarding each as worthy of her attention. Such a person has a kind of “divided self” – which, I will argue here, can be a perfectly good kind of self.12 It may seem that essential conflicts could be resolved if only “there were more hours in a day,” but I think this is often false, since having meaningful relationships with others is more a factor of what proportion of time you spend with them than the number of hours total. A child who has a parent’s attention for 5 per cent of the latter’s waking hours is not getting much attention, even if for some reason that could work out to a large number of minutes. Imagine a day 100 hours long, and an infant who spends sixty of those hours sleeping, thirty-five of them with a sitter, and five of them with his mother. It would not be strange for this infant’s mother to feel an endorsed desire to spend much more time with him. Analogously, what feels like a significant work accomplishment is partly relative to the other elements of one’s life. It would not be unusual for a person to feel that

Pluralism, Unity of Value, and “Delimiting Argument”

43

they hoped their family life would take up over half their time and that their work life would take up more than half their time. And such a person would be essentially inconsistent. Furthermore, even being able to be in more than one place at one time might not solve a person’s valuational dilemmas. If I value having dinner with my partner, I do not want him to be also upstairs working on his book at the same time. I want his full attention. There is a poignant example of this in the graphic novel Watchmen.13 There, Doctor Manhattan can be in as many places at once as he likes; he is able to divide himself up. At one point, his girlfriend discovers that while he is having sex with her, he is also in his lab, working on a project. She is enraged, and properly so: relationships require not just numbers of hours but proportions of one’s self to feel meaningful. Doctor Manhattan regrets his actions; he wants to love her. To me the moral of this story is that even the person who can be in two places at once cannot solve the problem of valuational inconsistency. Sometimes we value things instrumentally, for other goods that they bring about. And it may seem that if our inconsistency is merely instrumental, we haven’t got a genuine “inconsistency” at all. So, for example, one might say that if I value miserliness (A) only because I value having a pleasant retirement, but I value generosity (not-A) only because I value the admiration of my friends, I haven’t got an “inconsistency” in my values: what I really care about is my own comfort (X) and status (Y), and there is a possible world in which these wouldn’t conflict: the world in which, say, I am very wealthy. But even such instrumental cases can reflect genuine conflict, and understanding these cases depends both on how our ultimate valuational ends fit together, and on the nature of the relationship between the end and the means.14 For example, I may value capitalist competition (A) because of its good economic effects (X), but value its absence (not-A) because greater equality makes most people happier or more satisfied (Y). In this case, it may not be clear whether it is possible to have X and Y without A and not-A, and if it is not, we have the same kind of inconsistency as we do in having non-instrumental values for A and not-A: there is no way to be valuationally satisfied. So these are cases of inconsistency as well. Furthermore, many inconsistencies are basic and non-instrumental. To value a life of security and of risk, and to care that one’s friendly rival win and also that she lose, to want to be a stay-at-home mom and also have a career, are all non-instrumentally inconsistent ways of valuing.

44

Moral Reasoning in a Pluralistic World

In any case, since valuing is a vague sort of state, whose objects are not always clear, I’ll take a person’s valuational outlook to be essentially inconsistent when there is no possible way things could be that would satisfy their endorsed desires and attitudes entirely – no “personal valuational utopia.” Notice that a valuational utopia may be different from a desires utopia: if A longs to have sex with his ex-wife but wishes he didn’t want to, having sex with her may be in his desires utopia but not in his valuational utopia. Thinking this way in terms of dissatisfaction shows a further important point: that the availability of appropriate compromises does not indicate either kind of consistency. I may compromise between capitalism and socialism, and reap some of the good of each, and still be dissatisfied with the missed opportunities for greater economic good and dissatisfied with the inequality that does exist. I may compromise between saving my money and giving it away, and still be dissatisfied with the inability to do more of each. If I value staying home with my children and also my work, I will likely compromise; as long as I have endorsed desires going unsatisfied I am nonetheless inconsistent. If those endorsed desires are inevitably dissatisfied, I’m essentially inconsistent, and if I’m contingently dissatisfied because of the way things happen to be, then I’m practically inconsistent or merely conflicted. But in neither case does the appropriateness of the compromise remove the inconsistency.

5 against the delimiting argument: practical inconsistency The delimiting argument can be interpreted via either kind of inconsistency: we might say values appropriately delimit or “fit together with” one another when they are practically consistent or when they are essentially consistent. I argue that neither interpretation explains a need for delimiting in a way that would support rich coherence; in this section I focus on practical inconsistency and in the next on the more difficult case of essential inconsistency. Recall that the person who is conflicted in either sense will be unable to instantiate or honour his values fully, and that this leads to a leftover or remainder experienced by the person after choosing. It is because persons who value in such ways seem rational, and seem to appropriately regret the value that is lost, that this remainder is sometimes called “rational regret.” The appropriateness of remainders may

Pluralism, Unity of Value, and “Delimiting Argument”

45

be obscure for the case of essential inconsistency, but I take it to be fairly obvious in the case of practical inconsistency. Marie will naturally experience regret for the option she has been unable to pursue, and this makes sense, since she values both honesty and fidelity. It makes sense for the mother who simply can’t find a job that accommodates her values of home life and career to regret the value she cannot honour. Indeed, those who face terrible choices, such as that between killing in war and failing to protect others, seem more admirable for feeling regret rather than less. Even if the choice of what, ultimately, one ought to do is obvious, one may still regret what one was unable to do. This shows that people value appropriately when they value in ways that are practically inconsistent. Practical inconsistencies are extremely common, given the many overlapping concerns and roles that most of us take on in ordinary life. The rationality of regret in such cases shows there is no reason for a person to pursue practical consistency in his values. Many of us find ourselves in situations like Marie’s. In such cases, the fact that one is practically conflicted simply means that one values both alternatives. Practical inconsistency in such cases simply means that one values both, and that having selected one option as the best overall, we continue to value what, because of the circumstances, we could not have had. To interpret the delimiting argument in terms of practical inconsistency is to say that values “fit together” when they entail no practical conflicts. And the delimiting argument asks us to consider that people are better off, or more fully rational, when they do not have such conflicts. But this is implausible. Not only is there no reason for people to pursue this form of consistency, there are even reasons for people not to pursue this kind of consistency when their regret seems appropriate and admirable. If delimiting requires practical consistency, then we must see our values as non-overlapping in the practical sense: this means that rational regret in cases of conflict among values would not be rational or appropriate. This, I believe, flies in the face of what we ordinarily take to be rational, sensible, and good among valuers. Even critics of value pluralism allow that mere conflictedness in the sense of practical inconsistency is not a problem for people. Robert Talisse, in articulating why he thinks pluralism must be committed to more than just practical conflicts of values writes, “surely value pluralism must involve more than the claims that there are costs involved

46

Moral Reasoning in a Pluralistic World

even in doing good, and that we often must forego some goods in order to attain others. These are commonplaces of any mature person’s view of the moral world.”15 This means the delimiting argument, when interpreted in terms of practical consistency, is implausible, since it requires us to reject what Talisse so elegantly describes as part of what any mature person already knows. This suggests the delimiting argument must be interpreted in terms of essential inconsistency, and indeed, it has sometimes been understood in just this way.16 When it comes to rationality and well-being it may seem intuitive that this kind of inconsistency, in which values are inherently in conflict, is obviously bad. In contrast to the case of practical consistency, it is often thought that the question of why a person ought not value essentially inconsistently has an easy answer: to do so is irrational. A person essentially divided in what he values, one may say, will be troubled: he will fail to act decisively; he won’t commit to things that matter; he will be unable to make decisions that reflect his identity; he will always be dissatisfied with himself. The unity of the person demands integrity of the person’s values, showing why inconsistency and ambivalence are dangerous and irrational. However, contra these intuitions, I’ll argue that not only does rationality not demand practical consistency, it doesn’t demand essential consistency either.

6 why might valuing essentially inconsistently be bad for a person? I address next the question of what, from the point of view of the self, might be bad about essentially inconsistent valuations.17 Harry Frankfurt has been a forceful and articulate defender of the idea that some kinds of internal dividedness are deeply troubling, and destructive for a person, rendering him irrational or non-autonomous.18 Here I explain some of Frankfurt’s ideas generally before applying them, in sections 6a to 6d, to the problem at hand. Taking second-order desires to represent what a person values, Frankfurt writes, “if there is an unresolved conflict among someone’s second-order desires, then he is in danger of having no second-order volition; for unless this conflict is resolved, he has no preference concerning which of his first-order desires is to be his will. This condition, if it is so severe that it prevents him from identifying himself in

Pluralism, Unity of Value, and “Delimiting Argument”

47

a sufficiently decisive way with any of his conflicting first-order desires, destroys him as a person. For it either tends to paralyze his will and to keep him from acting at all, or it tends to remove him from his will so that his will operates without his participation.” Here the claim is that deep conflict is troubling in at least two ways: a person may be unable to act at all, and if he can act, he will be alienated from his action in a way that makes it “not his own.” For a simple, Frankfurtian example, consider a drug addict who both values the pleasure of taking drugs and also values sobriety, and who is faced with the decision of whether to take the drug. He may be unable to act because he cannot choose what to do, ceaselessly switching his preference between his end of pleasure and his end of sobriety. Even if he does manage to act the person will not identify with his action: it could have gone either way, it reflects nothing about him essentially, it is neither to his credit nor to his blame how he acts, because he is not really the author of his action, failing as he does to stand behind it. In later work Frankfurt goes beyond these kinds of cases to claim that more general valuational ambivalence is bad, and there he focuses more clearly on the distinction between mere conflict and ambivalence. What a person needs with respect to his valuations, Frankfurt says, is “wholeheartedness”; to achieve this, it is not necessary to rid oneself of conflicting desires – what is needed, rather, is that one identify with one feeling or attitude wholeheartedly, and come to regard others as outlaws. The conflict, then, is transformed into a conflict between the outlaw desire and the person who struggles not to have it.19 If ambivalence is a disease of the soul, this is the “cure.”20 To get clearer about the significance of and “cure” for general valuational ambivalence, let’s look at Frankfurt’s exchange on the subject with J. David Velleman.21 Velleman raises doubts about Frankfurtian unification using Freud’s example of the Rat Man, an obsessive patient whose case history Freud published in 1909. The Rat Man – Velleman explains – both loved and hated his father, and consequently had troubling symptoms such as “repeatedly doing and undoing an action, or thinking and contradicting a thought.”22 But Freud’s discussion, Velleman says, shows not that the problem is ambivalence, but rather that the problem is the repression of the hatred; the way the Rat Man repressed his hatred involved identifying himself with his love, and dissociating himself from his hatred. This is what made

48

Moral Reasoning in a Pluralistic World

the Rat Man so unhappy. And this repression seems oddly close to the “cure” for ambivalence recommended by Frankfurt.23 In a reply to Velleman, Frankfurt says that the real problem here is not the mere conflict of emotions, but rather the failure to take a stand with one side of the conflict against the other – this failure is what ambivalence truly is. To address this, the Rat Man need not get rid of his conflicting emotions, but should just stand decisively behind his love and against his hatred.24 What is wrong with ambivalence, he says, is shown by the Rat Man himself; of his waffling and self-contradiction, Frankfurt says, “that sort of self-defeating behavior and thought violates the elementary requirements of rationality.”25 This tells us more clearly what unity requires: one can accept both attitudes as part of oneself, as long as one stands decisively behind one of them and regards the other as an outlier, external to the self. Here Frankfurt emphasizes the idea raised earlier, that the person who is valuationally inconsistent will be unable to act. We also find here a new thought: that even if he can act, the person’s action will be ineffective, because he will vacillate, serving one then another of his values. Third, we see again the worry that when the divided person acts, the act will not be autonomous, will not reflect the self in the right kind of way. These concerns, I take it, might apply to various kinds of essential inconsistency. A fourth difficulty, in contrast, applies just to the special case of ambivalence: this is the problem that the ambivalent person will be dissatisfied not just with his life, but also with himself. 6a Valuational Inconsistency and the Inability to Act The “inability to act” charge is a common one in discussions of inconsistency of various kinds, but it shouldn’t be, because when it comes to action, conflict is unproblematic, and essential inconsistency is the same as mere conflict.26 Mere conflict of valuations does not generally render us unable to act; if it did, our wills would be constantly paralyzed – ineffective behaviour would be much more widespread than it actually seems to be. Contingent valuational conflicts arise very often: most people play various roles in life, which results in various commitments, which in turn often demand incompatible things. Many people have commitments to careers and to loved ones that conflict. Many also have various good works they value that they would pursue if they had more time or more money. That they do not means

Pluralism, Unity of Value, and “Delimiting Argument”

49

that their commitments to these conflict. Shall I volunteer at the school or at the soup kitchen? Or should I just go to my kid’s soccer game? If contingent conflicts like these rendered people unable to act effectively, very little would get accomplished. But when it comes to the ways that conflict might or might not impede action, the person with a contingent conflict is in the same situation as the person with an essential conflict: each, when it comes down to the moment of action, has various options that he wants to pursue, of which he must choose one. When one is deciding what to do, it makes no difference whether the conflict is essential or contingent. Indeed, the cases in real life that cause us the most daily trouble are often contingent cases, arising because of commitments to jobs, family, friends, and so on that we haven’t the resources to fulfil. For the case of general valuational inconsistency, recall the conflicted mother who may be merely conflicted or essentially inconsistent. Whether she is one or the other makes no difference with respect to the kinds of conflicts of choice that she experiences as arising out of the more general conflict. She is offered a job and must decide whether to take it or turn it down, or she must decide one morning whether to stay home with her sick child or drop him off at a friend’s house so she can get to her job, or she must choose how many hours her child will spend in day care. The everyday conflicts of life will arise for the conflicted mother in the same way whether her essential conflict concerns her incompatible life roles or her contingent conflict concerns two values that – because she lives in a child-unfriendly world – contingently conflict. From the perspective of the person, these two kinds of conflict would feel very similar. The abstract possibility of a world in which one’s valuations can be fulfilled has nothing to do with the experience, in our world, of being valuationally conflicted – and thus has no impact on whether a person can act effectively. If people have difficulty acting, it is not because they are conflicted. It is more apt to say that a person who has difficulty acting does so because she cannot choose or prioritize among the things she values. 6b Valuational Inconsistency and the Possibility of Vacillation Even if a person is able to act on a case-by-case basis, we may worry that his actions may themselves be “inconsistent” in the sense that he vacillates, serving one value then another. It is true that the person try-

50

Moral Reasoning in a Pluralistic World

ing to integrate inconsistent values into his life may make individual decisions that seem not to add up. The conflicted mother may make choices that go this way and that, sometimes choosing to focus on her family and sometimes choosing to focus on her career. The ambivalent child may act on different feelings at different times. If these people undermine the effects of their previous choices with their later ones, we might describe this as a “doing and undoing.” But often the scattered choices of a conflicted person reflect not a “doing and undoing” nor a will that is “inchoate and indeterminate,” but rather something highly sensible: a compromise. While it is possible for the conflicted person to vacillate in a self-undermining way, he need not; he may “vacillate” because he is seeking out the best way to stay faithful to two things he values, or to himself. A mother who chooses to attend her child’s play one week and chooses her own work the next week is, in a sense, “vacillating,” but she is not selfundermining, and she is behaving in the best way possible for someone who values her family and her work. If one feels deeply committed to two valuations, feeling each to be important, and feeling each to be part of one’s identity, it seems the only rational course of action is to try to accommodate each as much as possible. This is unexceptional in the sense that we seem to do it all the time, weighing concerns about family against those of careers, concerns about charity against the good of our loved ones, and various feelings we have about the people around us. This applies in the same way whether the conflict is essential or contingent. The working woman who also values being a stay-at-home mom will likely make use of the same kinds of compromises that the contingently conflicted mother will. Such compromises require people to make choices that build on, rather than undermining, their previous choices, but these compromises are equally available to the merely conflicted mother and the inconsistent one. A person who both loves and hates his father may also “vacillate” in a sense: torn between affection and anger, he may act generously and lovingly on some occasions, and coldly on others. This may well be the best course of action, allowing the child to express his divided feelings: it hardly seems better either to deny one’s affection and cut off ties altogether, or to force oneself to be loving on all occasions. I allow that there may well be reasons to refrain from harming the father, and to provide a kind of care for him; there may even be reason for the

Pluralism, Unity of Value, and “Delimiting Argument”

51

child to try to love his father more. But these are moral reasons, and do not follow from considerations about wholeheartedness; if they did, they would be as easily directed to supporting wholehearted hatred. They also do not demand that one never express one’s divided feelings. In practical terms, Frankfurt’s demand that we “stand behind” the love and against the hatred will only remove the danger of vacillation by requiring us to act lovingly on all occasions: simply regarding the hate as an outlier emotion will not itself be sufficient to address this unless one’s behaviour becomes consistently loving. But this does seem close to the “repression” Velleman finds troubling. What a conflicted person needs in order to avoid vacillation is a “valuational prioritization”: a considered judgment on how to value the things he values, how his preferences for these values fit together, and what proportion of his concern each ought to take up. This prioritization would be roughly analogous to the “moral field” introduced in chapter 1: it is a way of directing and prioritizing among our many concerns, here understood with respect to valuing generally and not just with respect to moral values. But as I’ve argued, having a prioritization does not entail either practical or essential consistency. It might be thought that prioritizing is somehow possible in the case of practical conflicts in a way that it is not for essential conflicts. Frankfurt, in an expression of this idea, says that “being wholehearted with respect to one ‘psychic element’ is compatible with being wholehearted with respect to another that contingently conflicts with it – at least, as long as one element has a higher priority.”27 But prioritizations of general valuations can be just as rationally formed between A and B essentially conflicting, and even between A and not-A, as between A and B contingently conflicting. The woman who values her career and her family life in a way that is inconsistent – because there is no possible world in which she will be satisfied – and the woman who values them in a way that is consistent may each form prioritizations in the same ways. The person who loves and hates his father may well love his father more, but still experience hatred as a significant part of his valuational life. The drug addict who values both the pleasure of the drug and sobriety may well value them in a prioritized way, caring about one more than the other. So my point stands that there is nothing worse about essential conflicts or ambivalence here.28 Of course, it is true that being inconsistent creates conditions under which one may vacillate. But so does the ordinary con-

52

Moral Reasoning in a Pluralistic World

flictedness of overlapping life concerns. So nothing about the danger of vacillation bears on valuational inconsistency itself. 6c Valuational Inconsistency and Autonomy Our third worry is now obviously relevant, since we may wonder: if a person’s behaviour is determined by factors external to himself, how can it truly be his? The concern here seems to be that the inconsistent person, in choosing to sometimes serve one value and sometimes serve another, is committed to neither – and thus his action fails to reflect any valuational commitment at all. As with the previous discussion, I will distinguish between the person who has and the person who lacks a prioritization. For the person who has one, I see no difficulty about autonomy, because whether the inconsistency is essential or practical, the valuationally inconsistent person may indeed act in a way that properly reflects the self: he may act in ways that show respect for each of his values in the proportion that he endorses. The compromises and choices this person deems appropriate will reflect his particular valuational stance toward the world. This seems as much reflection of the self as anyone could require, and as much autonomy. In fact, an inconsistent person may be more obviously acting autonomously than a consistent one, since an inconsistent person must reflect before acting, whereas a consistent person – especially an unconflicted one – may be acting simply on impulse. And recall that prioritizations are as available between inconsistent values as they are between merely conflicting ones. So there is no difficulty in these cases, and as before my general point stands: there is nothing especially bad about inconsistency over mere conflictedness. Consider again the conflicted mother. Any conflicted mother, with or without a prioritization, may find herself regularly basing her decisions on external and seemingly inconsequential factors: I’ll go to the PTA meeting tonight because it’s at a convenient location and the traffic is light, but I’ll skip the Halloween party planning because I must spend the time responding to emails I neglected during the day. For the person with a prioritization, there is nothing non-autonomous about such behaviour; in fact, it is the stuff of which sensible compromises are made. Assuming the overall pattern makes sense, the person making such choices is absolutely reflecting her “self” – she is

Pluralism, Unity of Value, and “Delimiting Argument”

53

reflecting that she cares about both her career and her family in some particular proportional way. Perhaps we may worry that even this prioritized person risks behaviour that does not match up with her values: she values her career, but finds she sacrifices it so often that she loses her job. We can imagine that the trivia of life always go one way, suggesting to her that she ought to stay home, go to the school event, etc. But as long as the person has a prioritization, I claim, the problem here is not conflictedness, but any of three others: perhaps she doesn’t really value her career all that much, and this is why she often opts in favour of family; perhaps she deliberates poorly, and is overly influenced in her decision-making by incidental things; or perhaps she has poor impulse control, and fails to act in accordance with what she has decided to do. Any of these difficulties would suffice to create the relevant problems, and would do so even in a person who was not valuationally conflicted but just easily distracted. That many conflicted people manage to work out such compromises shows that conflictedness is not the crucial issue here. Again, these conclusions hold in the same way for both the inconsistent mother and the merely contingently conflicted one, and similar considerations apply to the ambivalent person who loves and hates his father. The child who expresses each of these ambivalent feelings in some proportion reflecting his valuations is reflecting his self. So there is nothing especially bad here about either form of inconsistency. A brief consideration of some standard accounts of autonomy supports these intuitions.29 Of course, one way of understanding autonomy is grounded in the very notions of identification we have been working with here. On these views a person is autonomous if his relationship to his desires and values is one of appropriate essential identification, so that, for example, it is free from influences that are inappropriately external.30 By hypothesis, our inconsistent persons do identify with the desires, goals, and values they end up pursuing; it’s just that they also identify with other desires, goals, and values that are incompatible with them. In the case of mere conflict, it is clear that – assuming prioritization – conflictedness does not undermine a person’s autonomy; since, as I’ve argued, one may prioritize among “inconsistent” values as well as among merely conflicting ones, the same reasons show that – assuming prioritization – inconsistency does not undermine a person’s autonomy either.

54

Moral Reasoning in a Pluralistic World

A second way of seeing autonomy focuses on whether a person is properly reflective and able to articulate and respond to good reasons.31 Again, nothing about inconsistency prevents a person with a prioritization from reflecting and acting on the basis of good reasons. Just as in the case of mere conflict, such a person will typically develop her prioritization in response to certain features of the world, how her choices are structured, and so on, and just as in the case of mere conflict, such a person will typically act on or change her prioritization in response to reasons. 6d Valuational Inconsistency and Dissatisfaction with the Self For each of the three proposed difficulties I’ve considered – inability to act, vacillation, and failure of autonomy – I’ve argued that inconsistency is very like mere contingent conflictedness. The fourth worry, in contrast, is constructed to apply specifically to ambivalence. The idea here is that the ambivalent person will not just be inevitably dissatisfied, he will be dissatisfied with himself rather than with the external world. Consider again Frankfurt’s idea that contingent conflicts differ from essential ones in that they do not threaten the unity of the self, since “being wholehearted with respect to one element is consistent with assigning a higher priority to another.”32 But no one can be satisfied with being ambivalent, because the elements of such a person will work against one another; such a person can never be satisfied with himself, since he will always be restless: part of his self is working against other parts. Let me distinguish two related concerns here. One has to do with the abstract integrity of the self, the second with dissatisfaction with the self. Each of these has the greatest force when we consider ambivalence in particular. With respect to integrity, we may wonder: If I endorse and stand behind my desire for A, and I take it on as part of myself, then how could I possibly endorse and stand behind my desire for not-A? Doesn’t taking on not-A mean, in a sense, rejecting A? It is true that the ambivalent person lacks a kind of “integrity” that the non-ambivalent person has. Insofar as he takes on as his own each of his ambivalent feelings, the child who loves and hates his father may well experience his inner life and self as internally divided. This person is “of two minds,” as the saying goes. But as long as he is as able to avoid being self-undermining, what is bad about being of two

Pluralism, Unity of Value, and “Delimiting Argument”

55

minds? Imagine the child’s hatred is somewhat stronger than his love. A demand for wholeheartedness, considered on its own, seems to suggest this person would be better giving up his love. But this seems wrong.33 With respect to dissatisfaction, there are two ways to understand the problem. Recall from section 3 above that there are two subtly different ways of saying what an “endorsed” desire is: we may characterize it either in terms of self-recognition, or in terms of approval. On the first interpretation, the dissatisfaction worry is that the person will be troubled by taking on for himself care for opposite things; on the second, the dissatisfaction worry is that the person will be troubled particularly by his approval of two opposite things. Let me take these up in turn. If we take endorsement as self-recognition, although the ambivalent person will be inevitably dissatisfied, I see no reason for him to be dissatisfied with himself. Although Frankfurt mentions vacillation as one of the pitfalls of ambivalence, we’ve already seen that this is mistaken: vacillation in behaviour is no less likely with contingent conflicts than it is with essential ones. Especially if one’s contingent conflict involves values that are quite evenly matched, one may choose to serve one value and then another, though as I’ve also stressed this can be sensible compromise rather than vacillation. But if vacillation is not the specific pitfall here, what is? Velleman’s idea that self-acceptance is what matters for the ambivalent child suggests a way to understand a person who is ambivalent yet content: he acknowledges and accepts his own filial hatred as part of himself. Whether a person is willing to accept such hatred as part of himself would seem to me to have more to do with the other facts of the case than with ambivalence. If the child feels his hatred is justified, or appropriate, he may be willing to accept his own ambivalence; if he feels his hatred is arbitrary or undeserved, or if he feels hatred is a morally bad feeling, he may struggle to overcome it. But these sources are distinct from the love the child feels, and so are not tied to resolving ambivalent feelings. Frankfurt does recognize the possibility that conditions may be such that ambivalence is an appropriate response, but he suggests they will be rare, and he says that even in this case, one cannot desire to be ambivalent for its own sake. This, he claims, shows that we “wholeheartedly desire to be wholehearted.”34 Perhaps, if this is so, all it means is that conflicts between wholeheartedness and the desire to accom-

56

Moral Reasoning in a Pluralistic World

modate inconsistent values are always contingent. I suppose they may be, if it is contingent that I value the things I do. But put this way, the point seems to lead to little in the way of irrationality claims: as long as there can be conditions under which it is best to be ambivalent, it can be rational to be ambivalent, and in this case ambivalent may be the best way I can be. Perhaps one may object here that even if there are conditions under which I must accept my own ambivalence, I cannot hope that such conditions will obtain – and insofar as this means that I prefer wholeheartedness to ambivalence I am, in a sense, dissatisfied with my ambivalent self. I have no reason to value ambivalence for its own sake. But this would be an effective argument only if there were independent reasons to value wholeheartedness for its own sake. What would these reasons be? Wholeheartedness may seem intrinsically appealing when we think of it in terms of a comfortable certainty and lack of anxious doubt. But a person who decided to lie, to wage war, or to protect his children by harming others would be a worse person for being wholehearted in his decision. Even if they just “stand behind” the item they judge to be best, regarding the other as an outlier, these people seem less admirable the more wholehearted they are about their choices. We expect morally sensitive people to feel sorrow and regret over the harm they are causing, even when they act for the best overall. In these circumstances ambivalence is better than wholeheartedness.35 If this is right, the most one could say about wholeheartedness is that it is sometimes good and sometimes not good, which is just what I have claimed about ambivalence itself. When endorsement is taken as approval, the problem may seem more acute, since to approve of a desire is so closely bound up with having a moral belief or attitude toward the thing in question. Being for something morally, one is typically against those who are against it, and this suggests that a person who is ambivalent will be against himself. For example, imagine a person who loves animals and endorses – in the sense of approval – his desires to promote their welfare; for this person to also endorse with approval his desire to hunt would, indeed, seem odd. It seems that the animal-lover part of him should condemn and be alienated from – and certainly be dissatisfied with – the hunting-lover part of him. But while there is something odd about this person, I claim that this oddness derives from, and is dependent on, the assumption that

Pluralism, Unity of Value, and “Delimiting Argument”

57

if the first of these items is morally good, the other may not also be morally good.36 If one does make that assumption, then indeed the person has a problem – since he is approving of something he is himself committed to regarding as not worthy of approval. But this problem is not a problem of inconsistency. It is a problem of approving of what one regards as not good to approve. To see that inconsistency of valuations is not itself the difficulty here, note that a person who has good reasons for endorsing with approval both his desire to promote animal welfare and his desire to hunt, and who is in a state of uncertainty about what to do, need not be dissatisfied with himself at all. Imagine we’re talking about a young aristocrat in England in the late nineteenth century: he has grown up regarding fox hunting as a wonderful part of his cultural life, a link to his ancestors, and an activity that develops one’s courage and fortitude. He then develops a caring for animals and finds himself in a quandary. He may feel that both hunting and animal welfare are good things; he may be sad that they are mutually exclusive; he may be uncertain how to proceed. We may imagine he is truly ambivalent, endorsing his desire to hunt and his desire to refrain from hunting. As long as he regards both with approval, however, there is no reason for this person to be dissatisfied with himself. He may rather tend toward self-satisfaction, seeing himself as more morally sensitive than both the unconflicted hunter and those who shun hunting altogether. This person approves of what he finds worthy of his approval. I’m not claiming such cases are typical, just that they show that any dissatisfaction with the self derives not from inconsistency or ambivalence, but rather from the feeling – a feeling that may accompany inconsistency, but needn’t – that one approves of what one does not regard as worth approving of. To see this, notice that in non-moral cases no such dissatisfaction arises. Recall the example of friendly rivalry, in which a person both wants his friendly rival to win and wants his friendly rival not to win. Assuming the person regards both of these desires as morally neutral, there is no reason for her to feel dissatisfied with herself for having them. Perhaps one may object here that any serious moral person must be morally against not-A when he is morally for A, on grounds that moral seriousness requires taking a stand one way or another. Something like this is, I believe, correct, but is more fruitfully explored through principles and reasoning than through values – as I do in

58

Moral Reasoning in a Pluralistic World

later chapters. In chapters 4 and 5 I address the question of how and why one ought to aim to bring coherence to one’s moral judgments.

7 inconsistent values, compromises and priorities, and the values of others In chapter 1 I talked about value pluralism and about value field pluralism, raising the idea that while we care about similar kinds of things – harm prevention, fairness, loyalty, etc. – we care about them to varying degrees, in varying proportions, and directed at various objects. I referred to empirical research suggesting that the possibility of our having varying “value fields” was the best explanation for widespread moral disagreement. In this chapter, I’ve said that the pluralistic way we value puts a burden of proof on monism to give reasons for reinterpreting our values. One way of reinterpreting our values is through delimiting: that is, by seeing values as fitting together – as integrated – because their boundaries delimit one another. I’ve argued in this chapter against such interpretations, on the grounds that delimiting erases remainders and is not rationally required. This does not show that monism is inconsistent with the existence of remainders: as we’ve seen, even a single principle in favour of promise keeping can create remainders. But it does show that the most common strategy for integration – the delimiting strategy – does not succeed, and thus does not itself discharge the burden of proof. I’ve also argued here that we are able to prioritize among values even while regarding them as inconsistent, and that this is so for both practical and essential inconsistency. In practice, such prioritizations often manifest themselves as compromises: either of the two conflicted mothers may pursue a course that will allow her to honour both her valuation for family and her valuation for work. As I’ve been emphasizing, such compromises – even when appropriate – need not render a person consistent in either sense; these people may remain “inconsistent” in the sense that in all cases they will experience remainders and must sacrifice something of value. Note that appropriate prioritizations and compromises also do not typically require what is sometimes called a “lexical” ordering of one’s values, that is, an ordering on which some values take universal priority over others, with some values always prioritized above others in

Pluralism, Unity of Value, and “Delimiting Argument”

59

every situation.37 Conflicted mothers may prioritize their values without taking either of their values to always be the first to satisfy: typically, people find ways to honour their values proportionally, and directedly, without ordering them lexically. This picture of people as inconsistent valuers who compromise in ordinary life leads, I believe, to an illuminating thought about differences in value fields: people might have more similarity in their values than in the prioritizations they endorse, and when they do, retaining “inconsistency” rather than pursuing delimiting can facilitate mutual understanding. Two conflicted mothers might differ in their choices, compromises, and prioritizations, with one choosing to stay at home and the other choosing to work. Notably, these inconsistent people will understand and empathize with one another far more than people who simply value in accordance with their prioritizations. In contrast, the mother with a specific and confidently held sense of proportion, with delimited and non-overlapping values, may find other ways of valuing more foreign to her. The same thought carries over to the moral realm. The person who values liberty more than equality but acknowledges that his caring for these values conflicts will share much with the person who values equality more – and, crucially, they will share more if they are conflicted than they would if they valued in ways that were rationally delimited and thus consistent. Valuational inconsistency can thus bring shared understanding to those who value differently. In his discussion in favour of unification, Dworkin does at one point mention “another possibility” that “values don’t conflict just because they do, but because they work best for us when we conceptualize them so that they do.” He says of this that it “is a conceivable view, and perhaps someone might make it seem plausible.”38 I’ve argued here that it’s more than plausible – in fact we have strong and pressing reasons to conceptualize our values this way.

60

Moral Reasoning in a Pluralistic World

3 Epistemological Arguments and Rich Coherence

In chapter 1, I argued for a particular relationship between value pluralism and principle pluralism: that if values are genuinely plural, a theory with multiple principles engendering conflicts will be the best moral theory. In chapter 2, I gave some reasons for regarding our practice as genuinely pluralistic, arguing against the idea that a rational or autonomous person will have to find a way to rid themselves of conflict and ambivalence. The next question is whether we have reason, in our moral theorizing, to aim for or prefer theories with fewer and simpler principles. Here, I address this question through consideration of the norm of systematicity, which says that the general principles of a theory should be as few and as simple as possible. Systematicity brings together the ideas of unity, simplicity, and connectedness, and is thus rich coherence. It is often argued that we have some epistemic reason to pursue systematicity because a more systematized theory gives us better justification or increased understanding. For example, recall Marie’s dilemma about whether to keep her promise or to lie. A fully systematic theory like consequentialism says that there is only one principle in the case: act so as to bring about the best consequences overall, whether that means lying or promise breaking. This appears to justify the answer and to give us insight into why the answer is correct: she ought to choose the one act over the other because the one act brings about the best consequences. In this chapter, though, I argue that given certain facts about the pluralistic ways we value, these ways of supporting systematicity do not succeed: there is no reason to think that unified answers to questions about moral matters are better than pluralist answers to those

Epistemological Arguments and Rich Coherence

61

questions. Rich coherence methods are inappropriate for pluralist contexts, because they stack the deck against pluralist theories for no good reason. This conclusion follows not only from the stronger, normative assumption that we are right to value, as we do, pluralistically, but also from the weaker and less contentious descriptive claim that the way we value in practice is pluralistic.

1 understanding and explanatory justification: against systematicity In favour of rich coherence, it is sometimes said that in the moral domain we want to know not only what is so but also why it is so, and that in richly coherent theories some of our beliefs support or provide a best explanation for others.1 Shelly Kagan emphasizes the need for understanding when he writes that the “need for explanation in moral theory cannot be over-emphasized. We want our moral theory to help us understand the moral realm.”2 David Brink emphasizes the need for justification when he writes that “the degree of a belief system’s coherence is a function of the comprehensiveness of the system and of the logical, probabilistic, and explanatory relations obtaining among members of the belief system ... we are all familiar (I assume) with the greater explanatory coherence that some bodies of belief exhibit when compared with others and that we seek to achieve in our own beliefs.”3 The first of these ideas has to do with understanding and acting for the right reasons. In morality we not only want to know what to do, we also want to know why a given action is the right one. The belief that medical testing requires informed consent is widely shared. But even if we fully agree in this belief about what is morally required, we want to know why it is morally required. Understanding why informed consent is important is necessary for implementing the rule in the right way, for developing other appropriate ethical constraints on medical testing, and for helping to figure out what to do in difficult cases – such as the one concerning the double standard for testing. The second idea has to do more directly with justification, truth, and being right about what to believe or not believe. If the more general principle that we ought not infringe others’ autonomy rights is the best explanation for the more specific principle about consent of medical research participants, then they seem mutually justifying: the second supports and is explained by the first.

62

Moral Reasoning in a Pluralistic World

Understanding, however, is importantly dependent on justification because to give us insight explanations have to be good explanations; that is, we need some reason to think they are better, or more justified, than others. Unless we have reason to think the general prohibition against violating others’ autonomy rights is true and is the best explanation for the informed consent requirement, we have no way of knowing whether we’ve got the right reason. And thus we don’t have any better understanding than we did before. To know why something is so, we need not only a general principle to explain it, but some reason to think the explanation accurate and justified. Knowing when we have good explanations can be complex. This point can be seen in a simple analogy with explanations about the ordinary world. If I propose to explain why this cat is black by saying that all cats are black, then I obviously have neither increased justification nor improved understanding, because the “explanation” appealed to is false; even if the generalization involved were true, this wouldn’t be a good explanation. If I propose to explain why the economy was failing in 2009 by appealing to the general principle that the economy generally fails when the American Football Conference wins the Super Bowl (and the Steelers won that year), then I likewise have neither increased justification nor improved understanding. The correlation exists, but it doesn’t depend on any actual causal regularity. The causal mechanism it proposes isn’t one after all; the generalization is “accidental.” This difficulty is especially pressing in the context of evaluating systematicity in pluralist contexts. In theories with multiple principles, beliefs do “support” and “explain” one another: principles explain and are supported by judgments, derived principles explain and are supported by fundamental principles. The difference between more and less systematized theories is not that we fail to have explanations in the latter, but rather that the explanations in pluralist theories will be different, and will appeal to multiple fundamentals. To know that we have increased understanding and justification in systematized theories over less systematized ones, we need to know whether the principles in more systematized theories provide better explanations than the ones in less systematized theories. For example, suppose I judge that it is wrong to lie and wrong to harm others, and I am trying to figure out whether to explain both judgments by appeal to the general wrongness of causing harm, or

Epistemological Arguments and Rich Coherence

63

whether to explain the lying prohibition by appeal to the injustice of lying. A set of beliefs based on the former is more systematic, but in both cases we have beliefs that “explain” others – the question concerns which explanation is the best one. In this section I argue that because of value pluralism we have no way of knowing when and whether more systematized theories provide better explanations than less systematized ones, and that this undermines support for systematicity. I claim first that this follows from the normative thesis of value pluralism: if values are plural, explanations that appeal to more fundamentals will sometimes be better than those that appeal to fewer, in the simple sense of being more accurate. At the end of the section I show that the same conclusions follow when we’re agnostic about the normative thesis. If values are plural then clearly sometimes the best explanations are in less systematized theories, because if values are plural then by hypothesis the optimal theory will not be fully systematic. Systematicity says that the general principles should be as few and as simple as possible. That is, we should aim to reduce the overall number of principles. But if values are plural, principles grounded in very different kinds of considerations are not expressible in terms of some single principle – insisting on fewer, simpler principles would lead to error. Systematicity entails that more unified theories are better than less unified ones. But if values are plural, there are limits to the extent to which unifying our moral cares puts us on the right track. Systematicity entails that we should prefer theories that are more connected and less patchy to ones that are patchier. But if values are plural, the ideal theory may well be somewhat patchy. In the absence of knowledge about how many fundamental values there are, we have no way of knowing just when systematizing yields better explanations and when it does not. For a systematicity norm to be justified, it would have to be the case that the number of fundamentals we are right to appeal to is lower than the number of fundamentals we do appeal to in a moral theory. But on what basis could we know the number of fundamental values we are right to appeal to? In the context of conviction ethics, our only evidence about these matters comes from our considered convictions. But from this evidence we could never learn that the true number of fundamentals is lower than the number we appeal to in our considered convictions.

64

Moral Reasoning in a Pluralistic World

To see the argument in more detail, recall from chapter 1 that systematizing means two things: making trade-offs with intuitive acceptability and preferring more systematized theories when intuitive acceptability is equal. I take these up in turn. In a case where trade-offs are required, one starts with a principle pluralist theory and considers whether to adopt a more systematized theory that fits less well with one’s considered convictions. For example, suppose one starts with a theory that appeals to a value of non-maleficence and a value of justice, has principles describing obligations not to harm others and not to lie, and fits well with certain convictions one has about telling the truth even when it will be harmful. To say that trade-offs between intuitive acceptability are appropriate is to say that it would be better to adopt a new theory that explains the importance of non-maleficence and honesty derivatively, in terms of a third value or principle such as beneficence or respect for persons – even if that new theory does not fit as well with our convictions about when to tell the truth. But in the absence of independent knowledge that there is only one value at stake here instead of two, we have no reason to make any such trade-off. Indeed we generally have no reason to make any such tradeoff in the absence of independent knowledge that the number of fundamental values we ought to appeal to is lower than the number of fundamental values we do appeal to. The line of thought we are pursuing says that more systematic theories have beliefs that support and provide best explanations for others. But as I’ve argued, this claim is justified only if we have reason to think the explanations in more systematic theories are better. And we have no such reason. It is easy to see how a more systematized theory might seem, on the surface, to provide more and deeper explanations than a more pluralistic one, since in addition to explaining some derived principles in terms of other, fundamental ones, they seem to explain more fundamental principles in terms of fewer fundamental ones and thus seem to go further into the matter. Concern over the depth of explanation may motivate Kagan when he writes that even though explanations may “have to come to an end somewhere,” that is “no license to cut off explanation at a superficial level.”4 But just because explanations seem deeper in a more systematic theory does not mean that they genuinely are. If values are plural, explanations properly come to an end before a theory is systematized, and we do not know when it is right to stop.

Epistemological Arguments and Rich Coherence

65

In his arguments in favour of rich coherence, Brink emphasizes the real possibility that the “most coherent set of moral beliefs could be radically pluralistic or fragmented.”5 And in his defence of moral realism against the skeptical challenge from objections having to do with disagreement, he claims that some moral questions have no answer because of “incommensurability” and the possibility of “moral ties.”6 I take this to mean that there may be cases in which values can’t be compared, so that there is no single right answer about what to do overall, and cases in which our obligations are so evenly matched that there is no single right answer about what to do overall. It would follow that in making the necessary trade-offs between coherence and intuitive acceptability, we may find that the most coherent theory that is acceptable might not be very unified or systematic. But I’ve argued here that lack of support for systematicity doesn’t just mean the end result could be fragmented; it indicates the more radical conclusion that there is no reason to adopt a goal of systematicity in making trade-offs at all. Now consider the second aspect of systematicity: that we ought to prefer theories with fewer principles when no trade-offs are required, because we are deciding among theories that are equally intuitively acceptable. A special case in this category concerns the one in which we are wondering whether to replace our current set of beliefs with a more systematized one that is the same with respect to intuitive acceptability. For example, we might find a less pluralistic theory that ends in the same considered convictions as a more pluralistic one. Systematicity in these contexts acts as a kind of tiebreaker. It might seem in this context that as long as no trade-offs are required, we have nothing to lose: a theory could fit with our considered convictions and be fully systematized. In arguing for his ruleconsequentialism, Brad Hooker says we can systematize a Rossian list of duties by appeal to the single principle “an act is wrong if it is forbidden by the code of rules whose internalization by the overwhelming majority of everyone everywhere in each new generation has maximum expected value in terms of well-being (with some priority for the worst off).”7 Hooker says that other things being equal, a unified explanation of our obligations is better than a pluralist one for several connected reasons. The unifying principle “would be an interesting discovery”; the unified theory would contain “information that moral pluralism does not,” “would contain a deep connectedness omitted by

66

Moral Reasoning in a Pluralistic World

moral pluralism,” and “would be able to explain more on the basis of fewer assumptions.” In these senses, he says, “the theory with the deeper principle is more informative and integrated.”8 But even when used as a tiebreaker, I claim systematicity is unsupported. Generally, systematized theories that fit equally well with our considered convictions still give us different explanations for those convictions and different reasons for belief. By definition if one set of moral beliefs is more systematic than another it has fewer fundamental principles, and thus contains different and more systematic explanatory relations from that other. For the former to be better and more justified than the first, we would need some reason for thinking that the explanations in more systematic theories are better explanations. Again, we do not have such a reason. If the normative thesis of value pluralism is true, the right explanations and reasons for our moral beliefs will appeal to multiple fundamentals rather than a single one; the systematized theory would be worse, would not be a discovery, and would not contain more information than a pluralist one because it would contain explanations that were mistaken. The “brute curiosity” we have to know why things are true, which Hooker describes so vividly, would not be satisfied by systematized explanations. To see how this matters practically, notice that one use Hooker proposes for a systematized theory is that it will help us resolve disagreements.9 Of course, the fact that we have such disagreements already complicates the idea of fitting with all of “our” considered convictions: in a world of moral disagreement and value-field variation, it is not even clear what this means, since those convictions vary from person to person. At most, we might say that there is a person A such that there is a systematized theory that fits with all of A’s considered convictions. In this case, using systematicity as a tiebreaker would mean having grounds on which to say that A’s theory is better than a rival, more pluralistic theory – say, the more pluralistic theory that fits with B’s considered convictions. But in a context of pluralism, there is no reason to prefer A’s theory; its being systematized is a strike against it. Using A’s systematized theory to resolve disagreements would be a mistake. For example, in his discussion of the ethics of pre-marital sex Hooker argues that a prohibition on premarital sex made sense in circumstances in which “life expectancy was short, little social safety net existed, no reliable methods of birth control were available, and there was no reliable

Epistemological Arguments and Rich Coherence

67

mechanism for making both parents responsible for their children except marriage.”10 Though people did not appeal to the harms of premarital sex in explaining why they were against it, Hooker says that when people “could no longer see the harms associated with pre-marital sex,” the prohibition no longer made sense.11 This is intended to support the rule-consequentialist view: we implicitly adopt the moral beliefs that track the rules which, if internalized broadly, would lead to best consequences overall. Compare this systematized moral view to a pluralist one that recognizes the importance of harm prevention but sees sexuality primarily in terms of autonomy rights. These views might disagree on the appropriate forms of sexual freedoms in a range of contexts. It could be, for instance, that internalizing rules against casual sex would promote well-being overall; for the rule-consequentialist this would be reason to find it immoral, while for the relevant pluralist it would not be. To say that having a systematized theory will help us resolve disagreements is to say that we have reason to prefer the rule-consequentialist view, and thus endorse the relevant prohibition. But if values are plural, there is no good reason to prefer the former or endorse the latter.12 Similar considerations apply to Robert Audi’s proposed systematization of Rossian principles.13 Audi argues that Rossian duties can be unified though a Kantian approach that considers whether our action falls under a maxim we can universalize and whether our actions treat persons as mere means. Audi says that a more systematized theory is better, partly on grounds that it is less “dogmatic” and less of a “hodgepodge” than pluralist theories that are not systematized. He writes, “if we can realize the possibility this suggests – namely, a systematization of the Rossian duties by appeal to a more comprehensive set of grounds than Ross appealed to – then contrary to what the dogmatism charge implies, that systematization might provide both reasons for holding the principles and a source of correctives for false or merely apparent moral intuitions or moral principles. If intuitionism can be enriched by even a partial account of why Rossian principles hold, of what they have in common, of why one of them sometimes takes priority over another, of what sort of thinking helps to internalize them, and of related matters that go with a well-developed systematization, then an intuitionist ethical view will be much more plausible.”14 He concludes that “if the main ideas suggested in this chapter

68

Moral Reasoning in a Pluralistic World

are sound, we can retain the attractive features of a Rossian intuitionism and still extend the theory to a more comprehensive, better unified view.”15 In these passages we find two related ideas: first, that in the absence of systematicity, we must be “dogmatic” in our beliefs, and second, that a “hodgepodge” of principles is unacceptable. But I’ve tried to show here that we have no reason to think a pluralistic “hodgepodge” of principles would be improved by being systematized in this way: both pluralistic and systematic theories contain explanations, and in the context of conviction ethics at least, we have no grounds for thinking that the explanations in more systematic views are better. And thus, to prefer systematic theories for no good reason is to be dogmatic in a different way. If we had independent reason to believe that the explanations in systematic theories were better, we would have support for systematicity. As I’ve explained, we cannot get such a reason from conviction ethics. It would have to come from an assumption that value is ultimately unified or that single principle theories are best. Since these assumptions are incompatible with pluralism, rich coherence methods are inappropriate for use in pluralist contexts. So far I’ve been assuming the normative thesis of value pluralism. But I claim that the same considerations apply even if we are agnostic about the normative thesis. The arguments above rest on the fact that we do not know whether, and when, explanations in more systematic theories are better than those in others. But if this is so under the assumption that values are plural, it is also so under the assumption that we do not know whether values are plural. All that is required for this argument is uncertainty about value pluralism. The descriptive thesis is sufficient to put the burden of proof squarely on the defender of systematicity and monism. That we lack evidence about justification that is independent of our considered convictions is one reason I am skeptical of the “evidentiary consistency” criterion Sayre-McCord proposes in his later work as an element of coherence.16 There Sayre-McCord defends a rich method on epistemological grounds, focusing on evidential consistency, connectedness, and comprehensiveness. The important difference between this rich conception and the one discussed in his earlier work is that consistency is here understood as “evidential consistency” rather than as the resolution of conflicts. A set of moral beliefs is evidentially consistent

Epistemological Arguments and Rich Coherence

69

“if and only if the weight of the evidence provided by the various beliefs in the set don’t tell, on balance, against any of the others.” In this later work Sayre-McCord argues that “our beliefs, moral and otherwise, are justified only if, and then to the extent that, they cohere well with the other things we believe,” where coherence is understood in this “evidentiary” sense.17 This could be read as an argument on behalf of rich coherence and its connection to justification. But read this way I do not think it succeeds in challenging anything I have said here. First, appealing to “evidence” and evidential consistency in the characterization of coherence seems to presuppose that we already have an account of what counts as justificatory in the moral context. But in the context of conviction ethics, it is coherence itself that tells us what does and does not count as justificatory. And as I’ve argued, in the absence of assumptions about the nature of value we don’t know what form that coherence takes – that is, whether or not it takes a form that demands systematicity. For example, does a principle requiring us to lie to protect friends or keep promises in certain circumstances “count against,” in this evidentiary sense, a principle requiring honesty? The answer depends on whether values are unified or plural, on what kind of coherence gives us justification, and thus on whether we already have reason to prefer rich forms of coherence. Given this, evidentiary consistency cannot help with a justification of rich coherence in the context being considered here. Similar remarks hold for the inclusion of connectedness and comprehensiveness. We’ve seen that these, in their connection to systematicity, are connected to justification only in the presence of certain question-begging assumptions about the nature of value. So while I leave open the possibility that Sayre-McCord is right in drawing some general connection between coherence and epistemic justification, I conclude that the justification does not succeed in supporting a rich version of coherence over alternatives.18 In his canonical work on coherence and justification Norman Daniels suggests two ways of addressing the general question of how to distinguish when we do and do not have better explanations for our beliefs in coherent theories. If successful when applied to systematicity, these would indeed give us reason to think that systematic explanations are actually better. The first concerns the move to “wide equilibrium” – or coherence with a broader set of beliefs – and the second concerns analogies with coherence in the scientific realm.

70

Moral Reasoning in a Pluralistic World

I do not think either successful in the present context; I discuss them next.

2 systematicity, coherence, inference to the best explanation, and science Daniels suggests that the problem of distinguishing between accidental and justified generalizations – and thus determining when we have better explanations – can be partly addressed by moving to wide reflective equilibrium, in which we demand coherence not only among moral beliefs but among moral and non-moral beliefs together. Wide reflective equilibrium asks us to bring into coherence our moral judgments, moral principles, and background theories about matters such as the nature of persons and the role of morality in society. When it comes to justification, Daniels imagines that the background theories might function as a kind of useful “check” on the principles and judgments: coherence with the background theories would count in favour of a given set of moral principles and judgments, and lack of such coherence would count against them. This works as a justification only if the background theories aren’t simply restatements of the moral content. As Daniels puts it, to achieve “independent support,” “we should require that the background theories ... be more than reformulations of the same set of considered moral judgments involved when the principles are matched to moral judgments.”19 Daniels gives an example from Rawls: in using the original position to reason about mutually acceptable principles of justice we make use of shared beliefs about other matters, including theories of human nature and theories of the role of morality in society. These shared beliefs in “background theories” rest on moral judgments, but the moral judgments in the background theories are independent from those we are aiming to evaluate. In this case, the background theories talk of fairness and personhood; the results have to do with rights and entitlement, which are distinct. Whatever the general benefits of wide reflective equilibrium, I claim it will not help us with determining whether and when more systematic explanations are better explanations – epistemologically – than more pluralistic ones. This is because wide equilibrium functions as justification only if we assume the kind of systematicity – or at least, connectedness – that is up for debate here. To say that an

Epistemological Arguments and Rich Coherence

71

explanation is a good one because it links up the judgments in one moral domain with the judgments in another rests on an assumption that values are not plural in the sense described; it assumes that they are connected or that a common ground explains both. These judgments are just what are in question here, because if values are plural there are domains that are distinct and disconnected. Indeed if values are plural the ideal theory may be somewhat patchy, having parts that are internally connected but not connected to one another. This means it would be question begging against value pluralism to assume that wide reflective equilibrium can show us why we ought to favour more systematic explanations over less systematic ones. With respect to larger questions of epistemological justification, Daniels – like others – suggests we might understand moral coherence as part of coherence reasoning more generally. Indeed, later in his paper Daniels says that to draw conclusions about whether morality is objective in the right way, and not just a restatement of various prejudices we’ve learned from our peers, we might consider an analogy with natural science: “we may be able,” he writes, “to piggyback a claim about objectivity in ethics onto the analogous claim we are assuming can be made for science.”20 There are two ways to understand coherence and its relationship to justification in the scientific context: probabilistically and explanatorily. Some informal ways of talking about coherence and justification seem to rely on a probabilistic interpretation even though this is not made explicit. For example, coherence seems related to probabilistic justification when we say things like: “if the murderer was seen to be wearing a minister’s collar, and the suspect is a minister, this increases the likelihood that the suspect did it.” The bioethicists Beauchamp and Childress say that calling on a “wide body” of moral experience will help with justification, and offer an analogy to a courtroom: “If sufficient numbers of independent witnesses converge to agreement in recounting the facts of a story, the story gains credibility beyond the credibility of any one individual who tells it ... The greater the coherence in a story that descends from initially credible premises and convergent testimony, the more likely we are to believe it and accept the story as right.”21 This might seem promising, on the grounds that it tells us about the connection between coherence and truth given the independence of various particular judgments – which might give us a way to talk

72

Moral Reasoning in a Pluralistic World

about the added justification from independent background theories. But the probabilistic approach, I claim, is not useful here, because it is based on the way having a range of independent observations increases the likelihood that an explanation fitting all of them will be true. When we look into how and why coherence is related to justification probabilistically, we see that the relationship depends on a rich understanding of probabilistic reasoning that we don’t have in the moral realm, such as the ability to assign conditional probabilities – the probability of A given B. In the moral realm, this would mean having a way to understand concepts such as “the likelihood that this murder is wrong given that killing innocent persons is wrong” in such a way that doesn’t reduce these to 0 or 1. But what could this be? If we define murder as the killing of innocents, it is 1. If we do not, we don’t know what the probability is; it seems impossible to say what this might mean. For this reason, I set the probabilistic approach aside. The alternative way of seeing coherence and justification is explanatory. In the science case, the explanation of the way that coherence can be justifying explanatorily is understood roughly along the following lines: (1) coherent theories are successful; (2) the best explanation of that success is truth; and so (3) coherent theories are likely to be true.22 We might consider something analogous in the moral case, saying, for example, that if a moral theory is successful, in some way open to further specification, the best explanation of that success is the truth of the theory. But when we try to use this argument to justify systematicity, we run into the problem that systematic moral theories have not been successful in the necessary way. The closest thing we have to successful systematized theories are utilitarianism and neo-Kantian theories focused on respect for persons. But in areas in which it is important to apply the right theory in practice, such as bioethics, these theories are not used. For example, both of the current dominant bioethics approaches – that of Beauchamp and Childress and that of Gert, Clouser, and Culver – are pluralistic in the sense of making use of multiple principles. This suggests that to some applied ethicists, at least, no richly coherent theory seems sufficiently intuitively acceptable to replace these pluralistic theories. Indeed, of the possibility of finding fully systematizing moral theory, Beauchamp writes, “every attempt to carry it out in moral philosophy – through rules or other constraints that, in effect, are absolute – has failed.”23

Epistemological Arguments and Rich Coherence

73

Some individual moral beliefs have been remarkably successful. In a different context, David Brink lists several items that he says virtually everyone now condemns morally: slavery, racial discrimination, rape, and child abuse.24 We might question how widespread this agreement is, but even where there is widespread agreement – about the wrongness of slavery and child abuse, the rightness of telling the truth, and so on – there is little agreement about why these things are right and wrong. There is much philosophical controversy over whether a unified theory might explain our shared beliefs, and if so, which is the right one, and there is substantive disagreement over the correct explanations of these beliefs themselves. For richly coherent theories that are consequentialist, this lack of agreement about reasons is perhaps at its most visible and dramatic in discussions of slavery. We agree that it is wrong. The consequentialist must explain why it is wrong on grounds that the benefits do not outweigh the harms. But even if it is true that the benefits never outweigh the harms, and even if it could be shown that the benefits of enslaving people never outweigh the harms, this explanation does not fit with what many people take to be the correct explanation, which is based on the injustice of ever enslaving someone – aside from benefits and harms. Evidence for the fact that people take the latter explanation to be the right one is found in the fact that in his criticism of utilitarianism Rawls is able to say that that view is mistaken on grounds that “it permits one to argue, for example, that slavery is unjust on grounds that the advantages to the slaveholder do not counterbalance the disadvantages to the slave and to society at large.”25 Rawls suggests that his readers will share his feeling that this should not be permitted: that is, the explanation for why slavery is wrong is not the right one. For richly coherent theories that are based on respect for persons, disagreement over explanations is more likely to arise in “murderer at the door” type cases, in which we can bring about much good by doing a small act that does not strictly seem respectful of persons. In the given example, the murderer is asking where your family members are so that he can kill them, and the moral conviction most of us share is that one ought to lie in this case. The disagreement, again, concerns why we ought to lie. In neo-Kantian theories like that of Christine Korsgaard, respect for rational autonomy explains why one must always exactly meet one’s “perfect duties” such as telling the truth and keeping promises, and why

74

Moral Reasoning in a Pluralistic World

one has leeway in meeting one’s “imperfect duties” such as beneficence (she calls this “mutual aid”). In Korsgaard’s view, the reason one may lie to the murderer at the door has to do with the murderer being a deceiver: applying the Formula of Universal Law, we get the result that one may lie to deceivers “to counteract the intended results of their deceptions,”26 and applying the Formula of Humanity we get the result that “you do not have to passively submit to being used as a means.”27 As long as it is permissible to lie, it can be one’s imperfect obligation in the circumstances to protect others, because of imperfect duties of mutual aid.28 That there is disagreement over this explanation is shown by the fact that for anyone who admits duties of beneficence as basic duties, the explanation will be different, because it will appeal directly to the bad consequences of telling the truth. This would include consequentialists, but also pluralists. To support systematicity inductively, there would have to be widespread agreement not only on certain moral beliefs but also on the unified reasons for those beliefs. This we do not have. So richly coherent theories have not tended to succeed, and unified explanations even of shared beliefs have not been widely adopted. Interestingly, analogous difficulties arise in the science case. First, as the longstanding debate over scientific realism shows many successful and coherent scientific theories have turned out to be false. Second, scientific theories are not themselves generally coherent in the rich sense: in solving problems we often appeal to different theories in different contexts. These difficulties with the explanatory connection will concern us here in the following sense: if the best defences of the coherence-truth connection in the science case rest on details that aren’t analogous in the moral case, then the piggybacking strategy fails. I’ll argue that this is just what happens.

3 many coherent scientific theories have turned out to be false Our questions, then, are to what extent the piggybacking strategy succeeds and to what extent it supports systematicity. In this section I consider defences of coherence relevant to a particular difficulty about success: how can we say coherence leads to truth when so many highly coherent theories – such as the theory of phlogiston or of ether – have turned out to be false?29 Again, my aim is not to evaluate the science context, but merely to examine whether and how the best

Epistemological Arguments and Rich Coherence

75

defences of the coherence-truth connection in the science case carry over to the moral case. What I take to be the strongest defence of the coherentist position against this concern in science is Paul Thagard’s. On Thagard’s account it’s not enough to have simple coherence; rather, explanatory coherence leads to (approximate) truth when it involves “deepening.”30 Roughly, a theory is deepened when we are able to find the underlying causal basis for a hypothesis. So, for example, the germ theory of disease has been deepened over the years by an understanding of just how viruses infect cells, replicate, and so on. Thagard says that all deepened theories have survived, and that the best explanation of this fact is that deepened theories are (approximately) true. Thagard points out that this conclusion fits with our understanding of nature as having various levels all interacting to produce particular phenomena. When we deepen a theory, we discover which mechanism at a lower level produces and is related to the mechanisms we already have an understanding of at a higher level, and the discovery of such a mechanism makes the truth of the theory much more likely. It’s not obvious how to carry these ideas over to the moral case, because they rest on a certain picture about theory survival, data, mechanism, and explanation: to survive, a theory must explain the data, and to explain the data is to describe a mechanism that produces the relevant phenomena.31 It is with respect to these causal mechanistic structures that deepening is understood: we find underlying features of reality at one level that function to produce effects at a different level.32 How should this idea be understood in the moral domain? The most plausible way, I believe, is to say that moral deepening involves a kind of uncovering of a general, fundamental, moral idea that explains and brings together a set of moral judgments. So, for example, moral beliefs about the general wrongness of lying would be deepened by the Kantian thought that one must always treat other persons as ends-in-themselves, and never as mere means. Moral beliefs about not causing unnecessary harm would be deepened by the utilitarian thought that one should do what brings about the best consequences for everyone. Moral beliefs that we have the liberty to conduct our lives as we see fit would be deepened by ideas about our rights to self-ownership and self-determination. But in contrast to the scientific context, moral theories deepened in these ways often do not survive. The belief that lying is always wrong,

76

Moral Reasoning in a Pluralistic World

regardless of the circumstances, is “deepened” by Kantian thoughts about never using a person as a mere means. The moral thought that the state may use eugenics to bring about a stronger, better race of humans is “deepened” by utilitarian considerations. The idea that we have no real obligation to help those in need is “deepened” by libertarian principles entailing that the only duty is that of respecting the liberties of others. In each case, a narrower and more specific moral belief is explained by a more general and more fundamental one. But none of these “deepened” theories has survived. We laud those who prevent crimes by lying to the criminal; eugenics is regarded as a clear moral violation; and it is widely thought that we have some genuine obligations to the poor and suffering, even if we disagree about the extent of these. I claim that the reason for this disanalogy concerns the pluralistic quality of the way we value – the descriptive fact of pluralism, not the normative thesis. If we value various kinds of things, it is not surprising that what may lead us to one set of beliefs when we deepen in one way may lead to a different set of beliefs when deepened in another. In the science case, we seem to have a unified world we are trying to describe – and, indeed, causal explanations are a unified form of explanation – but nothing like that holds in the value case. It might seem that deepening does play a role in our moral practice by helping moral theories take hold and survive. I’ve described the way beliefs changed regarding forced sex with a spouse: it wasn’t considered “rape,” and it wasn’t considered wrong, partly because rape was a violation of chastity, marriage was considered a “unity,” and women lacked independent autonomy rights in the sexual domain. In her discussion of this change as an example of the power of moral theorizing, Martha Nussbaum says that the judgments against marital rape were able to take hold in society partly because of the way they were embedded in fundamental beliefs about respect for autonomy and not using persons as mere means.33 But to me it seems the shift in values that precipitated this change is more complex than is suggested by a “deepening” interpretation.”34 In the article Nussbaum cites, the legal historian Rebecca Ryan discusses several kinds of changes leading to the decision to outlaw marital rape: the rise of individualism made dependency relations awkward to theorize; the shift away from seeing women as servant-like made it impossible to deny women individual rights of their own; forced sex

Epistemological Arguments and Rich Coherence

77

came to be seen as a violation of those individual rights rather than as a mere violation of chastity.35 It is true that part of this story concerns the expansion of the autonomy rights to a group including women, and I discuss expansion later on. But there’s a difference between this and deepening. For one thing, there is a difference in order: a deepened theory is already accepted, then is understood better through deepening; the idea that marital rape was wrong was not accepted until the expansion of individual rights made denying its wrongness seem inconsistent and contradictory. Indeed, while Nussbaum interprets the relevant themes explicitly in terms of Kantian moral “theory,” Ryan talks more broadly about individualism, equality, and feminism, and she emphasizes that even after it was obviously inconsistent to treat forced sex in marriage as legal, the law did not change until feminists brought enormous practical pressure to bear during the twentieth century. In any case, whether or not I am right in my interpretation these kinds of examples do not help with systematicity, and do not seem best understood epistemologically. To say that a patchwork of somewhat deepened or expanded moral ideas is better than a patchwork of narrower ones does not say anything about the importance of pursuing as few, simple, general principles as possible. Indeed, it would tell us nothing about how to choose between a patchwork with many fundamentals and one with fewer fundamentals, each of which exhibit similar degrees of expansiveness. But the essence of systematicity is that we ought to prefer the former. And to give us epistemological support, it would have to be that deepened theories tend to survive in general. The examples I’ve given of rejected “deepened” theories show this isn’t so.

4 science as a patchwork of laws Recall that the piggybacking strategy tries to buttress claims about the coherence-truth explanatory connection in ethics by appealing to analogies with science. Our questions are to what extent this strategy succeeds and to what extent it supports systematicity. In section 3 I focused on defences of the connection related to success; here I focus on defences of the connection having to do with difficulties related to the possible disunity of scientific descriptions of the world. The strategy against which I am arguing here, then, is to take defences of unity

78

Moral Reasoning in a Pluralistic World

or non-patchiness in our scientific descriptions, and to extend them by analogy to claims about unity of value and thus the appropriateness of systematicity. The best places to look for such defences is in responses to the work of Nancy Cartwright, since she has articulated with care the claim that science is disunified, non-coherent, a patchwork of disconnected laws.36 Our question is the extent to which her critics’ claims in favour of unity can be carried over. In explicating her view, Cartwright discusses an example of attempting to predict the eventual location of a dollar bill being blown about by the wind.37 “Mechanics,” she says, “provides no model for this situation,”38 since in practice it tells us only about the force of gravity. It is tempting to think that there is, in principle if not in practice, some model of mechanics that would allow us to model the force of the wind. It would follow that mechanics applies to the situation even if it does not produce for us a model of it. But Cartwright considers this an unjustified belief, based on an unsubstantiated “fundamentalist” idea that scientific theories must be universal in scope. We have no evidence for this fundamentalism, and if we were clear-headed, she suggests, we would see that what science gives us is a patchwork of distinct theories, some of which apply in some domains and some of which apply in others. The successes of a given theory show only that that theory is “true in its domain, not that its domain is universal.”39 There is, then, no reason to believe in the unity-of-science picture, in which multiple theories are reducible to a single one. Our best description of the world is inevitably a patchwork of various theories not reducible to one another.40 But we can see right away that the matter of universal applicability central to Cartwright’s idea is not relevant here, since the universal applicability of laws doesn’t really concern the issue of systematicity and rich coherence at all. If we have a range of theories which are nonreducible to one another but which all apply universally, the analogue to that in the moral case is a range of fundamental sources of value or basic principles none of which is reducible to one another but all of which apply universally. So criticisms of the dappled worldview that rest on showing that scientific theories do apply universally will not help support systematicity.41 A different response to Cartwright attempts to show that what seem like distinct theories really are not, because they are reducible to a single theory. Sheldon Smith interprets Cartwright’s claims about the dollar bill example as hinging essentially on a claim of distinct-

Epistemological Arguments and Rich Coherence

79

ness: point-particle mechanics cannot model the effect of the wind; fluid dynamics cannot model gravity; since these are distinct, no single theory can model the situation, even in principle.42 In an effort to “chip away”43 at the examples that motivate claims of a dappled world, Smith says that in this case at least the claim of distinctness is too quick, since here the equations of fluid dynamics can be understood as part of classical continuum mechanics (or CCM), which in turn is based on Cauchy’s laws, and point-particle mechanics is derived from Cauchy’s laws with all “contact forces set equal to zero.”44 That is, although fluid dynamics and point-particle mechanics seem distinct, they derive from the same set of general principles. In Smith’s view, this way of seeing the case does support a mild version of the Cartwrightian view of nature as described by a “patchwork” of laws, since it is true that where there are different materials at hand one may well need to make use of different sub-theories. But crucially, we may see this patchwork as having an underlying unity, since the sub-theories can be understood as parts of a more general theory. If the situation were analogous in the moral context, this might seem like support for systematicity, because if what seem like distinct theories are really sub-theories of a third more general one, this is a way of appropriately systematizing that is compatible with a certain patchiness at the level of concepts and fundamentals. The explanations in the more general theory would be more systematic than those in the sub-theories considered separately, and the evidence of dappling in examples would be removed. But such a line of thought would not succeed. First, for particular examples to have the necessary significance, the burden of proof would have to be on those arguing for patchiness and disunity: it is because the dollar bill example is thought to be evidence for a dappled world that the existence of an alternative analysis of it makes a difference. But we’ve seen already that the burden of proof here is on the rich coherentist, to cite positive evidence for systematizing. Second, as we’ve seen in section 2 the important examples of irreducible plurality in the moral case have not been amenable to this kind of unification; our most widely used theories in bioethics, for example, are pluralistic. Of course, it is open to the rich coherentist to insist that this is merely a failure of progress or education, and to say that if we were thinking more clearly and had a better system of moral education, we might converge on some highly systematic and unifying theory from

80

Moral Reasoning in a Pluralistic World

which our other moral judgments will be seen to follow. But this mere possibility gives us no reason to pursue systematicity now in moral reasoning, nor any reason for an individual agent to seek out or prefer a more systematic moral theory to a less systematic one. The force of this is apparent if we consider the dialectical role of moral coherence discussed in chapter 1: that other things being equal, saying that one theory is more coherent than alternatives should be a way of recommending it. Suppose a supporter of a newly discovered more unified theory encounters an interlocutor who values pluralistically. The dialectical role requirement means that the unified theory’s being more systematic should be a way of recommending it. But if we rest our claims about systematicity on theories we might develop in the future, why would it be? There would be nothing we could say in support of it.

5 coherence in mathematics Some philosophers, such as Michael Smith, make analogies with mathematics.45 Indeed, a piggybacking strategy with mathematics instead of science might seem particularly apt, since mathematics seems to have a kind of unified nature, and since in some ways moral reasoning is more like mathematical reasoning than like the empirical reasoning we do in science. Smith pursues such a thought in the slightly different context of desires. Smith says that we have a tendency toward coherence in our desires that is constitutive of rationality. Maximally coherent desire sets are “general” and “unified,” and it is because such desire sets make more sense than alternatives that we ought rationally to pursue them. I’ve argued elsewhere that in the desire case at least, there is no relevant tendency or pressure toward coherence; desires are not the sort of thing that ought to be coherent.46 But we might adapt Smith’s idea to the context of moral principles and judgments. In illustrating his idea about coherence, Smith makes use of the following example. Suppose someone has desires about the proper distribution of goods. If she desires that Adam, Bob, and Charles each get X, and that David get less than X, she may make her desire set more coherent by asking what it is that Adam, Bob, and Charles all have in common. If that quality – say, being in desperate need – is one that David has too, she should come to desire giving X to David, too.47 In

Epistemological Arguments and Rich Coherence

81

this example, her new desires are more coherent and thus “make more sense” than the set of desires she started with. Whatever we think about the desire case, this kind of example is intuitive in the case of moral beliefs: one’s moral beliefs do seem to make more sense when they apply similar criteria to various people, and it is easy to think that if one believed Adam, Bob, and Charles ought to get X, and David is in the same circumstances as they, one ought to believe that David ought to get X. But it appears that this doesn’t illustrate any norm having to do with “rich” coherence. It illustrates the norm of case consistency. Believing David ought to get what others do when he is in the same relevant circumstances is judging similarly cases that are similar in morally relevant ways; the reasoning is compatible with the existence of multiple norms applied consistently as well, and does not rest on finding as few simple principles as possible. At some points, however, Smith suggests that the important item here is not the weaker norm of consistency but the stronger one of “generality.” We can understand this more fully through his exchange with Geoffrey Sayre-McCord. In a critical discussion, Sayre-McCord points out that making one’s desires more general does not always seem to make them make more sense. He asks us to consider: suppose he enjoys coffee ice cream and thus desires to eat it. Would his desire set make more sense if he developed a desire to eat all kinds of ice cream? It seems silly to say that it would. In reply Smith says that the difficulty here is not generality per se, but rather that the case has been generalized in the wrong way: what would make sense is to say that one should eat what one enjoys eating. “Adding this desire,” he says, “will indeed ensure that [the person’s] overall desire set makes more sense, and so make it less liable to reasoned criticism because if the new desire were added he would no longer be liable to the charge of making arbitrary distinctions.”48 As we’ve seen, though, it is difficult to know when one has generalized in the right way, and because of this, we do not know when more systematized theories are better than less systematized ones. With respect to these kinds of deeper questions about justification and why coherence is related to justification, Smith, like Daniels, pursues a kind of piggybacking strategy – except that Smith makes his analogy with “a priori beliefs” rather than science. Coherence, he says, isn’t so much explained as assumed in the a priori case: there is pres-

82

Moral Reasoning in a Pluralistic World

sure toward making our a priori beliefs coherent, and perhaps we can say the same about moral beliefs. When considering a priori reasoning mathematics forms an obvious starting point, and Smith suggests that mathematics provides an apt analogue for ethics.49 But I claim first that simple arithmetic is not the right thing to compare with here, because coherence in simple arithmetic can be understood simply as logical consistency: the elements of simple arithmetic, after all, can be captured in purely logical terms. Since our concern here is with rich coherence, logical consistency is not what we’re after. A better analogy would be with reasoning about the infinite. Indeed, infinite set theory seems to involve just the kind of convergence process that we’re interested in, and it works in a highly specified and straightforward way: there are axioms, which lay down generalities about what sets exist and what they are like, and from these we deduce particular set-theoretic conclusions. In set theory we have axioms from which we prove facts about sets. As in the moral case, there is a kind of “reflective equilibrium” at work because one thing that guides us in axiom selection is our intuitions about particular cases.50 This raises the question of just what kinds of coherence – beyond logical consistency – are important in the mathematical domain, and why. The most striking aspect of the answer in the present context is that there is virtually no pressure for systematicity: reducing the number of basic principles from several to one isn’t something set theorists feel moved to do. It is true that from a practical point of view, a very long list of axioms would be unworkable, but this is so in the moral case as well and does not support monism over pluralism. On a standard conception of set theory, there are nine basic axioms (some of which are axiom schemas), and while it is true that there is a single basic conception relation, that of membership in a set, these axioms are not seen as problematic for being multiple. So systematicity and rich coherence do not themselves get support from this particular analogy.

6 generality, consistency, and systematicity It might seem that support for systematicity can be found indirectly, through generality. Recall that to be sufficiently general enough to be

Epistemological Arguments and Rich Coherence

83

a “general principle” means being formulated with no names or indexicals. But some principles are more general than others, and highly general principles apply the same single criterion to a large range of cases. So, for example, the principle against lying to one’s siblings is less general than the principle against lying to one’s family members, which is in turn less general than the principle against lying. It may be thought that the more general our principles are, the fewer of them there will be since each will cover a wider range of cases. This, I believe, is how people have often thought of generality: the hope is that the more general one’s principles, the fewer of them one wants or needs. This would give us an indirect argument for systematicity. It is generality that matters, not systematicity; but generality, together with the demand that we avoid principles that offer conflicting directives, leads naturally to reducing the number of basic principles one endorses. But this line of thought assumes that at most one principle can apply to or cover any given case, because otherwise, we may have multiple, highly general, conflicting principles, and there would be no support for systematicity. Therefore, it rests on our having a norm against conflict. But as I’ve already argued in chapter 1, there is no reason to have such a norm. The principles against lying and in favour of promise keeping are both quite general, but they both apply to Marie’s case. A principle in favour of promoting the collective good is very general, but it applies to many cases, such as the testing case, alongside other principles. That principles can conflict means that, in the absence of other arguments for systematicity, the generality of principles does not tend toward our wanting or needing fewer of them. Notice, too, that this conclusion is supported by the reflections on the science and mathematics cases above: when we reason about the external world or the mathematical one, we do take multiple principles to cover the same cases, as when we say that gravity and laws of optics or the axioms of extensionality and pairing both apply to the same cases at the same time. Nothing about such situations undercuts the sense that both principles are apt. I discuss the status of “generality” as a norm further in chapter 5; my point here is that even if there were a requirement to generalize principles, this would not give us a way of supporting systematicity.

84

Moral Reasoning in a Pluralistic World

4 Case Consistency

I’ve argued that the way we value is pluralistic, that if values are plural the best moral theory will be a theory with multiple principles engendering conflicts, and that there is no epistemological reason to systematize our theories by seeking out as few simple principles as possible. This means that rich coherence is inappropriate for pluralist contexts. In this chapter and the next, I develop an alternative conception that I call “pluralist coherence.” The central norm of pluralist coherence is case consistency, or judging in accordance with morally relevant similarities and differences. Here I explain case consistency: in sections 1 through 4 I explain its nature, and in sections 5 and 6 I address two objections to basing a coherence method on it. The first objection to basing a coherence method on case consistency is that it is too complex and open-ended to be useful, and the second objection is that it allows for too much arbitrariness in our moral judgments. It might seem that the best way to be consistent is simply to rank principles, for example in a lexical ordering: judgments from higherranked principles would always override judgments from lowerranked ones. But in section 7 I explain why case consistency does not require a ranking of principles and, in general, the pursuit of consistency will not lead to a ranking.

1 what is case consistency? A person’s moral beliefs are case consistent insofar as she judges in accordance with morally relevant similarities and differences: she judges morally relevantly similar cases similarly unless she can cite a

Case Consistency

85

morally significant difference between them. A simple and paradigmatic example of case inconsistency is that induced by framing effects. When the same thing is described in two different ways, people judge differently. This is case inconsistency, because the only difference between two cases is the words used to describe them, and virtually everyone considers this a morally insignificant difference. In a recent paper on the effects of indignation on judgment, Daniel Kahneman and Cass Sunstein discuss examples of effects that cause people to judge differently cases that are identical in all the facts, just because of the way the cases are presented. For example, when a tax policy is described in terms of exemptions it seems unfair, but when described in terms of benefits seems fair.1 When cases are presented to juries much depends on how the case is described, and people judge one way under one description and another under another simply because of the way their emotions are engaged. Using the term “coherence” much as I understand “case consistency” here, Kahneman and Sunstein summarize by saying, “because of how indignation operates, it is extremely difficult for people to achieve coherence in their moral intuitions.”2 A set of principles and judgments can be case inconsistent in at least two ways. First, the principles can be structured so that they apply in different ways to cases that have no morally significant difference between them. “Telling a lie at home is wrong” and “Telling a lie outside is permissible” are case-inconsistent principles, if one regards “being at home” versus “being outside” as a morally insignificant difference. Second, in cases of conflict, judgments about one’s overall obligation in conflict cases can be made in case-inconsistent ways. Consider the case in which one must lie to keep a promise. If I judge that I ought to lie to keep promises when the promisee is blond, but ought to break them and tell the truth when the promisee is brunette, then I have a set of principles and judgments that is case inconsistent, if one regards the hair colour distinction as introducing a morally insignificant difference. The interpretation of consistency as case consistency fits with our use of the word “consistent” in ordinary contexts. A person who makes a special exception for herself or her friend is frequently charged with being inconsistent. Inconsistency in the sense of moral inconsistency is most commonly associated with the failings of hypocrisy and

86

Moral Reasoning in a Pluralistic World

making special exceptions where none are warranted, rather than with moral conflict. The pursuit of case consistency forms the basis for a commonly used method in moral reasoning. One of the most common ways we criticize the views of others is by pointing out inconsistencies of this kind in their views: we frequently accuse others of judging similar cases differently with no explanation of what the moral distinction is. We assume that this sort of criticism is intended to show a problem with the person’s moral outlook; frequently it is accepted and acknowledged as such, and the person against whom the charge is levelled responds by either trying to show that his distinction is morally significant or changing some of his moral judgments. Eliminating case inconsistencies in one’s moral beliefs is thus already considered a widely acceptable way of moral reasoning and improving one’s moral views.3 Finally, case consistency is compatible with value pluralism in a way that rich coherence is not. A demand that one person judge similarly cases that are morally similar does not require anyone to unify their values or systematize the principles they endorse. It requires only finding a way to prioritize among values and principles in ways that are consistent from one case to another. As I explain below, case consistency need not – and typically should not – be understood as requiring a ranking or lexical ordering of principles. It may be appropriate to lie to keep a secret in some circumstances and not in others; as long as there are morally significant differences between the cases, obeying the one principle in one case and the other in the other is no threat to case consistency.

2 the nature of case consistency: preliminaries To explicate in more detail the nature of case consistency, I start with a few preliminary observations. The first concerns the difference between chronological coherence and coherence at a given time. In his characterization of coherence as it functions in value pluralistic contexts Alan Goldman says: “one must not judge a case differently from a previously settled case unless one can cite a morally relevant difference between them, a difference that can be shown to make a difference in other cases as well.”4 I think this is mostly right, but I pro-

Case Consistency

87

pose an amendment to the idea of “previously settled” cases. It is not being consistent with previously settled cases that matters: what we want is to say what it is for a set of beliefs to be case consistent at a given time. This is because it is essential that a person can remain coherent even while she changes her moral beliefs. When a person’s judgments and principles fail to fit, coherence requires only that either one’s principles or one’s judgments be altered; it does not tell us which. Indeed, agnosticism about this is implicit in the metaphor so often associated with coherence methods: we work back and forth between our principles and our judgments. Encountering new situations, reacting to them in certain moral ways, and judging them differently from how we expected to is a normal part of moral change. What we need is an account not of consistency through time, but of how to say whether a set of judgments is case consistent at a specific moment. A second observation has to do with the nature of the requirement. It may seem that an objection to case consistency is that there is no general obligation to do the same thing in different circumstances. As has been pointed out, we do appropriately treat cases differently even when there may be no morally significant differences.5 For example, in the context of perfect and imperfect duties, George Rainbolt discusses the following kinds of cases. If a person gives money to one charity and not to another, she is clearly not violating moral coherence, and if a person risks her life to save one drowning person but fails to save another, to call her action morally wrong seems absurd. The drowning example suggests something about “supererogatory” acts, or acts that are morally admirable without being morally required. Doing what is above and beyond the call of duty in one case does not require anyone to do what is above and beyond the call of duty in other cases, or risk violating case consistency. The charity example suggests something about “imperfect” obligations, or those we have substantial leeway in deciding how to fulfil. The fact that we fulfil our obligation in one way at one time creates no obligation to fulfil it in the same way another time. Rainbolt says that these kinds of categories, as well as the existence of simply “permitted” acts, shows that there is no general requirement to “treat like cases alike,” and suggests that this casts doubt on norms like that of case consistency.6 However, these kinds of cases show not that we should reject case consistency, but rather that the most apt interpretation of case consis-

88

Moral Reasoning in a Pluralistic World

tency in the moral context is not in terms of “treating” but in terms of “judging.” One must judge similarly in similar circumstances – judge, that is, what kinds of acts are “required,” “forbidden,” “permitted,” and “supererogatory.” If we judge both saving A and saving B to be morally admirable without being morally required then we are case consistent, and an agent who saves only A has not failed to be case consistent – she has simply done something above and beyond her obligations. In this way the fact that we have leeway in making moral choices need not count against case consistency; rather, we need only be consistent in the cases we judge to permit such leeway. A third observation concerns morally relevant similarities and differences. In his 1963 book Freedom and Reason R.M. Hare argues that moral judgments must be “universalizable” in the sense that they must apply in the same way to all cases that “we admit to be exactly or relevantly similar.” This, of course, resembles my consistency requirement. But since, as Hare acknowledges, no cases are really exactly the same, the idea of morally relevant similarities is a crucial one. How should it be understood? Hare’s answer appeals to the idea of interchangeability: “All we have to do is imagine an identical case in which the roles are reversed.”7 If an action has positive implications for me and negative implications for another, I should evaluate the action by imagining whether I would accept the negative implications if our roles were interchanged. This, Hare says, will force us to “count as morally relevant only those properties which [one] is prepared to count as morally relevant even when other people have them,” and thus rules out “special pleading” for one’s own case.8 I claim, though, that this kind of interchangeability test does not go far enough to determine case consistency. First, case consistency requires judging morally cases that are morally similar, regardless of the relative benefits to ourselves and others. Suppose I take a positive stand on a particular policy on universalist grounds, willing to endorse it whether I stand to gain, or, interchanging myself hypothetically with another, whether I stand to lose. Then suppose I hear the policy described in a different way and take a negative stand that is also universal. My judgments may be consistent with interchangeability, but they are not case consistent. So case consistency requires a more complex analysis. Second, as I discuss in more detail below in section 4, interchangeability assumes, without argument, a certain

Case Consistency

89

form of relevance: that all persons are morally relevantly similar. This shows that to understand the role relevance plays in case consistency, we must go beyond interchangeability.

3 the nature of case consistency: “moral factors” The natural direction suggested by the preceding discussion is that we start with a “moral factor” view like that suggested by some work – in quite different contexts – by Shelly Kagan and David Brink.9 In describing the general idea, Kagan says that “the moral status of an act is a function of the various (genuinely relevant) factors that come into play in the given situation.”10 In the context of moral conflict, Brink says “to determine [overall] obligations, we must do moral factor addition.”11 In the moral factor view, different moral considerations present different moral factors that must be taken into account in deciding what is best to do overall. A full account might include various kinds of moral factors that represent not only duties but also permissions, rights, and ideals, but I’ll just focus on obligations. These come in various strengths, which we might call “valences.” For obligations, valences are like degrees of oughtness – they reflect the strength of the obligation in question. An act that must be done has a high valence; an act that one ought to do but is of minimal importance has a low valence. In cases in which the principles entail a set of obligations, what one must do “overall” is meet the obligation with the highest valence. If there is no conflict, this trivially entails simply meeting one’s single obligation. If there is a conflict, this entails acting on the obligation that is most pressing or important overall. In the case of obligations, the metaphor of trumping seems more apt than that of a function of factors, though we might want such a function for the more general case. This much of the story is familiar from the Rossian picture we’ve already encountered: we have a multiplicity of principles that can generate conflict, and an idea that when obligations conflict, we ought to figure out which is our most pressing or “actual” obligation. But now we add case consistency, which requires assigning the same valences to obligations that are morally similar in significant ways. Certain kinds of duties of honesty have high moral valence, and case consistency requires that similar kinds of duties of honesty have high

90

Moral Reasoning in a Pluralistic World

moral valence in other situations. The duty to keep certain kinds of promises has high moral valence, and case consistency requires that duties to keep similar kinds of promises have high moral valence as well. With respect to the question of how we assign valences to obligations, recall that the coherence method is here understood as arising from within conviction ethics – that is, from within the general approach in which we start moral reasoning from within a set of moral judgments we already have. It is natural to say that part of the “moral judgment” we start with is a judgment about how strong any moral requirement or prohibition in question is. In any moral system in which there are conflicts among principles, we ought to take these convictions as telling us both what ought to be done or avoided, and how much it ought to be done or avoided. This is intuitively plausible, given that we often make moral judgments about how strong our obligations are.12 The method of coherence in the context of conviction ethics does not require that we know ahead of time all the valences of our obligations, because part of what we are doing in “working back and forth” among our principles and judgments is seeking out a case-consistent – and thus coherent – way of judging how strongly our obligations apply. The method requires only that we begin with some rough judgments of valences – just as we begin with various judgments of other kinds. In a pluralistic context, one way to work out what valences some moral set of judgments is committed to is to “read off” the valences from the judgments we make in cases of conflict. That is, if in one situation we judge honesty much more important than the demands of friendship, we can infer that in that kind of situation, we ought to assign a higher valence to the demands of honesty and a lower one to the demands of friendships. In his work on conflict in bioethical reasoning, Henry Richardson has suggested that “reading off” the relative valences of moral duties from judgments in cases of conflict is a justificatory dead-end, on grounds that it is intuitive and ad hoc: the judgment about prioritizing merely restates one’s judgments about what is best overall, and thus cannot do any justificatory work.13 I discuss this objection more fully in section 5 below, but notice here that at least in the context of conviction ethics, reading off introduces no new difficulties in this direction over those inherent in the method, which is already com-

Case Consistency

91

mitted to taking our moral convictions as a good starting place for moral reflection. If the judgment that telling a certain lie would be wrong can be a part of that starting place, then so can the judgment that telling a certain lie would be very wrong – indeed, so wrong that the demands of friendship do not counterbalance it.

4 the nature of case consistency: morally significant differences The main difficulty with case consistency is, of course, knowing what are, and are not, morally significant similarities and differences. I claim here that judgments of moral significance are themselves also moral judgments, and thus on the conviction ethics model must also be accounted for as other moral judgments are.14 To see this, notice first that to say what constitutes a proper generalization and what constitutes a proper exception to a generalization always requires a set of beliefs about relevance. Donald Hubin explains this in the context of desires by pointing out that the fact that something will cause me pain, rather than you, is certainly significant from the prudential point of view, though it may or may not be from a moral point of view.15 To ask simply whether the distinction is significant makes no sense; we must ask whether it is significant from some set of beliefs about relevance. In the moral case, it is the moral point of view that matters. And beliefs about moral relevance have moral content; they are beliefs about what matters and why. Thus they are naturally considered alongside other moral beliefs. It follows that on the conviction ethics model they are just some of the beliefs we are trying to bring into coherence with one another. For example, if I tell a harmful lie to my great-aunt while standing on a footstool and I judge that my action is morally reprehensible, this is clearly not support for a moral rule against talking to my greataunt, or against talking to her while standing on a footstool. The case isn’t generalized in the right way: the cases covered by the principle are similar in morally insignificant ways. It may be support for a moral rule against lying, or it may be support only for a moral rule against telling harmful lies; which it is depends on other elements of our moral point of view. Conversely, if I tell a lie to my great-aunt to save a life and I judge that my action is morally permissible, I do not

92

Moral Reasoning in a Pluralistic World

know whether my judgment violates the standards of case consistency until I know whether “done in order to save a life” is the kind of thing that creates a morally significant difference between cases. As we saw briefly in the discussion of Hare in section 2, it is sometimes thought that the basis of consistency is interchangeability. And it might seem that because of this, something like the golden rule can be supported by pure reason. Hare himself says that the requirement of universality can be deduced from the logic of moral language alone, on grounds that judging an action to have a property rationally requires judging it to have that property in all similar cases. Others have had similar ideas – saying, for example, that “whatever rule any agent or judge applies to other persons he must also apply, or be willing to have applied, to himself, and conversely,”16 that “what is right for one person must be right for any similar person in similar circumstances,”17 that any moral agent must see himself as “merely one person among others ... [while others are] persons in just as full a sense,”18 or that rational moral demands “cannot depend on the egocentric vantage point of the reasoner.”19 Even here, though, these judgments rest on an understanding of just which similarities and differences between persons are morally significant and which are not. Imagine Daniella has some particular obligation to give aid to David in virtue of something specific: perhaps David is in some terrible trouble, or perhaps David is her son. If the obligation arises because David is in terrible trouble, we would not want to say that Daniella is bound by coherence to give aid to everyone else in the world, whether they are in trouble or not. If obligation arises because David is her son, we would not want to say that everyone else in the world is bound to give aid to David, whether they are related to him or not. Rather, it seems a proper interpretation of the interchangeability principle would go something like this: in the first case, regardless of who you are if you, too, are in some terrible trouble Daniella must help you, and in the second case, if you, too, are Daniella’s son, she must help you. But once we are creating such hypotheticals, only information about moral relevance can tell us where we may begin and end with them. While it would be appropriate to say “he is not in need; I don’t bear these obligations” or “I’m not his mother; I don’t bear these obligations,” it does not seem appropriate to say, “I’m an especially intelligent and sensitive person; I don’t bear these obligations,” or “He

Case Consistency

93

only has one leg; I don’t bear these obligations.” This is because being or not being in need, and being or not being a person’s parent, are morally significant differences between persons in this context, while being or not being intelligent and sensitive and being or not being disabled do not. The reason that “telling a lie at home is wrong” and “telling a lie outside is permissible” are case-inconsistent principles for most of us is that most of us take location to be morally insignificant – we believe that where a person is located doesn’t matter to the moral reasons against telling lies. Case consistency would require a person who agreed with that assessment and who endorsed both principles either to change her principles or to change her judgments of what is morally significant. It does not, however, tell us which of these to do. In some places Hare suggests that interchangeability, rather than resting on judgements about similarity, can inform us about similarity. For example, to determine whether “the fact that I have a mole in a particular place on my chin entitles me to further my own interests at others’ expense,” Hare says I should “suppose that my mole disappears, and that my neighbour grows one in the very same spot on his chin.”20 I should learn that having the mole is not morally significant from the fact that I do not think this justifies my neighbour’s special entitlement. Hare’s views are embedded in complex metaethical views in such a way that a detailed discussion will have to wait until chapter 5 (see section 8). But it is worth noting here that in the present context, conclusions about special entitlement cannot follow from such thought experiments because conclusions about special entitlement are necessary to carry out the thought experiments. Consider the question of whether being of royal lineage entitles a person to pursue her interests at others’ expense. Monarchists, basing their judgments on the moral significance of royal birth, will say yes. Republicans, basing their judgments on the moral insignificance of royal birth, will say no. The interchangeability test relies on, and thus cannot determine, judgments of moral relevance. With respect to conflict, case consistency is an appropriate way of understanding what it is to have a “principled” way of judging in the context of multiplicity. In the context of multiplicity, having a principled way of judging plausibly means having a way of appealing to multiple criteria in ways that makes sense – that is, we do not privi-

94

Moral Reasoning in a Pluralistic World

lege one principle in one context and another in another for no good reason. But being able to cite a morally significant difference between cases is a good reason for privileging one principle in one context and another in another. Indeed, to say that there can be morally significant differences between cases is to say that there are such reasons. In their defence of a pluralist approach in bioethics (the “principlism” of Beauchamp and Childress), John-Stewart Gordon, Oliver Rauprich, and Jochen Vollman present a (hypothetical) case involving an elderly woman who is quite sick and wants her feeding tube removed so that she can die peacefully.21 The doctor treating her refuses, on grounds that his duties of non-maleficence toward her include making sure that she is fed. In judging what is best to do in this case overall, the fact that the woman has already lived a long life is relevant, as is the fact that she wants to die. As the case is presented, she is mildly depressed; clearly her state of mind is relevant to this decision. Things that we believe ought to be irrelevant include whether it would be more convenient for her neighbours if she were to die sooner rather than later. Again, these judgments about moral relevance are judgments about what matters and why; they are thus moral judgments we are trying to bring into coherence with other judgments. Our judgments about moral relevance are typically embedded in our overall moral systems and ways of valuing in complex ways. For example, consider the change mentioned already from thinking of sex in terms of chastity, purity, and appropriateness – and thus denying the possibility of marital rape – to thinking of it in terms of consent and autonomy and thereby showing marital rape to be morally similar to non-marital rape. In the former way of valuing, the fact that a woman is married to a particular man creates a distinction morally relevant to whether we judge a crime to have been committed against her in forced sex. This distinction is morally relevant because it distinguishes appropriate or sanctioned sex from that which is not. A moral view that takes autonomy as fundamental judges that the distinction is irrelevant. What matters is the woman’s autonomous consent, and whether or not she is married to the man who rapes her is not a morally significant difference. Notice that to judge whether a person ought to lie to keep a promise, we typically need to know several morally relevant features of the case before we can decide. What is the nature of the promise? What is the

Case Consistency

95

nature of the lie? Who will be harmed in breaking the promise or lying? How much harm will be caused? What were the circumstances under which the promise was made? What is the relationship among the people concerned? Of course, people differ in their judgments about what are and are not morally relevant similarities and differences. This is a cause of much moral disagreement. Because of this variation, this conception of case consistency allows us to judge a person’s moral beliefs for case consistency from the point of view of her own beliefs about relevance and from the point of view of someone else’s beliefs about relevance. When we undertake the pursuit of moral coherence as individuals subjecting our beliefs to coherence demands, we rely on our own standard of relevance; when we criticize others we might want to use a standard that is essential to us but external to them. This means that a set of beliefs can be internally case consistent but case inconsistent from the point of view of another set of moral beliefs. If not otherwise specified, I’ll use “case consistent” to mean “internally case consistent.” My proposal differs from the one based on Kantian unification proposed by Audi. Recall from chapter 3 that Audi proposes adjudicating conflicts by considering first, whether our action falls under a maxim we can universalize, and second, whether our actions treat persons as mere means. Asking whether a certain adjudication can be universalized can be a good practical strategy for the pursuit of case consistency: it tests whether one would endorse applying the same weights for different people. Assuming we take being a person as a morally significant difference (as many of us do), this is a good test of case consistency. Universalizing, however, can be complicated. Kantian-inspired tests do not ask about the morally significant consequences of universalizing, but rather about logic of it. This is why Audi says that breaking promises to take care of sick children would not “undermine” the practice of promising. But what is it about the sick children situation that shows that this is so? If the answer to this question has to do with the reason for breaking the promise being a good one, or being a reason the promisee can understand and share, then the universalization is implicitly appealing to considered convictions about what outweighs what and when. In this case the reasoning rests implicitly on already having a principled prioritization. Similar remarks apply to the question of whether breaking a promise in the given case “uses” the promisee. How

96

Moral Reasoning in a Pluralistic World

can we judge whether it does? If this judgment rests on whether the promise is broken for a good reason, or an acceptable one, rather than a poor or unacceptable one, then again the reasoning rests implicitly on already having a principled prioritization. Recall now from chapter 1 the debate over the double standard for medical testing in developing countries. In her treatment of this issue, the ethicist Ruth Macklin frames the problem in terms of morally relevant distinctions, just as an appeal to case consistency would suggest.22 She describes the various values at stake, including respect for autonomy and persons, justice, and the promotion of the collective good.23 She acknowledges that participants in the debate share these values and share an interest in providing helpful treatments for diseases such as malaria that primarily affect people in developing countries.24 She explains how these shared values nonetheless issue in a double standard in which people in developing countries face greater risks in tests. She asks whether the difference between the developed countries and the developing countries is “morally relevant” in a way that would explain how the double standard does not violate consistency.25 In the end, she judges that justice requires doing away with the double standard after all: the cases are not different in the right ways to justify the differential treatment.26 The coherence norm in play here is case consistency. The proposal so far is open to two immediate objections. First, it may be said that the pursuit of case consistency as here explained is too complex and open-ended to be useful. Second, it may be said that coherence based on case consistency risks too much arbitrariness in our moral judgments – arbitrariness that could be avoided by rich coherence. After all, anyone can make anything case consistent by introducing and endorsing the right kinds of morally significant differences. I address these two objections in sections 5 and 6, respectively.

5 is case consistency a useless norm? It might be thought that this account of case consistency is too complex and open-ended to be useful. Can valences of moral obligations and beliefs about moral relevance ever be clear and specific enough that case consistency would help us figure out what to believe? But in fact a norm of case consistency forms the implicit basis for a range of work in applied ethics. I’ve already mentioned some reasoning

Case Consistency

97

on the “double standard” problem and how it makes use of a norm of case consistency. As a more representative sample of work in applied ethics, consider three papers on abortion that are so widely admired as to be included in virtually every anthology and taught in a vast array of courses on moral problems: Mary Anne Warren’s “On the Moral and Legal Status of Abortion,”27 Don Marquis’s, “Why Abortion Is Immoral,”28 and Judith Jarvis Thomson’s “A Defense of Abortion.”29 Warren responds to the traditional argument that abortion is killing an innocent human being and is therefore wrong. She argues that “being a human being” is not a morally relevant similarity between cases, while “being a person” is. One reason she gives for this is that we would judge the moral status of new beings we encountered not by their genetic similarity to us but by their consciousness, their abilities to communicate, and other like qualities. She thus deploys a strategy of “case consistency” to recommend a particular set of morally significant similarities and differences. Don Marquis’s “future-like-ours” argument asks us to judge similarly cases that are morally similar and argues that having a certain kind of future is a morally relevant similarity. His strategy is as follows. Starting with the obvious fact that it is wrong to kill “us,” he asks why. The answer he gives is that killing us deprives us of a future. Since killing a fetus deprives it of a future as well, abortion is generally wrong. Again, this deploys a norm of case consistency, asking us to acknowledge a morally relevant similarity and to judge consistently in accordance with it. In her much-discussed “violinist analogy,” Judith Jarvis Thomson presents the following thought experiment. Imagine you are kidnapped by the Society for Music Lovers, and attached to an ailing violinist. You are keeping him alive, and if you detach yourself he will die. Thomson argues that in such a case we would generally judge that one is not morally required to stay attached: it would be kind and perhaps decent to do so, but one is not violating anyone’s rights in getting up and walking away. Analogously, in certain circumstances a woman who is pregnant may also “detach herself” from her fetus, even if it means that the fetus will die. Again, case consistency explains the reasoning of the article: we are presented with a claim of morally significant similarity, and an absence of morally relevant differences, and asked to judge similar cases similarly.

98

Moral Reasoning in a Pluralistic World

None of these standard arguments rests on hypotheses about rich coherence or a single source of value. All are compatible with the possibility of value pluralism and principle pluralism; and all ask us to follow a certain understanding about moral relevance. Thomson’s argument, in particular, seems explicitly pluralistic in that it asks us to consider how we weigh different moral elements, such as a person’s autonomy and the needs of dependent beings, against one another in different settings. Since these are paradigm cases of arguments considered useful in applied ethics, they show that case consistency is not a useless norm. One might object that these are not actually useful arguments, on grounds that they have not “settled” the abortion debate or induced consensus, and that this is so because they appeal to differing basic judgments about significance and about valences. But my point here is not that they are useful for creating consensus, but rather that they are useful for showing us justified ways that certain judgments fit more coherently with other judgments. That is, the open-endedness and complexity that arise in deploying case consistency are no impediment to its being useful in this sense. Furthermore, when it comes to consensus and the relationship of theory to public discourse, these arguments are at least as useful as those grounded in unified and systematic theories, which have also failed to produce consensus on the abortion question. Evidence for this claim is that while both sides commonly engage with the arguments about personhood, rights, and so on, non-philosopher pro-choice advocates seldom appeal to prochoice conclusions from utilitarians such as Singer, and non-philosopher pro-life advocates seldom feel a need to challenge or respond to such conclusions, either. This lack of broad engagement with utilitarianism shows that in the abortion debate, a theory’s being unified and systematic does not make it more useful. That these three arguments are widely admired does not itself show that their methods are justificatory. What it shows is that the norm of case consistency is not too open-ended or complex to be useful in practice: these are cases in which it is used in practice to good effect. The fact that we appeal in these judgments to different basic assumptions about significance and weights raises again the difficulties mentioned above about justification and non-arbitrariness. I discuss these further in section 6.

Case Consistency

99

Henry Richardson argues that assigning weights to conflicting obligations – a method sometimes called “balancing” – is unjustifying and concludes that balancing ought to be replaced by “specification.” According to Richardson there are two kinds of balancing: intuitive and justified. Intuitive balancing should be rejected, he says, and the considerations involved in justified balancing are better expressed through the framework of specification. Justified balancing occurs when we can give “articulable reasons” for why one obligation seems to outweigh another in a given case, and intuitive balancing occurs when we cannot give such reasons. The problem with intuitive balancing is that when we use our intuitions to judge how conflicting oughts should be weighed against one another, we are not really reasoning morally – there is no way of justifying the outcome. Suppose we try to argue that in a given conflict, in which we ought to do A and we ought to do B, we ought to do A overall, and suppose we try to justify our choice using balancing, reasoning that A overrides B because A is a stronger obligation. In that case, he says, we aren’t really justifying our choice at all; we are simply “reading off” the weights from our judgment that in the given case, what we ought to do overall is A.30 But in the context of conviction ethics there is nothing improper about inferring the valences or weights of obligations or concerns from the judgments we make in cases of conflict. The method of pluralist coherence also commits us to preferring “justified” over “intuitive” balancing, but allows that reasons can take the form of convictions that one concern outweighs another in a given case. As an illustration of this, consider an example Richardson uses. Imagine a defence lawyer in a rape case. The client tells his lawyer he is guilty, but insists on offering as defence that the victim consented; investigation turns up a bitter ex-boyfriend of the victim who would be willing to testify that she had consented to sex with many people in the past. The lawyer suspects that sexist members of the jury, believing that sexually experienced women have bad characters and can’t be trusted, would be caused by such testimony to doubt the victim. Should the lawyer use the testimony of the ex-boyfriend and crossexamine the victim about her sex life? Richardson says that on the one hand, it seems that lawyers ought to use all lawful means to pursue their clients’ interests, but on the other, it seems wrong to cross-examine rape victims about their sex lives, especially if the intent is to play

100

Moral Reasoning in a Pluralistic World

into sexist stereotypes about sexually active women being untrustworthy.31 Richardson says that this lawyer faces a conflict of norms: it would be wrong to cross-examine the victim this way, yet it is also required insofar as the lawyer ought to do all he can to get his client acquitted. In Richardson’s suggested analysis using specification, we reason first that lawyers are only required to do what is ethical in pursuing their client’s interests, and second that rape victims need special protection for their allegations, “just as the accused is due the adversarial protections of the presumption of innocence.”32 Richardson presents the described reasoning as “abbreviating” the sophisticated casuistical discussion of David Luban in his Lawyers and Justice: An Ethical Study.33 In his text Luban offers several reasons why a lawyer ought not cross-examine a rape victim in the suggested way. As we see in the example, in rape cases, it is the victim who is on trial: as Luban says, “the defendant is confronted by the state, but the victim is confronted by the millennia-long cultural tradition of patriarchy.”34 Because of this there is a moral symmetry between the accuser and the accused not usually present in criminal trials. The process must ensure that women will not generally be afraid to make rape accusations, and such cross-examination would undermine this goal. Luban says that these considerations apply in the same way whether or not the accused is guilty: “balancing the defendant’s rights against the rape accuser’s rights in order to determine the moral bounds of zealous advocacy must be done without considering either the defendant’s guilt or the accuser’s innocence.”35 Richardson presents Luban’s process as an example of the kind of reasoning that goes into justifying a particular specification – a particular way of narrowing the principle so as to avoid conflict with other principles and show what ought to be done overall. But I claim that with respect to methodological norms of justification Luban’s process is most aptly seen as a process of pursuing case consistency in a context of conflict. Luban argues that there is a certain symmetry between the accuser and accused that is not typically present but is in cases of rape, and says this creates a symmetry of rights; he argues that the interest we share in making sure rape victims come forward is morally significant in weighing on the side of not allowing such cross-examinations; he writes of balancing the rights of the accuser and the accused. In such cases, the process of finding reasons is the

Case Consistency

101

same as that of pursuing pluralist coherence among conflicting moral considerations. At least in the context of conviction ethics it is hard to see how there could be other ways of providing reasons. In this context, then, the ways we have of justifying relative weights are compatible with reading off the weights of duties and obligations in some cases.36

6 non-arbitrariness A second possible objection claims that the account I’ve given of coherence as case consistency is toothless – especially when it comes to morally significant differences – because it allows anyone to justify anything he wants as long as he is willing to stand behind the differences in question. Indeed if one is prepared to defend, say, racial distinctions as morally significant differences one may be case consistent in judging people differently on the basis of race, and someone who was willing to say that morally irrelevant things like first names or days of the week were morally significant differences would be coherent in believing a wide range of seemingly arbitrary beliefs. We might put the objection this way: what we hoped for in “coherence” was a way of coming to know and justify moral beliefs in a systematic way. Forms of coherence that are based on case consistency don’t help with this, because they give us no guidance on what is, and what isn’t, a relevant or an arbitrary distinction.37 My response to this objection has three parts. First, any method of coherence in the context of conviction ethics must rest on a set of assumptions about morally significant differences and similarities, and there is a wide range of views on what these should be. Second, systematizing and rich coherence are worse off in the relevant sense because they, too, require us to make arbitrary assumptions about what does and doesn’t make for a morally significant similarity or difference, and the assumptions they must make diverge more deeply from the pluralistic way we value. Third, because morally significant differences are moral beliefs one must be prepared to endorse and defend, the problem of cooked-up distinctions like first names and days of the week is not a problem that arises much in practice. First: it may seem that systematized sets of moral beliefs avoid the difficulty of having to make assumptions about morally significant similarities and differences, but this is not so. The requirement of

102

Moral Reasoning in a Pluralistic World

interchangeability of perspectives mentioned above rests on an assumption that persons are morally equal. But this is an assumption about what differences and similarities are morally significant. It presupposes, for example, that “both being persons” is a morally significant similarity and that all kinds of differences between persons are not morally significant differences. But not everyone shares such a view of moral significance. Peter Singer has a different one: as he emphasizes, his “Principle of Equal Consideration of Interests,” which says that all interests should be taken into account in the same way, conflicts with the idea that “being persons” is itself a morally relevant similarity.38 We may certainly use the equal consideration principle to justify taking persons into account differently as long as we are taking their interests into account in the same way. So this principle rests on the moral standard of relevance that the interests of all are morally similar rather than the personhood. Indeed, there is no consensus on where such similarity begins and ends. From a richly coherent Kantian point of view the fact that a person happens to have a preference for or interest in something does not suffice to create a morally significant difference between persons; for Singer it does. Bernard Williams writes of the special commitments one has to people and causes that can be, in themselves, morally significant.39 Bioethicists Gert, Culver, and Clouser admit a whole class of “morally relevant” considerations that can make differences in cases.40 And in a recent discussion of the varieties of “impartial” reasons, Amartya Sen describes the duties that arise for us in virtue of the particular and asymmetrical powers some of us have: to right an injustice, to protect animals, to care for a child.41 Because these duties arise from relationships and abilities specific to individuals, their existence rests on a particular understanding of moral signifiance. Second: when it comes to the question of arbitrariness, systematicity and rich coherence are in a worse position than views based on case consistency. This is because the assumptions about moral significance one must make to systematize require saying that there is only one kind of morally significant similarity, whereas any pluralistic way of valuing – as we typically value – requires saying that there are multiple kinds of morally significant similarities and differences. As we just saw, in Singer’s view the only relevant similarity is that between

Case Consistency

103

one interest and another, and the only difference is that between interests and anything else. But implicit in the pluralistic way we value is the existence of various distinctions and similarities that can be morally relevant. A person who values more than one irreducible value or endorses more than one non-derived principle has a set of beliefs about moral relevance according to which cases can be morally significant or different in more than one way. This is because a theory with multiple non-derived principles will have more than one way that cases can significantly differ morally: other principles simply may, or may not, apply in the given cases. A person who endorses just two principles, “Don’t lie” and “Keep your promises,” allows at least two ways cases can be different in morally significant ways: they may differ with respect to honesty or they may differ with respect to integrity. Any form of rich coherence asks us to reject some of these as morally significant differences in favour of a single morally significant difference. But on what grounds? To reject “several” for “one” without a good reason is not only arbitrary in the sense of making certain assumptions about significance, it is also – and more disturbingly – arbitrary in that those assumptions do more violence to our ordinary ways of valuing for no good reason. This is the sense in which I mean that rich coherence is worse than case consistency from the point of view of concerns about arbitrariness. For an example that illustrates this point, consider again marital rape. We’ve seen that there was a shift from seeing the matter in terms of marriage making a morally significant difference (so that a man could not rape his wife) to seeing the matter in terms of autonomy and consent (so that forced sex in marriage is rape). In the understanding of the situation based on autonomy rights, whether the man wants to force his wife to have sex with him, or would get pleasure out of it, or has a preference or interest for it is not relevant to the morality of the situation. But in a rich coherentist view based on the idea that “interests” form the only morally relevant similarity and difference, we would have to say that this interest is relevant to the morality of the situation, even if only in a small way. But this seems mistaken: the interest of a man in raping his wife is not relevant at all. This shows that even a richly coherent utilitarian view asks us accept some judgments and reject others about what is and is not morally significant.42

104

Moral Reasoning in a Pluralistic World

Third: in contexts of reasoning together about what is right, the possibility of proposed distinctions such as first names or days of the week is not a significant one. This is because to be case consistent one must not only introduce or state a morally significant difference, one must be willing to stand behind or endorse the distinction in question as morally significant. This means that when in practice we appeal to a norm of case consistency, there are factors that affect what distinctions we are willing to appeal to and that influence to what degree our arguments will be successful with others. As I understand it here, in endorsing and defending a morally significant distinction or similarity one is endorsing and defending a moral claim. While it might be possible to introduce highly arbitrary distinctions such as “having a first name beginning with J,” or “happening on a Tuesday,” to make one’s views seem morally case consistent, in practice it is much more common for people to defend their views by appeal to considerations that reflect what they are willing to stand behind as caring about, however strange or appalling it may seem to others to care about that set of qualities or distinctions. When we encounter views we disagree with it is common to find judgments of moral relevance that are problematic not in being unrecognizable and baffling, like “having a first name beginning with J,” but rather in being embedded in moral systems we reject on moral grounds. Defenders of the race-based slavery of the American South, for example, far from defending their actions by appeal to skin colour as a morally significant difference sought instead and with great energy to justify their judgments and actions in terms that were embedded in an internally consistent value system. In reasoning influenced by Aristotle, it had long been believed that some people were naturally fitted to be slaves; the question for defenders of slavery was always how to explain and justify the choice of some people rather than others. The historian David Brion Davis says that people with distinctive physical features were “made to order for ... Europeans who struggled for centuries to find markers that would help justify class polarities and also help to identify, at some distance, people who could be classified as ‘natural slaves.’”43 These “natural slaves” were said to be unintelligent and to lack reflection and self-control; for this reason, it was said, things were better for everyone – including slaves themselves – under a system of benevolent masters. These were beliefs that slavery supporters were willing to stand behind and endorse, and they were

Case Consistency

105

beliefs abolitionists succeeded in undermining. So the justification of slavery was embedded in a value system reflecting beliefs we reject, rather than resting on judgments about moral significance that proponents introduced with no attempt at justification. From the perspective of case consistency, people who are willing to defend racial differences as morally significant in justifying slavery are not “inconsistent” or “incoherent” in their moral views, and their problem is not a failure of reason. From within our way of valuing and judging we can say that they are making an evident moral mistake: it is wrong to enslave anyone and race is not a morally significant difference in the context of this judgment. A virtue of this way of seeing things is that race can function as a morally significant difference in other contexts, so that people who distinguish on the basis of race for other reasons – as those supporting, say, affirmative action – are not case inconsistent or lacking in reason in making this distinction. It is true that case consistency itself cannot help us distinguish between moral values and other values: if a person is willing to stand behind a peculiar value and endorse an idiosyncratic set of morally significant similarities and differences based it, there is nothing case consistency can say about why this person is mistaken. This means case consistency must rest on, and cannot ground, a theory of what counts as a moral value. But insofar as this is a limitation it is one shared by the method of conviction ethics generally, since that method always allows a person to take their own considered convictions as apt starting places for moral reasoning.

7 prioritization is not lexical ordering It may seem that in contexts of multiplicity, prioritization is best understood in terms of a ranking or lexical ordering of principles. For example, Mill talks of a “determinate order of precedence” among principles,44 and the prioritized theory most familiar to many of us is found in Rawls’s principles of justice: Rawls uses a ranking (he calls it a “lexical ordering”) of his two principles, the “liberty principle” and the “difference principle,” to address the question of priority.45 Let me say that a ranking of principles is any prioritization that, like Rawls’s, demands that we consider the top ranked principles to always engender obligations of greater importance than lower ranked ones. That is, in a conflict between A and B, whichever obligation is

106

Moral Reasoning in a Pluralistic World

entailed by a higher-ranked principle will become our actual or overall obligation. I claim that case consistency is a fundamental norm that may sometimes lead to rankings, but typically will not. Notice first that ranking our principles does not on its own solve the priority problem. Since single principles like “keep promises” generate conflicts, putting such a principle on a ranked list would still leave us with the difficulty of determining what to do overall in such circumstances. So even rankings leave us wondering what to do overall in cases of conflict. Case consistency is also a more fundamental norm than a ranking requirement in two ways. Even ranked principles can issue in judgments that are not case consistent, and in this context ranking doesn’t help with justification. If principles are formulated so as to introduce judgments that are not case consistent – e.g., keep promises but only on Tuesdays – then ranking principles will lead to case-inconsistent sets of judgments. Furthermore, even though properly formulated ranked principles lead to case consistency many case-consistent theories do not contain ranked principles. Typically, a theory with ranked principles will not be as good a fit with our considered convictions as a case-consistent theory whose principles are not ranked. This is because most of us weigh moral considerations against one another in ways that are incompatible with rankings. Judging case consistently often requires privileging one principle in one kind of conflict and another in another. Recall Marie’s dilemma, in which she has promised to keep a secret and must decide whether to keep her promise or lie. This involves a conflict between principles against lying and principles of promise keeping. It is easy to imagine conflicts between these in which promise keeping is clearly paramount: if, under the same conditions, Marie’s niece confesses to having once smoked a cigarette once years ago, and Marie has promised discretion about it, she ought to keep her promise: her status as confidante of her niece requires it. If, however, the niece confides that a younger friend of hers is showing signs of anorexia, and the friend’s parents ask about it, Marie ought to tell the truth. The reason judging the one way in one case and the other in the other does not violate case consistency is that there are morally significant difference between the cases: the quality of the lie, the consequences of telling the truth, the relationship between Marie and the people in question. In the same way, imagine that in the case in which Anthony ought to pay back to his neighbour what he has borrowed,

Case Consistency

107

his child gets sick and he needs the money for a hotel room near the hospital. In this case we judge Anthony ought to spend the money on the hotel room. But if his child is not sick and Anthony wants the money for a tutor so the child can get into a better school, it seems Anthony ought to pay it back. In the same way, I need not keep a promise to repay a rich man the nickel I borrowed if doing so requires me to be away from the child I have to care for, yet I must keep a promise to pay a significant sum if it takes me away from my child for ten minutes and it’s easy to arrange a sitter.46 Again, this is because the cases are different in morally significant ways even though the same two principles are at stake. So on what grounds might we prefer ranked principles? Perhaps it might be thought that ranking is a good norm because unlike case consistency it does not require us to make use of beliefs about moral significance and valences in justification. But in the context of coherence methods and conviction ethics, how else can a ranking be justified? Rawls’s thought experiment justifying the principles of justice and making use of the original position is justificatory only in the presence of certain beliefs about both: that “being persons” is a morally relevant similarity, that other similarities and differences are not morally relevant, that mutual advantage is a priority, that the results of the experiment fit with certain considered convictions we hold about the relative importance of liberty and distribution. All prioritizations must make use of similar assumptions. Leaving aside the difficult question of whether Rawls is right to prioritize as he does in the case of justice, the assumptions most of us make in figuring out how to prioritize do not end in rankings of principles in morality in general.47

108

Moral Reasoning in a Pluralistic World

5 Pluralist Coherence

I’ve argued in chapters 1 through 3 that rich coherence is inappropriate for use in pluralist contexts. In chapter 4 I began the task of developing and defending an alternative: pluralist coherence. There, I argued on behalf of taking case consistency – judging in accordance with morally relevant similarities and differences – as the central notion in this alternative conception of coherence. In this chapter I develop pluralist coherence further. I explain how moral dilemmas and the expansion of values are connected to a pluralistic kind of moral progress; I discuss the normative question of why a person ought to aim to be coherent in the sense of pluralist coherence; I explicate pluralist coherence further through a few examples; and I end by comparing the method to those of some other approaches.

1 remainders and the reality of dilemmas Recall from chapter 1 that a set of multiple principles grounded in various conflicting values leads to conflicts: Marie ought to keep her promise, and she ought to tell the truth, and she cannot do both. I explained there that we can render a set of multiple principles consistent in either of two ways: we can use specification to introduce exception clauses – rather than “keep your promises,” the principle might be “keep your promises, unless” – or we can take the principles to generate pro tanto obligations which disappear when they are overridden by others – if the duty to keep the promise is weightiest overall, the duty to tell the truth simply disappears.

Pluralist Coherence

109

In this section I consider the status of moral conflict and its relation to the formal existence of moral dilemmas. In chapter 1 I argued against explaining away conflict, and for retaining conflict. The reasons for the latter had to do with the moral remainder – the moral phenomenon that persists in cases of conflict and gives rise to a feeling of regret or distress. I emphasized the fit between the remainder and our moral experience: when we encounter a dilemma we often experience such feelings of regret; the remainder explains why this feeling arises and why it feels like it has a moral component, a sense that there was something not just sad but morally regrettable in one’s having acted as one did. Matters concerning the nature and status of the remainder are complex, however, because they concern the question of the reality of dilemmas. A dilemma is a situation in which one ought to do A and ought to do B and cannot do both. Saying that dilemmas are real is typically understood as saying that both obligations remain in force in the ordinary way, despite the impossibility of meeting both of them. This means that when a person encounters a dilemma, she has two obligations and can only meet one. There is, therefore, an obligation that the person failed to meet. For example if Marie judges that overall she ought to keep her promise then she has failed to tell the truth, and thus has done something she ought not to have done. There is one simple and well-known argument for the conclusion that dilemmas should not be taken to be real. The person who encounters a dilemma fails to meet an obligation; thus, on the face of it, it would seem appropriate for them to feel guilty or for us to blame them. But assuming she judges well what her overall obligation is, this person has clearly done her best: it’s not her fault that she couldn’t meet both obligations. It feels inappropriate to blame those who have done their best. So dilemmas cannot be real.1 They ought to be explained away, by either pro tanto or specification. Elsewhere I have called this objection to the reality of dilemmas the “stringency objection,” since it suggests that dilemmas make demands on an agent that are excessive and strict.2 In section 2 below, I discuss ways of addressing the stringency objection. But first let me elaborate here the positive argument for taking dilemmas to be real. This argument focuses on the role the remainder plays in our moral lives. To understand this role, recall first that a set

110

Moral Reasoning in a Pluralistic World

of principles is practically consistent when there is a possible world, or a way that things could be, in which conflicts do not arise. The principles in favour of promise keeping and telling the truth are practically consistent in this sense, because though they engender conflicts in our world, in circumstances like Marie’s there is a possible world in which such conflicts never arise. This is the world in which the relevant circumstances never come up. Perhaps, for example, one is never asked about the secret and so encounters no need to lie. In her arguments in favour of taking dilemmas to be real, Ruth Marcus argues that when there is such a possible world, we have an obligation to try to bring it about. We have a special obligation, over and above our others, to try to make our world closer to a world in which there are no conflicts. We ought to “arrange our lives and institutions,” as Marcus puts it, so that dilemmas are less likely to arise.3 Intuitively, the idea is to aim to make the world so that more of our obligations can be met, because fewer of them conflict. Perhaps the simplest kinds of examples involve planning for resources: if you know you need to provide, say, for two different children over the next year, then obviously you should try to have enough to provide for both now and in the future. It is not appropriate to fail to plan, have resources enough for one only, and then to claim that because of a dilemma, one can meet only one of one’s two obligations. To explain dilemmas away in these cases would require us to say of the parent, “he ought to feed Jack and he ought to feed Jane but he cannot feed both, so either through specification or pro tanto, only his most pressing duty is his overall duty.” There would be nothing this person ought to do that he failed to do. But if he cannot feed both because he failed to even try to arrange for feeding both, then clearly there is something he failed to do, and that something is plausibly described as failing to arrange things so that conflict would not arise. Similarly, in Marie’s case, an obvious way to remove the dilemma is to persuade the daughter to be honest with her father. This may not be the best course of action in all circumstances since there may be other values and dangers in play, but in the right circumstances it is a way of proceeding that honours both fidelity and honesty, and in that way it is good. The special obligation explains why. In his recent bioethics work, Joseph DeMarco perceptively ties the existence of the remainder to the duty we have to try to promote what we value, even when values conflict. He calls this the “mutuality” prin-

Pluralist Coherence

111

ciple: “Act to establish the mutual enhancement of all basic moral values” – that is, if you experience a conflict between obligations, try to arrange things so that the conflict does not arise in the future.4 He illustrates mutuality with a bioethics example.5 Suppose there are parents who want to refuse medical treatment for their child on grounds that the treatment involves a blood transfusion, which they are against on religious grounds. There is then a conflict between principles of respect for parents’ wishes for their children and principles that require doctors to save the lives of their patients whenever possible. The technique of specification, he says, would indicate treating this case by specifying the first principle: instead of requiring respect for parents’ wishes, it can be specified to require respect for parents wishes only when those wishes do not involve abusing those children or violating their rights. In this case, we know what to do overall and there is no conflict. Though DeMarco doesn’t address the pro tanto approach, we can see it leads to a similar outcome: if the obligation to protect patients’ lives overrides that of respecting the parents’ wishes in certain kinds of cases and becomes one’s only real obligation, then the conflict disappears and the latter obligation fails to be in force. DeMarco argues that mutuality gives a better analysis of the situation, and I agree. Mutuality entails that in addition to forming an overall judgment about the matter, we ought to consider developing treatments that would allow us to honour both principles, in this case by seeking treatments for the same illness that do not involve transfusions. Because this allows us to honour more of our values, it is a better approach; taking the conflict to be real shows why this is not a case one can resolve without a remainder showing that we have obligations to prevent future similar conflicts as much as possible.6 It might be objected to this analysis that the outcome is mistaken on grounds that if we have limited resources, we ought not focus them on developing such treatments, so mutuality entails a false conclusion. But the point about limited resources doesn’t undermine mutuality: it is just a manifestation of the fact that we have other ends that conflict with, and possibly override, the ones in question. Notice here that mutuality also leads to an intuitive conclusion about the elderly woman described in chapter 4, who wants her feeding tube removed. That case presents a conflict between the obligation of obeying the patient’s wishes and that of prolonging her life. A better solution, if it were available, would be to reduce the patient’s suffering and improve her spirits so that she

112

Moral Reasoning in a Pluralistic World

wanted to live, thus removing the conflict.7 If similar conflicts could be prevented in the future, that would be best overall. In a vivid real-life example of conflict and the role of mutuality, in 2012 Florida lifeguard Tomas Lopez was fired for leaving his post to save a drowning person. The guidelines for his job prohibited him from saving people outside the company’s protected zone of the ocean, where the man was drowning. The guidelines said to call 911 in such a situation, but Lopez said he was sure that if he had done that the man would have drowned. Lopez was in a conflict of obligations: he ought to have obeyed the rules of his position, and he ought to have saved the drowning man. Clearly, we should try to arrange our lives and institutions so that a lifeguard is as rarely as possible put in the position that saving a drowning person requires violating his work responsibilities. The company supervisor explained to a TV news reporter that “we have liability issues and can’t go out of the protected area,” which illustrates how creating the conditions in which dilemmas are less likely to arise can be complex.8 The mutuality principle shows clearly why there is a responsibility to try. As this example shows, the role the remainder plays in mutuality applies not only when the alternatives seem quite evenly matched and it is puzzling what one ought to do overall, but also when one alternative clearly overrides the other. The lifeguard’s most pressing obligation was to save the drowning person: as he said himself, “I have no doubts I did the right thing ... I believe I did what was right, and that if someone needs help you’re going to go help them, regardless if you’re a lifeguard or not.”9 But it would have been better had he not had to choose between this and doing his job. Recall Anthony, too, who ought to pay back his neighbour and ought to be with his son and who cannot do both. Even if Anthony clearly ought to be with his son overall, we ought to try to arrange things so the conflict does not arise: it may not be possible, but certainly things are better when he can meet both of these obligations. When we succeed in arranging our lives and institutions so that conflict does not arise, this is a form of practical moral progress. It is moral progress because the more we obey mutuality, the more we are able to meet our obligations and honour our values. It is practical because it involves not a change in our values and obligations, but a change in how well we honour those values and meet our obligations. For example, in the medical testing case, we’ve seen the way concerns of justice,

Pluralist Coherence

113

benevolence, liberty, and respect for autonomy conflict. The mutuality principle entails that we ought to aim to bring about conditions under which such conflicts are minimized. Reducing inequality would make studies less likely to be coercive or exploitative; eradicating disease would reduce the need for benevolence. This is progress not in the sense of improving our moral opinions, but in the sense of making the world a better place. Traditional thinking about moral progress, sometimes informed by Karl Marx and G.W.F. Hegel, understands that notion with an emphasis on comparing cultures across historical eras. The kind of progress that emerges here, however, should be understood as occurring within a particular culture or cultural tradition (or on an even smaller scale), over a short span of time. In a 1977 essay, “Moral Progress,” Ruth Macklin works mostly in the former framework of comparing cultures.10 But part of her characterization of progress is particularly apt here. Her “principle of humaneness” says that “one culture, society, or historical era exhibits a higher degree of moral progress than another if the first shows more sensitivity to (less tolerance of) the pain and suffering of human beings than does the second, as expressed in the laws, customs, institutions, and practices of the respective societies or eras.”11 The principle explicitly defines moral progress partly in terms of our behaviour, since it judges our institutions, practices, etc. If we act to avoid dilemmas, we meet more of our duties to kindness, which satisfies Macklin’s criterion for progress. Under Macklin’s definition, moral progress does not have to include a change in moral principles. By defining progress in terms of our cultural practices, Macklin allows that an increase in our abilities to live up to those principles in a practical way suffices to indicate progress. This is what I have argued the remainder account provides for: a way to understand why and how we are motivated to change our communities so that we can better live up to those principles, by honouring more of our values and meeting more of our obligations.

2 in defence of dilemmas: answering the stringency objection Now, recall the stringency objection, that taking dilemmas to be real requires us to blame those who have done their best. In particular, this objection says, attributing to a person the same degree and same kind

114

Moral Reasoning in a Pluralistic World

of fault or failure as that we attribute to someone who simply failed to do what she ought to have done makes no sense.12 Philippa Foot points out that moral necessity is a shield against charge of wrongdoing just as physical or mental necessity is. To say “I couldn’t help it” is appropriate whether the reason one is prevented has to do with physical and mental constraints like being in the wrong place at the wrong time to help or moral constraints like having other overriding obligations. If Marie cannot tell the truth because she has made a promise that prevents it, she may truthfully claim that she cannot tell the truth. Blame and the charge of moral failure would be unjustified. Blame and charge of moral failure may seem not only unjustified but also inappropriate and cruel. There was outrage over the firing of the lifeguard, and it certainly seems inappropriate to say that he is blameworthy for failing in his job duties despite having saved a life in a tough situation not at all of his own making. And consider cases like that in William Styron’s novel Sophie’s Choice, in which Nazis give Sophie a choice of which of her two children to save, while the other will be killed. If she does not choose they will both die. Patricia Greenspan points out that taking Sophie’s dilemma to be real – she ought not give up one child and she ought not give up the other – requires us also to say that no matter what she does, Sophie has broken the obligation against causing her child’s death, and thus ought to feel guilty and is properly blameworthy. But this seems too harsh. Sophie did the best she could. Of course she feels guilty. Shouldn’t our moral theory tell her that her guilt is not warranted, rather than adding to it?13 As a first step in addressing this objection, note that we needn’t interpret the remainder as indicating wrongdoing in the way the objection does. In her original work, Marcus emphasizes the link between remainders and motivation: in case of a dilemma – the effect of the unmet obligation – whether it is best described as guilt, remorse, or regret, motivates us to try to arrange things in the future so that dilemmas are less likely to arise. The reality of dilemmas, in turn, explains why there is a remainder: this results from the unmet obligation. Dilemmas, in this view, are real whenever there is conflict – whether or not one obligation overrides the other. But to say that the remainder is real we needn’t say that it takes on the particular form of being identical with that of an ordinary unmet obligation. For the remainder to play its role, it need only represent a

Pluralist Coherence

115

negative moral fact, of a kind to be further specified. That role is partly motivational, but it is also epistemic: sometimes it is when we encounter conflicts that we learn that our world is structured or arranged poorly, and there is something we ought to change. In addition, that role is more forward-looking than backward-looking: what matters is that we explain the duty to arrange things for the future, not the guilt or blameworthiness of the person in the dilemma. This means we learn from real cases about the conditions that create or give rise to dilemmas; the obligation is to prevent such conditions in the future and does not require guilt or blameworthiness for having failed to do so in the past. Further, notice that often the arrangement of our lives and institutions is a necessarily collective action. This means the relevant responsibility is collective as well. Obviously, the lifeguard, the doctor in the blood transfusion case, and Sophie are not individually responsible for bringing about the conditions under which their conflicts become less likely. This would require action on the part of whole groups of people, including large bureaucratic organizations and government agencies. It might seem that this role for the remainder makes the most sense when principles are practically consistent. If there is a possible world in which our principles engender no conflicting obligations, the mutuality principle tells us to arrange our lives and institutions so that we are closer to that particular world. If there is no such possible world, how can we do this? But, as in the discussion in chapter 2 of conflict in values, there is not much difference between practical and essential conflict. If there is no possible world in which conflict does not arise, this means only that we must aim to minimize it knowing we cannot rid ourselves of it altogether. Indeed, as in the values discussion, we may sometimes not even know whether our principles conflict essentially or not. So the mutuality principle applies to practical and essential inconsistencies in the same way.

3 against the specification of principles In this section I revisit the question of specification. Recall from chapter 4 that Richardson argues in favour of specifying principles on grounds that reading off relative weights of obligations is ad hoc and non-justifying. I’ve said there why I think this isn’t a problem in the

116

Moral Reasoning in a Pluralistic World

context of conviction ethics. Here I make the further claim that even when we’ve determined what is right to do overall, we have reason not to specify our principles. Specification involves narrowing, qualifying, or interpreting norms so that we can see more accurately how they apply in the given case. Recall that in the example of the lawyer from chapter 4, specification entails that instead of endorsing a principle demanding that lawyers pursue their client’s interests by all lawful means, we should endorse a principle that lawyers ought to pursue their clients’ interests by all lawful and ethical means. In another example Richardson gives, instead of saying that generosity and tolerance are overridden by outrage and punishment when the act that is committed is “beyond the pale,” which would be an expression of balancing the norms, we ought to endorse a new norm. The new norm, instead of endorsing generosity and tolerance, endorses being generous and tolerant, even to transgressors, except when the transgression is beyond the pale.14 Specification allows us to adopt a new norm that expresses our reasons and thus gives us a way of justifying our choice directly. The justification is applicable to future cases in a straightforward way. The virtue of specification might seem to be found in the practical way it helps us with consistency: once we’ve pinpointed the relevant considerations, a specified principle helps us appeal to the same kind of considerations in future cases. A specified principle tells us how to judge in future cases so as to be consistent with our judgments of past cases in a practical way.15 However, specification is not an apt formal expression of this ideal, because of its implications about the remainder. I’ve argued that the remainder plays an important role in moral practice, in showing us that we ought to try to change our lives and institutions so as to minimize the possibility of conflict. Expressing the outcome of our process in terms of specification is inconsistent with this view of conflict: specified principles no longer conflict, because they build in exception clauses. For example, the new legal norm does not conflict with any duty to cross-examine, and the norms of generosity and tolerance, when specified with the exception clause, no longer generate any conflict at all. No conflict, no remainder, no explanation of the need to change our world in the relevant ways. Richardson suggests otherwise, arguing that a conflict can be seen as genuine even while the principles are specified. This is because “on the model of specification … the general but non-absolute norm,

Pluralist Coherence

117

understood as stating an actual obligation, can stand alongside the specified version that averts the particular conflict at issue.”16 So, for example, if principles A and B entail conflicting obligations X and Y, we can specify A or B so that the conflict is avoided and there is a single obligation, while at the same time retaining the unspecified principles A and B. As Jonathan Dancy has argued, however, it is hard to see how specification and conflict can co-exist.17 As Richardson says, for the specified norm to be consistent with the general norm the general norms must be formulated as non-absolute – that is, with clauses like “Generally one ought to.” But if they are formulated this way, there is no genuine conflict generated. And if there’s no conflict generated by the principles then there’s no explanation of any remainder. Sometimes it is appropriate to narrow principles. For example, the idea that one is not obligated to keep promises made under duress is plausibly captured by saying that the principle of promise keeping is better expressed by a principle in favour of keeping only promises not made under duress. This is an appropriate narrowing and has the form of specification – of saying more clearly where the principle does and does not apply. But in this case the reason for narrowing is not related to conflict or dilemmas: there are not two things a person ought to do. And thus a remainder would serve no purpose. Rather, we do want to narrow the range in which the principle applies. In cases of conflict and dilemma, though, we ought to keep the general principles and not specify them. And with respect to mutuality, one possible world in which the conflict would not arise would be one in which there was no point to cross-examining the victim about her sex life, because it was widely believed to be irrelevant to her testimony. In this possible – and less sexist – world, whether a woman consented to sex on previous occasions would suggest nothing about her character and trustworthiness. So such cross-examination would be useless, and the lawyer could fully pursue his clients’ interests without the conflict ever arising. It would be better to try to bring about such a world, and the remainder is an indicator of this.18

4 expansion, generality, and moral progress It is sometimes suggested that generality ought to be part of moral coherence: Kappel, for example, puts it on a list of coherence ele-

118

Moral Reasoning in a Pluralistic World

ments.19 Sometimes the word “general” means simply what I am calling case consistent: applying the same criterion to cases unless one can cite a morally significant difference between them. Clearly pluralist coherence endorses this sort of generality. But another way of understanding generality is as I’ve characterized it here: as having to do with applying a single criterion to a large range of cases. While principles must be “general” in the sense of not appealing to indexicals and names, some principles can be more general than others. For example, the belief that one ought to keep one’s promises to family members is a general principle, but the belief that one always ought to keep one’s promises is a more general principle. I argued in chapter 3 that there is no special connection between generality and systematicity: though it might seem that the more general the principles are the fewer of them there can be, this isn’t so, because more than one principle can apply to any given case. I claim further here that generality in this sense doesn’t belong as a norm of pluralist coherence reasoning on grounds that generalizing a principle sometimes improves it, but sometimes makes it less plausible or even morally troubling. I’ve mentioned one example already: that of promises under duress. The principle in favour of promise keeping is plausibly improved by being narrowed, and by requiring us to keep only promises not made under duress. Other examples come easily to mind. Principles requiring respect for property rights are plausibly improved by restricting their range to include some kinds of goods – e.g., material consumer goods – and not others, such as organs and infants.20 Principles allowing consensual sex are plausibly restricted in certain situations of highly unequal power relations, for instance an employer may be properly restricted from seeking sex with an employee even if the sex is consensual.21 That these principles are intuitively improved by being narrowed suggests that taking coherence to include a norm of generality is not wise. There are, I believe, two reasons generality has seemed attractive. First, it is confused with systematicity: the search for general principles is conflated with the search for few, simple principles. But as we’ve seen, these are distinct. Second, generality can be interpreted not as “applying to many cases” but as simply meaning “non-arbitrary.” This makes it obviously attractive, but I’ve argued in chapter 4 that case consistency is no worse than other coherence methods when it comes to arbitrariness. So this kind of generality is already covered.

Pluralist Coherence

119

I propose that pluralist coherence should include a norm not of generality but of expansion, endorsing the idea that there is a connection between moral progress and the expansion of our circle of concern. Expansion means that when we move beyond parochial considerations of family, community, nation, or even “human” only to bring more beings into the range of our moral beliefs, this is a kind of moral progress, since our care and concern extends to more rather than to fewer. In The Expanding Circle, Peter Singer writes movingly of our coming to realize that moral recognition ought to extend past our selves, past our families, past our communities, and even past our species, so that all conscious beings can be objects of moral concern. Singer expresses this idea through his preference, utilitarianism, a richly coherent theory. But expansion of moral concern is compatible with value pluralism, with principle pluralism, and with the pursuit of case consistency as described here. Nothing prevents us from extending multiple obligations and rights so that they include more persons and other beings. When consideration is extended through a value to apply to or encompass beings in a larger rather than smaller circle, this is a form of moral progress. It is a relative form of moral progress, because it is from within a moral point of view that we can judge the extension to be better. We can see such a phenomenon at work in the debate over medical testing in the developing world. In the past, people simply carried out tests whenever they wanted without much concern for subjects. The expansion of autonomy rights and consideration of harm meant later subjects had to be treated with respect – respect that was formalized in guidelines for medical testing. As long as the effects on subjects are theorized in terms of harms and benefits and autonomy, the double standard may not seem to matter, but the expansion of considerations of justice renders such a standard problematic. There is an apt particular example of expansion in Macklin’s discussion. Macklin mentions a manuscript reviewer who says that the concept of “distributive justice” applies only within a social group and not “between countries.”22 Macklin disagrees, arguing that we often apply the concept in more open-ended way; she endorses an “expanded” view of distributive justice and argues that double standards violate that notion. Two things are worth nothing about expansion in the context of pluralist coherence. First, expansion and pluralist coherence are both consistent with partiality for one’s own group. Macklin draws a con-

120

Moral Reasoning in a Pluralistic World

nection between the reviewer who restricts the concept of distributive justice and the view of a New York Times letter writer responding to an op-ed piece by the economist Jeffrey Sachs. Sachs had written that “compassion” gave us reason to help fight AIDS in Africa, and the writer’s response was to say that “states should not be expected to value the lives of noncitizens as much as citizens.”23 If we are indeed weighing various values against one another to figure out what to do in complex cases, it is consistent with expansion and pluralist coherence to say that while distributive justice is applicable in the given case, citizenship creates morally significant differences and we have special obligations to those we share citizenship with.24 Second, expanding our circle of concern in contexts of multiplicity we can expect to encounter more conflicts rather than fewer for the simple reason that more principles will be applicable to more cases. The more we care about others and apply to them a value of beneficence, the more we extend a duty to promote the well-being and autonomy of all, the more conflicts we’ll encounter, in part because of simple limits of resources. But such conflicts are to be expected, and are dealt with as all other conflicts: through prioritizing and trying to improve matters so that similar conflicts will not arise in the future. This means that a kind of moral complexity is to be expected with moral progress and is not a sign that something has gone wrong. Again, the expansion of our circle of concern in the medical testing case obviously leads to more conflicts rather than fewer, but this is compatible with progress.

5 why be coherent? Why ought we to be coherent in the sense of “pluralist coherence”? I’ve noted that Kahneman and Sunstein understand “coherence” much as I do “case consistency” here; they say that “because of how indignation operates, it is extremely difficult for people to achieve coherence in their moral intuitions.”25 Given that it is so difficult, why ought we aim to do it? I’ve already said something about remainders and expansion. When we try to bring about a world in which conflicts are minimized, we meet more of our obligations and thus honour more of our values. If I ought to do A and I ought to do B, then things are better from the moral point of view if I can do both rather than being caught in a

Pluralist Coherence

121

dilemma. And expanding moral concern to more beings also means promoting more of what we already take to be good. The more difficult matter concerns case consistency. Why ought we judge consistently, and how does the relevant reason not commit us to rich coherence and systematicity? One simple answer is the one already mentioned in chapter 4: we want our moral beliefs to be principled, not arbitrary. In discussing the implications of non-arbitrariness there, I argued that rich coherence and systematicity are no better from this point of view than case consistency. There is a clear sense in which case consistency is a reflection of non-arbitrariness: to be justified in judging cases differently, we need an explanation of why there is a morally significant difference between them; to judge cases differently without being able to cite such a distinction is to admit that we have no good reason for why it is appropriate to judge one way in one case and another in another. So non-arbitrariness is one kind of answer to the normative question that explains a need for case consistency and not for rich coherence. A deeper way of considering the matter is as follows. We might ask: why ought we judge consistently? There are two ways of interpreting and understanding this question. In the first, simpler interpretation the question means: “Given that a person has a set of judgments about morally relevant similarities and differences, why ought she judge in accordance with this set of judgments?” If we think of beliefs about moral relevance as giving us a standard of evaluation for judging moral similarity and difference, a standard we appeal to in understanding which cases must be judged similarly, this form of the question concerns why we ought to appeal to that particular standard. In the second, more complex interpretation, the question means: “Why ought a person have a set of beliefs about moral similarity and difference? Why have a ‘standard’ like this at all? Why not simply judge each case as it comes along, without regard to one’s judgments about other cases?” The first question has an almost trivial answer: to have beliefs about morally significant similarities and differences is to have beliefs about how to judge cases; to ignore or defy these beliefs without modifying them is a form of irrationality – it is to commit oneself to taking a certain point of view only to ignore or defy that point of view. You cannot have beliefs about a moral standard of evaluation which

122

Moral Reasoning in a Pluralistic World

you then ignore; it’s in the nature of a standard that you use it in forming other beliefs. The second question is more difficult. I’ll propose a three-part answer. First, it is commonly said that the moral point of view is essentially a universal point of view. To be moral a set of beliefs must be universal in some sense, because otherwise those beliefs are more like a set of prejudices or disconnected opinions. Pluralist coherence interprets this universality as case consistency. The principle that one ought to keep promises is not general in the sense of recommending what to do in every situation, but it is universal in the sense that it says that everyone who has made the appropriate kind of promise ought to keep it. Second, case consistency is a pragmatic necessity for social, interactive, and interdependent beings like us, because we cannot engage in moral debate with those who are not willing to subject their moral beliefs to consistency requirements. When we encounter those who judge on different bases in different cases, and who are unwilling to bow to consistency demands, we cannot have a dialogue with them about the reasons for their moral judgments since those reasons may well change from case to case to case in ways we cannot understand. For example, imagine an interlocutor who judges some abortions to be permissible and others to be impermissible, but will not explain what the difference is. When asked, he does not offer any explanation; he is simply content to judge some cases one way and some cases another without considering any policy on the matter.26 Clearly this person is not case consistent, and because of this we cannot engage in any meaningful dialogue with this person about whether and when abortion should be permissible. For us to be able to reason together, we must be willing to defend certain morally significant differences and judge in accordance with them. The ability to engage with others about moral beliefs in this way does not require rich coherence; it requires merely that one be willing to stand behind what one takes to be significant differences between cases. This sense in which consistency is a “pragmatic” necessity is different from one Dancy discusses in his work on moral particularism. Moral particularism is the view that sensitivity to moral reasons does not require general principles, because different situations present many different factors that can be relevant. Dancy argues that principles are not necessary for making people’s behaviour more pre-

Pluralist Coherence

123

dictable or for the pursuit of agreement. People are already fairly predictable, he says, and are not made more so when they believe moral principles. And to judge difficult moral disagreements by a lawyerly appeal to precedent and rules, he says, would commit us to a picture in which disagreement subverts the whole moral enterprise.27 So principles are not pragmatically necessary. The sense in which I claim pragmatic usefulness is not in making behaviour predictable. It does have to do with resolving disputes, but as I’ve suggested already and argue further below, the usefulness of case consistency does not rest on our coming to agree. The usefulness of case consistency has to do with understanding others’ reasons when we do disagree, not because we demand agreement through shared judgments applied to all cases in the same way, but for the reason just mentioned, that we cannot understand or engage with the moral point of view of someone who is unwilling to stand behind the morally significant distinctions they endorse. Of course as I explained in chapter 4, case consistency in this sense is also not tied to precedent since a person’s moral views can change but remain case consistent at all times. Third, judging in accordance with standards about like cases is required by values such as integrity, justice, and fairness. Integrity requires us to stand behind our judgments: making a moral distinction without being willing to stand behind the distinction as based on a morally significant difference is a failure of a kind of integrity.28 Fairness requires that one judge similar cases similarly unless we can cite a good reason to judge them differently. To say that there is a morally significant difference between cases is to say there is a good reason to judge them differently, and without such good reasons it is unjust to condemn, or praise, or punish, or reward one person and not another. Of course, I have just been arguing that these values are just some among many and can be overridden, so this particular kind of moral explanation of case consistency may not yield the kind of universal force to the norm of case consistency that we’re looking for. However, in practice many agents do care about fairness in just the sense described here, and they do feel that to judge cases differently without being able to cite a moral distinction between them is somehow hypocritical in a way that violates moral norms.29 It is worth noting here how this explication of normativity shows a difference between pluralist coherence and the view of moral reasons

124

Moral Reasoning in a Pluralistic World

associated with particularism. It is true that my view shares with that of the particularist the idea that different factors arise and manifest themselves differently in different situations. Because there is no ranking of principles, and we have to judge valences based on various moral factors, it might seem that we need to make separate and distinct judgments about virtually all cases just as the particularist does. But particularists reject the importance of general principles altogether. Recall that in pluralist coherence, to count as a general principle rather than a particular judgment, a principle must be formulated without names and indexicals. In a deeper sense, the particularist rejects the need to apply to one case that which applied to others – a need that is at the heart of the norm of case consistency. With respect to this matter Dancy writes: “The second prong of the particularist attack is to ask why we should suppose that a feature that counts in favour in one case must count the same way wherever it appears. To this question, I think, no real answer has been produced. Generalists tend to point out that if one claims that a feature counts in favour here and against there, one has something to explain. But the particularist is happy to admit this. It is true that if a feature counts in favour in one case and against in another broadly similar case, there must be an explanation of how this can be. That explanation will presumably be given by pointing to other differences between the cases. In the second case, perhaps, something that is required for the feature to count in favour is in fact absent, though it was present in the first case. Such explanations must be available, and they can be found. None of this does anything to restore a generalist conception of how reasons function.”30 The pluralist coherentist differs from the particularist over the interpretation of the normative “must” in the sentence “it is true that if a feature counts in favour in one case and against in another broadly similar case, there must be an explanation of how this can be.” Pluralist coherence and case consistency interpret this “must” as an obligation: to say that a reason functions in one case and not in another, one must be willing to defend the distinction in a certain way. Since this interpretation seems to me to form a demand for non-particularity, I doubt it is the interpretation Dancy intends; the passage above suggests more of a predictive must – there must be a reason somewhere, and we could look for it if we wanted to. In this way, then, pluralist coherence is not like particularism.31

Pluralist Coherence

125

This explication of the normativity of coherence also differs from Goldman’s account, which focuses on rationality. Recall from chapter 4 that Goldman’s idea of coherence in pluralist contexts was that “one must not judge a case differently from a previously settled case unless one can cite a morally relevant difference between them, a difference that can be shown to make a difference in other cases as well.” I’ve already explained why I think the focus shouldn’t be on “previously settled cases,” since what we’re interested in is not coherence through time but rather coherence at a given time. With respect to why we ought to be coherent, Goldman argues that coherence is rationally required because of the way we honour our values through time. To the question of why, and how, rationality demands that we treat morally similar cases similarly, Goldman gives the following argument. Previous moral judgments, he says, reflect the way we value: particular persons care about things in certain prioritized ways, and that prioritized sense of value manifests itself in their judgments. A morally significant difference is one that reflects that sense of valuing – as valuing human life, say, creates exceptions when an act is done to save a life. Further, Goldman explains, a difference may count as “morally significant” only if it is a difference that functions in various contexts; the claim that a difference is morally significant only in a single circumstance would violate our sense of non-arbitrariness. The link to rationality, then, is that to judge in ways that fail to accord with one’s past judgments would be to undermine one’s self. “It is a truism that, if I believe two contexts for action cannot be distinguished by reasons for acting differently in them, I must, if rational, act the same way in both. (Of course in nonmoral contexts, my reason for acting differently may be just to vary things a bit.) Since moral judgments guide moral actions, I must, if I aim to be moral, judge two situations in the same way if I cannot distinguish between them morally. Otherwise, my judgments will be self-defeating in just the way that actions at cross-purposes are self-defeating. I will defeat an aim that I am committed to pursuing in just those circumstances at hand. Actions must be consistent not only with the purposes behind them, but also with other actions and the purposes and values they express, as these are prioritized in a stable ordering.”32 On this view, the connection between rationality and coherence is that fail-

126

Moral Reasoning in a Pluralistic World

ures of coherence lead us to promote one value on one occasion and defeat it on another – which makes no sense. But the loyalty argument conflates moral arbitrariness and changes in moral outlooks. Suppose Marie initially values honesty above all and believes that people should always tell the truth. She then encounters her niece, and comes to realize how important her discretion is, and over time comes to think that keeping her promises to her niece can be more important than telling the truth to her brother. She then modifies her moral beliefs accordingly, so that she judges that in similar cases it is obligatory for people to keep promises of discretion to people they love, even if that entails lying. As long as she modifies her judgments in an appropriate way, she is coherent in the sense described here. This shows that the loyalty argument is wrong to conflate moral non-coherence and moral change: in this version of the story Marie changes but is coherent. People do change their values and priority rankings over time, and it seems their doing so is often for the best, serving goals such as community, friendship, and intimacy. A person raised to value in accordance with one system of priorities may well adopt a new one on moving to a new country, or marrying into a new family, or simply seeing things differently as a result of life experience. Such change is not irrational; indeed, it seems things would be much worse overall if people did not change their priority rankings to more closely resemble those of others near to them.

6 pluralist coherence and moral disagreement The picture I’ve been laying out – of valuers who often share values but direct and prioritize among them differently, and who use pluralist coherence as a model for moral reasoning – gives us several ways of understanding moral disagreement. First, we may disagree with someone because one of us is not case consistent: though we share values, principles, and priorities, one of us is judging differently in the absence of a morally significant distinction we’re able to endorse. In the discussion of framing effects, I noted that we frequently make judgments in ways that are influenced by the words used to describe the situation. If two people disagree about, say, an economic policy because the policy was described to them in different ways, they may be in the situation in which they share values and principles but don’t see which policy is consistent with those values and principles. In

Pluralist Coherence

127

cases such as these, the pursuit of case consistency will tend toward consensus. If Kahneman and Sunstein are right, as I think they are, to say that emotions such as indignation make case consistency very difficult for us to achieve, there may be many cases of everyday disagreement that take this form. Second, we may disagree with someone because while we share values and principles with him, we prioritize among those values and principles differently. Two people may both care about honesty and fidelity, and may share a belief in principles against lying and in favour of protecting one’s friends. But if they prioritize among these differently they will disagree over what should be done overall in various cases, as Marie and her brother do. In this situation, two people can both be internally case consistent and share principles, but will nonetheless disagree, and the pursuit of case consistency will not produce consensus. Notably, however, the exchanges of reasons that their disagreement prompts may produce other kinds of important knowledge: in learning that we share values and principles, we may see our interlocutors as less foreign, and in understanding our difference as one of prioritization we may be more willing to compromise. Third, we may disagree with someone over what does and does not count as a morally significant distinction. Again, in this situation people may be internally case consistent on their own terms but case inconsistent from each other’s point of view; they will thus disagree. This is the kind of disagreement that produces the most bitter and persistent moral disagreements in public life. A person who believes that sex is only appropriate in heterosexual marriage endorses as morally significant distinctions that will not be regarded as morally significant by someone who theorizes sex through the value of personal autonomy; the disagreements that arise will be intractable because the disagreement is so basic. In this situation, people can share values but they direct or interpret those values in such a way that they endorse different principles – and thus also judgments. These ways of explaining moral disagreement show how reason is useful, but in contextually limited ways, in dealing with moral diversity and resolving moral disagreement. In the first kind of disagreement reason can work to resolve disagreements by showing interlocutors that they are mistaken in failing to judge like cases alike. Any person’s set of moral beliefs can be improved by being made more case consistent, and reason plays a role in justification and persuasion

128

Moral Reasoning in a Pluralistic World

when we engage with those with whom we roughly share values. If you and I share a set of basic moral beliefs, but one of us is judging “inconsistently” because of distorting effects or because one of us is self-servingly making distinctions we are not willing to stand behind as morally significant, then reason can explain how that person ought to change her mind. But in the second and third kinds of disagreements, consensus happens only when people change either their prioritizations or their senses of what does and does not count as a morally significant difference. These kinds of changes, I believe, proceed through social, cultural, and emotional changes. In these situations moral persuasion requires making use of strategies such as trying to see the world from someone else’s perspective, appeals to emotions, narrative, art, and so on. But even though reason in these contexts needn’t secure consensus, it is nonetheless useful. Often it is useful for understanding that we have an intractable disagreement and why. Sometimes the exchange of arguments shows us that we have one or another form of disagreement that we didn’t know we had. I discuss this further in the case of the double standard in medical testing in the next section.

7 examples Some examples may clarify these views. The work of applied ethicists working on the medical testing problem shows an implicit appeal to a method very like that of pluralist coherence. In chapter 4 I described the way Ruth Macklin frames the problem in part by talking about conflicting values such as respect for autonomy and persons, justice, and the promotion of the collective good, as well as the way she considers the question of whether the difference in the double standard – that of living in a relatively rich country or a developing one – is morally relevant in the right way to justify differential treatment. And as I mentioned in chapter 1 Macklin also considers whether we would enrol underinsured Americans in trials similar to those proposed for developing countries, arguing that we generally would not. This question is a good one to ask from the point of view of case consistency, since it seeks to determine what the significant differences are that we can stand behind. Anyone who wants to say that the trials are acceptable because the people in the placebo control group are no worse off than they would have been otherwise should apply this standard to

Pluralist Coherence

129

people in his own country too; if he is unwilling to do so, then this suggests inconsistency on his part.33 This inconsistency could be resolved by anyone willing to stand behind “being in our country” and “being in another country” as a morally significant difference. I mentioned above the Times letter-writer who said that countries ought to value citizens’ lives more than those of non-citizens. Such a view can be internally coherent from the point of view explicated here, but only if its proponent is willing to stand behind and endorse this departure from taking lives to be of equal value. In his essay “Testing Our Drugs on the Poor Abroad,” Thomas Pogge proposes several provocative analogies meant to disrupt our assurance in certain simple assumptions. From the point of view of harms and benefits it might be thought that the relevant question is simply whether the benefits outweigh the harms. If this is the only relevant question, it might seem we could justify not only the AIDS trials described earlier, but also the even more controversial – and eventually cancelled – “Surfaxin” trials, which intended to test a drug meant for use in the developed world by using a placebo control in the developing world. Surfaxin was a new version of an existing product that helps premature infants breathe. If a placebo-controlled trial were to be conducted in a developing country, half the infants in the test would get treatment that they would otherwise have no access to. Since the infants getting the placebo are no worse off than they would have been, and the infants getting the new drug are being benefitted greatly, the benefits outweigh the harms. Pogge points out, however, that resulting net benefits do not always make an action justifiable. A rich eccentric who wants to spatter paint on random people does not make her action justified if she turns the prank into a two-person lottery in which one person gets the paint and the other gets $30,000. This is because the harms and benefits are detachable: she could just give away the money and not do the prank at all. And likewise in the Surfaxin case: anyone could, after all, give all the infants the drug known to be effective and not produce any harms at all. So the Surfaxin trial would be wrong. Another analogy, meant to show that informed consent is not the only morally relevant feature, involves a filmmaker who wants to film the slow deaths of some sailors lost at sea. She offers them a deal of a coin flip, followed by rescue if heads and filmed death if tails. They consent to her offer. But clearly she has acted wrongly, and this shows

130

Moral Reasoning in a Pluralistic World

that consent, even when informed, is not sufficient either. These analogies both appeal implicitly to a norm of case consistency: if the cases are similar in morally relevant ways, we ought to judge them similarly. Several commentators on the testing issue make the point that a proper response to the problem of medical testing in developing countries must take into account the circumstances, and particularly the inequality, that creates the conditions for the dilemma. Pogge, for example, says that when labs carry out such tests they are taking advantage of “social injustice,” and Andrew Siegel says that our failures to obey the duties of beneficence make us all complicit in any wrongs done to the test subjects in these cases.34 The pluralist coherentist’s emphasis on the remainder makes the same point: in addition to trying to do the right thing overall in the immediate circumstances, we should be seeking to do the things that will render conflict less likely to arise. There are several ways we might try to bring about a world so that the conflicts in question do not arise. Obviously, if there are fewer diseases to cure, these conflicts arise less often. But the particular problem of the double standard also arises because of the degree of inequality between developing countries and richer countries. It is only because of this inequality that justice conflicts with beneficence in the case. So pursuing policies that would lessen inequalities would also help create a world in which the relevant conflicts are less likely to arise. Taking pluralist coherence as a model for moral reasoning would show three further things in this context. First, arguments based on the pursuit of case consistency through consideration of morally relevant similarities and differences are not, in virtue of this, less rigorous, logical, rational, or hard-headed than those that apply a unified and systematized moral theory such as utilitarianism or neo-Kantianism. Second, the debate over whether double standards are ever justified is properly and fruitfully conducted as a debate over what to do morally, and need not involve questions about relativism, absoluteness, and the nature of moral norms. Finally, the mutuality principle might offer a set of actions we can agree on even in the context of disagreement: we may not share judgments about what to do overall, but we may share a judgment that there is a conflict. As another example of how pluralist coherence functions, consider arguments and debates about abortion. As I explained in chapter 4, some canonical arguments about abortion can be viewed as implicit-

Pluralist Coherence

131

ly relying on case consistency. The most illustrative example of this is Judith Jarvis Thomson’s famous violinist argument; as I explained in chapter 4, this argues that just as we have the right to detach ourselves from an ailing violinist, so a pregnant woman has the right to detach herself from a fetus, even if that fetus is a person and is dependent on her for life. This argument proceeds by asking us to agree that two cases are similar in morally significant ways: the violinist attachee and the pregnant mother. It then asks us to draw the same conclusion in the one case that we draw in the other. Nothing about the argument hinges on systematizing or unifying: indeed, it is part of the framework of the argument that there are multiple moral values at stake: our rights to self-determination, the violinist and fetus’s rights to life, our duties to help others. The claim of the argument is that if these are to be weighed against one another in a certain way in a range of violinistlike cases, they should be weighed against one another in the same way in the abortion case. Much of the discussion and criticism of Thomson’s argument has focused on these claims of morally relevant similarities and differences, and in particular on the ways that the situation of the kidnappee is and is not like the situation of the pregnant woman. Is the fetus like the violinist? Is the relationship the same? Is having sex like being kidnapped? Sometimes, if not always? Maybe in the case of rape? This method of analysis again reflects an implicit case consistency norm: we proceed by showing morally relevant similarities and differences and to draw conclusions having to do with appropriate prioritization. Crucially, pluralist coherence says that one moral consideration can “override” another without the former cancelling out or erasing the latter. This means that in the case of abortion, prioritizing among the principles having to do with fetus’s value and those having to do with women’s autonomy does not require us to understand those as “delimiting” one another or to say that the conflict was “dissolved.” Margaret Little points out that in the abortion debate we often talk as if the fetus’s rights, rather than “outweighing” a woman’s rights, actually “circumscribe” them: if and when the fetus’s right to life is paramount, we talk as if the woman’s rights have disappeared.35 But the reality of the remainder in pluralist coherence shows clearly why talking this way is mistaken: values should not be interpreted

132

Moral Reasoning in a Pluralistic World

as delimiting or circumscribing one another but rather as genuinely conflicting. A virtue of pluralist coherence is that it allows us to preserve the sense that the values in question may genuinely conflict, while explaining the fact that different judgments must be justified in terms of morally significant differences between cases. There is, of course, deep disagreement over which cases of abortion – if any – are appropriately classed as involving “conflicts,” partly because there is deep disagreement over which fetuses – if any – have moral status. In her original paper on dilemmas, Marcus discusses the case of abortion, claiming that what most views share is the mistaken idea that “there is, despite uncertainty, a resolution without residue; that there is a correct set of metaphysical claims, principles, and priority rankings of principles which will justify the choice.”36 To see abortion as a case of conflict means that in some cases, no matter what is chosen overall, and even if that choice is justified, there may be a remainder. In these cases, the mutuality principle indicates that we ought to minimize the likelihood that conflict will arise, which in this case means taking steps to try to minimize the likelihood of unwanted pregnancies – something widely thought to be a good idea. I claim that there is a payoff to the use of pluralist coherence in this context. For a range of views on the morality of abortion itself, there are advantages to interpreting at least some kinds of cases as presenting occasions of moral conflict: that is, as conflicts between the value of a fetus and the duties we owe to it and the values of autonomy and the rights we have to pursue and direct our lives as we think best and to preserve our own health and well-being. Advantages to pro-life views of such an interpretation are obvious: it is more plausible, and will be more persuasive to more people, to say that the rights of the pregnant person are overridden than to say that they are non-existent. To say they are non-existent requires embedding opposition to abortion in a set of values that treats women as not fully persons, which is both extreme and unlikely to move people who are undecided. An advantage to pro-choice views of such an interpretation is that conflict allows us a plausible way to say consistently both that it makes sense that the decision to have an abortion might be an agonizing one and to say that the decision is the right one. This helps explain the commonly held moral feeling that even if abortion is permissible it would be wrong for another person to pressure someone to choose it. And of course the remainder explains the common-sense preference

Pluralist Coherence

133

for birth control over abortion. This is a way of avoiding the likelihood of conflict in the future, and thus rests on the mutuality principle; it is thus compatible with saying that the value of autonomy is overriding in relevant cases. Pluralist coherence asks each person to weigh the various moral considerations, obligations and rights at stake in this kind of case in the same way in other kinds of cases, but even when interlocutors do this well, disagreement can persist. For example, a person who is against abortion on grounds that it is never permissible to harm human life, because this duty always outweighs others, has prima facie reason to now judge similarly cases involving other dependent persons. A person who values personal autonomy in other realms, taking respect for autonomy to outweigh other duties, has prima facie reason to accord respect to a woman’s personal autonomy in abortion cases as well. A natural implication is that people who value differently, with different priorities, will draw different conclusions about cases, even when each of them is internally case consistent. People who care less about personal autonomy will judge differently from those who care about it more and prioritize it more highly. People who share the values of fetal life and autonomy may differ in their moral beliefs about cases because they prioritize among these values differently.37

8 pluralist coherence compared to other approaches It is worth considering how pluralist coherence compares to other approaches that have proposed pursuing coherence or consistency in pluralist contexts. Though the method of pluralist coherence bears certain affinities with that associated with the two dominant approaches in bioethics – those of Beauchamp and Childress and of Gert, Culver, and Clouser – there are important differences, largely having to do with the treatment of moral conflict. Beauchamp and Childress explicitly endorse coherence as an appropriate method for discovering and justifying moral beliefs, and they implicitly endorse a framework like that of “conviction ethics,” in which we treat our moral convictions as having prima facie plausibility and seek the most coherent theory that is sufficiently intuitively acceptable. They emphasize that moral reasoning must contain both top-down and bottom-up reasoning: that is, we must use principles to generate and evaluate judgments but also

134

Moral Reasoning in a Pluralistic World

use cases and judgments to generate and evaluate principles.38 Though they make use of the phrase “prima facie” in describing the obligations the principles commit us to, and insist that in dilemmas we are not obligated to do both of the incompatible actions, they acknowledge the importance of the remainder.39 This shows that in using the expression “prima facie” to describe obligations, they are not committing themselves to the Rossian idea that when one obligation overrides another the second somehow fails to exist as a real obligation. So far this is similar to the approach I am offering here. There are nevertheless important differences. Beauchamp and Childress understand judgement in cases of conflict through the idea of “balancing”: this is “the process of finding reasons to support beliefs about which moral norms should prevail.”40 Beauchamp and Childress emphasize that balancing is not purely intuitive but involves considering reasons for assigning weights in certain ways. In their example, a physician who has promised to take her son to the library may encounter a conflict if an emergency case requires her attention. Beauchamp and Childress say that in balancing this physician will engage “in a process of deliberation that leads her to consider how urgently her son needs to get to the library, whether they could go later to the library, whether another physician could handle the case, and so on.”41 If a life hangs in the balance, and only her expertise will save the patient, her overriding obligation will be to stay at work. In a way this is similar to determining weights and valences as described in chapter 4. But crucially, there is no mention of a norm like case consistency. Indeed, the structure of their discussion seems to imply that coherence and balancing are unrelated: coherence tells us about the principles; balancing tells us about conflict. This is partly manifested in the way they understand the distinction between balancing and specification. Specification, which as we’ve seen is generally the specification of principles with added clauses saying how and when they are to be applied, is characterized here as being “concerned primarily with [the] scope (i.e., range)” of principles.42 In their example illustrating specification, psychiatrists encounter a conflict of duties when asked to do forensic evaluations of patients incapable of consenting to the evaluation. A specification of the informed consent principle in this case says: “Respect the autonomy of persons who are the subject of forensic evaluations, where consent is not legally required, by disclosing to the evaluee the nature and purpose of the evaluation.”43

Pluralist Coherence

135

Interestingly, Beauchamp and Childress associate specification with the application of a kind of consistency, writing: “Specification requires that a moral agent extend norms by both narrowing their scope and generalizing to relevantly similar circumstances. Thus, ‘respect the autonomy of competent patients when they become incompetent by following their advance directives’ is a rule suited for all incompetent patients with advance directives.” They then claim that not all cases can be handled this way, because some cases require individualized attention. They write: “However, it often seems that the responses of caring moral agents, such as physicians and nurses, are specific to the needs of this patient or this family in this circumstance. Numerous considerations must be weighed and balanced and any generalizations that could be formed might not hold even in related cases ... balancing allows for a due consideration of all norms bearing on a complex, very particular circumstance.”44 This suggests an interpretation in which specification applies when things are relatively straightforward and we know how to invoke the appropriately relevant factors, and balancing is necessary when cases present new, unusual, or particularly complex aspects.45 My understanding of these concepts diverges from this conception both in the way the distinction between balancing and specification is drawn and in the role of consistency. I’ve argued here that the important difference between methods of weighing considerations against one another and methods based on specification involves not consistency but rather the nature of the remainder. And with respect to consistency, pluralist coherence asks us to always aim to be case consistent by weighing moral obligations similarly in similar kinds of circumstances. This may be more easily done in some cases than in others, but remains a worthy aim in all. If a case is particularly unusual or complex, this does not mean the pursuit of consistency is useless there – it just means that pursuit will be unusual or complex. For practical reasons we may not be able to be as specific in such cases, but the same norm of case consistency nonetheless applies. Finally, although Beauchamp and Childress are committed pluralists their perspective on the nature of coherence is not always clear. On the one hand, they seem unworried by the fact that their theory is pluralistic in the sense of having multiple principles. They write: “Many writers in biomedical ethics seem to think that we would

136

Moral Reasoning in a Pluralistic World

rightly have more confidence in our principles and considered judgments if only we could justify them on the basis of a comprehensive ethical theory. However this outlook may have the cart pulling the donkey: We should have more confidence in an ethical theory if it could be shown coherent in a comprehensive way with the norms and considered judgments comprising the common morality. If an ethical theory were to reject any of the four clusters of principles defended in this book, for example, we would have a sound reason for healthy skepticism about the theory, not for skepticism about the principle(s) ... In [our] theory (if it is a ‘theory,’) there is no single unifying principle or concept.”46 On the other hand, they sometimes seem to address their pluralist approach as a second-best or non-ideal – a kind of compromise between the ideally coherent theory and the theory they endorse. In this vein, they write, “can the common morality itself be made coherent? If one argues, as we do, that a heap of obligations and values unconnected by a first principle comprises the common morality, is there any hope of coherence, short of so radically reconstructing norms that they only vaguely resemble the norms in the common morality?”47 In this latter passage, the implication seems to be that it would be better to have a “unified” coherence, and that a “heap” of duties falling under various values and disconnected from one another is somehow the best we can do in aspiring to that. Of course, I disagree that there is anything non-ideal about the form of the theory having multiple principles. Though they endorse a plurality of principles, Gert, Culver, and Clouser emphasize the “systematic” quality of their account. Their system includes principles, moral ideals, and information about what is and is not morally relevant, and includes also a single criterion for determining when an exception to a rule is appropriate.48 The specific principles they recommend they understand as reflecting a “common morality” implicitly endorsed by many of us. Instead of conflict, they talk of exceptions. When it comes to understanding how principles apply in an appropriately general way, they say that to determine whether an exception should be allowed, we should consider the harms that would result from allowing all similar cases to form publicly acceptable and publicly known exceptions. For instance, in deciding whether it would be appropriate to harass or deceive a stroke victim to get him to do his exercises, we would consider what harms

Pluralist Coherence

137

would result from rules publicly allowing people to always harass or deceive stroke victims in this way.49 This method of judging in cases of conflict represents a theoretical departure from the one recommended here: considering the harm that would result from a public rule is different from using the method of weights and valences discussed in chapter 4. The use of this procedure leads to a “case consistent” set of moral beliefs only if one takes only harm prevention to be morally significant in the relevant conflictual cases. From the point of view of pluralist coherence in the context of conviction ethics, it seems plausible that many people would not take this to be the only morally significant consideration. Many of us consider our right to honest information from our doctors as introducing a very strong consideration in itself, and would say that even if somehow publicly allowing lies in certain circumstances would be better overall from the point of view of health, one still has a right to know the truth; the harms associated with making exceptions aren’t relevant. This suggests that case consistency applied in the context of conviction ethics would lead to a different way of finding overall obligations in cases of moral conflict. Next let’s consider Paul Thagard’s view. Thagard embeds a coherence approach to ethical reasoning in a broader context of coherence reasoning generally.50 In ethics, Thagard says, there are four kinds of coherence: deductive, explanatory, analogical, and deliberative. Deductive coherence concerns the relation of entailment between principles and judgments: the principle that one ought to keep one’s promises and the judgment that I ought to keep this promise are deductively coherent; since this is a symmetric notion they thus support one another. Explanatory coherence concerns the way non-moral facts play a role in factual reasoning. Analogical coherence asks us to judge cases similarly that seem to have similar moral aspects to them. All three of these kinds of coherence are broadly compatible with the approach of pluralist coherence, in which we work back and forth between principles and judgments, trying to judge consistently from one case to another – though as I’ve been emphasizing, from the point of view being developed here, the question of what one takes to be an apt analogy necessarily hinges on substantive moral judgments about what are and are not morally significant similarities and differences. Deliberative coherence is the most interesting element in the present context, because it has implications for cases in which goals and

138

Moral Reasoning in a Pluralistic World

actions are and are not appropriately compatible. Thagard says that when an action facilitates a goal there is a positive constraint between them, and when actions are incompatible – in the sense that one cannot do both – there is a negative constraint between them. In general, if goals are unrelated and disconnected deliberative coherence does not posit a negative constraint among them or require us to find a standpoint from which they can be unified. Some goals are desirable for intrinsic reasons, and nothing prevents these from being multiple.51 So Thagardian coherence, like pluralist coherence, does not aim for unifying goals. But coherence does affect some goals. If goals are facilitated by mutually incompatible actions, then a negative constraint arises between them. For example, suppose I have a goal of always eating the desserts my friend makes and a goal of losing weight, and those goals are facilitated by incompatible actions because my friend makes a lot of desserts and I cannot both eat and refrain at the same time. In this case, there is a negative constraint between my goals. The constraint is a soft constraint, in the sense that the good of rendering coherent is taken into consideration alongside other desiderata; the constraints are pressures not requirements. One way to interpret this in the language of principles would be to say that the goals of obeying a set of multiple principles will have soft negative constraints among them whenever incompatible actions facilitate obeying those principles. And incompatible actions facilitate obeying principles just when the duties the principles recommend conflict, i.e. when the principles generate dilemmas. In a case of moral dilemma, if stopping my car and making phone calls facilitates helping an accident victim, and continuing to drive facilitates my keeping my appointment, then since I cannot stop and drive at the same time the set containing only one of the two goals of helping or keeping the appointment will be more coherent than the set containing both goals. There is thus some pressure toward abandoning one of these goals. This constitutes the most substantive difference between the Thagardian view of coherence and my own; as I’ve argued, pluralist coherence does not have a norm recommending avoidance of either practical or essential conflict among principles. This manifests itself in two specific ways. First, in pluralist coherence, when principles do recommend incompatible actions, taking the resulting dilemma to be real

Pluralist Coherence

139

means we retain both actions as goals without positing a negative constraint among them – even a soft constraint. Second, in cases of conflict the important norm is that of prioritizing, rather than rendering coherent. Thus, that the goals of always telling the truth and always keeping one’s promises lead to conflicts does not make them less than fully coherent; when they engender conflicts we prioritize one over the other, but there is no pressure to give one of them up. In this way, pluralist coherence emphasizes the positive roles for moral conflict, while focusing the coherence aspect on judging in accordance with relevant similarities and differences. Finally, let’s look at the views of R.M. Hare, who also emphasizes judging cases similarly to one another. Though he is widely known as a utilitarian, in his 1963 book Freedom and Reason Hare lays out a moral conception that might seem to bear certain theoretical similarities to my pluralist approach. Hare emphasizes that reason requires only that our judgments be universalizable in the sense of applying in the same way to identical or relevantly similar cases. And as with pluralist coherence, he says that people may be committed to a range of values, so that reason cannot rule out the possibility of different people holding different but internally consistent moral views. But there are some deep differences. First, Hare takes as his starting point the use of moral language: his project is to draw conclusions about morality from the logic of moral predication and the meanings of moral words alone. He thus rejects completely the conceptual framework taken for granted here, that our considered moral convictions are good starting places for moral thinking.52 Second, in part because of this conceptual difference, when Hare talks of “universalizing” he does not mean applying a moral judgment in the same way to relevantly similar cases. Instead he means considering whether we would desire, or “accept,” that a kind of act be done in relevantly similar circumstances. As with the golden rule, this means we ought to treat others as we would want to be treated. Hare suggests that this question appropriately hangs on interests: we should ask ourselves whether the satisfaction from an action outweighs any disutility we would experience if we were other persons affected by it. Because my approach begins with moral convictions, however, it asks us to consider how we would morally judge that a kind of act be done in relevantly similar circumstances. To see more clearly the difference, consider Marie’s case. Here Hare’s universalization requires Marie to

140

Moral Reasoning in a Pluralistic World

consider her situation from her brother’s point of view: would she accept being lied to in his place? Marie should ask herself whether the satisfaction of her interest from lying would outweigh the disutility of being lied to in her brother’s position. This is a very different question from the one posed by case consistency, which has to do with the moral aspect of lying in relevantly similar cases. Third, Hare’s understanding of moral relevance is different from that associated with pluralist coherence. Hare proposes using hypothetical interchangeability to infer what should, and should not, count as a morally significant difference; I explained in chapter 4 section 2 why in the present context these inferences cannot work. But in addition, the application of the interchangeability test rests on judgments about moral similarity that the pluralist coherentist may or may not endorse: e.g. that “being a person” is a morally relevant similarity.53 A final, and perhaps most interesting, point concerns what Hare calls “fanaticism.” Recall that universalization for Hare appropriately concerns asking how an action affects the interests of people affected, where we take ourselves to be interchangeable with others. This focuses on interests, not on other considerations or moral convictions. In Freedom and Reason, Hare does admit the possibility of a commitment to a moral ideal that would cause a person to consider matters beyond interests. This is fanaticism, and Hare speaks of it in the most negative terms: people become “really dangerous” when they pursue ideals at the cost of others’ interests. The examples he considers include racism, fascism, and “the sanctity of property rights.”54 But from the point of view of pluralist coherence, the distinction Hare makes between reasonable or neutral moral views and “fanaticism” cannot be drawn: in my view all moral judgments rest at bottom on considered convictions, which themselves have moral content – so that everyone with moral beliefs would be, in some sense, like Hare’s “fanatics.” Even if we grant, for the sake of argument, that beliefs resting on universalized judgments about interests are in some sense “neutral,” it would still be the case that the fanatics would include anyone – like Rawls – who judges that justice is sometimes due to people regardless of utilitarian considerations. A pluralist coherentist naturally rejects conclusions like Hare’s on grounds that from her perspective no non-moral arguments can show that judgments based on values like justice are somehow less rational or justified.55

Pluralist Coherence

141

Later in life Hare deepened his commitment to utilitarianism and rejected the possibility of rational commitment to ideals. In his 1981 book Moral Thinking, Hare argued that only consideration of interests, and thus only utilitarianism, was consistent with rationality. This move was partly based on the new idea that interchangeability could be applied to moral commitments as well as to interests, so that universalization meant not only the consistency that arises from putting oneself in the other’s place, but also, in different stages of moral reasoning, taking equal account of rival ideals. Even setting aside the differences already mentioned, the pluralist coherentist sees no reason to follow this: case consistency does not require that a person committed to justice and a person committed to expediency give the ideals of the other consideration. As Peter Singer says in a critical discussion of Hare, this essentially treats moral convictions as kinds of desires.56

142

Moral Reasoning in a Pluralistic World

6 Conclusion

In this chapter I give a brief overview of my arguments, then draw out three implications. The first concerns philosophical methodology and addresses both a striking disconnect between theoretical ethics and applied ethics and also the peculiar pedagogical experience of introducing applied ethics students to moral reasoning by exposing them to multiple, incompatible, unified theories. The second counters the idea that pluralism makes moral reasoning and progress somehow impossible. The third proposes a more optimistic interpretation of our public moral discourse than is usually offered, by showing how entrenched moral disagreement can persist even when interlocutors are reasoning well. I end with some brief concluding thoughts about how moral change can be facilitated.

1 overview The question I began with was how moral reasoning is best understood in a world with varying moral outlooks, a diversity of moral views, and substantive moral disagreement. How should we understand coherence as a method of moral reasoning in pluralist contexts – that is, in contexts in which people value various things in various ways and to varying degrees? I’ve argued that traditional conceptions of coherence, in which we aim to find as few simple principles as possible, are inappropriate for use in our pluralist context. They stack the deck against pluralist theories for no good reason; in the absence of independent reasons for preferring unified theories, the adoption of traditional conceptions of coherence is based on reasoning that is

Conclusion

143

question begging. This does not show that there cannot be any such reasons, and I do not claim to have shown that unified views are untenable. But I have tried to show the most common strategies in favour of unification do not succeed, and in particular that coherence itself is not a reason for unification. I’ve argued in favour of an alternative conception, that of pluralist coherence. Pluralist coherence is based on case consistency, and adds to this a mutuality principle – that we ought to try to arrange our lives and institutions so that dilemmas do not arise – and a norm encouraging the expansion of our circle of concern. We have reason to understand our values – such as liberty and equality – as genuinely distinct and conflicting, rather than as unified by a “super-value” or “delimited” by one another. Rationality and autonomy do not require a form of unity that rejects essential conflicts in values or ambivalence. Compromises among values are necessary: what is needed are appropriate compromises, but these do not entail unification or delimiting. Though it is sometimes said that beliefs in richly coherent sets support or explain one another, and hence are better justified, I’ve argued that this is not so. What we need to know in the context of justification is which explanations are better and worse than others; but in the absence of assumptions about the unity of value that would be question begging in the present context, we have no reason to think the explanations in more systematized theories are better than those in more pluralistic ones. Analogies with coherence in science and mathematics also do not support a norm of systematicity in the moral domain. In the alternative I’ve proposed – pluralist coherence – the central norm is that of case consistency. In the context of principle pluralism this entails two things. First, when judging how strong our obligations are in cases of conflict, we ought to assign similar strengths to similar obligations in similar contexts. For instance, if we judge that the obligation to tell certain kinds of lies is appropriate to protect family members in certain circumstances, we ought to judge similar lies to be justified in similar circumstances. Second, the form of our principles ought not introduce different judgments where there are no morally significant differences. The principles “telling a lie at home is wrong” and “telling a lie outside is permissible” are case inconsistent as long as one regards “being at home” versus “being outside” as a morally insignifi-

144

Moral Reasoning in a Pluralistic World

cant difference. I discussed the example of framing effects: instances in which we judge cases differently only because different words are used to describe them. Since virtually no one would be willing to say that the words used in descriptions should create morally significant differences between cases, case consistency requires us to judge similarly cases that are the same except for the way they are described. Judgments about moral significance are properly seen as moral judgments, because they are judgments about what matters and why. In the case of an elderly woman who is sick and wants to die, we tend to think that her age and wishes are relevant, as is the question of whether she is depressed. Whether it would be more convenient for her neighbours for her to die now is not relevant. These judgments of moral relevance are moral judgments and thus are understood as other moral convictions are: we begin with some judgments and try to bring them into coherence. Importantly, to reasonably judge differently we must be willing to stand behind a judgment as being morally significant. Therefore part of why we ought to be consistent in this sense has to do with the value of integrity in the sense of standing for something. Against the use of case consistency as coherence norm it may be objected that it gives us no way of determining, in a non-subjective way, what is and is not a relevant distinction or similarity. This means that using the method of pluralist coherence gives us no way to rid our judgments of arbitrariness. I’ve argued for a three-part response to this. First, judgments of arbitrariness are always relative to some standard of evaluation: what is arbitrary from one point of view is not arbitrary from another. Second, rich coherence views are worse off than pluralist coherence from the point of view of arbitrariness. Finally, because we must stand behind the judgments of moral relevance we endorse, the problem of people introducing “distinctions” they know to be completely arbitrary is not one that arises much in practice. Pluralist coherence adds to case consistency two elements: in cases of conflict and dilemma, moral remainders tell us to try to change our lives and institutions so that dilemmas are less likely to arise, and expansion recommends making larger our circle of moral concern. These lead to two different kinds of moral progress compatible with pluralism and disagreement. Because we endorse multiple values, direct our cares differently, and prioritize among them differently we will come to different

Conclusion

145

judgments both about how strong our obligations are and about what counts as a morally significant difference or similarity. Two things follow from this. First, anyone’s set of moral beliefs is improved by being made “internally” case consistent – or consistent on its own terms. In the case of framing effects, since virtually no one would be willing to say that the words used in descriptions should create morally significant differences between cases, case consistency requires us to judge similarly cases that are the same except for the way they are described. But second, two different sets of moral beliefs can each be internally case consistent without agreeing with one another, and without being case consistent from each other’s point of view. So, for example, a person who values honesty above all may be case consistent and judge that it is never permissible to lie to protect one’s friends, while a person who values honesty less may be case consistent and judge that it is sometimes permissible to lie to protect one’s friends. A person who believes that being a man or woman is the kind of thing that can create a morally significant difference will be case consistent in endorsing principles assigning them different duties and obligations; a person who does not believe there are such morally significant differences must judge cases similarly regardless of sex.

2 implications: unified theories and philosophical methodology If I am right about coherence and pluralism, a first implication concerns methodology. In chapters 4 and 5 I gave several examples of reasoning in applied ethics that implicitly made use of a method much like that of pluralist coherence: arguing on behalf of a certain set of morally significant similarities and differences and pointing out why certain ways of forming compromises among conflicting values were more appropriate than others. We might classify such work as part of applied ethics. But as I argued in chapters 1 and 3, work on moral reasoning itself – work we might classify as part of metaethics – typically endorses conceptions of coherence that are rich and systematizing. It is common for philosophers taking a theoretical approach to prefer systematized theories to pluralist ones. Even defenders of Rossian theories argue for those theories partly on the grounds that they are, or can be made to be, more systematic than they seem.

146

Moral Reasoning in a Pluralistic World

On the face of it this suggests a disconnect between methods considered theoretically appropriate and methods that are actually used. It might seem that one way of bridging this disconnection is through the idea of trade-offs. Keeping in mind that coherence methods always require making trade-offs between coherence desiderata and intuitive acceptability, we may say that while systematicity is desirable, there is no systematic theory that is sufficiently intuitively acceptable. Beauchamp and Childress imply a thought along these lines when they argue for a pluralist view but also list norms such as simplicity in their coherence desiderata.1 This implies that applied ethicists use pluralist approaches not because they are best but because they are a best approximation to the ideal. But I’ve argued here against this trade-off status for systematicity insofar as we don’t have reason to prefer systematic theories even in this sense. That the trade-off interpretation is unsuccessful matters practically because, as I explain here, even an unmet preference for unified and systematic theories leads to problematic methodology in moral reasoning. Consider again the example of abortion. From the point of view of rich coherence, systematicity, and unity the way to reason about abortion is to figure out the most unified theory that fits best with our considered moral convictions and then apply it to figure out what to believe about abortion. Of course, the immediate problem is that there is deep and widespread disagreement, especially among philosophers, about which unified theory is best. And different unified theories, of course, yield different results. Preference utilitarians like Singer say that we ought to maximize the preference satisfaction of all beings affected by our actions. Singer has argued that on these grounds abortion is generally permissible: up until around eighteen weeks a fetus cannot feel any pain or pleasure and so has no interests at all; even after that, because the fetus has no rationality, autonomy, or self-consciousness its interests will generally be outweighed by those of a pregnant woman who wants to choose abortion.2 In cases in which overall preference satisfaction is best promoted by abortion, this is a morally good choice – that is, not just morally permissible but morally good. Some Kantian theories emphasize the importance of being able to apply universally the principle under which one is acting. In “A Kantian Argument against Abortion,” Harry Gensler argues that since

Conclusion

147

none of us would consent to having been aborted, we ought not abort others. Other Kantian theories emphasize the importance of valuing rationality and autonomy. In “Animality and Agency: A Kantian Approach to Abortion,” Lara Denis argues that because women have a duty of respecting themselves as a rational free agents, and fetuses are not rational free agents, both the reasons for abortion and the ways in which it affects a woman’s valuing nature are significant. It follows that some abortions are morally permissible, and some are even required – for example to save a woman's life. The multiplicity of unified views and their implications is awkward for anyone who wants to defend the usefulness of systematized moral theories in figuring out what is right to do. In a review of Robert Audi’s Moral Value and Human Diversity, a book that – among other things – attempts to show what major moral theories have in common, Diane Jeske expresses the awkwardness this way: “if the supposed ‘experts’ cannot reach agreement, [non-experts] will conclude, we have even more reason for regarding moral questions as lacking objective answers” and even more reason to think that is no rational way of adjudicating competing positions.3 In university courses on contemporary moral issues this difficulty manifests itself as a practical problem. It is common in such courses to begin by introducing theories to be used as the course goes along. That is, a range of different theories is presented, some unified, and then we refer to those theories in discussing the issues themselves. This method raises questions. If there are many moral theories that seem plausible, doesn’t this suggest a state of deep indecision on the part of the experts? If even moral theorists cannot agree on how to proceed, what hope is there for the rest of us? In controversial cases like abortion we are likely to disagree about which richly coherent theory fits best with our considered convictions, since our considered convictions vary and are often strongly felt. In their bioethics work Gert, Culver, and Clouser call this way of using theories the “anthology” approach, saying it comes to the same thing as being an “antimoral theory” altogether.4 One way of trying to address this difficulty goes something like this. Moral theories help us think systematically about moral difficulties, so in that sense they are useful. While it is true that there are multiple systematic theories, we find in practice that some moral theories

148

Moral Reasoning in a Pluralistic World

are better suited to thinking about some problems than others since they seem to focus on what is essential to the difficulty. For example, when thinking about policy questions that concern many people, it may seem natural to think in utilitarian terms since utilitarianism gives us a way of evaluating different proposals that affect a lot of people. But it may seem natural to think of an issue like sex work in terms of Kantian respect for autonomy since this focuses on the importance of not using another person as a mere means, as clients or others may seem to “use” these workers – especially if there is deception or coercion involved. At the start of his moral issues textbook Mark Timmons includes a “Moral Theory Primer” that addresses the issue along these lines. Suggesting that utilitarianism may be well-suited to thinking about torture and “ticking bomb” cases but not to other moral questions, Timmons says that if we set aside the competing answers problem “we might then view [the theories in question] as providing different ways of diagnosing and thinking about a moral problem, where in some cases the best approach is utilitarian, whereas in others it is a virtue ethics perspective, and still in others, some other moral theory best gets at what is morally most important to consider. In other words ... some practical moral questions are best approached from, say, the perspective of act utilitarianism, others not.”5 But my proposal of pluralist coherence suggests that this approach is deeply mistaken, because it ignores – even flagrantly violates – the norm of case consistency. Essential to pluralist coherence is the aim of finding a way to prioritize among what we value in a way that can be applied from one case to another in a relatively stable way, and judging cases similarly unless we can cite morally significant differences between them. This means aiming to apply the same range of prioritized values to a range of different problems, rather than applying a different theory to illuminate each problem. For example, if Kantian respect for autonomy and persons seems a good explanation for some of our judgments, and thus to be an appropriate object of value, then it is an appropriate object of value in all cases – including ones like policy questions where its implications may not be obvious. Anyone endorsing a prioritization in which respect for individual rights trumps other values may not then appeal to utilitarianism to figure out when it is permissible to torture people.

Conclusion

149

What each moral reasoner ought to strive for is a pluralistically coherent moral point of view that honours various values in appropriate ways and applies those values across the board. This shows the problem with the anthology approach: instead of applying a single prioritized set of values across the board it encourages us to apply one value to one case and another to another. It applies different values to different cases with no consideration of whether those cases differ in morally significant ways. Rejecting the anthology approach does not require that we ignore moral theories, since the values and ways of thinking they introduce are useful. The question, however, would not be which one is right but rather how to integrate the concerns they present. The important question would be how to make sensible compromises that can apply broadly. In the passage cited above Timmons goes on to say that some moral theories seem like the right tools for some jobs and not others because they seem to capture what is morally significant about the situation: while utilitarianism’s counting up may suit “ticking bomb” cases, it may not suit environmental issues or situations where individual rights and duties seem most relevant.6 But from the point of view of pluralist coherence, this way of seeming appropriate or most useful should function as a red flag, a warning against doing what the anthology approach suggests. Situations that lend themselves to counting up and comparing harms and benefits frequently distract us from principles of protection for individual rights that seem so important to us in other settings and thus ought to matter to us in this setting as well. If respect for individual autonomy matters in other cases, it matters in the ticking bomb cases as well; it can then be overridden, but it does not disappear just because adding harms and benefits comes so naturally to us in the given context. This is not the place for a full discussion of pedagogy, but I will note some advantages to using pluralist coherence in teaching moral issues. The first is that the norm of case consistency is itself a useful teaching tool: it is intuitive, its reasoning style is deployed in many applied ethics arguments, and it allows students to see ways that their own views may be problematic even if no unified ethical theory is the right one. If a student confronts the fact that her views about abortion are internally inconsistent with her views about the destruction of fetuses created for IVF treatments, then reflects on the matter, this is a pedagogical success.

150

Moral Reasoning in a Pluralistic World

More broadly, using pluralist coherence in moral issues courses would allow students to see moral inquiry less as a passive task of theory identification and application and more as the construction of a coherent personal ethical outlook. Because the method begins with convictions, students would be encouraged to see their existing beliefs as relevant to the process. The teaching process would encourage students to examine their own beliefs for internal coherence, articulating what they do and do not see as morally relevant similarities and differences, confronting challenges to those beliefs from the arguments in the readings, and seeing what coherence commits them to about other matters in virtue of their beliefs. For example, students who appeal to a general principle against killing humans to explain their own conclusions about abortion would have to confront what those beliefs would entail about euthanasia, torture, and war, potentially challenging their previous beliefs on those matters. In addition to its intellectual value, such an activity would be an exercise in the development of a student’s ethical self. Finally, using pluralist coherence in teaching moral issues can help students understand that some moral disputes are intractable for deep reasons, and not merely because one group of people is unreasonable. Seeing how disagreements can arise because of different values and prioritizations – and analyzing these using pluralist coherence – can show students how alternative views can be internally coherent and nonetheless differ from their own. Given how little time most people spend seriously engaging with the ideas of those who value very differently from themselves, a student who sees an alternative view as reasonable instead of merely unthinking can also represent a pedagogical success.

3 implications: pluralism and the impossibility of moral reasoning and moral progress The second implication concerns the idea that there is an incompatibility between pluralism and reasoning – that pluralism somehow makes the use of reasoning impossible. Richard Posner suggests this, saying that people who believe that reason can help us in contexts of moral diversity are relying on a false idea that people share moral premises and only disagree when someone has made an empirical or

Conclusion

151

logical mistake. He writes: “The hope of philosophers who stress the moral diversity within every society is that members of the same society will be reasoning from the same moral premises, so that if they disagree on some moral question this must mean that one of them has made a logical or empirical mistake, in which event there may be a demonstrably correct answer to the question, at least for that society. The hope is forlorn. Moral contestants, even when they are members of the same society, typically do not agree on all the premises from which they argue. And so moral pluralism provides no leverage for moral critique, but if anything reinforces the lesson of moral relativism.”7 Here I take Posner to be saying that the possibility of a fruitful use of reason in morality is dependent on agreement – agreement that we do not have because of moral diversity and pluralism. Furthermore, he says, even when we acknowledge that we are being inconsistent this needn’t move us to change: we can be “logically” consistent and “behaviourally” inconsistent, for example if we continue to eat meat after acknowledging the suffering of animals.8 Thus the hope of reason’s being useful in moral debate and moral practice is futile. He illustrates his claim with the example of abortion, writing that “abortion is moral in cultures that have liberal attitudes toward sex or that have adopted a feminist ideology, but it is immoral in ones that want to limit sexual freedom, promote population growth, or advance religious beliefs in the sanctity of human life. These cultures coexist in the United States, and their respective adherents do not share enough moral common ground to reason to agreement.”9 This line of thought attempts to connect two things: persistent disagreement in which people’s moral beliefs are highly correlated with their social world and the uselessness of moral reasoning. But from the point of view of pluralist coherence, this form of disagreement is compatible with reasoning – and even explainable and expected. In the view presented here, such disagreement and correlation is exactly what we would expect to find. People who direct their values differently and thus have a different sense of what counts as a morally significant difference will differ in their judgments about abortion. But this does not mean that they are not reasoning. I say more about this in section 4. In his attack on the use of moral reasoning in general and on moral philosophers in particular Posner singles out Judith Jarvis Thomson’s violinist analogy for special scorn. In criticizing Thomson’s argument

152

Moral Reasoning in a Pluralistic World

as a use of the kind of “moral reasoning” he thinks unjustifiable, Posner argues that the relationship between a mother and a fetus is not like the relationship between two strangers. But from the point of view of pluralist coherence, to say this is not to criticize moral reasoning, but rather to engage in it on precisely the recommended terms. It is to argue for a morally significant difference, one that moral philosophers have also argued for.10 Indeed in another context Posner goes so far as to praise the modest back and forth he finds in certain applied ethics work: “When used modestly in specialized fields of applied moral theory, such as bioethics, [reflective equilibrium] can produce a commonsensical kind of policy analysis, illustrated by a book by James Childress on bioethics.” He then charges that bioethics in this style “is a field broken off from philosophy.”11 From the point of view being developed here, any division between bioethics and the rest of philosophy disappears: the modest back and forth associated with this kind of applied reasoning is absolutely of a piece with the metaethical demands for coherence and a full role for reasoning in the moral realm.12 When commonsensical policy analysis takes this form it is an instance of – not opposed to – moral reasoning. The methodology Posner praises here seems to be the same as the methodology Posner is so critical of in Thomson: that of reasoning back and forth between general theoretical principles and intuitions about particular situations. Posner also suggests that pluralism is a problem for moral progress. But as we’ve seen, the pluralist view of coherence with remainders and expansion allows us to say at least a few useful things about moral progress without making assumptions about moral absoluteness and objectivity.13 Creating conditions under which dilemmas are less likely to arise gives us a kind of practical moral progress, as does the expansion of our circle of concern. The kind of practical moral progress that results from improving the world so that dilemmas are less likely to arise does not require anything like moral absolutes or objectivity, and expansion is compatible with multiple conflicting values and principles. Indeed, even in cases in which we disagree over prioritizations in conflicts we can sometimes agree on steps to avoid the relevant conflicts in the future, and on ways to extend protections, rights, and moral care to a larger group.

Conclusion

153

4 implications: moral psychology and the bitterness and futility of moral debate Most of Posner’s criticism focuses on the status of reasoning itself, not on whether people engage in it properly. The third implication I want to discuss involves not whether moral reasoning is possible at all, but rather the degree to which people engage in it. This matter concerns the degree to which people do use reason in figuring out what to believe morally in practice, and the claim that it is because that degree is low that there is so much “futility and bitterness” to moral debate in public discourse. Some moral psychologists have said that empirical data about the way we form and defend moral judgments show that reason doesn’t really play a role in that practice: when we give “reasons” for our moral beliefs they are at best “post hoc” window dressing on beliefs already formed. Observing people making and defending moral judgments, for example, Jonathan Haidt says that typically reason plays virtually no role whatsoever.14 We like to think of ourselves as reasoning out what to believe from general principles. But that’s not what we do: instead, we make snap judgments, and only if pressed look for reasons to back them up. For example, Haidt says that people who were told a story about a brother and sister who had sex once, were happy they did it, and used two methods of birth control, nonetheless judged immediately and intuitively that what the siblings had done was wrong; when pressed to explain why, they searched for reasons to justify the initial judgment. Even when they couldn’t find any good ones, they persisted in their judgment. In explaining his “social intuitionist” model, Haidt says that people think more like lawyers than judges, looking for reasons to support a belief already formed rather than using available evidence to determine what the right belief is. The model is intuitionist because it posits that moral judgments are based on quickly formed and unreflective intuitions; it is social because it posits that we form our moral beliefs and attitudes largely in response to others in our social world, rather than through private reflection or reasoned judgment.15 Sometimes, people with beliefs they cannot support with reasons will continue in their belief, despite having no reasons – they are, in Haidt’s phrase, “morally dumbfounded.”16

154

Moral Reasoning in a Pluralistic World

Of course, the mere fact of reliance on snap judgments does not tell us the whole story about how people do and do not use reason in the moral domain. We make snap judgments in ordinary life all the time in many domains in which we have excellent reasons for our beliefs. For example, we frequently make snap judgments about how other people are likely to behave in given situations, saying immediately and without complex thought that they will or will not do certain things and behave in certain ways. These judgments could easily be supported by appeal to the ways people usually behave, but no one relies on reasoning at the moment to arrive at the judgment. They make quick intuitive judgments, and if pressed, search for reasons to back them up. But if we are relying implicitly on reason in our judgments in this way, then it should be possible, in principle, to come to see that our initial judgments were mistaken. That is, it should be possible for a person to use reason to discover that her initial intuition about a judgment was mistaken, or did not fit with or contradicted other beliefs she holds, and to decide on that basis that her initial judgment was wrong. Though Haidt acknowledges this as a real possibility, he argues that it is a rare occurrence. That is, while we can use reason to figure out that an initial judgment doesn’t fit with our other considered convictions, and thus to change it, we just don’t. Haidt puts it this way: “The literature on everyday reasoning suggests that this ability [to change one’s mind as a result of reasoning] is common only among philosophers, who have been extensively trained and socialized to follow reasoning even to very disturbing conclusions.”17 This would mean that for the vast majority of us, reason is playing virtually no role in moral belief formation. Haidt’s title, “The Emotional Dog and Its Rational Tail,” provides an apt metaphor for his point: we think that reason is in charge, and is the dog wagging the emotional tail, but in fact, the emotions are in control. In response to Haidt, David Pizarro and Paul Bloom point out that there are ways of “educating” our moral intuitions.18 Involuntary responses often contain some degree of cognitive appraisal: hearing a person whistling on the street will cause you no fear, but hearing a person whistling when you are alone in your house in the middle of the night will scare the dickens out of you. Our point of view affects how we react to things, and it is possible to change our point of view. Bloom and Pizarro suggest that when we adopt the perspective of another person, for example by putting ourselves in their place, we

Conclusion

155

gain the perspective that grounds moral beliefs – such as that requiring we treat others the way we would like to be treated. Furthermore, putting yourself in certain situations changes the way you react to things: students who took classes on racism taught by an AfricanAmerican professor were less likely to have implicit negative attitudes toward African Americans.19 And daily life presents us with innumerable conflicts requiring our deliberation, such as those between family demands and work obligations (like those I discussed in chapter 2).20 Some evidence suggests that people in moral quandaries, such as young women trying to decide whether to end their pregnancies through abortion, do engage in reflective and deliberative processes. In reply to Bloom and Pizarro, Haidt says that while people can do these things, they rarely do.21 People seldom seek out evidence that contradicts what they already believe, and seldom go out of their way to put themselves into new situations that will change their moral views. He asks his readers to remember how often in the past year they have “agonized” over a moral issue, and how often they have made unreflective immediate decisions about what to do. He guesses that the number of agonized decisions will be less than 1/100th of the number of unreflective ones, and points out that even what feels like deliberation can be just ad hoc searching for reasons for beliefs on two conflicting sides. In the resulting picture, some people might be able to use reason in moral thinking, but in the vast majority of cases they cannot and do not. Haidt says that this unwillingness to use reason to form new beliefs is what causes such “bitterness, futility, and self-righteousness” in moral discourse in cases of disagreement. In debates over abortion, politics, and so on, he says, we believe we’ve used reason to figure out what to think, we produce arguments that we think are good, and we “expect the other side to be responsive to such reasons” – but this expectation is based on an illusion that people are using reason to figure out what to believe in the first place.22 Since this is happening on both sides, the arguments we exchange with one another in moral discourse typically add nothing in the way of real justification or persuasion. People behave more like fans at a shouting match than like rational interlocutors, and this produces the bitterness and futility of moral arguments in cases of disagreement. In one way, Haidt’s theme echoes Posner’s concern that people make moral judgments based more on what community they find

156

Moral Reasoning in a Pluralistic World

themselves in than on the results of a reasoning process, and that reasoning is not playing a role. The difference is that where Posner seems to be saying that reasoning is impossible in the abstract because there is no foundation for it, Haidt seems to be saying that whatever we think about the abstract question of justification, people just don’t engage in it in practice. But in my view this interpretation of the empirical results is overly pessimistic, because the particular bitterness and futility we see in moral debate are compatible with every interlocutor using pluralist coherence and case consistency to come to a set of beliefs that is coherent in the sense being explicated here. Notice first that the lack of agonizing does not, in itself, show much. We do not agonize over everyday decisions involving long-term goals either: we go to work or to class, make broccoli for dinner, practice the piano, and so on without much agonizing but certainly with good reasons. The lack of agonizing is not because we aren’t using reason; it’s because we already know what we think about a wide range of subjects. And this is possible in the moral realm as well. It is in cases of conflict – especially when there are multiple conflicting considerations to be weighed and we aren’t sure how to do so – that we must stop to try to figure out what we think. This is supported by Pizarro and Bloom’s observation that people do stop to deliberate, especially when faced with an important and difficult decision such as whether to have an abortion. When it comes to public life, much informal moral discourse proceeds in a way that exemplifies the pursuit of case consistency. For example, in the case of abortion, the most obvious pro-life and prochoice arguments proceed in this way. Pro-life proponents say that abortion is murder, murder is wrong, so abortion is wrong, and thus they seek to show that there is a relevant moral similarity, and no morally relevant differences, between unborn fetuses and ordinary persons who must not be killed. Opponents typically respond to this argument by citing what they take to be morally significant differences between abortion and murder, sometimes based on the status of the fetus. Prochoice proponents say that people have rights to autonomy over what happens to their bodies, abortion is an exercise of that right, so abortion is morally permissible. Opponents typically respond by trying to show how this exercise of autonomy is not like other permissible forms of bodily autonomy, but like other impermissible ones.

Conclusion

157

These are cases of people appealing to a norm of a case consistency, but differing in values and in judgments about morally significant differences. Recall that from the point of view of pluralist coherence, some entrenched disagreements arise because people direct and prioritize among their values differently, and disagree about what is and is not a morally significant difference. If this explanation of disagreement in terms of pluralism is right, and if the pursuit of case consistency is an apt way of understanding what “reason” is in the moral domain, then what we would expect to find is that in these latter cases – we might call them “value-based disagreements”– interlocutors cannot agree, not because one or both of them is failing to use reason, or relying excessively on emotion or intuition, but because their disagreement transcends moral reason and instead tracks the person’s value field. Pluralist coherence says that judgments about moral significance are moral judgments, and that coherence means applying the same value field to a variety of cases. For this reason, we should expect that those who disagree about abortion will, if coherent in the sense being explicated here, either change their views about abortion or come to disagree about other moral matters as well, and that people’s beliefs about abortion would correlate with other things about them – and especially with their values. In fact there is evidence for both of these phenomena. People who are unwilling to change their beliefs about abortion do change their beliefs about other moral matters to fit their abortion beliefs. Chris Kaposy’s work on abortion points out how often “analogical” arguments are used in the abortion debate, and how often they lead to irresolvable “stalemates,” because people can, and do, revise their judgments about other moral matters to make them fit.23 Analogical arguments are instances of reasoning via case consistency, and Kaposy shows that people form or alter their beliefs about other matters, such as euthanasia, to make them fit with their judgments about abortion.24 Though it will not bring about agreement, this is one way that pluralist coherence recommends bringing beliefs into coherence. Empirical evidence also supports the common observation that people’s opinions about abortion seem well-correlated with other things about them: with their beliefs about other moral matters, with

158

Moral Reasoning in a Pluralistic World

their values, with their beliefs about feminism and the role of women, with their attitudes about sex, and so on. In a 1982 sociological study, Donald Granberg reports finding “very large differences ... between respondents affiliated with MCL, a ‘pro-life group,’ and those affiliated with a ‘pro-choice’ group as to their values, attitudes, and beliefs.”25 More recently, Michael Emerson’s 1996 study showed statistical correlations between abortion attitudes and “world view.”26 The picture that emerges from this is very different from Haidt’s pessimistic diagnosis. It shows how reasoning – reasoning well and in good faith – is compatible with certain kinds of bitterness and futility in moral arguments. Futility can happen not because people are not responding to reasons, but because different assumptions about moral relevance – grounded in different ways of directing and prioritizing values – lead to particular instances of intractability. Even among those who agree in valuing human life and respect for autonomy, there can be entrenched disagreements arising from differences in prioritization: differences about what outweighs what and when. In the pluralistic context, reason can help us determine which views are consistent, and help us determine what other moral beliefs fit with ones we already hold, but cannot help us determine how values should be directed and prioritized. Bitterness is natural in such a context, because though we are giving what seem to us good arguments our interlocutors never come to agree with us. In the “social intuitionist” approach that Haidt develops, what matters for moral judgment in practice isn’t reason but rather the influence of social groups. In his reply to Pizarro and Bloom, Haidt emphasizes the role of social interaction in “triggering” new appraisals: “ever since Plato wrote his Dialogues, philosophers have recognized that moral reasoning naturally occurs in a social setting, between people who can challenge each other’s arguments and trigger new intuitions.”27 But pluralist coherence explains why this social element to moral belief is compatible with the use of moral reason: social and cultural interactions inform the way people direct and prioritize among the things they care about. It follows that reasoning is compatible with entrenched disagreement and with the relationship Haidt observes between the moral and the social. It is also compatible with what Haidt describes as “moral dumbfounding” – the tendency we have, when asked to justify certain ethical beliefs, to say “I don’t know; I can’t explain it; I just

Conclusion

159

know it’s wrong.” If reasoning aims to bring consistency to considered moral convictions, it is not surprising that some of those convictions will be not easy to explain in terms of a more general principle. They are the basic elements of a moral point of view.

5 concluding thoughts If I am right about value pluralism and pluralist coherence, two final thoughts about our moral lives suggest themselves. First, being a “moral exemplar,” a “moral ambassador,” and a “moral artist” are just as important as engaging in public debate about contentious moral issues. If some moral changes require social and cultural shifts in thought, bringing about moral change will require going beyond arguments. This could take many forms. One that seems promising is standing as a representative of the values one endorses. This means living publicly in ways that honour those values. Proponents of gay rights, gay marriage, and queer activism in general have recently used such techniques to great effects: alongside a coherent argumentative position, activists offer themselves as examples of how living a queer life is living a good life. They are ambassadors and exemplars. Another approach that seems promising is the use of narrative and art to convey moral messages. That these move people morally does not mean people are not using reason. Second, systematicity is not only unjustified, it can also be dangerous since pursuing it requires us to alienate ourselves for no good reason from many of the things we value and care about.28 As I’ve emphasized here, even among those who explicitly endorse pluralist theories systematicity is commonly regarded as an ethical ideal – perhaps impossible but always worth pursuing. What makes systematicity seem attractive is the hope of avoiding arbitrariness: by relying on a single principle or overarching value we try to ensure a kind of rational way of approaching things that does not require us to make moral judgments and distinctions we would not know how to support or argue for. But I’ve argued here that this particular route to non-arbitrariness is a based on an illusion. At least in the context of the method that takes our considered convictions as good starting points for moral reasoning, systematized views also require us to make basic judgments about what is, and is not, a morally relevant similarity or difference. These judgments are no less arbitrary than those the plu-

160

Moral Reasoning in a Pluralistic World

ralist makes about how to weigh various moral considerations against one another. In the absence of such justifications we have no reason for doing what systematicity requires, and no reason to regard it as a moral ideal. In fact, we have important reasons not to systematize. Systematicity requires us to reject whichever basic commitments to values and moral judgments cannot be supported through the unifying value or principle we settle on.29 Sometimes those basic commitments reflect the moral aspects of ourselves that we should be most proud of: they reflect disparate values such as honesty, fairness, loyalty, and benevolence, and also our particular relationships and attachments to people, institutions, causes, and ideals. To systematize is to reject the vast majority of these in favour of one. The danger is that we’ll discard much of what we value most in life.

Notes to pages 000–000

161

Notes

introduction 1 2 3 4 5 6 7

8

9 10 11 12 13

See Ryan, “Sex Right.” This is just one factor among many; see Davis, Inhuman Bondage. Quoted in Baum, “New York Legislature.” Brink, Moral Realism, 250. Sunstein and Kahneman, “Cognitive Psychology.” They cite Thomas Schelling as the source of this example. Ross, Right and the Good. I use “moral theory” here to mean a set of moral judgments and principles. In his paper “An Unconnected Heap of Duties?” David McNaughton defends Ross against the charge that his theory is too unsystematic and unified by arguing that Ross’s theory is more systematic and unified than is often thought. More precisely, “An act is wrong if it is forbidden by the code of rules whose internalization by the overwhelming majority of everyone everywhere in each new generation has maximum expected value in terms of well-being (with some priority for the worst off).” Hooker, Ideal Code, Real World, 32. Audi, Good in the Right. Singer, The Expanding Circle. Beauchamp and Childress, Principles of Biomedical Ethics, 396. Haidt, “Emotional Dog”; see especially 823. For a discussion of pluralism as a moral point of view, see Wolf, “Two Levels of Pluralism.”

162

Notes to pages 12–16

chapter one 1 There is an enormous literature on these cases and this topic. I make particular use here of two sources: Macklin, Double Standards, and Hawkins and Emanuel, eds, Exploitation and Developing Countries. There was also some controversy over the question of trials in developing countries before the late 1990s and before these particular studies; an editorial by Marcia Angell in the New England Journal of Medicine addressed a similar range of difficulties in 1988 (Angell, “Ethical Imperialism?”). This editorial sparked much discussion of the issues. 2 Article 29 of the 2000 Declaration of Helsinki, added that year, says: “The benefits, risks, burdens and effectiveness of a new method should be tested against those of the best current prophylactic, diagnostic, and therapeutic methods. This does not exclude the use of placebo, or no treatment, in studies where no proven prophylactic, diagnostic or therapeutic method exists.” Later this passage was amended. The current (2013) version has, as Article 33: “The benefits, risks, burdens and effectiveness of a new intervention must be tested against those of the best proven intervention(s), except in the following circumstances: where no proven intervention exists, the use of placebo, or no intervention, is acceptable; or where for compelling and scientifically sound methodological reasons the use of any intervention less effective than the best proven one, the use of placebo, or no intervention is necessary to determine the efficacy or safety of an intervention and the patients who receive any intervention less effective than the best proven one, placebo, or no intervention will not be subject to additional risks of serious or irreversible harm as a result of not receiving the best proven intervention. Extreme care must be taken to avoid abuse of this option.” For discussion, see Carlson, Boyd, and Webb, “Revision of the Declaration.” 3 For more information, see Corbie-Smith, “The Continuing Legacy.” 4 Angell, “Investigators’ Responsibilities,” 968. 5 “Contextual” disagreement is related to what Robert Audi calls “circumstantial relativism”: that what you ought to do in a situation depends on the circumstances. As he says, this form of moral diversity is “not highly controversial” (Audi, Moral Value, 25). 6 Mackie, Ethics. 7 These facts about the pluralistic ways we value raise difficult questions. Are we right to value pluralistically? How do we prioritize among different values? Does a proper prioritization require adopting a more general point of view, from which the multiple values can be unified? I address these in chapter 2.

Notes to pages 16–22

163

8 See Mason, “Value Pluralism.” 9 See, e.g., the work of Jonathan Haidt and Craig Joseph; for further references see note 14 below. 10 Pinker, “Moral Instinct.” 11 This question hangs on that of whether, and when, beneficence outweighs respect for persons. For example, genetic tests pose little risk to participants, often taking the form of a simple blood test, and the benefits to scientific knowledge can be great. May genetic material ever be used in ways that go beyond what has explicitly been consented to? For discussion of this question, see Macklin, Double Standards, 149–50. 12 Macklin raises the question of justice as equity or equality in Double Standards, chapter 2: “Maintaining Ethical Standards in Research.” 13 See Mason, “Value Pluralism.” Though some take pluralism to be, by definition, committed to claims about incommensurability or incompatibility, I’ll use the term “pluralism” here to mean simply that there are a multiplicity of values not reducible to one another or to some single “super-value.” I say more about this in chapter 2. 14 See, for instance, Haidt and Joseph, “Moral Mind,” Haidt and Joseph, “Intuitive Ethics,” Shweder et al., “‘Big Three’ of Morality,” and Shweder, Jensen, and Goldstein, “Who Sleeps by Whom.” 15 Haidt and Joseph, “Moral Mind.” 16 Nagel, “Fragmentation of Value.” 17 Ibid., 132. 18 Ibid. 19 Ibid., 134. 20 This is the criterion Don Loeb describes as “universality” in his “Generality and Moral Justification.” R.M. Hare also distinguishes “generality” from other concepts such as universality and understands it in a similar way. See his Freedom and Reason, 38–9, and Moral Thinking, 41. I say more about Hare’s views in chapters 4 and 5. 21 The latter of these is more general than its predecessor because it applies a single criterion to a larger range of cases; I discuss generality further in what follows. 22 Sayre-McCord, “Coherence and Models.” 23 For more on the use of considered convictions in coherence reasoning, see Scanlon, “Rawls on Justification.” 24 Various interpretations of “possibility” lead to various ways of understanding this; here I’ll take possibility roughly along the lines of conceivability as proposed by Yablo, “Is Conceivability a Guide?” 25 Marcus draws the distinction in her “Moral Dilemmas and Consistency.” In

164

26 27

28 29 30

31 32 33 34 35

36

37

38

39 40

Notes to pages 22–8

her framework, only essential conflicts are evidence of “inconsistency”; I use the terms “essential” inconsistency and “practical” inconsistency instead. The status of conflict and dilemmas raise more complex issues, which I consider below. See, for example, Sayre-McCord, “Coherentist Epistemology,” and Klemens Kappel’s characterization in his “Meta-Justification” (Kappel is arguing against the possibility of such meta-justification, but he characterizes coherence using similar basic elements to those I describe here). Brink, Moral Realism, 250. Kagan, Limits of Morality, 13. Hooker, Ideal Code, Real World, 21. I discuss these authors further in chapter 3. Dworkin, Rawls, and Sayre-McCord also defend versions of coherence that go beyond consistency (see Dworkin, Taking Rights Seriously, Rawls, Theory of Justice, and Sayre-McCord, “Coherentist Epistemology”). Though he focuses on desire rather than belief, the coherence that Michael Smith appeals to is a rich form as well; see his Moral Problem. Kappel makes this point about trade-offs and coherence desiderata in “Meta-Justification,” 132. See Arneson, “Broadly Utilitarian Theories.” Andrew Siegel raises a question like the latter in “Kantian Ethics.” Kappel, “Meta-Justification,” 132. Of course one may create single principles trivially using “and.” I take it that this is a non-interesting way of unifying our principles – and indeed, a theory with conjunctive principles would not, insofar as the principles were complex, be systematic. Ross, Right and the Good. Kagan argues that “pro tanto” captures more accurately than “prima facie” what Ross is trying to articulate, and I agree (Limits of Morality, 17). For this terminology see, e.g., Richardson, “Specifying Norms.” For a discussion (and rejection) of specification in the area of legal rights, see Finklestein, “Two Men.” Sayre-McCord, “Coherence and Models,” 171. In his later paper, “Coherentist Epistemology” Sayre-McCord takes consistency to be “evidential consistency.” I discuss this approach in chapter 3. Sidgwick, Methods of Ethics, 341. For thoughts along these lines see Davidson, “How Is Weakness” and Lemmon, “Deontic Logic”; for discussion of these see Marcus, “Moral Dilemmas and Consistency.”

Notes to pages 28–39

165

41 Ruth Marcus points out the importance of such examples in her “Moral Dilemmas and Consistency.” 42 Sidgwick, Methods of Ethics, 341. 43 Paul Bloom makes a related point informally when he says that moral deliberation is not restricted to specialists: “most people are regularly forced to ponder dilemmas such as the proper balance of work and family” (“How Morals Change”). 44 Marcus, “Moral Dilemmas and Consistency,” Williams, “Ethical Consistency” and “Consistency and Realism.” 45 Against this line of thought, Philippa Foot suggests the emotion could be irrational but not “discreditable” (“Moral Realism,” 382). 46 Sayre-McCord, “Coherence and Models,” 171–2; he mentions Mill as raising the same concern. 47 Ibid. A third possibility is that we specify conditions for the application of various rules. This isn’t relevant here since it’s not a way of systematizing – the principles would be fewer but more complex. 48 Rawls, Theory of Justice. 49 One might interpret Rawls as giving a single principle along the lines of “just actions are those in accordance with a set of rules that we would choose behind a veil of ignorance.” But Rawls does not present his view this way: he distinguishes between his two “principles” and the background justification he gives for them. In any case, only the possibility of a justifying method that is not systematizing is necessary for my argument. 50 Sayre-McCord, “Coherence and Models,” 172.

chapter two 1 Berlin, Crooked Timber. 2 Berlin, Liberty. 3 See Berlin, Crooked Timber and Liberty. For some discussion of these metaphors see Talisse, “Value Pluralism.” 4 See, e.g., Anderson, Value in Ethics. 5 See, e.g., Griffin, Well-Being. 6 Williams, “Ethical Consistency,” and Stocker, Plural and Conflicting Values. 7 Hurka, “Monism, Pluralism.” 8 Dworkin, Justice for Hedgehogs. See chapter 6, “Moral Responsibility.” See also his “Do Liberal Values Conflict?” 9 For discussion of how this ambiguity in the notion of identification bears

166

10 11

12

13 14

15 16

Notes to pages 40–6

on theories of autonomy, see Christman, “Constructing the Inner Citadel,” 113. For a related and illuminating discussion with examples, see Schroeder and Arpaly, “Alienation and Externality.” I consider below the question of whether ambivalence presents special difficulties beyond that of A and B conflicting essentially. Notice that the valuationally inconsistent person need not be logically inconsistent: even faced with the fact that he values or cares about A and not-A, the person may say that A is good in a way and that not-A good in another, taking no stand on which – if any – is good overall. Or he may say that A is good and not-A is good, simply allowing that good and bad need not be logically opposite predicates. To say that A is good in a way and notA is good in another is not necessarily to say that there are aspects of A that are good and others that are not-good (I discuss this kind of “instrumentality” below); I mean here instead that the predicates good and bad may be modified. See the discussion in Greenspan, “A Case of Mixed Feelings.” Greenspan gives a version of the friendly rival case based on simple desires rather than endorsed desires; I discuss the desire case fully in my “On Essentially Conflicting Desires.” Of course, the merely conflicted and the essentially inconsistent will each have to prioritize what they care about, choosing one thing over another. One might wonder at this point: as long as I am able to choose, or prefer, A to B, am I really “inconsistent”? My answer is yes; I discuss this matter below. Moore, Watchmen. Brink, writing in the context now of moral conflict rather than coherence more generally, suggests that because cases of valuing A and not-A are usually, deep down, cases of valuing A in virtue of aspect X and not-A in virtue of aspect Y, cases of real ambivalence are rare. But, as I argue in “On Essentially Conflicting Desires,” not all cases are of this form, and even ones that are can be properly seen as cases of ambivalence. In any case, as I explain here, the relative consistency of X and Y is still under consideration; thinking in terms of essential versus contingent conflicts – rather than ambivalence versus non-ambivalence – makes clear that the “aspect” treatment does not address the relevant issues. See Brink, “Moral Conflict.” Talisse, “Value Pluralism,” 95. I think it is fair to interpret Talisse as appealing implicitly to this sort of interpretation – that is, as arguing that it is when values “inherently” conflict that regret is irrational – though he doesn’t explicitly appeal to the idea of essential inconsistency (Talisse, “Value Pluralism”). In a different context

Notes to pages 46–51

17

18

19 20 21 22 23

24 25 26 27

167

(and not making a delimiting argument), Martha Nussbaum also distinguishes “essential” conflicts as problematic for moral theory in a way that contingent ones are not, writing: “In general, a moral theorist is not likely to defend a picture of the human good in which the distinct and plural goods recognized by the theory are intrinsically in conflict with one another; most moral theorists have been attached to consistency in that sense. But there is absolutely nothing to prevent the moral theorist from acknowledging that two goods that are in principle compatible, for example one’s attachment to one’s children and one’s attachment to one’s work, may come into conflict in particular contingent circumstances.” Nussbaum, “Why Practice Needs Ethical Theory,” 245. Cheshire Calhoun discusses “integrity” in work related to my theme (“Standing for Something”) and while I share her general view that integrity is a kind of social virtue, and that inconsistency and ambivalence are sometimes required for a life well lived, my focus in this paper is somewhat different. In particular, Calhoun does not distinguish “inconsistency” from mere conflictedness, as I do below, and does not try to assess whether there is a relationship between inconsistency and rationality. J. David Velleman also suggests that valuational inconsistency is a problem for a person; of a person who yearns both for wealth and success and for a life of reflection and self-cultivation, Velleman says, “there is something irrational about being so conflicted, about holding on to goals that cannot be jointly attained.” Velleman, “Willing the Law,” 30. Frankfurt, “Identification and Wholeheartedness,” 172. Frankfurt, “Faintest Passion,” 100. Velleman, “Identification and Identity,” and Frankfurt, “Reply.” Velleman, “Identification and Identity,” 101. Velleman admits that there is an important difference: the Rat Man “expelled [his hatred] from his self-awareness or self-understanding, whereas Frankfurt’s view implies that he should have consciously rejected it and thereby expelled it from the self.” But this rejection, he says, doesn’t seem any better. What the Rat Man should have done, on Velleman’s view, was “to accept his filial hostility as part of himself; to accept himself as ambivalent toward his father.” Velleman, “Identification and Identity,” 103. Frankfurt, “Reply,” 126. Ibid., 127. I argue an analogous point in the case of desires in my “On Essentially Conflicting Desires.” Frankfurt, “The Faintest Passion,” 103.

168

Notes to pages 51–61

28 I discuss the relationship to the self in the next subsection, on autonomy. 29 In characterizing these views and suggesting that they are “standard,” I follow Christman, “Constructing the Inner Citadel.” 30 See, e.g., Dworkin, “The Concept of Autonomy” and also the works of Frankfurt himself. Of course, if we assume that autonomy requires wholeheartedness, these people are not autonomous, but in that case, neither are people who are merely conflicted and have clear prioritizations. In any case, simply assuming that autonomy requires wholeheartedness would be question begging in the present context. Views such as these, based on identification, face an important challenge with respect to whether we should understand “identification” along the lines of self-recognition or approval. I take no stand on this matter here; my remarks apply to each in the same way. 31 See, e.g., Haworth, Autonomy. 32 Frankfurt, “The Faintest Passion,” 103. 33 In “Standing for Something,” Calhoun argues that integrity ought to be understood not as a value of unity, but as the social value of “standing for something.” Ambivalence and dividedness are no impediment to integrity in this sense. I discuss the significance of this kind of integrity for pluralist coherence in chapter 5. 34 Frankfurt, “The Faintest Passion,” 106. 35 As we’ve seen, moral dilemmas provide rich examples of cases in which a person seems to be more admirable for feeling less wholehearted about his choice; we want and expect people in some dilemmas to feel a kind of psychological pain or sorrow about their choice. I discuss dilemmas and regret further in what follows. 36 Of course, such a person may act in ways that are self-undermining, say by protesting the very activities he engages in. As I’ve argued above, however, while ambivalent people may act in self-undermining ways, so may people who are merely conflicted. 37 The term “lexical” is used because of the analogy with the way all As come before all Bs in the dictionary. 38 Dworkin, Justice for Hedgehogs, 120.

chapter three 1 Daniels mentions this as a strategy in the moral context but does go on to acknowledge difficulties, explaining that “more has to be said about how some parts of the system of beliefs ‘support or explain others’ than [has been] provided,” Daniels, “Reflective Equilibrium.”

Notes to pages 61–9

2 3 4 5 6 7 8 9 10 11 12

13

14 15 16 17

18

169

Kagan, Limits of Morality, 13. Brink, Moral Realism, 103. Kagan, Limits of Morality, 13, 14. Brink, Moral Realism, 251. Emphasis in the original. Ibid., 202. Hooker, Ideal Code, Real World, 32. Ibid., 21. Emphasis in the original. Ibid., 4, 175. Ibid., 176. Ibid. In a discussion of Hooker’s single-principle consequentialism, Jeffrey BrandBallard argues that the advantage of monism over pluralism is the same as the advantage of pluralism over particularism: we need general principles for pragmatic reasons, to reduce our cognitive burden in figuring out what to do. He makes this argument in the context of defending Hooker against the claim that the single principle rule-consequentialism provides is, if truly fitting with our considered convictions and a Rossian list of principles, otiose, an idle wheel (emphasizing here the arguments of Alan Thomas, “Consequentialism”). My point here is, of course, different from that of the idle wheel, and concerns the possibility that the explanations in more systematic theories are incorrect explanations. See Brand-Ballard, “Why One Basic Principle?” Audi, Good in the Right. Audi presents his theory as a version of “intuitionism” – that is, as committed to the idea that we know some moral facts because they are self-evident. I leave aside, here, the question of how self-evidence, intuition, and “convictions” are related. Ibid., 83; see also 66 for more on dogmatism. Ibid., 159. Sayre-McCord, “Coherentist Epistemology.” Sayre-McCord says that with this formulation he is “steering clear” of certain “extreme” views, such as that “coherence itself provides evidence that a system of beliefs is likely to be true,” ibid., 152. The conception asks that we seek coherence not only within our set of moral beliefs but also between our moral beliefs and other beliefs. Dale Dorsey argues that coherence must go beyond consistency, because two consistent but unrelated normative sentences such as “Calley acted wrongly” and “I ought to eat an egg salad sandwich for lunch” do not form a coherent set in the appropriate way even though they are consistent with one another (Dorsey, “Coherence Theory,” 498). But it seems to me that the characteriza-

170

19 20 21 22

23 24 25 26 27 28 29

30

Notes to pages 70–5

tion of coherence he does offer does not support norms like systematicity among principles. Dorsey says that a set of beliefs about the evaluation of actions and states of affairs is coherent when in addition to being consistent each of its members is “derivable by a system of deductive logic or inductive generalization,” with all other sentences in the set and requisite non-normative sentences available as premises. But this means that a set with multiple principles that are fundamental in my sense is fully coherent as long as those principles are supported by judgments about particular cases. To make our two unrelated sentences about Calley and egg salad part of a coherent set, we needn’t relate them to one another; it is sufficient to simply add sentences regarding the morality of killing in war and sentences regarding what foods one ought to eat. This is fully compatible with the existence of multiple basic principles, perhaps including something like “don’t eat meat” and “don’t kill innocent persons.” Thus, though he doesn’t emphasize the pluralist possibilities it opens up, Dorsey’s view of coherence seems to be broadly in sympathy with the one I am proposing here rather than a richer form. Daniels, “Wide Reflective Equilibrium,” 259. Ibid., 279. Beauchamp and Childress, Principles of Biomedical Ethics, 386. Laurence Bonjour’s approach in general epistemology is along these lines (see his Structure of Empirical Knowledge). I’m agnostic here about whether we need “approximate” truth rather than truth here, and what role it would play. Beauchamp, “Principlism,” 185. Brink, Moral Realism, 208. Rawls, “Justice as Fairness,” 188. Korsgaard, “Right to Lie,” 330. For a more complex analysis of the Kantian view of such cases, see Varden, “Kant and Lying.” Korsgaard, “Right to Lie,” 338. Ibid., 340. Proposed in the seventeenth century, the phlogiston theory postulated an element called “phlogiston” that was released in combustion. The theory turned out to be false; we now understand the phenomena in question in terms of oxidation. In the late nineteenth century, “luminiferous ether” was suggested as the medium through which light propagated; that theory was later abandoned in favor of explanations based on quantum theory and relativity. Thagard, “Coherence, Truth.” Thagard also discusses “broadening,” which is when a theory explains new facts. But deepening is most crucial to a theory’s success, and I focus on it here.

Notes to pages 75–9

171

31 Ibid., 38. 32 It is true that we can discuss the mechanisms that produce certain moral beliefs in people, and this descriptive story may well be causal. I may well say that my belief that lying is wrong explains, psychologically and causally, my belief that you ought not tell this lie, and that my confidence in the wrongness of lying increases when I find more individual judgments of cases in which lying is wrong. Such a theory may be “deepened” by a better understanding of neuroscientific phenomena, and such deepened theories of moral psychology may be more likely to be true. But such a descriptive theory does not help us understand the importance of moral coherence, since insofar as there are agents who are morally incoherent there are presumably true theories about why they are. This descriptive matter is not what we are after. 33 Nussbaum, “Why Practice Needs Ethical Theory.” She cites the excellent legal history Ryan offers in “Sex Right.” 34 Nussbaum is not addressing anything like “deepening”; she just offers the case as one in which moral theory, in the form of Kantianism, plays an important role. 35 Ryan, “Sex Right.” 36 I take the patchwork formulation from Nancy Cartwright’s chapter heading, “Fundamentalism versus the Patchwork of Laws,” in her Dappled World. 37 Cartwright, Dappled World, 27. She attributes the example to Austrian philosopher and sociologist Otto Neurath. 38 Ibid. 39 Cartwright, Dappled World, 27. In the quoted passage she is discussing mechanics particularly, but I think it is fair to represent her view of all theories this way. 40 Again, I take no stand here on whether these conclusions are right in the science case; my concern is with criticisms of the claims of a patchwork of laws and with possible analogies to the moral context. 41 See, e.g., Sklar, “Dappled Theories,” 437–8. Sklar points out that when details of one theory seem irrelevant for certain explanations – as, I take it, pointparticle mechanics seems irrelevant for predicting the location of the dollar bill – we can often explain, with the theories we have, just why it seems irrelevant: it’s not that it “doesn’t apply” in the case at hand, it’s just that for certain intelligible reasons what it tells us isn’t useful there. 42 Smith, “Models and the Unity of Classical Physics.” 43 Ibid., 461. 44 Ibid., 464.

172

Notes to pages 80–90

45 Smith, Moral Problem and Ethics and the A Priori. He focuses on coherence of desires, not coherence of beliefs; I’m adapting the line of thought to be relevant here. For a survey of analogies between morality and mathematics in general, see Clarke-Doane, “Moral Epistemology.” 46 Marino, “Moral Rationalism.” 47 Smith, “In Defence of The Moral Problem,” 269. 48 Ibid. I say more about arbitrariness in chapter 4. 49 Smith, “Realism,” 408. He says there: “after all, something like such a convergence in mathematical practice lies behind our conviction that mathematical claims enjoy a privileged rational status. So why not think that a like convergence in moral practice would show that moral judgments enjoy the same privileged rational status?” 50 I am grateful to Penelope Maddy for helpful discussion of the mathematical case. For more on axiom choice in set theory, see Maddy, Defending the Axioms.

chapter four 1 2 3 4 5

6 7 8 9

10 11 12

Sunstein and Kahneman, “Cognitive Psychology.” Ibid., 103. I give some examples in chapter 5. Goldman, Practical Rules, 161. Rainbolt, “Perfect and Imperfect Obligations,” and Gillespie, “On Treating Like Cases Differently”; for general discussion see Murphy, “Justifying Departures from Equal Treatment,” and Marmor, “Should Like Cases Be Treated Alike?” Rainbolt, “Perfect and Imperfect Obligations,” 253. Hare, Freedom and Reason, 33. Ibid., 107. Kagan, “The Additive Fallacy,” and Brink, “Moral Conflict.” Kagan is not endorsing the factor view but merely describing it in order to show that factors aren’t “additive.” Brink is writing about the nature of moral conflict, not about how to determine what to do. Kagan, “The Additive Fallacy,” 14; Kagan goes on to argue that function is not simply additive. Brink, “Moral Conflict,” 217: Brink defends the additive approach in fn. 5. In his discussion of Ross, McNaughton proposes that principles have “weights” of a kind built in; in judging what to do overall, some weaker ones begin with a disadvantage, like a handicap, that they would have to

Notes to pages 90–9

13 14

15 16 17 18 19

20 21 22 23 24 25 26 27 28 29 30

173

overcome. For example, he says Ross took duties of nonmaleficence to be stronger, generally, than those of beneficence. From the point of view being outlined here, a general weighting such as this could certainly be case consistent; how intuitively acceptable it was for a person would depend on the person’s judgments about what to do overall in a range of cases (McNaughton, “An Unconnected Heap of Duties?”). Henry Richardson, “Specifying, Balancing,” 297. Gert, Culver, and Clouser propose that common morality includes beliefs about what features of cases are “morally relevant,” in addition to other things like moral ideals. See Gert, Culver, and Clouser, Bioethics and “Common Morality.” Hubin, “Irrational Desires.” This is Alan Gewirth’s “principle of appetitive-reciprocal consistency,” from his “Categorical Consistency in Ethics,” 289. Singer, Generalization in Ethics, 362. Nagel, The Possibility of Altruism, 88. Pinker, “The Moral Instinct.” Pinker goes on to say that “unless I am Galactic Overlord, I have to state my case in a way that would force me to treat you in kind. I can’t act as if my interests are special just because I’m me and you’re not, any more than I can persuade you that the spot I am standing on is a special place in the universe just because I happen to be standing on it.” This defence of interchangeability is odd, because it seems to rest on the contingent circumstance of our being roughly equal in powers to hurt others. Hare, Freedom and Reason, 107. Gordon, Rauprich, and Vollmann, “Applying the Four Principle Approach.” Macklin, Double Standards, 44–5. Ibid., 214–23. Ibid., 24ff. Ibid., 445. Macklin, Double Standards; see particularly 260. Mary Anne Warren, “On the Moral and Legal Status of Abortion.” Marquis, “Why Abortion Is Immoral.” Thomson, “A Defense of Abortion.” Richardson says: “This leaves me with the suspicion that what is actually happening is that Gert, Culver, and Clouser are noticing which moral conclusions we actually come to and then reading off from those judgments the assertion that we judge one harm to be greater than another.” Richardson, “Specifying, Balancing,” 296–7. He’s talking about comparing “harms”

174

31

32 33 34 35 36

37

38 39 40 41 42

43

44

Notes to pages 100–5

because he is addressing the point to Gert, Culver, and Clouser, who propose dealing with conflict in terms of harm (see chapter 5, section 8). In his text, Richardson expresses the reasons against cross-examination by saying “it is wrong to defame someone’s character by knowingly distorting their public reputation” (282). I’ve altered this explanation slightly for clarity. I’ve also included more details of the example than Richardson does; these are drawn from the legal ethics book that is the original source of the example (Luban, Lawyers and Justice, 150). Further discussion of this book follows. Richardson, “Specifying Norms,” 283. Ibid., 283. Luban, Lawyers and Justice, 151. Ibid., 152. In chapter 5 I consider and argue against Richardson’s second claim that when balancing is justified the results are better expressed through specification. In this direction Kagan writes: “An adequate justification for a set of principles requires an explanation of those principles – an explanation of why exactly these goals, restrictions, and so on, should be given weight, and not others. Short of this, the principles will not be free of the taint of arbitrariness which led us to move beyond our original ad hoc shopping lists,” Kagan, Limits of Morality, 13, emphasis in the original. Singer, Practical Ethics, 20. See, e.g., his contribution to Smart and Williams, Utilitarianism. Gert, Culver, and Clouser, Bioethics. Sen, Idea of Justice. Since Singer does not use the method of conviction ethics this should not be taken as a stand-alone criticism of his view. The point here is that within that method, rich coherence is no better off than case consistency with respect to determining what is and is not an arbitrary distinction. Davis, Inhuman Bondage, 53. Davis also mentions the “curse of Ham,” which refers to the idea that since God had punished Ham with becoming a slave in the Bible, his descendants were naturally and properly treated as such. These attempts at justifying slavery marked a different line of thought from that Davis attributes to Augustine, who Davis says “interpreted society’s need for slaves as part of a universal human depravity, as the way the world is and must be accepted, as distinguished from the City of God,” Inhuman Bondage, 44. Mill, Utilitarianism, 3.

Notes to pages 105–12

175

45 Rawls, Theory of Justice. 46 Rainbolt discusses a similar example, “Perfect and Imperfect Obligations,” 251. 47 Macklin, discussing the disagreement over low-risk genetic studies, writes that the disagreement is understandable because “there is no fixed order for ranking the two applicable ethical principles: respect for persons, which underlies the informed consent requirement, and beneficence, which requires a favorable balance of benefits over harms in research.” Double Standards, 150. Robert Veatch also says that a full ranking of principles “is generally believed implausible.” He goes onto say that “the appropriate manner for the resolution of conflict among competing principles is the specification of norms in a mixed strategy that involves both some lexical ordering and some balancing,” “Resolving Conflicts,” 211, 217.

chapter five 1 For development of this idea see Foot, “Moral Realism”; for discussion see my “Moral Dilemmas.” 2 See my “Moral Dilemmas.” 3 Marcus, “Moral Dilemmas and Consistency.” Of course, some ways of avoiding conflict are better than others. Some ways of avoiding a particular conflict might seem to bring about a world worse in other ways; this just means there is a further conflict to be considered and a different alternative should be sought. It might seem a problem that for an individual, avoiding conflict is best achieved by staying home and doing nothing. This would conflict with norms about the good of involvement in life’s activities, and perhaps duties to oneself. Below, I emphasize the ways in which avoiding conflict is a collective, not individual, matter. 4 Joseph DeMarco, “Principlism and Moral Dilemmas,” 102. For a discussion of dilemmas in the context of virtue ethics, see Russell, “What Virtue Ethics Can Learn.” 5 DeMarco cites Gillon, “Four Scenarios,” for the original example. 6 Henry Richardson has argued that specification is consistent with taking conflict to be real; I discuss this below in section 3. 7 Thanks to Rosalind Abdool, who pointed this out to me. 8 Zarrella and Mungin, “Florida Lifeguard.” In response to public outcry he was offered his job back. He turned it down. Several other lifeguards quit in response to the incident. 9 Zarrella and Mungin, “Florida Lifeguard.”

176

Notes to pages 113–22

10 Macklin, “Moral Progress.” 11 Ibid., 371–2. This is only part of her definition. 12 For development of this idea see Foot, “Moral Realism”; for discussion, see my “Moral Dilemmas.” 13 Greenspan, “Moral Dilemmas and Guilt.” For discussion, see Marino, “Moral Dilemmas.” 14 I’m paraphrasing here. Richardson says that instead of the principle “be generous and tolerant” we might move to “be generous and tolerant towards all persons even when they have transgressed, but towards their behaviour only when that behaviour is within the pale.” “Specifying, Balancing,” 300. 15 Beauchamp and Childress sometimes seem to associate specification with having the right criteria. I discuss this further below, in section 8 of this chapter. 16 Richardson, “Specifying Norms,” 298. 17 Dancy, Ethics Without Principles, 128. 18 As with the discussion of delimiting in chapter 2, the point about remainders in this context is directed specifically against specification. It’s not that unified views are incompatible with the existence of remainders, but rather that given plurality and the arguments against unity in chapter 3, remainders ground reasons not to specify principles. 19 Kappel, “Meta-Justification,” 132. 20 See, e.g., Anderson, Value in Ethics and Economics, and Satz, Why Some Things Should Not Be for Sale. 21 For an insightful discussion of why such restrictions can be appropriate and how laws ought to be framed to protect sexual autonomy, see Schulhofer, Unwanted Sex. 22 Macklin, Double Standards, 69. 23 Patterson, “AIDS and the Global Conscience.” The letter is cited and discussed by Macklin, Double Standards, 231. 24 The same letter writer, however, goes on to elaborate his thought in ways that suggest the state ought not value the lives of citizens at all, writing: “if [pursuing well-being for citizens] requires helping other nations for national security purposes ... then foreign aid is a worthwhile endeavor. But if it is merely charity ... then such aid is foolish.” This disregard is different from partiality; expansion of values would tend toward the opposite conclusion. Patterson, “AIDS and the Global Conscience.” 25 Sunstein and Kahneman, “Cognitive Psychology,” 103. 26 In Swann’s Way, Proust describes the character of Odette in such terms: “In

Notes to pages 123–31

27 28

29

30 31

32 33 34 35

177

vain, however, did Swann expound to her thus all the reasons that she had for not lying; they might have succeeded in overthrowing a general system of mendacity, but Odette had no such system; she was simply content whenever she wished Swann to remain in ignorance of anything she had done, not to tell him of it,” 413–14. Dancy, Ethics without Principles; see chs 5 and 7, section 5. The kind of integrity that matters here is not that associated with unity, but rather like that characterized as a “social virtue” by Calhoun in “Standing for Something.” She writes: “Persons of integrity treat their own endorsements as ones that matter, or ought to matter, to fellow deliberators,” Calhoun, “Standing for Something,” 258. In “Moral Compromises” Theo van Willigenburg also argues that having integrity is compatible with making compromises, particularly in a context of pluralism. In his “Coherence and Models” Sayre-McCord says that our basic intuitive notion of fairness (required in order not to give a circular defence of coherence) does not support coherence because “we do not think people are behaving unfairly if they act without possessing a fully articulated coherent theory.” My argument in terms of judging like cases alike is distinct from this; it is true that we don’t blame people for not having a fully worked out theory. But we do think people are behaving unfairly if they fail to judge like cases alike from, say, prejudice or fondness, and we do blame them if they continue to do so after the inconsistency is pointed out to them. Dancy, “Moral Particularism.” It is also true that whether pluralist coherence fits with a particularist view of the nature of rightness and wrongness hardly matters. I’ve tried to talk about how one may improve one’s moral view by becoming more case consistent, and how we might use case consistency to reason together in cases of disagreement. It is possible to understand these recommendations less in the spirit of a challenge to particularism and more in the spirit of “here’s how to reason using moral coherence if particularism is true.” Anyone not convinced that pluralist coherence makes substantive enough claims about generality is welcome to read these efforts in that spirit. Goldman, Practical Rules, 161–2. Goldman’s views have changed since; see his Reasons from Within: Desires and Values. Macklin, Double Standards. This is just one of several lines of thought in the text. Pogge, “Testing Our Drugs,” 112; Siegel, “Kantian Ethics,” 192. Little, “Abortion, Intimacy.”

178

Notes to pages 132–7

36 Marcus, “Moral Dilemmas and Consistency,” 131. 37 I say more about the abortion example and implications for moral practice in chapter 6. 38 Beauchamp and Childress, Principles of Biomedical Ethics, ch. 10. 39 They write: “It is misleading to say that we are obligated to perform both actions in these dilemmatic circumstances. We should discharge the obligation that, in the circumstances, we judge to override or to compromise what we would have been firmly obligated to perform were it not for the conflict,” Principles of Biomedical Ethics, 11. But they also say that “moral residue [what I am calling the “remainder”] results because an overridden prima facie obligation does not simply go away when overridden,” and that “regret and a sense of moral residue ... are the natural result of the fact that an overruled prima facie obligation does not mean that the obligation winds up counting for nothing,” Principles of Biomedical Ethics, 16. These remarks are consistent with the view of the remainder I outlined above. 40 Beauchamp and Childress, Principles of Biomedical Ethics, 20. 41 Ibid. 42 Ibid. 43 Ibid., 17. They say this is “roughly the provision recommended” in the “Ethical Guidelines for the Practice of Forensic Psychiatry and the Law,” as revised October 1991. 44 Ibid., 20–1. 45 Beauchamp and Childress also say: “To say that a problem or conflict is resolved or dissolved by specification is to say that norms have been made sufficiently determinate in content that, when cases fall under them, we know what ought to be done,” ibid., 19, and that the goal of merging them “is too streamlined to handle all situations of balancing,” ibid., 20. 46 Ibid., 388. 47 Ibid., 396. They also include “simplicity” as a desideratum for a moral “theory” (335), but they are later vague about how much it matters to have a “theory” in the strict sense and whether their collection of principles forms one (397). 48 Gert, Culver, and Clouser, Bioethics. Because of this single rule, the resulting view has been described as fundamentally monistic (Brand-Ballard, “Consistency, Common Morality”). I gave some reasons in chapter 2 for why such rules do not render pluralist views monistic. 49 See Gert, Culver, and Clouser, “Common Morality,” 313. They say there that when we consider publicly allowing exceptions, “it is clear that deception will result in more harm than would harassment.”

Notes to pages 137–52

179

50 In the following discussion I refer to both Thagard’s book Coherence in Thought and Action and his earlier paper “Ethical Coherence.” 51 Thagard, Coherence in Thought and Action, 128. 52 There isn’t space here for a discussion of whether Hare succeeds in his attempt to avoid relying on moral judgments. For an argument that he does not, see Nagel, “The Excessive Demands of Impartiality.” 53 In some places Hare includes animals and asks us to consider whether we would accept the results of universalized actions that affect them. He doesn’t take a final stand on whether the narrower (person) or broader (persons and animals) category is the significant one; for discussion see Singer, “Reasoning toward Utilitarianism.” 54 Hare, Freedom and Reason, 176. 55 Because these difference are grounded in deep foundational ones, I cannot show here that Hare is conclusively mistaken, but for discussion of the positive role ideals play in moral life see Alan Gettner, “Hare and Fanaticism.” 56 Singer, “Reasoning Toward Utilitarianism,” 149. Singer goes on to suggest Hare might appeal to the possibility of moral beliefs not being objective to justify the equal consideration of ideals in a different way. Singer also points out in this essay that Hare’s own project of justifying morality by appealing to the meanings of moral words falters here, since “if we ask ourselves whether ordinary people consider that a moral judgment must give equal weight to all ideals, irrespective of their content, the answer would surely go against Hare,” 149.

chapter six 1 2 3 4 5 6 7 8 9 10

Beauchamp and Childress, Principles of Biomedical Ethics, 335; see also 388. Singer, Practical Ethics, ch. 6: “Taking Life: The Embryo and Fetus.” Jeske, Review of Moral Value and Human Diversity. Gert, Culver, and Clouser, Bioethics, viii. Timmons, Disputed Moral Issues, 32. Ibid., 32–3. Posner, Problematics, 22. Ibid., 52. Ibid., 23. Little emphasizes that gestation involves an “intimate” relationship, not an impersonal one, in her “Abortion, Intimacy.” 11 Posner, Problematics, 50–1. He cites here Childress, Practical Reasoning in Bioethics.

180

Notes to pages 152–8

12 Posner’s claim is an overstatement in other ways as well: Beauchamp and Childress offer their principles as closely connected to established ethical theories and to other principles developed by moral philosophers, Principles of Biomedical Ethics, 333–4. 13 For an account of moral progress that does not rest on “moral realism” see Wilson, “Moral Progress.” 14 Haidt, “Emotional Dog.” See also Pizarro and Bloom, “Intelligence of the Moral Intuitions,” and Haidt, “Emotional Dog Does Learn New Tricks.” I discuss these texts below. 15 Haidt, “Emotional Dog and Its Rational Tail,” 815. 16 Ibid., 817. 17 Ibid., 829. 18 Pizarro and Bloom, “Intelligence of the Moral Intuitions.” 19 Ibid., 195. The students were of various racial backgrounds; for details see Rudman, Ashmore, and Gary, “‘Unlearning’ Automatic Biases.” 20 Ibid.; see also Bloom, “How Morals Change.” 21 Haidt, “Emotional Dog Does Learn New Tricks.” 22 Ibid., 823. 23 Kaposy, “Two Stalemates.” 24 Ibid., 7. 25 Granberg, “Comparison of Pro-choice and Pro-life Activists.” In his study, Granberg asked people who identified as pro-choice and pro-life to rank their values. As would be predicted by the value-field variation explanation of moral disagreement, he found that while there were some shared values, these were prioritized differently (see especially 89). However, I should note that one of the values that showed the steepest difference in relative prioritization was “salvation”; it is not clear to me what this means as a moral value. This does complicate my use of this as evidence. 26 Emerson, “Through Tinted Glasses.” Emerson’s discussion is somewhat difficult to interpret from a philosophical point of view because of the way he classifies moral outlooks. He distinguishes on one end of a spectrum those who believe morality is “absolute” in the sense that the rules specify behaviours that must be followed in every circumstance and at the other end “relative/utilitarian” reasoners who believe a few general principles apply to different situations in different ways. This crosscuts typical philosophical distinctions, since utilitarianism in philosophy is not associated with relativism except in the most trivial sense that what is right to do depends on the circumstances. Utilitarianism is an “objective” moral view in the sense

Notes to pages 158–60

181

that it entails that there is a right answer to the question of what to do in a given set of circumstances. 27 Haidt, “Emotional Dog and Its Rational Tail,” 820. 28 In one way this echoes Hare’s idea that the moral “fanatics” are the most dangerous people. As we saw in chapter 5, however, for Hare fanatics are those committed to any ideals at all. The warning here is against being committed to unified ideals, those based on single values, at the expense of others. 29 Bernard Williams emphasizes the ways that utilitarianism alienates us from our personal cares and values and thus undermines our integrity; see his contribution to Smart and Williams, Utilitarianism.

ii

preface

This page intentionally left blank

Bibliography

Anderson, Elizabeth. Value in Ethics and Economics. Cambridge: Harvard University Press, 1993. Angell, Martha. “Ethical Imperialism? Ethics in International Collaborative Clinical Research.” New England Journal of Medicine 318 (1988): 1083–6. – “Investigators’ Responsibilities for Human Subjects in Developing Countries.” New England Journal of Medicine 342 (2000): 967–9. Arneson, Richard. “Broadly Utilitarian Theories of Exploitation and Multinational Clinical Research.” In Exploitation and Developing Countries: The Ethics of Clinical Research, edited by Jennifer Hawkins and Ezekiel Emanuel, 142–74. Princeton: Princeton University Press, 2008. Audi, Robert. The Good in the Right: A Theory of Intuition and Intrinsic Value. Princeton: Princeton University Press, 2004. – Moral Value and Human Diversity. Oxford: Oxford University Press, 2007. Baum, Geraldine. “New York Legislature Passes Gay Marriage Bill.” Los Angeles Times, 25 June 2011. http://articles.latimes.com/2011/jun/25/nation/ la-na-new-york-gay-marriage-20110625. Accessed 22 August, 2011. Beauchamp, Tom. “Principlism and Its Alleged Competitors.” Kennedy Institute of Ethics Journal 5 (1995): 181–98. Beauchamp, Tom, and James Childress. Principles of Biomedical Ethics. 6th ed. New York: Oxford University Press, 2009. Berlin, Isaiah. The Crooked Timber of Humanity. Princeton: Princeton University Press, 1998. – Liberty. New York: Oxford University Press, 2002. Bloom, Paul. “How Morals Change.” Nature 464 (2010): 490.

184

Bibliography

Bonjour, Laurence. Structure of Empirical Knowledge. Cambridge: Harvard University Press, 1985. Brand-Ballard, Jeffrey. “Why One Basic Principle?” Utilitas 19 (2007): 220–42. – “Consistency, Common Morality, and Reflective Equilibrium.” Kennedy Institute of Ethics Journal 13 (2003): 231–58. Brink, David. Moral Realism and The Foundations of Ethics. Cambridge: Cambridge University Press, 1989. – “Moral Conflict and Its Structure.” The Philosophical Review 103 (1994): 215–47. Calhoun, Cheshire. “Standing for Something.” Journal of Philosophy 92 (1995): 235–60. Carlson, Robert V., Kenneth M. Boyd, and David J. Webb. “The Revision of the Declaration of Helsinki: Past, Present and Future.” British Journal of Clinical Pharmacology 57 (2004): 695–713. Cartwright, Nancy. The Dappled World: A Study of the Boundaries of Science. Cambridge: Cambridge University Press, 1999. Childress, James. Practical Reasoning in Bioethics. Bloomington: Indiana University Press, 1997. Christman, James. “Constructing the Inner Citadel: Recent Work on the Concept of Autonomy.” Ethics 99 (1988): 109–24. Clarke-Doane, Justin. “Moral Epistemology: The Mathematics Analogy.” Noûs 48 (2014): 238–55. Corbie-Smith, Giselle. “The Continuing Legacy of the Tuskegee Syphilis Study: Considerations for Clinical Investigation.” The American Journal of the Medical Sciences 317 (1999): 5–8. Dancy, Jonathan. “Moral Particularism.” The Stanford Encyclopedia of Philosophy (Fall 2013), edited by Edward N. Zalta, http://plato.stanford.edu /archives/fall2013/entries/moral-particularism. – Moral Reasons. Oxford: Blackwell, 1993. – Ethics without Principles. Oxford: Clarendon Press, 2004. Daniels, Norman. “Wide Reflective Equilibrium and Theory Acceptance in Ethics.” Journal of Philosophy 76 (1979): 256–82. – “Reflective Equilibrium.” The Stanford Encyclopedia of Philosophy (Winter 2013), edited by Edward N. Zalta. http://plato.stanford.edu/archives /win2013/entries/reflective-equilibrium/. – Justice and Justification: Reflective Equilibrium in Theory and Practice. Cambridge: Cambridge University Press, 1996. Davidson, Donald. “How Is Weakness of the Will Possible?” In Moral Concepts, edited by Joel Feinberg, 93–113. Oxford: Oxford University Press, 1970.

Bibliography

185

Davis, David Brion. Inhuman Bondage: The Rise and Fall of Slavery in the New World. Oxford: Oxford University Press, 2006. DeMarco, Joseph. “Coherence and Applied Ethics.” Journal of Applied Philosophy 14 (1997): 289–300. – “Principlism and Moral Dilemmas: A New Principle.” Journal of Medical Ethics 31 (2005): 101–5. Denis, Lara. “Animality and Agency: A Kantian Approach to Abortion.” Philosophy and Phenomenological Research 76 (2008): 117–37. DePaul, Michael. “Two Conceptions of Coherence Methods in Ethics.” Mind New Series 96 (1987): 463–81. Dorsey, Dale. “A Coherence Theory of Truth in Ethics.” Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition 127 (2006): 493–523. Dworkin, Gerald. “The Concept of Autonomy.” In The Inner Citadel: Essays on Individual Autonomy, edited by John Christman, 54–62. New York: Oxford University Press, 1989. Dworkin, Ronald. Taking Rights Seriously. Cambridge: Harvard University Press, 1977. – Law’s Empire. Cambridge: Belknap Press of Harvard University Press, 1986. – “Do Liberal Values Conflict?” In The Legacy of Isaiah Berlin, edited by Mark Lilla, Ronald Dworkin, Robert B. Silvers, and Aileen Kelly, 73–90. New York: New York Review Books, 2001. – Justice for Hedgehogs. Cambridge: Harvard University Press, 2011. Emerson, Michael. “Through Tinted Glasses: Religion, Worldviews, and Abortion Attitudes.” Journal for the Scientific Study of Religion 35 (1996): 41–55. Finklestein, Claire. “Two Men and a Plank.” Legal Theory 7 (2002): 279–306. Foot, Philippa. “Moral Realism and Moral Dilemma.” Journal of Philosophy 80 (1983): 379–98. Frankfurt, Harry. “Freedom of the Will and the Concept of a Person.” Journal of Philosophy 68 (1971): 5–20. – “Identification and Wholeheartedness.” In The Importance of What We Care About, 159–76. Cambridge: Cambridge University Press, 1988. – “The Faintest Passion.” In Necessity, Volition, and Love, 95–107. Cambridge: Cambridge University Press, 1999. – “Reply to J. David Velleman.” In The Contours of Agency, edited by Sarah Buss and Lee Overton, 124–8. Cambridge: MIT Press, 2002. Gensler, Harry. “A Kantian Argument against Abortion.” Philosophical Studies 49 (1986): 83–98.

186

Bibliography

Gert, Bernard, Charles Culver, and K. Danner Clouser. Bioethics: A Return to Fundamentals. New York: Oxford University Press, 1997. – “Common Morality versus Specified Principlism: Reply to Richardson.” Journal of Medicine and Philosophy 25 (2000): 308–22. Gettner, Alan. “Hare and Fanaticism.” Ethics (1977): 160–4. Gewirth, Alan. “Categorical Consistency in Ethics.” The Philosophical Quarterly 69 (1967): 289–99. – Reason and Morality. Chicago: University of Chicago Press, 1980. Gillespie, Norman. “On Treating Like Cases Differently.” The Philosophical Quarterly 25 (1975): 151–8. Gillon, Raanan. “Four Scenarios.” Journal of Medical Ethics 29 (2003): 267–8. Goldman, Alan. Practical Rules. Cambridge: Cambridge University Press, 2001. – Reasons from Within: Desires and Values. Oxford: Oxford University Press, 2010. Gordon, John-Stewart, Oliver Rauprich, and Jochen Vollman. “Applying the Four-principle Approach.” Bioethics 25 (2009): 293–300. Granberg, Donald. “Comparison of Pro-choice and Pro-life Activists: Their Values, Attitudes, and Beliefs.” Population and Environment 5 (1982): 75–94. Greenspan, Patricia. “A Case of Mixed Feelings: Ambivalence and the Logic of Emotion.” In Explaining Emotions, edited by Amélie Rorty, 230–41. Berkeley: University of California Press, 1980. – “Moral Dilemmas and Guilt.” Philosophical Studies 43 (1981): 117–25. Griffin, James. Well-Being: Its Meaning, Measurement and Moral Importance. Oxford: Clarendon Press, 1986. Haidt, Jonathan. “The Emotional Dog and Its Rational Tail: A Social Intuitionist Approach to Moral Judgment.” Psychological Review 108 (2001): 814–34. – “The Emotional Dog Does Learn New Tricks: A Reply to Pizarro and Bloom.” Psychological Review 110 (2003): 197–8. Haidt, Jonathan, and Craig Joseph. “Intuitive Ethics: How Innately Prepared Intuitions Generate Culturally Variable Virtues.” Dædalus 133, no. 4 (2004): 55–66. – “The Moral Mind: How Five Sets of Innate Intuitions Guide the Development of Many Culture-specific Virtues, and Perhaps Even Modules.” In The Innate Mind, vol. 3, edited by Peter Carruthers, Stephen Laurence, and Stephen Stich, 367–91. New York: Oxford University Press, 2007. Hare, R.M. Freedom and Reason. Oxford: Clarendon Press, 1963.

Bibliography

187

– Moral Thinking: Its Levels, Method, and Point. Oxford: Clarendon Press, 1981. Hawkins, Jennifer, and Ezekiel Emanuel, eds. Exploitation and Developing Countries: The Ethics of Clinical Research. Princeton: Princeton University Press, 2008. Haworth, Lawrence. Autonomy: An Essay in Philosophical Psychology and Ethics. New Haven: Yale University Press, 1986. Hooker, Brad. Ideal Code, Real World. New York: Oxford University Press, 2000. Hooker, Brad, and Margaret Little, eds. Moral Particularism. New York: Oxford University Press, 2001. Hubin, Donald. “Irrational Desires.” Philosophical Studies: An International Journal for Philosophy in the Analytic Tradition 62 (1991): 23–44. Hurka, Thomas. “Monism, Pluralism, and Rational Regret.” Ethics 106 (1996): 555–75. Jeske, Diane. Review of Robert Audi, Moral Value and Human Diversity. Notre Dame Philosophical Reviews, 11 November 2008. http://ndpr.nd .edu/news/23845-moral-value-and-human-diversity/. Accessed 8 March 2015. Kagan, Shelly. “The Additive Fallacy.” Ethics 99 (1988): 5–31. – The Limits of Morality. Oxford: Clarendon Press, 1989. Kaposy, Chris. “Two Stalemates in the Philosophical Debate About Abortion and Why They Cannot Be Resolved Using Analogical Arguments.” Bioethics 26 (2012): 84–92. Kappel, Klemens. “The Meta-justification of Reflective Equilibrium.” Ethical Theory and Moral Practice 9 (2006): 131–47. Korsgaard, Christine. “The Right to Lie: Kant on Dealing with Evil.” Philosophy and Public Affairs 15 (1986): 325–49. Lemmon, John. “Deontic Logic and the Logic of Imperatives.” Logique et Analyse 8, no. 29 (1965): 39–71. Little, Margaret. “Abortion, Intimacy, and the Duty to Gestate.” Ethical Theory and Moral Practice 2 (1999): 295–312. Loeb, Don. “Generality and Moral Justification.” Philosophy and Phenomenological Research 56 (1996): 79–96. Luban, David. Lawyers and Justice: An Ethical Study. Princeton: Princeton University Press, 1988. Mackie, J.L. Ethics: Inventing Right and Wrong. New York: Penguin, 1977. Macklin, Ruth. “Moral Progress.” Ethics 87 (1977): 370–82. – Double Standards in Medical Research in Developing Countries. Cambridge: Cambridge University Press, 2004.

188

Bibliography

Maddy, Penelope. Defending the Axioms: On the Philosophical Foundations of Set Theory. New York: Oxford University Press, 2011. Marcus, Ruth. “Moral Dilemmas and Consistency.” The Journal of Philosophy 77 (1980): 121–36. Marino, Patricia. “Moral Dilemmas, Collective Responsibility, and Moral Progress.” Philosophical Studies 104 (2001): 203–25. – “On Essentially Conflicting Desires.” The Philosophical Quarterly 59 (2009): 274–91. – “Moral Rationalism and the Normative Status of Desiderative Coherence,” Journal of Moral Philosophy 7 (2010): 227–52. – “Ambivalence, Valuational Inconsistency, and the Divided Self.” Philosophy and Phenomenological Research 83 (2011): 41–71. – “Moral Coherence and Value Pluralism.” Canadian Journal of Philosophy 43 (2013): 117–35. – “Moral Coherence and Principle Pluralism.” Journal of Moral Philosophy 11 (2014): 727–49. Marmor, Andrei. “The Rule of Law and Its Limits.” Law and Philosophy 23 (2004): 1–43. – “Should Like Cases Be Treated Alike?” Legal Theory 11 (2005): 27–38. Marquis, Don. “Why Abortion Is Immoral.” Journal of Philosophy 86 (1989): 183–202. Mason, Elinor. “Value Pluralism.” The Stanford Encyclopedia of Philosophy (Fall 2011 Edition), edited by Edward N. Zalta, http://plato.stanford.edu /archives/fall2011/entries/value-pluralism/. McConnell, Terrance. “Moral Dilemmas.” The Stanford Encyclopedia of Philosophy (Fall 2014 Edition), edited by Edward N. Zalta, http://plato.stanford .edu/archives/fall2014/entries/moral-dilemmas/. McNaughton, David. “An Unconnected Heap of Duties?” The Philosophical Quarterly 46 (1996): 433–47. Mill, John Stuart. Utilitarianism. Indianapolis: Hackett, 1979. First published 1861. Moore, Alan. Watchmen. Illustrated by Dave Gibbons. New York: DC Comics, 1995. Murphy, Jeffrie. “Justifying Departures from Equal Treatment.” The Journal of Philosophy 81, (1984): 587–93. Nagel, Thomas. The Possibility of Altruism. Oxford: Clarendon Press, 1970. – “Fragmentation of Value.” In Mortal Questions, 128–41. Cambridge: Cambridge University Press, 1979. – “The Excessive Demands of Impartiality.” London Review of Books 4, no. 12 (1 July 1982).

Bibliography

189

Nussbaum, Martha. “Why Practice Needs Ethical Theory: Particularism, Principle, and Bad Behavior.” In Moral Particularism, edited by Brad Hooker and Margaret Olivia Little, 227–55. Oxford: Clarendon Press, 2000. Patterson, Greg. “AIDS and the Global Conscience.” The New York Times, 10 February 2001, http://www.nytimes.com/2001/02/15/opinion/l-aids-andthe-global-conscience-930644.html. Accessed 12 July 2014. Pinker, Steven. “The Moral Instinct.” The New York Times Magazine, 13 January 2008. Pizarro, David, and Paul Bloom. “The Intelligence of the Moral Intuitions: A Comment on Haidt (2001).” Psychological Review 110 (2003): 193–6. Pogge, Thomas. “Testing Our Drugs on the Poor Abroad.” In Exploitation and Developing Countries: The Ethics of Clinical Research, edited by Jennifer Hawkins and Ezekiel Emanuel, 105–41. Princeton: Princeton University Press, 2008. Posner, Richard. The Problematics of Moral and Legal Theory. Cambridge: Belknap Press of Harvard University Press, 1999). Proust, Marcel. In Search of Lost Time, Volume 1: Swann’s Way. Translated by C.K. Scott Moncrieff. New York: Modern Library, 1992. Rainbolt, George. “Perfect and Imperfect Obligations.” Philosophical Studies 98 (2000): 233–56. Rawls, John. A Theory of Justice. Cambridge: Belknap Press of Harvard University Press, 1971. – “Justice as Fairness.” The Philosophical Review 67 (1958): 164–94. Raz, Joseph. “The Relevance of Coherence.” Boston University Law Review 72 (1992): 273–321. Richardson, Henry. “Specifying Norms as a Way to Resolve Concrete Ethical Problems.” Philosophy and Public Affairs 19 (1990): 279–310. – “Specifying, Balancing, and Interpreting Bioethical Principles.” Journal of Medicine and Philosophy 25 (2000): 285–307. Ross, W.D., The Right and the Good. Edited by Philip Stratton-Lake. Oxford: Clarendon Press 2002. First published 1930. Rudman, Laurie A., Richard D. Ashmore, and Melvin L. Gary. “‘Unlearning’ Automatic Biases: the Malleability of Implicit Prejudice and Stereotypes.” Journal of Personality and Social Psychology 81 (2001): 856–68. Russell, Daniel. “What Virtue Ethics Can Learn from Utilitarianism.” In The Cambridge Companion to Utilitarianism, edited by Ben Eggleston and Dale E. Miller, 258–79. Cambridge: Cambridge University Press, 2014. Ryan, Rebecca. “The Sex Right: A Legal History of the Marital Rape Exemption.” Law and Social Inquiry 20 (1995): 941–1001.

190

Bibliography

Satz, Deborah. Why Some Things Should Not Be for Sale: The Moral Limits of Markets. New York: Oxford University Press, 2010. Sayre-McCord, Geoffrey. “Coherence and Models for Moral Theorizing.” Pacific Philosophical Quarterly 66 (1985): 170–90. – “Coherentist Epistemology and Moral Theory.” In Moral Knowledge?, edited by Walter Sinnott-Armstrong and Mark Timmons, 137–89. New York: Oxford University Press, 1996. Scanlon, T.M. “Rawls on Justification. ” In The Cambridge Companion to Rawls, edited by Samuel Freeman, 139–67. Cambridge: Cambridge University Press, 2003. Schroeder, Timothy, and Nomy Arpaly. “Alienation and Externality.” Canadian Journal of Philosophy 29 (1999): 371–87. Schulhofer, Stephen. Unwanted Sex: The Culture of Intimidation and the Failure of Law. Cambridge: Harvard University Press, 1998. Sen, Amartya. The Idea of Justice. Cambridge: Belknap Press of Harvard University Press, 2009. Shweder, Richard, Lene Arnett Jensen, and William M. Goldstein. “Who Sleeps by Whom Revisited: A Method for Extracting the Moral Goods Implicit in Practice.” New Directions for Child and Adolescent Development 67 (1995): 21–39. Shweder, Richard, Nancy Much, Manamohan Mahapatra, and Lawrence Park. “The ‘Big Three’ of Morality (Autonomy, Community, and Divinity), and the ‘Big Three’ Explanations of Suffering.” In Morality and Health, edited by Allan Brandt and Paul Rozin, 119–69. New York: Routledge, 1997. Sidgwick, Henry. The Methods of Ethics. Indianapolis: Hackett, 1981. This is a republication of the 7th (1907) edition published by MacMillan and Company. Siegel, Andrew. “Kantian Ethics, Exploitation, and Multinational Clinical Trials.” In Exploitation and Developing Countries, edited by Jennifer Hawkins and Ezekiel Emanuel, 286–313. Princeton: Princeton University Press, 2008. Singer, Marcus. “Generalization in Ethics.” Mind 64 (1955): 361–75. – Generalization in Ethics: An Essay in the Logic of Ethics, with the Rudiments of a System of Moral Philosophy. New York: Knopf, 1971. Singer, Peter. “Reasoning Toward Utilitarianism.” In Hare and Critics: Essays on Moral Thinking, edited by Douglas Seanor and N. Fotion, 145–59. Oxford: Clarendon Press; New York: Oxford University Press, 1988 – The Expanding Circle: Ethics, Evolution, and Moral Progress. Princeton:

Bibliography

191

Princeton University Press 2011. First published 1981. – Practical Ethics. 3rd ed. Cambridge: Cambridge University Press, 2011. Sklar, Lawrence. “Dappled Theories in a Uniform World.” Philosophy of Science 70 (2003): 424–41. Smart, J.J.C., and Bernard Williams. Utilitarianism: For and Against. Cambridge: Cambridge University Press, 1973. Smith, Michael. “Realism.” In A Companion to Ethics, edited by Peter Singer, 399–410. Oxford: Blackwell, 1991. – The Moral Problem. Oxford: Blackwell, 1995. – Ethics and the A Priori. Cambridge: Cambridge University Press, 2004. Smith, Sheldon. “Models and the Unity of Classical Physics: Nancy Cartwright’s Dappled World.” Philosophy of Science 68 (2001): 456–75. Stocker, Michael. Plural and Conflicting Values. Oxford: Clarendon Press, 1990. Styron, William. Sophie’s Choice. New York: Bantam, 1980. Sunstein, Cass, and Daniel Kahneman. “Cognitive Psychology of Moral Intuitions.” In Neurobiology of Human Values, edited by Jean-Pierre P. Changeux, Antonio Damasio, and Wolf Singer, 91–105. Berlin, New York: Springer-Verlag, 2005. Swanton, Christine. “Profiles of the Virtues.” Pacific Philosophical Quarterly 76 (1995): 47–72. Talisse, Robert. “Value Pluralism and Liberal Politics.” Ethical Theory and Moral Practice 14 (2011): 87–100. Thagard, Paul. “Ethical Coherence.” Philosophical Psychology 11 (1998): 405–22. – Coherence in Thought and Action. Cambridge: MIT Press, 2000. – “Coherence, Truth, and the Development of Scientific Knowledge.” Philosophy of Science 74 (2007): 28–47. Thomas, Alan. “Consequentialism and the Subversion of Pluralism.” In Morality, Rules and Consequences: A Critical Reader, edited by Brad Hooker, Elinor Mason, and Dale E. Miller, 179–202. Edinburgh: Edinburgh University Press, 2000. Thomson, Judith Jarvis. “A Defense of Abortion.” Philosophy and Public Affairs 1 (1971): 47–66. Timmons, Mark. Disputed Moral Issues: A Reader. 2nd ed. New York: Oxford University Press, 2011. Van Willigenburg, Theo. “Moral Compromises, Moral Integrity and the Indeterminacy of Value Rankings.” Ethical Theory and Moral Practice 3 (2000): 385–404.

192

Bibliography

Varden, Helga. “Kant and Lying to the Murderer at the Door ... One More Time: Kant’s Legal Philosophy and Lies to Murderers and Nazis.” Journal of Social Philosophy 41 (2010): 403–21. Veatch, Robert. “Resolving Conflicts Among Principles: Ranking, Balancing, and Specifying.” Kennedy Institute of Ethics Journal 5 (1995): 199–218. Velleman, J. David. “Identification and Identity.” In The Contours of Agency: Essays on Themes from Harry Frankfurt, edited by Sarah Buss and Lee Overton, 91–123. Cambridge: MIT Press, 2002. – “Willing the Law.” In Practical Conflicts: New Philosophical Essays, edited by Peter Baumann and Monika Betzler, 27–56. Cambridge: Cambridge University Press, 2004. Warren, Mary Anne. “On the Moral and Legal Status of Abortion.” Monist 57 (1973): 43–61. Williams, Bernard. “Ethical Consistency.” Proceedings of the Aristotelian Society, Supplementary Volumes 39 (1965): 103–38. – “Consistency and Realism.” Proceedings of the Aristotelian Society, Supplementary Volumes 40 (1966): 1–22. Wilson, Margaret. “Moral Progress Without Moral Realism.” Philosophical Papers 39 (2010): 97–116. Wolf, Susan. “Two Levels of Pluralism.” Ethics 102 (1992): 785–98. Yablo, Stephen. “Is Conceivability a Guide to Possibility?” Philosophy and Phenomenological Research 53 (1993): 1–42. Zarrella, John, and Lateef Mungin. “Florida Lifeguard Says He’s Been Offered His Job Back.” CNN US, 6 July 2012. http://www.cnn.com/2012 /07/05/us/florida-lifeguard-fired/index.html. Accessed 12 July 2014.

Index

abortion, 3, 4, 5, 12, 97–8, 122, 151, 155; analyzed through pluralist coherence, 130, 156–8; and delimiting, 131–2; future-likeours argument, 97; and mutuality, 133; and personhood, 97, 98; pro-choice views on, 98, 132, 156; pro-life views on, 98, 132, 156; and remainders, 131–2; unified theories applied to, 146–7; violinist argument, 97, 131, 151–2 accidental generalization, 62, 70 ambivalence, 38, 40, 46, 47–8, 55-6, 168n33 Angell, Marcia, 13 anthology approach to theories, 147, 149 Anthony example, 26, 30, 31, 106–7, 112 applied ethics, 96–7, 142, 145, 152 arbitrariness, 4, 31, 55, 81, 118, 121, 126, 159; and case consistency, 84, 101–4, 144 Aristotle, 104 art, 128, 159 Audi, Robert, 7, 67, 95, 147, 162n5 autonomy: and abortion, 4, 131,

132–3, 147, 156–7; in medical testing case, 96, 113, 119, 120, 128; personal, 11, 38, 52–4; sexual, 18, 67, 76, 94, 103; as a value, 8, 16, 24, 26, 61, 62, 73, 77, 98, 113, 134, 135, 148, 149 balancing, 99, 116, 134–5; justified versus intuitive, 99 Beauchamp, Tom, 71. See also Beauchamp and Childress Beauchamp and Childress, 8, 33, 71, 72, 94, 133–6, 146, 180n12 beneficence. See benevolence benevolence, 4, 5, 7, 16, 25, 29, 104, 113, 130, 160, 163n11 Berlin, Isaiah, 34–5 blood transfusion, refusal of, 111 Bloom, Paul, 154–6, 158 Bonjour, Laurence, 170n22 Brand-Ballard, Jeffrey, 169n12 Brink, David, 5, 23, 61, 65, 73, 89, 166n14 Calhoun, Cheshire, 167n17, 177n28 Cartwright, Nancy, 78–9 case consistency, 6, 7, 11, 84–107,

194

Index

127, 135, 137, 143–5; compatible with value pluralism, 86; different from treating like cases like, 87–8; difficult to achieve, 120; in examples, 128–33; and expansion, 119; and framing effects, 85, 143; and interchangeability, 88–9, 139–41; internal, 95, 104, 127, 129, 133, 139, 145, 149, 150; and Michael Smith, 81; normativity of, 120–6; pragmatic aspects of, 122–3; as a principled way of judging, 93, 121; and specification, 99–101; too complex and open-ended, 84, 96–101; versus ranking or lexical ordering, 84. See also arbitrariness; morally relevant similarities and differences; pluralist coherence causal mechanism, 62 chastity, 4, 18, 76–7, 94 child abuse, 73 Childress, James, 152. See also Beauchamp and Childress coercion, 17, 18, 24, 113, 148 coherence: chronological, 86, 123, 125–6; moral (see moral coherence); in scientific theories, 70–1, 74–80, 83 collective action, 7, 115, 175n3 collective good, 16, 17, 26, 83, 96, 128. See also harm prevention complexity, moral, 120. See also pluralist coherence, too complex and open-ended comprehensiveness, 22, 23, 61, 67–8, 69, 136 compromise, 4, 7 44, 50, 52–3, 58–9, 127, 143

conflict. See consistency and inconsistency. See also dilemmas; Marie example; medical testing case; prioritization; pro tanto obligations; remainders; specification; valence; value pluralism conjunction of principles using “and,” 164n35 connectedness, 6, 8, 22, 23, 32, 60, 63, 65, 68, 69, 70–1, 136 consent: informed, 24, 61, 62, 129–30, 134; sexual, 18, 94, 99, 103, 117 consequentialism, 24, 60, 65–7, 73, 74, 169n12. See also utilitarianism consistency and inconsistency: case (see case consistency); essential, of principles, 21–2, 27, 32–3, 115, 138; essential, and the self, 42–3; essential, of values, 39–44, 46–58; practical, of principles, 21–2, 27, 29, 32–3, 110, 115, 138; practical, and the self, 41–2; practical, of values, 39–46; valuational inconsistency and autonomy, 52–4; valuational inconsistency and dissatisfaction with the self, 54–8; valuational inconsistency and inability to act, 48–9; valuational inconsistency as merely instrumental, 43; valuational inconsistency and vacillation, 49–52. See also delimiting; evidentiary consistency conviction ethics, 21, 63, 68, 69, 90, 91, 99, 101, 105, 107, 116, 133, 137 convictions. See judgments Dancy, Jonathan, 117, 122–4 Daniels, Norman, 69–71

Index

Davis, David Brion, 104 deception, 18, 24, 74, 136–7, 148. See also honesty; lying Declaration of Helsinki, 162n2 delimiting of values, 36–59, 143; in abortion context, 131–2; characterized, 37; interpreted using essential inconsistency, 46; interpreted using practical inconsistency, 44 DeMarco, Joseph, 110–11 Denis, Lara, 147 desires: conflicting, 46–7, 53, 166n14; and endorsement (see valuation); second-order, 46; dialogue, 122 difference principle (Rawls), 105. See also justice, distributive dilemmas, 36, 106, 138, 168n35; and abortion, 132; Beauchamp and Childress on, 134, 178n39; and blame, 109; characterized, 27; and deliberative coherence, 138; and disagreement, 4; and expansion, 121; in moral life, 10, 30; and moral progress, 112, 144, 152; and mutuality principle, 110–13, 143; and plural coherence, 126–8, 133; and prioritization, 5; reality of, 27, 108–13; and specification, 117; stringency objection to, 109, 113–15. See also mutuality principle; remainders disagreement: and abortion, 132, 151; about reasons, 73–4; about moral relevance, 95; basic versus contextual, 14–15; characterized, 12–15; compatible with reasoning, 9, 126–8, 142, 150–2, 155–7;

195

entrenched, 142, 158; and expansion, 8; intractable, 127, 128, 133, 150, 158; and moral progress, 7, 112–13, 117–20, 144; and particularism, 123; and pedagogy, 150; and pluralist coherence, 126–8, 133; real versus apparent, 14; related to prioritization, 16–17, 58, 127, 145, 157–8; related to value pluralism, 4, 16–17, 58, 95, 126–7, 145, 151, 157–8; and ruleconsequentialism, 66–7; and skepticism, 3; and unified theories, 146–7. See also moral discourse, public dishonesty. See honesty disunity. See unity diversity. See disagreement Dorsey, Dale, 169n18 double standard. See medical testing case Dworkin, Ronald, 37, 59 economic policy, 126 economic prosperity, 44, 62 emotion, 31, 48, 85, 127, 128 equality and inequality, 16, 17, 35, 44, 130; economic, 7–8. See also justice, distributive essential inconsistency. See consistency and inconsistency ethical imperialism, 3, 14 eugenics, 76 euthanasia, 150, 157 evidentiary consistency (SayreMcCord), 68–9 expansion, 8, 77, 108, 117–20, 121, 143, 144, 152, 176n24 exploitation, 13, 14, 17, 113

196

Index

fairness: as a value 4, 6, 8, 16, 18, 34, 58, 160; and case consistency, 123; and medical testing case, 17; in wide reflective equilibrium, 70 fidelity, 6, 7, 16, 17, 25, 26, 29, 36, 39, 45, 110, 127. See also loyalty; Marie example framing effects, 6, 85, 126, 144, 145 Frankfurt, Harry, 46–58 generality, 20, 22, 28, 75–6, 80–1, 82–3, 117–19 generosity, 41, 43, 116, 116 Gensler, Harry, 146 Gert, Culver, and Clouser, 72, 102, 133, 136–7, 147 Gewirth, Alan, 175n16 goals, 137–9, 156, 167n18 golden rule, 92, 139 Goldman, Alan, 86, 125–6 Gordon, John-Stewart, 94 gratitude, 6, 25 guilt, 109, 114–15 Haidt, Jonathan, 9–10, 18, 153–6, 158 Hare, R.M., 88, 93, 139–41, 163n20 harm prevention: in consequentialism, 66–7, 73, 75; in Gert, Culver, and Clouser, 136–7; and libertarianism, 24; and lying, 24, 62, 95; in medical testing case, 119, 129; obligations of, 25, 26, 64; and pluralist coherence, 149; as a value, 6, 17, 18, 19, 34, 58. See also collective good HIV transmission study, 12–14. See also medical testing case honesty: in conflict with other val-

ues, 7, 15, 17, 21, 26, 29, 39, 126, 127, 145; as derivative, 64; and evidential consistency, 69; and moral relevance, 103; and remainders, 36, 45, 110; and specification, 27; and valence, 89–90; as a value, 4, 6, 8, 16, 160. See also lying Hooker, Brad, 7, 23, 65–6 Hubin, Donald, 91 Hurka, Tom, 36 hypocrisy, 85 imperfect duties, 74, 87 incommensurability, 35, 65, 163n13 incompatibility. See incommensurability inconsistency. See consistency and inconsistency indignation, 85, 120, 127 inequality. See equality and inequality injustice. See justice integrity, 54, 123, 144, 177n28 interchangeability, 88–9, 92–3, 102, 140–1 intuitions. See judgments Jeske, Diane 147 judgments, 20–1 justice: and case consistency, 123; distributive, 105, 107, 119–20; and lying, 63; in medical testing case, 17, 26, 96, 112, 119, 128, 130; Rawlsian theory of, 32, 70, 105, 107; and slavery, 73; as a value 6, 16, 18, 25, 102, 140, 141. See also equality and inequality; fairness Kagan, Shelly, 23, 61, 64, 89

Index

Kahneman, Daniel, 85, 120, 127 Kantian and neo-Kantian moral philosophy, 24, 67, 73–4, 76, 77, 95, 130, 146–7, 148 Kappel, Klemens, 25, 117 Korsgaard, Christine, 73–4 lexical ordering, 32, 58–9, 84, 86, 105–7 libertarianism, moral, 24, 76 liberty and freedom, 16, 17, 34, 35, 37, 59, 75, 105, 107, 113, 143 liberty principle (Rawls) 105 Little, Margaret, 131 logical consistency, 21, 166n11 loyalty, 4, 8, 16, 18, 34, 58, 126, 160. See also fidelity Luban, David, 100 Lying, 4, 6, 7, 29, 69, 83, 95, 110, 127, 145; and evidentiary consistency, 69; and moral coherence, 20, 22, 83, 91; and moral relevance, 85–6, 94, 75–6; and remainders, 56; and specification, 27; and unified theories, 62–4, 73–4. See also honesty; Marie example Mackie, J.L., 15 Macklin, Ruth, 96, 113, 119–20, 128 Maddy, Penelope, 172n50 Marcus, Ruth, 31, 110, 132 Marie example: and conflicting values, 17, 37, 39; and interchangeability, 139–40; introduced, 15; and moral coherence, 20; and moral conflict 21, 22, 26, 28, 83, 108, 109–10; and prioritization, 106, 126, 127; and remainders,

197

30–1, 36, 45; and unified theories, 24, 60 Marquis, Don, 97 mathematics, 80–2 McNaughton, David, 161n7, 172n12 medical testing case, 12–14, 15, 17, 24, 26, 61, 96–7, 120; analyzed through pluralist coherence, 128–30; and expansion of values, 119; mutuality applied to, 112 Mill, J.S., 105 miserliness, 43 monism, principle, 25, 68, 82 169n12 monism, value, 34, 36, 37, 38, 58 moral ambassador, 159 moral artist, 159 moral coherence: characterized, 20–5; dialectical role of, 25; and intuitive acceptability, 21, 23, 64, 133, 146; and justification, 23, 28–9, 31–2, 60, 61–70, 71–2, 81, 101–5, 106–7, 116; pragmatic aspect of, 122–3, 169n12; relation to conviction ethics 21; working back and forth, 4, 87, 90, 133, 152. See also comprehensiveness; connectedness; generality; simplicity; systematicity; unification; unity; universality; universalization moral discourse: bitterness in, 9, 10, 127, 153–6; public, 9–10, 98, 142 moral dumbfounding, 10, 153, 158 moral exemplar, 159 moral factors, 89 moral particularism, 122–4 moral progress, 5, 7–8, 9, 79, 108, 112–13, 117–20, 144–5, 150–2

198

Index

moral significance. See morally relevant similarities and differences moral understanding, 60–2, 122–3 morally relevant similarities and differences, 6, 11, 81, 84–96, 102–3, 107, 121–2, 125, 128, 159; and abortion, 131; characterized, 91–6; and conviction ethics, 91; and interchangeability, 93–4, 140; location, 93, 128, 143–5; and pedagogy, 150 mutuality principle, 110–13, 133, 143; and abortion, 132; and collective action, 115; and moral progress, 130; and specification, 117. See also dilemmas Nagel, Thomas, 19–20 narrative, 128, 159 non-maleficence. See harm prevention Nussbaum, Martha, 76–7, 166–7n16 objectivity, 10, 15, 71, 147, 152, 179n56 overall obligations, 29–30, 32, 89, 94 overridingness. See prioritization pedagogy, 9, 147–50 persuasion, 7, 127–8 philosophical methodology, 8, 142, 145–50 Pinker, Steven, 16, 173n19 Pizarro, David, 154–6, 158 pluralist coherence, 9, 108–41, 143–5; and abortion, 130–3; and applied ethics, 128; applied to medical testing case, 128–30; comparison with view of

Beauchamp and Childress, 133–6; comparison with view of Gert, Culver, and Clouser, 136–7; comparison with view of Hare, 139–41; comparison with view of Thagard, 137–9; examples, 128–33; and moral disagreement, 126–8; and moral particularism, 123–4; normative status of, 120–6; and partiality, 119–20; and remainders, 108–13; role of case consistency in, 84, 122; role of expansion in, 119; role of generality in, 118–19; and universality, 122 Pogge, Thomas, 129–30 Posner, Richard, 9, 150–2, 155–6 practical inconsistency. See consistency and inconsistency principle pluralism, 25, 60, 98, 119, 143 principles: characterized, 20; derived versus fundamental, 25 prioritization, 4, 5, 133, 139, 148; and abortion 131, 132, 157–8; and case consistency, 86, 95–6; changing through time, 126; creating disagreement, 7–8, 9, 16–17, 58, 127–8, 133, 144–5, 150, 157–8; and pedagogy, 148–9, 150; of principles/obligations, 16–17, 95–6; and systematicity, 28–33; and value pluralism, 25–6; of values, 11, 35, 37, 49, 51–4; versus ranking or lexical ordering, 6, 58–9, 84, 86, 105–7; versus Rossian judgment, 7 pro tanto obligations, 27, 28–9, 30, 31, 108, 109, 110

Index

promise-keeping: in conflict with other obligations, 15, 21–2, 27, 60, 83, 103, 114, 126, 139; and deductive coherence, 137; and deliberative coherence, 139; and evidentiary consistency, 69; and generality, 83, 118, 122; and moral relevance, 85, 94–5; in neoKantian theories, 73, 95; principle of, 4, 6, 20; and prioritizing, 17, 106–7; and remainders, 30, 108, 109, 110; single principle generates conflict, 28–9, 36, 58; and specification, 117; valence of obligations, 90; and value pluralism, 19, 26, 37, 39 Proust, Marcel, 176–7n26 queer activism, 159. See also samesex marriage racial distinctions, 101, 105 racism, 73, 104–5, 140 Rainbolt, George, 87 ranking. See prioritization; see also lexical ordering rape, 3–4, 18, 73, 76–7, 94, 99–100, 103, 117, 131 rationality, 122; and moral coherence, 80, 121, 125, 141; and valuational consistency, 11, 34, 38, 46, 48, 56, 143. See also Kantian and neo-Kantian moral philosophy Rauprich, Oliver, 94 Rawls, John, 32, 70, 73, 105–7, 140, 165n49 “reading off” objection (Richardson), 90, 99–101, 115, 173n30 reflective equilibrium, 69–71, 82, 152

199

regret, 30, 45, 56, 109, 114; rational, 35, 36, 43, 44, 45, 56, 166–7n16. See also remainders relativism: circumstantial, 162n5; moral, 10, 130 remainders: and abortion, 131–3; Beauchamp and Childress on, 134–5; epistemic aspect of, 115; in medical testing case, 130; and monistic theories, 36, 176n18; and moral progress, 113, 144, 152; motivational aspect of, 115; and mutuality principle, 111, 112, 144; nature of, 114–15; and normative status of coherence, 120; in relation to values, 35–6, 38, 44, 58; relation to moral dilemmas, 31, 108–9, 110–13; and specification, 116–17; in uneven dilemmas, 30, 112 respect for persons, 16, 17, 24, 26, 64, 73, 75–6, 95, 96, 128, 148, 163n11 rich coherence, 5, 11, 44, 84, and abortion, 98, 146–7; and arbitrariness, 96, 101–5, 121, 144; characterized, 23; and conflict, 29, 32 and dialogue, 122; inappropriate for pluralist contexts, 60–1, 65, 68–9, 86, 108; in mathematics, 82; in scientific theories, 78. See also systematicity Richardson, Henry, 90, 99, 173n30, 115–17 rights: and abortion, 3, 97, 98, 131–2, 156; of children, 111; and expansion, 119, 152; as factors, 89, 133; gay, 159; and generality, 118; and informed consent, 61–2;

200

Index

libertarian, 24, 75–6; in Nagel’s fragmentation of value, 19; in rape trials, 100; of sexual autonomy, 4, 67, 76–7, 103; and torture, 148–9; in wide reflective equilibrium, 70; in work of Hare, 140 Ross, W.D., 6, 25, 26–7, 67, 89, 134, 145 Ryan, Rebecca, 76–7 same-sex marriage, 4, 159 Sayre-McCord, Geoffrey, 21, 28–33, 68–9, 81. See also evidentiary consistency self-improvement, 6, 25 self: and autonomy (see consistency and inconsistency); dissatisfaction with (see consistency and inconsistency); divided, 42; ethical, 150; self-defeating behaviour and vacillation, 49–52, 125. See also autonomy, personal sex, 18, 41, 43, 44, 66–7, 76, 94, 99, 117, 118, 153; sex work, 12, 148; monogamy, 16, 41. See also chastity; rape Sidgwick, Henry, 28 Siegel, Andrew, 130 simplicity, 22, 23, 24, 60, 146, 178n47 Singer, Marcus, 173n17 Singer, Peter, 8, 98, 119, 146 Sklar, Lawrence, 171 slavery, 4, 73, 104 Smith, Michael, 80–2 Smith, Sheldon, 78 social intuitionist model, 158 Sophie’s Choice, 114 specification, 27, 109, 110, 115–17, 134–5

Stocker, Michael, 35 Sunstein, Cass, 85, 120, 127 supererogatory acts, 88 Surfaxin trials, 129 systematicity, 11, 12, 121, 143; and abortion, 146–7; and case consistency, 86; characterized, 23; as dangerous, 159–60; epistemological status of, 60–83; and generality, 118; relation to conflict and dilemmas, 28–33. See also moral coherence; rich coherence Talisse, Robert, 45–6 taxes, 6, 17, 22, 85 Thagard, Paul, 75–7, 137–9 Thomas, Alan, 169n12 Thomson, Judith Jarvis, 97, 131, 151–2 ticking bomb cases, 148, 149 Timmons, Mark, 148–9 torture 148, 150 Tuskegee experiments, 13 unification, 8, 26, 11, 59, 79, 95, 143 unified moral theories, 9, 10, 24, 60, 63, 65, 67, 80, 98, 130, 142–3, 145–50 unity: as part of coherence, 6–7, 22–3, 33, 37, 60, 78, 146, 136; of reasons, 74; of science, 77–80; of the self, 40, 46, 47, 48, 54, 143, 168n33; of value, 5, 11, 20, 34–59, 68–9, 143 universality, 78, 92, 122, 123, 146, 163n20. See also universalization universalization, 4, 7, 67, 88, 95, 139–40. See also universality utilitarianism, 5, 24, 76, 130, 140–1,

Index

146, 148, 180n26. See also consequentialism valence, 89–90, 96, 98–9 107, 124, 134, 137 valuation: characterized, 38–9; as endorsed desire, 38–9, 55; valuational utopia, 44. See also consistency and inconsistency value field variation: and abortion, 151, 157–9, characterized, 16; as good, 59. See also prioritization value pluralism, 45, 58, 159; and abortion, 98; and case consistency, 86; characterized, 16; descriptive thesis of, 18–20; and expan-

201

sion, 119; normative thesis of, 18–20, 25, 63, 68; and principle pluralism, 25; and systematicity, 63–6, 68, 71 Van Willigenburg, Theo, 177n28 Veatch, Robert, 175n47 Velleman, J. David, 47–8, 51, 55, 167n18 volition, 46 Vollman, Jochen, 94 war, 150 Warren, Mary Anne, 97 Watchmen, 43 wholeheartedness, 47, 51, 55–6 Williams, Bernard, 31, 35, 102