British Ethical Theorists from Sidgwick to Ewing (The Oxford History of Philosophy) 9780199233625, 0199233624

Table of contents :
Cover
British Ethical Theorists from Sidgwick to Ewing
Copyright
Dedication
Preface
Contents
Bibliographical Abbreviations
Henry Sidgwick
Hastings Rashdall
John McTaggart Ellis McTaggart
H.W.B. Joseph
H.A. Prichard
Bertrand Russell
G.E. Moore
E.F. Carritt
W.D. (Sir David) Ross
John Laird
C.D. Broad
A.C. Ewing
Introduction: British Ethical Theorists from Sidgwick to Ewing
1 Minimal Concepts
1.1 Thin Concepts
1.2 Deontic Concepts
1.3 Value-Concepts
1.4 Concepts and Claims
2 `Ought´ and `Good´
2.1 `Ought´ vs. `Good´
2.2 Reducing `Ought´ to `Good´
2.3 Reducing `Good´ to `Ought´: The Fitting-Attitudes Analysis
2.4 The Fitting-Attitudes Analysis: Objections
3 Kinds of Goodness and Duty
3.1 Intrinsic Goodness
3.2 Prima Facie Duty
3.3 Objective vs. Subjective Duty
4 Non-Naturalism
4.1 Moral Realism
4.2 The Autonomy of Ethics: The Open-Question Argument
4.3 The Open-Question Argument: Objections
4.4 Responses to Subjectivism
5 Intuitionism
5.1 Intuition and Self-Evidence
5.2 Sidgwick´s Conditions
5.3 Certainty and Inference
5.4 Levels of Intuition
6 Moral Truths: Underivative and Derived
6.1 False Derivations
6.2 Duty and Self-Interest
6.3 Theorizing Morality: Two Reasons
6.4 Theory vs. Anti-Theory
6.5 Degrees of Theory
6.6 Inherent Explanations
7 Consequentialism vs. Deontology
7.1 Sidgwick: Against Deontology
7.2 Sidgwick: For Consequentialism
7.3 Prichard, Carritt, Ross: Defending Deontology
8 Act-Consequentialism, Pluralist Deontology
8.1 Consequentialism: Act- and Indirect
8.2 Pluralist Deontology: Consequentialist Overlaps
8.3 Pluralist Deontology: Conflicts of Duty
8.4 Pluralist Deontology: Elaborations
9 Non-Moral Goods
9.1 What Is Pleasure?
9.2 Hedonism: For and Against
9.3 Aesthetic Appreciation
9.4 Knowledge and Achievement
9.5 Aggregating Goods
10 Moral Goods
10.1 Virtue: For and Against
10.2 Forms of Virtue
10.3 Personal Love
10.4 Moral Desert
11 Self-Benefit, Distribution, Punishment
11.1 Promoting Your Good
11.2 Distribution: Intrinsic and Instrumental Goods
11.3 Criminal Punishment
12 Historians of Ethics
12.1 Ancient Ethics
12.2 The British Moralists
12.3 Kant´s Ethics
Bibliography
Index

Citation preview

T H E O X F OR D H I S T O R Y O F P H I L O S O P H Y

British Ethical Theorists from Sidgwick to Ewing

TH E O X F O R D HI S TO R Y O F P H I L O S O P H Y

The Lost Age of Reason: Philosophy in Early Modern India 1450–1700 Jonardon Ganeri Thinking the Impossible: French Philosophy Since 1960 Gary Gutting The American Pragmatists Cheryl Misak

British Ethical Theorists from Sidgwick to Ewing Thomas Hurka

1

3

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries # Thomas Hurka 2014 The moral rights of the author have been asserted First Edition published in 2014 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2014937977 ISBN 978–0–19–923362–5 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

In memory of Dennis McKerlie

Preface This book’s origins lie in an ethics seminar I took in 1975 in my last term as an undergraduate at the University of Toronto. Taught by Wayne Sumner, the seminar excited me in part because, whereas I had mostly studied history of philosophy, its readings were often very current. The subject was utilitarianism, but Wayne fairmindedly included material on competing views. Here, however, his choices were more classical. For an alternative to the utilitarian account of the good we read the last chapter of Moore’s Principia Ethica, and for a deontological rival chapter 2 of Ross’s The Right and the Good. To me these readings were the highlight of the course. Moore’s ideal or perfectionist account of the good resonated with what I had read in my historical courses and with views I had grown up with, and his principle of organic unities expressed more clearly ideas in the Hegel I had (unaccountably to me now) done special courses on. I also admired Ross’s sharpness in framing the consequentialist/deontological debate and defending his side of it, and found the philosophical styles of both writers congenial. The seminar converted me to ethics as a branch of philosophy, but when I went to Oxford to do graduate degrees in it there was little call for further study of Moore or Ross. Their non-naturalist metaethics was still widely regarded as decisively refuted, and in any case the primary focus then was on normative ethics, which was enjoying a revival in the wake of works like A Theory of Justice and Anarchy, State, and Utopia. The revival was largely self-contained, seeing itself as starting afresh rather than needing much guidance from writers earlier in the century. Sidgwick, who had just been rediscovered, was seen as an exception, and in him I found many of the same merits as in Moore and Ross; when Rawls listed Rashdall as a perfectionist, I read him and found his work, too, engaging. So my interest in the philosophers of this period continued. But in the late 1970s there was for the most part little profit for a graduate student in studying them and certainly no encouragement to do so. Philosophy goes in cycles, and what one generation finds not worth reading another thinks has too long been undervalued. So in recent years the moral philosophers of the late nineteenth and early twentieth centuries have slowly been regaining prominence. Sidgwick’s status as the greatest of the classical utilitarians has only solidified, and one now more often sees references to Moore on organic unities, Ross on prima facie duties, Prichard on ‘why be moral?’, Broad on self-referential altruism, and Ewing on ‘good’. It therefore occurred to me that a book on an era that made those and other contributions might now meet with interest. The history of philosophy involves facts, and it took me some considerable time to identify and organize the thoughts of nine sometimes very productive philosophers on all the subjects they addressed. Both that work and the writing that followed were

viii

preface

supported by research grants from the Social Sciences and Humanities Research Council of Canada, as well as a Guggenheim Fellowship in 2006–7 and a Killam Research Fellowship from the Canada Council in 2011–13; during those last two years the bulk of the book was written. I am grateful to all three institutions for their generous and patient support. Some material in the book was published previously in ‘Moore in the Middle’, Ethics 113 (2003), ‘Underivative Duty: Prichard on Moral Obligation’, Social Philosophy and Policy 27, no. 2 (2010), and ‘Sidgwick on Consequentialism vs. Deontology: A Critique’, Utilitas 26 (2014). I thank these journals and their publishers for permission to reprint this material. Rob Shaver and Wayne Sumner read drafts of the complete manuscript and gave me doubtless too-uncritical comments on them. I am grateful to them as well as to David Phillips and another, anonymous referee for Oxford University Press. I also benefited from discussions on more specific topics with, among others, Roger Crisp, Jonathan Dancy, Stephen Darwall, Brad Hooker, Derek Parfit, Arthur Ripstein, Holly Smith, Philip Stratton-Lake, Sergio Tenenbaum, and Mark Timmons. My wife Terry Teskey helped me edit the book’s many successive drafts; I relied as always on her sound judgement and decisiveness. And I have a long-term debt to my former colleague at the University of Calgary, Dennis McKerlie. Likewise an Oxford graduate student in the 1970s, he shared my interest in the philosophers of this book’s period, and we had many conversations both about them and about moral philosophy more generally. He was an ideal philosophical interlocutor, often seeing things more clearly than I could; his own work, which he never promoted as much as other philosophers do, was first rate. Regrettably, health problems prematurely ended his philosophical career and then his life; this book is dedicated to his memory.

Contents Bibliographical Abbreviations

xi

Introduction: British Ethical Theorists from Sidgwick to Ewing

1

1. Minimal Concepts 1.1 1.2 1.3 1.4

Thin Concepts Deontic Concepts Value-Concepts Concepts and Claims

2. ‘Ought’ and ‘Good’ 2.1 2.2 2.3 2.4

‘Ought’ vs. ‘Good’ Reducing ‘Ought’ to ‘Good’ Reducing ‘Good’ to ‘Ought’: The Fitting-Attitudes Analysis The Fitting-Attitudes Analysis: Objections

3. Kinds of Goodness and Duty 3.1 Intrinsic Goodness 3.2 Prima Facie Duty 3.3 Objective vs. Subjective Duty

4. Non-Naturalism 4.1 4.2 4.3 4.4

Moral Realism The Autonomy of Ethics: The Open-Question Argument The Open-Question Argument: Objections Responses to Subjectivism

5. Intuitionism 5.1 5.2 5.3 5.4

Intuition and Self-Evidence Sidgwick’s Conditions Certainty and Inference Levels of Intuition

6. Moral Truths: Underivative and Derived 6.1 6.2 6.3 6.4 6.5 6.6

False Derivations Duty and Self-Interest Theorizing Morality: Two Reasons Theory vs. Anti-Theory Degrees of Theory Inherent Explanations

22 22 25 33 40 44 44 50 52 58 65 65 69 78 86 86 93 98 101 108 108 112 117 119 128 128 131 135 138 141 147

x contents 7. Consequentialism vs. Deontology 7.1 Sidgwick: Against Deontology 7.2 Sidgwick: For Consequentialism 7.3 Prichard, Carritt, Ross: Defending Deontology

8. Act-Consequentialism, Pluralist Deontology 8.1 8.2 8.3 8.4

Consequentialism: Act- and Indirect Pluralist Deontology: Consequentialist Overlaps Pluralist Deontology: Conflicts of Duty Pluralist Deontology: Elaborations

9. Non-Moral Goods 9.1 9.2 9.3 9.4 9.5

What Is Pleasure? Hedonism: For and Against Aesthetic Appreciation Knowledge and Achievement Aggregating Goods

10. Moral Goods 10.1 10.2 10.3 10.4

Virtue: For and Against Forms of Virtue Personal Love Moral Desert

11. Self-Benefit, Distribution, Punishment 11.1 Promoting Your Good 11.2 Distribution: Intrinsic and Instrumental Goods 11.3 Criminal Punishment

12. Historians of Ethics 12.1 Ancient Ethics 12.2 The British Moralists 12.3 Kant’s Ethics

Bibliography Index

150 150 158 165 172 172 178 183 186 194 194 198 203 205 211 216 216 221 228 232 237 237 244 252 259 259 267 272 281 299

Bibliographical Abbreviations Henry Sidgwick EEM

Essays on Ethics and Method, ed. M.G. Singer, 2000 (where possible Sidgwick’s shorter writings are cited from this volume)

EP

The Elements of Politics, 1891

GSM

Lectures on the Ethics of T.H. Green, H. Spencer, and J. Martineau, ed. E.E. Constance Jones, 1902

M

Henry Sidgwick: A Memoir, eds A. Sidgwick and E.M. Sidgwick, 1906

ME

The Methods of Ethics, 7th ed., 1907; earlier editions abbreviated ME1, ME2, etc.

OHE

Outlines of the History of Ethics for English Readers, 1886

PE

Practical Ethics: A Collection of Addresses and Essays, 1898

Hastings Rashdall ‘CAV’

‘The Commensurability of All Values’, 1902

E

Ethics, 1913

ICE

Is Conscience an Emotion?, 1914

‘LC’

‘The Limits of Casuistry’, 1894

‘PS’

‘Professor Sidgwick’s Utilitarianism’, 1885

‘PTP’

‘The Philosophical Theory of Property’, 1913

TGE

The Theory of Good and Evil, 2 vols., 1907

John McTaggart Ellis McTaggart ‘EHS’

‘The Ethics of Henry Sidgwick’, 1906

‘IV’

‘The Individualism of Value’, 1908

NE

The Nature of Existence, 2 vols., 1921 and 1927

SDR

Some Dogmas of Religion, 1906

SHC

Studies in Hegelian Cosmology, 1901

xii

bibliographical abbreviations

H.W.B. Joseph SPE

Some Problems in Ethics, 1931

H.A. Prichard MW

Moral Writings, ed. J. Macadam, 2002 (all Prichard’s ethical writings are cited from this volume)

Bertrand Russell ‘EE’

‘The Elements of Ethics’, 1910

G.E. Moore ‘A’

‘An Autobiography’, in P.A. Schilpp, ed., The Philosophy of G.E. Moore, 1942

‘CIV’

‘The Conception of Intrinsic Value’, in G.E. Moore, Philosophical Studies, 1922

E

Ethics, 1912

EE

The Elements of Ethics, ed. T. Regan, 1991 (contains the text of lectures given in 1898)

‘GQ’

‘Is Goodness a Quality?’, 1932

‘MME’ ‘Mr. McTaggart’s Ethics’, 1903 ‘NMP’ ‘The Nature of Moral Philosophy’, 1922 ‘P’

‘Preface to the Second Edition,’ in T. Baldwin, ed., Principia Ethica, revised edition, 1993 (written for a proposed second edition around 1921 but left unfinished and not published in Moore’s lifetime)

PE

Principia Ethica, 1903

‘R’

‘A Reply to My Critics’, in Schilpp, ed., The Philosophy of G.E. Moore, 1942

E.F. Carritt ‘AG’

‘An Ambiguity of the Word “Good”’, 1937

EPT

Ethical and Political Thinking, 1947

IA

An Introduction to Aesthetics, 1949

‘MP’

‘Moral Positivism and Moral Aestheticism’, 1938

TB

The Theory of Beauty, 1914

TM

The Theory of Morals, 1928

bibliographical abbreviations

xiii

W.D. (Sir David) Ross A

Aristotle, 1923

‘BO’

‘The Basis of Objective Judgments in Ethics’, 1927

‘EP’

‘The Ethics of Punishment’, 1929

FE

Foundations of Ethics, 1939

‘IME’

‘Is There a Moral End?’, 1928

KET

Kant’s Ethical Theory, 1954

‘NM’

‘The Nature of Morally Good Action’, 1929

RG

The Right and the Good, 1930

John Laird SMT

A Study in Moral Theory, 1926

C.D. Broad ‘A’

‘Autobiography’, in P.A. Schilpp, ed., The Philosophy of C.D. Broad, 1959

CE

Broad’s Critical Essays on Moral Philosophy, ed. D. Cheney, 1971 (where possible Broad’s shorter ethics writings are cited from this volume)

EHP

Ethics and the History of Philosophy: Selected Essays, 1952

EMP

Examination of McTaggart’s Philosophy, 2 vols., 1933 and 1938

FT

Five Types of Ethical Theory, 1930

‘R’

‘A Reply to My Critics’, in P.A. Schilpp, ed., The Philosophy of C.D. Broad, 1959

‘RSE’

‘Symposium on the Relations Between Science and Ethics’, 1941

‘SAP’

‘Are There Synthetic A Priori Truths?’, 1936

A.C. Ewing ‘BET’

‘Recent Developments in British Ethical Thought’, 1957

‘BVG’

‘Blanshard’s View of Good’, in P.A. Schilpp, ed., The Philosophy of Brand Blanshard, 1980

DG

The Definition of Good, 1947

E

Ethics, 1953

MP

The Morality of Punishment, 1929

‘PKE’

‘The Paradoxes of Kant’s Ethics’, 1938

xiv

bibliographical abbreviations

‘RI’

‘Reason and Intuition’, 1941

‘RSE’

‘Symposium on the Relations Between Science and Ethics’, 1941

‘SN’

‘A Suggested Non-Naturalistic Analysis of Good’, 1939

ST

Second Thoughts in Moral Philosophy, 1959

‘U’

‘Utilitarianism’, 1948

VR

Value and Reality, 1973

Introduction: British Ethical Theorists from Sidgwick to Ewing The subject of this book is a group of British ethical theorists active in the last quarter of the nineteenth century and the first half of the twentieth, principally Henry Sidgwick, Hastings Rashdall, John McTaggart Ellis McTaggart, H.A. Prichard, G.E. Moore, E.F. Carritt, W.D. Ross, C.D. Broad, and A.C. Ewing. The group disagreed about some important topics in ethics. Several were consequentialists, believing the right act is always the one with the best consequences; others affirmed deontological duties that make some acts with the best outcome wrong. Sidgwick thought the only intrinsic good is pleasure; others recognized additional goods such as knowledge and moral virtue. Among the latter, some thought it good if a person gets what he deserves while others did not. But they shared other important views, first in metaethics. They were all nonnaturalists, believing moral judgements can be objectively true rather than, say, just expressing emotions, and have a distinctive subject matter, so they are neither reducible to nor derivable from ones about natural science, theology, or metaphysics; no ‘ought’ follows from an ‘is’. They were also intuitionists, believing we can know the moral truth by direct insight or intuition, either about principles or about particular cases. In addition, they shared a general approach to normative ethics. They agreed about which are the fundamental concepts for moral thought and which should be set aside or treated as derivative. They also shared, to differing degrees, a commitment to theorizing morality. They believed that whenever a particular moral judgement is true there is a general principle that makes it true, and a central task of normative ethics is to identify the ultimate such principles. A further view derived from their metaethics. Because of their non-naturalism, they believed that some moral judgements, the most explanatory ones, are underivatively true, in the sense that there is no further explanation, either non-moral or moral, of why they are true. If we ask why pleasure is good or why we have a duty to promote others’ good there may be no answer: it just is and we just do. Truths like these cannot be explained, and to try to explain them, as many philosophers have, is a mistake. This view shaped much of their moral theorizing. Because they thought claims about, for example, other-regarding duty can be underivatively true, they felt

2

british ethical theorists from sidgwick to ewing

no need to derive them from principles that use different concepts and concern some other, supposedly more fundamental topic. Instead, their explanations often cited principles that are continuous with the claim being explained, using similar concepts but at a higher level of abstraction. Because these principles tend to be too close to a more specific conclusion to persuade someone who does not already accept it to do so, the group’s work often focussed more on the inner structure and workings of our moral thought than on providing independent justifications of it. These shared views make the group a distinctive school in the history of moral philosophy, pursuing the subject differently than earlier writers such as Aristotle, Hobbes, and Kant and than many present-day ones. Unfortunately, I do not have a satisfactory name for them. They are sometimes called ‘non-naturalists’, but that refers just to one aspect of their metaethics and ignores their equally important commonalities in normative ethics. ‘Intuitionist’ has both metaethical and normative implications, but in this context the latter are misleading. An intuitionist normative view contains a plurality of duties but no formal rules for weighing them against each other;1 the term therefore does not fit consequentialists such as Sidgwick, Rashdall, McTaggart, and Moore. ‘Underivativist’ would highlight a key shared assumption of the school but is an ugly neologism. I will therefore speak only of the ‘Sidgwick-toEwing school’. The book has three main aims, of which one is just to report or recover the school’s ideas. Some of these are reasonably well known. Sidgwick is now widely regarded as the greatest of the classical utilitarians, and his Methods of Ethics has been the subject of several books.2 The first chapter of Moore’s Principia Ethica is for many the locus classicus for the defence of non-naturalism, and many also know his principle of organic unities. Prichard’s ‘Does Moral Philosophy Rest on a Mistake?’ is famous for rejecting the ‘Why be moral?’ question and Ross’s The Right and the Good for introducing the concept of a prima facie duty. But other aspects of the same writers’ work are less well known, such as Moore’s extreme indirect consequentialism, Prichard’s views on subjective vs. objective rightness, and Ross’s account of what is good; others in the school, such as Rashdall, McTaggart, Carritt, Broad, and Ewing, are barely known or read at all. So the book’s first aim is simply to describe their ethical views and the rich vein of thought they contain; one thing that will emerge is how often what have been regarded as new discoveries of recent moral philosophy were known and widely discussed in this earlier period. A second aim is to demonstrate the school’s unity, by elaborating and illustrating the commonalities mentioned above. That they shared important views has often been recognized, but there has not been a detailed explanation of what those are or how they differ from those of rival approaches such as Aristotle’s and Kant’s.

1 2

For this usage see Rawls, Theory of Justice, p. 34; Urmson, ‘Defence of Intuitionism’, pp. 111–12. Schneewind, Sidgwick’s Ethics; Phillips, Sidgwickian Ethics.

introduction

3

The final aim is to evaluate the school’s views. If they disagreed on some topic, such as consequentialism vs. deontology or hedonism vs. perfectionism, who had the better arguments? If they shared a position, how persuasively did they defend it? The book will not settle all these issues; whether non-naturalism is true, for example, is beyond its scope. But it will try to assess the school’s views and will often treat them sympathetically. In particular, it will argue that their general approach to normative ethics, with its focus on structure rather than external justifications, is more illuminating than the more grandiose projects preferred by many other philosophers. Because of its focus on a unified school, the book will not give a complete history of British moral philosophy in the relevant period. In the late nineteenth and early twentieth centuries a rival and even more prominent approach to ethics was grounded in Idealist metaphysics and promoted by T.H. Green, F.H. Bradley, Bernard Bosanquet, and others. Their ideas will be discussed insofar as they generated critiques of our school and responses from them, but not in their own right; the same holds for naturalist and Kantian views. In the 1930s a challenge arose to nonnaturalism from the emotivism of A.J. Ayer and Charles L. Stevenson, later followed by the prescriptivism of R.M. Hare; these too will be discussed only as they drew responses from our school. The focus will only be on one, though an important, strand in British moral philosophy from around 1875 to 1960. The focus will also be primarily on the nine theorists listed above rather than on everyone who shared their general view. No utilitarians other than Sidgwick will get serious attention, and lesser ideal consequentialists such as Bertrand Russell, John Laird, W.A. Pickard-Cambridge, and H.W.B. Joseph will be mentioned only briefly, when a claim of theirs illustrates some broader consensus; the same is true of lesser deontologists such as J.L. Stocks. The book’s nine main figures were the most influential and philosophically interesting of the school, and there is more than enough to explore in them. The focus on a unified school has also shaped the book’s organization. Its treatment is not chronological, with a chapter on each member and covering all his views. It is thematic, with chapters on the different topics the school addressed and discussing all or most of their views on each. This format is better suited to highlighting their commonalities, topic by topic, and to analysing their differences with each other. Chapters 1 to 3 discuss their views on the moral concepts: which they took to be basic, how they saw the relations between the basic concepts, and how they understood specific forms of them such as intrinsic goodness and prima facie duty. Chapters 4 and 5 examine their metaethics, first their non-naturalism and then their intuitionistic moral epistemology. Chapter 6 discusses their general conception of moral theory, while Chapters 7 and 8 concern the theory of the right, one on the general debate between consequentialism and deontology and the other on the specific versions of those views defended by, principally, Sidgwick and Moore in the one case and Ross in the other. The next two chapters concern the theory of the good, Chapter 9 the non-moral goods of pleasure, aesthetic appreciation, and knowledge

4

british ethical theorists from sidgwick to ewing

and Chapter 10 the moral goods of virtue, personal love, and desert. Chapter 11 discusses duties to oneself, economic distribution and criminal punishment, and Chapter 12 the school’s views of earlier moral philosophers such as Plato, Aristotle, Butler, and Kant. This last chapter rounds out the book, since their criticisms of predecessors often reflected the school’s positive views and set them off by contrast. Like other philosophical movements, the school had a rise, an ascendancy, and then a decline. Its opening work was Sidgwick’s Methods, the first edition of which appeared in 1874. With its concern for precise statement and rigorous argument it reads very differently from earlier writing on ethics and is arguably the first philosophical work in a distinctively ‘analytic’ style. It was well received, as its six subsequent editions attest, and was followed in the 1880s and 1890s by some similar writing, including articles of Sidgwick’s defending his views and his Elements of Politics of 1891, and a series of articles by Rashdall. But those decades were dominated in many British universities by Idealist views with a very different flavour; the school was still emerging. The last years of the 1890s and the first decade of the new century saw a greater number of relevant works: in addition to articles by Rashdall, McTaggart, and Moore there were later publications by Sidgwick, such as his posthumous Lectures on the Ethics of T.H. Green, H. Spencer, and J. Martineau (1902), as well as McTaggart’s Studies in Hegelian Cosmology of 1901, which contained substantial discussions of ethics, Moore’s Principia Ethica of 1903, Rashdall’s Theory of Good and Evil of 1907, and Russell’s ‘The Elements of Ethics’ of 1910. The year 1912 saw Moore’s Ethics and Prichard’s ‘Does Moral Philosophy Rest on a Mistake?’, and 1913 and 1914 Rashdall’s Ethics and Is Conscience an Emotion?, and Broad’s first articles on ethics. The school was on the rise, and its period of ascendancy was the 1920s and 1930s. Those decades saw the publication of, again alongside numerous articles, the two volumes of McTaggart’s The Nature of Existence in 1921 and 1927, the second discussing value theory, Laird’s A Study of Moral Theory in 1926, Carritt’s Theory of Morals in 1928, Prichard’s lectures on ‘Duty and Interest’ and ‘Duty and Ignorance of Fact’ in 1928 and 1932, Ewing’s The Morality of Punishment in 1929, Ross’s The Right and the Good and Broad’s Five Types of Ethical Theory in 1930, Joseph’s Some Problems in Ethics in 1931, and Ross’s Foundations of Ethics in 1939. (Since the latter mostly refined the views of Ross’s first book, Broad gave it ‘the affectionate and accurate nickname of “The Righter and the Better” ’.3) Already, however, new winds were blowing, in the form of general philosophical movements: the logical positivism brought to Britain by Ayer in Language, Truth, and Logic of 1936 and the ordinary-language philosophy of Ludwig Wittgenstein in Cambridge and, differently, J.L. Austin in Oxford. To the younger philosophers in these movements the school’s approach to ethics was boring and old-fashioned. In a

3

Broad, Critical Notice of FE, p. 239.

introduction

5

1944 reference letter Isaiah Berlin said H.L.A. Hart, whom he thought too influenced by Joseph, would probably not be ‘a really distinguished epistemologico-logical highbrow’ but would produce work like ‘the solid pedestrian tramp of Ewing or Broad’; Hart himself worried that he was ‘a hack like Ewing’.4 This view solidified after the Second World War. In 1945 Sidgwick, Rashdall, and McTaggart were long dead, Prichard was seventy-four, and Moore, Carritt, and Ross were either over seventy or nearing that age. The school did continue to produce work. In 1947 Ewing published The Definition of Good, expanding the ideas in ‘A Suggested Non-Naturalistic Analysis of Good’ of 1939, and Carritt published Ethical and Political Thinking. Moral Obligation, a posthumous collection of Prichard’s writings on ethics, appeared in 1949, while in 1954 Ross published Kant’s Ethical Theory. Broad wrote articles on ethics from the late 1940s through 1964 and Ewing also remained active, publishing, alongside many articles, the trade-market Ethics in 1953 and Second Thoughts in Moral Philosophy in 1959, while Value and Reality, parts of which discussed ethics, appeared after his death in 1973. But these works were now far out of the mainstream, which was dominated by the new movements or descendants of them. And British philosophers of the 1950s were dismissive of the school, rejecting their views with brusque arguments that often rested on misrepresentations. Broad and Ewing experienced this attitude most directly, both because they were still active and because in Cambridge it took an especially virulent Wittgensteinian form. Broad avoided philosophical presentations partly because he did not want ‘to spend hours every week in a thick atmosphere of cigarette-smoke, while Wittgenstein punctually went through his hoops, and the faithful as punctually “wondered with a foolish face of praise” ’ (‘A’ 61). Ewing, who did attend and was treated with contempt by Wittgenstein,5 once said ‘I wish they’d read my books’; his obituarist said, ‘It is a hard fact that the majority of philosophers placed little value on his work’.6 In 1956 Broad commented ruefully, ‘though philosophies are never refuted, they rapidly go out of fashion, and the kind of philosophy which I have practised has become antiquated without having yet acquired the interest of a collector’s piece’ (‘R’ 830). The dismissiveness continued in the 1960s. That decade saw a trio of histories of twentieth-century moral philosophy, all focussed on metaethics and crediting Moore, inaccurately, with having revolutionized the subject with his defence of nonnaturalism in Principia.7 But they did not see Moore’s view as defensible; they thought it had deep flaws that later approaches such as emotivism or neo-naturalism would try to overcome. It had set the agenda for twentieth-century ethics, but by showing what an acceptable view cannot be.

4

5 Lacey, Life of Hart, pp. 117–18, 115. Edmonds and Edinow, Wittgenstein’s Poker, pp. 67–8. Grice, ‘Alfred Cyril Ewing’, pp. 500–1. 7 Mary Warnock, Ethics Since 1900 (1960); G.J. Warnock, Contemporary Moral Philosophy (1970); Hudson, Modern Moral Philosophy (1970). 6

6

british ethical theorists from sidgwick to ewing

In the 1970s and 1980s interest returned to normative ethics, but still with little attention to our school except perhaps to Sidgwick. As time has passed, however, the relevance to current work of, for example, Moore on what is good or Prichard and Ross on duty has been more noticed and references to them have become more frequent; there has also been a revival of non-naturalist metaethical views not far from theirs. This increasing interest makes, I hope, a comprehensive survey of their ideas now timely. If the school shared important assumptions, a natural question is how far this is because later members were influenced by earlier ones and adopted their views. It is not easy to say, partly because they did not cite others’ work as often as philosophers do today and partly because sometimes thinkers reason independently to similar conclusions. Here a relevant example is Franz Brentano. Moore read his Origin of Our Knowledge of Right and Wrong in translation right after finishing Principia and found in it ‘opinions far more closely resembling my own, than those of any other ethical writer with whom I am acquainted’ (PE xi).8 But Brentano, whose book was originally given as a lecture in Vienna in 1889, was not influenced by any British philosophers; he arrived at his views on his own.9 In addressing the question of influence it is useful to divide the school into two main groups. Four of them—Sidgwick, McTaggart, Moore, and Broad—were educated and spent all or most of their academic careers at Cambridge. Rashdall, Prichard, Carritt, and Ross had the same relationship to Oxford, while Ewing was educated and did his first writing at Oxford but from 1931 taught at Cambridge. There were close personal connections among the Cambridge members, all of whom were at Trinity College. Sidgwick taught both McTaggart and Moore as undergraduates, while McTaggart taught Moore and later Broad. Sidgwick was on the committee that awarded McTaggart a six-year fellowship in 1891. (He is reported to have said of McTaggart’s writing sample on Hegel, ‘I can see that this is nonsense, but what I want to know is whether it is the right kind of nonsense’ (Moore, ‘A’ 21).) He was also on the committee that turned Moore down for a similar fellowship in 1897, but in the following year told the philosopher who succeeded him to make sure Moore won, which he did. It was at McTaggart’s urging that Broad won a similar fellowship, and Broad later was McTaggart’s literary executor, overseeing the posthumous publication of the second volume of The Nature of Existence and writing his two-volume Examination of McTaggart’s Philosophy. Sidgwick, McTaggart, and Moore were all members of the secret Apostles discussion society; though Sidgwick was no longer active in Moore’s day, McTaggart was, and the young don and the precocious undergraduate met often outside lectures.

8

See also Moore, Review of Brentano, Origin of the Knowledge of Right and Wrong. Brentano, Origin of Our Knowledge. This book discusses other British philosophers such as Bain, Bentham, Grote, Hume, and the two Mills, but they are ones whose views Brentano, like Moore, rejected. 9

introduction

7

As an Oxford undergraduate Prichard was in a circle influenced by Rashdall,10 and he later had close contacts with Carritt and Ross. He was Carritt’s undergraduate tutor and for many years an active participant, with Carritt though apparently not with Ross, in a weekly Philosophical Tea that had been initiated by John Cook Wilson and involved short papers followed by discussion; Carritt said he talked philosophy with Prichard nearly once a week during term and corresponded with him on philosophical issues through the university mail.11 Prichard also corresponded with Ross and played golf with him; the two were good friends, and seem, with Carritt and perhaps for a time Ewing, whom Carritt taught as an undergraduate, to have interacted regularly while developing their ethical views. As for philosophical influence, Broad attributed McTaggart’s idiosyncratic version of Idealism to the combination of a youthful passion for Hegel with, among other things, ‘the teaching of Sidgwick and the continual influence of Moore and Russell’ (EHP 75). In ethics Sidgwick’s influence may be reflected in McTaggart’s casual acceptance of non-naturalism, which had been a comparative novelty in Sidgwick, and in his consequentialism and commitment to the measurement of value; his perfectionist theory of the good, however, departed from Sidgwick.12 Moore reported that as a student he was not attracted by Sidgwick’s personality and found his lectures dull (‘A’ 16); according to John Maynard Keynes, he thought Sidgwick a ‘wicked edifactious person’.13 But Moore allowed that he ‘gained a good deal’ from Sidgwick’s published works, ‘especially, of course, his Methods of Ethics’, (‘A’ 16), and there are important similarities between the two men’s views, both in metaethics and in their belief that though the correct moral principle is act-consequentialist, in everyday life we should follow simpler common-sense rules; clearly some ideas were transmitted. Moore said his greatest undergraduate influence was McTaggart (‘A’ 18–19), but apart from a brief conversion to Idealism he was most impressed by McTaggart’s insistence on clarity, which was shared by Sidgwick and other Apostles. Moore’s account of what is good was perfectionist like McTaggart’s but again it is hard to assess influence; though both placed high value on personal love, for example, they had very different views of what love is. The final Cambridge figure, Broad, named the dominant influences in his undergraduate days as Russell and Moore, the latter partly through Principia and despite his not then being in Cambridge; he was also influenced by McTaggart, though again more by his method than by his doctrines (‘A’ 49–50). He greatly admired Sidgwick, devoting by far the longest chapter of Five Types to him and calling his Methods of Ethics ‘on the whole the best treatise on moral theory that has ever been written’ (FT 143). 11 Matheson, Life of Rashdall, p. 54. Carritt, ‘Professor H.A. Prichard’, p. 146. In a book on McTaggart, P.T. Geach dismissed his ethics as not worth discussing because it was influenced by Moore’s PE, which he thought ‘a profoundly confused book’ (Truth, Love and Immortality, pp. 174–5). But the claim about influence here is dubious: McTaggart was defending the independence of ‘ought’ from ‘is’ and a version of ideal consequentialism before PE, for example in SHC of 1901. 13 Keynes, Letter to Strachey, quoted in Harrod, Life of Keynes, p. 114. 10 12

8

british ethical theorists from sidgwick to ewing

Rashdall’s role in the Oxford group is hard to assess, because the later members all rejected his consequentialism. But when Prichard discussed that view in ‘Mistake’ he took Rashdall rather than Sidgwick or Moore as its chief representative (MW 9), while Carritt called The Theory of Good and Evil one of the two works outside ‘the well-known classics’ he was most influenced by (EPT vi). The larger influence on Prichard was Cook Wilson and his philosophical realism, but Carritt reported that whereas Cook Wilson did not think this view required any change in moral philosophy, Prichard did and for that reason concentrated on the subject.14 His defence in ‘Mistake’ of a non-Kantian deontology was largely original, coming as it did when consequentialism was the dominant philosophical view, and he was also the main influence on Carritt and Ross. Many and even most of the views the three shared— about prima facie duties, the independence of rightness from motives, instrumental goodness, and more—were first proposed by Prichard; though one of the other two may have stated them more clearly, they derived ultimately from him. Carritt and Ross readily acknowledged their dependence on Prichard (TM vi, EPT v; RG v), and it was also recognized in Oxford; thus Ayer thought Prichard the most influential Oxford philosopher of the 1920s and 1930s, called Carritt ‘philosophically a paler Prichard’, and said Ross’s books ‘took a position similar to that of Prichard’.15 What about influences between the universities? There was, first, some mutual hostility. Sidgwick wrote sharply critical reviews of works on ethics by Bradley and Green (EEM 185–9, 193–4, 247–58; GSM 1–131) and Bradley responded with a pamphlet attacking Sidgwick’s Methods.16 McTaggart said that on meeting Bradley he felt ‘as if a Platonic Idea had entered the room’ (Moore, ‘A’ 22), but his later Idealist views were very different from Bradley’s, as was his approach to ethics. In his earliest days Moore too was a Bradleyan, but by the time of Principia he was criticizing Idealist views in both metaphysics and ethics. Broad did the same, saying Bradley was as inferior to Sidgwick in ethical and philosophical acumen as he was superior to him in literary style (FT 144) and making two brilliantly vitriolic remarks about Green. ‘Even a thoroughly second-rate thinker like T.H. Green, by diffusing a grateful and comforting aroma of “ethical uplift”, has probably made far more undergraduates into prigs than Sidgwick will ever make into philosophers’ (FT 144).17 And of a paper of Prichard’s criticizing Green: Seldom can the floor have been more thoroughly wiped with the remains of one who was at one time commonly regarded as a great thinker and who still enjoys a considerable reputation

15 Carritt, ‘Prichard: Personal Recollections’, p. 147. Ayer, Part of My Life, pp. 77, 95, 308. Bradley, ‘Mr. Sidgwick’s Hedonism’. 17 Broad was probably responding to an 1884 incident when the economist Alfred Marshall, in an internal Cambridge debate with Sidgwick, contrasted the hundred men who, ignoring examinations, hung on Green’s lips to learn the truth about human life with the handful of exam-takers taking notes from Sidgwick (Sidgwick, M 394). 14 16

introduction

9

in some circles. A large part of the lectures is occupied with disentangling the strands of clotted masses of verbiage, in which inconsistency and nonsense are concealed by ambiguity.18

From the other direction Prichard and Joseph vehemently opposed the use in philosophy of formal methods, especially mathematical logic, as promoted in Cambridge by Russell and others; Prichard’s correspondence also contains hostile comments about Moore’s ethics. But there was mutual influence and admiration between some Oxford and Cambridge members of the school. Rashdall dedicated The Theory of Good and Evil to ‘My Teachers Thomas Hill Green and Henry Sidgwick’, by which he meant his two greatest philosophical influences (TGE vi–vii). We can see his aim as to combine Sidgwick’s methodology, in particular his concern for clarity and his explicit commitment to consequentialism and the measurement of value, with a theory of what has value that is perfectionist like Green’s. He admired Sidgwick, who in turn considered Rashdall one of his ablest critics (ME xi–xiii, EEM 45–6). There was also mutual regard between Rashdall and McTaggart, the latter of whom read and commented on Rashdall’s book in draft (TGE ix) and is often cited in it. Moore initially wrote an ungenerous review of The Theory of Good and Evil but later recommended it to readers of his Ethics as the first work on ethics by a living writer they should read,19 while Rashdall thought Principia ‘brilliant’ (E 63n2). In The Right and the Good Ross said his main obligation was to Prichard, but he also mentioned Moore’s writings as ones from which, despite disagreements, he had ‘profited immensely’ (RG v–vi); certainly his conception of intrinsic value was close to Moore’s. Ross likewise admired Broad, whose account of prima facie duty he borrowed in Foundations of Ethics even though it differed from his own in The Right and the Good (FE 51–3, 79–82). Broad in turn wrote warm reviews both of Prichard, ‘a man of immense ability whom I have always regarded as the Oxford Moore’ (CE 14), and of Ross,20 while Ewing wrote extensively and respectfully of all three. It seems that whereas early members worked to some extent independently, later ones, while not quite seeing themselves as a distinctive school, recognized their commonalities and applauded and shaped each other’s work. Who were these philosophers? Before examining their ideas let us look briefly at their lives and characters. Henry Sidgwick (1838–1900), who is the subject of a recent major biography,21 was educated at Rugby, where he was friends with T.H. Green, and then at Cambridge, where he was much influenced by his membership in the Apostles. He later said the society’s Saturday evening debates seemed the most ‘real’ part of his life at Cambridge and called his membership the ‘strongest corporate bond I have known’; it also pointed him toward ‘the life of thought’ (M 34–5). After firsts in mathematics and classics he was elected to a fellowship at Trinity in 1859, but resigned it ten years 18 19 20

Broad, Critical Notice of Prichard, Moral Obligation, p. 557. Moore, Review of TGE; E 254 (in the Williams and Norgate 1912 edition). 21 Broad, Critical Notices of FE and Moral Obligation. Schultz, Henry Sidgwick.

10 british ethical theorists from sidgwick to ewing later because he could no longer subscribe sincerely to the Thirty-Nine Articles of the Church of England as was required. He explained his decision in a pamphlet titled The Ethics of Conformity and Subscription and later defended a similar view about the clergy in a journal article. Rashdall replied to this article, supporting a more relaxed policy largely on the ground that no one takes the words in a religious formula literally; here the utilitarian layman was stricter about truth-telling than the ordained perfectionist.22 Despite his resignation he continued to lecture at Cambridge, and he resumed his fellowship when the law that led him to renounce it was abolished; in 1883 he became Professor of Moral Philosophy. By then The Methods had made him an esteemed figure in philosophy, and he was also connected to the political and religious establishment through his wife Eleanor, whose brother was Arthur Balfour, later Conservative Prime Minister, and his sister, whose husband became Archbishop of Canterbury. But he never enjoyed the more widespread acclaim of his schoolmate Green, whose work he thought less careful than his own and whose success he somewhat resented. In his journal he wrote that a remark about seeing ‘the splendid zeal with which missionaries rush on to teach what they do not know’ ‘represents my relation to T.H.G. and his work’ (M 395). Though he was generally conservative on economic policy and voted Conservative, he was a liberal within the university. He promoted women’s education, first organizing lectures and examinations for women and then with his wife founding Newnham College, of which she was the second Principal and where the two of them lived from 1894. He also campaigned to modernize the Cambridge curriculum, including by abolishing a compulsory examination in Latin and Greek and introducing technical subjects such as engineering, but with only mixed success. His other main non-philosophical activity was the Society for Psychical Research, which he co-founded in 1882 and which sought empirical evidence for life after death by conducting seances and interviewing purported psychics such as Madame Blavatsky. These ‘ghostological’ investigations had a philosophical basis: because he thought only the existence of a God who rewards virtue in the afterlife can avoid a fundamental contradiction in our practical thought, he needed to know whether such a God exists. The later generation found his fussing about religion depressing; Keynes said Sidgwick ‘never did anything but wonder whether Christianity was true and prove that it wasn’t and hope that it was’.23 He was a lively conversationalist; in later years he grew fond of gardens and flowers and ‘used to walk about meditating in [the Newnham] garden, stroking his beard on the underside and holding it up against his mouth, which was a characteristic gesture of his’ (Broad, EHP 60). He suffered from sleeplessness and depression, the latter of which he largely hid from others, and in early 1900 was diagnosed with cancer. In May he gave his final philosophical presentation, a critique of Green’s 22 Sidgwick, Ethics of Conformity and ‘Ethics of Religious Conformity’; Rashdall, ‘Professor Sidgwick on the Ethics of Religious Conformity’. 23 Quoted in Harrod, Life of Keynes, pp. 116–17.

introduction

11

metaphysics to the Oxford Philosophical Society where he was ‘in brilliant form’ (EHP 61), and died in August of that year. Hastings Rashdall (1858–1924) was not only a philosopher but also a historian and a clergyman and theologian; he had three simultaneous careers. After a disappointing second in Greats at Oxford in 1881 he remained there for two years reading philosophy and theology and writing an essay on the medieval universities, for which he won the Chancellor’s prize. He later expanded that essay into his threevolume Universities of Europe in the Middle Ages of 1895, which is a standard historical work and still in print today. After briefly teaching outside Oxford he returned in 1888, as a fellow first of Hertford College and later of his undergraduate college, New. Ordained a priest in 1886, he would preach sermons in Oxford and elsewhere, often on theological subjects, where he had very liberal views. He denied that there will be retributive punishment in the afterlife, said an Anglican clergyman need not believe in the virgin birth, and denied that Jesus ever claimed to be divine. His major theological work was a set of Bampton Lectures on the atonement given in 1915 and published in 1919 as The Idea of Atonement in Christian Theology; he argued that Christ’s life reconciles sinning humanity with God only in that he is an exemplar by knowing whom we can improve our moral lives. When Warden Spooner of New College was asked about Christian Socialists in Oxford, he said there were only Rashdall and himself, ‘and I’m not very much of a Socialist, and Dr Rashdall isn’t very much of a Christian’. But Rashdall was very religiously minded and said that made his position difficult: ‘You see I am on the left wing of the Church and the right wing of the philosophers’.24 In metaphysics he was a ‘personal idealist’, believing reality is entirely spiritual but contains many individual souls rather than one Absolute Spirit; his view here was close to McTaggart’s. From 1885 on he published a series of articles on ethics and by 1897 was trying to expand them into a book. Unfortunately the process took ten years, so The Theory of Good and Evil did not appear until four years after Moore’s Principia, which took a similar ideal consequentialist line. At first Rashdall’s book received considerable attention, but over time Moore’s came to be seen as the canonical presentation of their shared theory even though on many topics Rashdall’s discussion was fuller. He was a lively writer, with a gift for striking examples. He was also a vigorous controversialist, as shown in his aggressive attacks on Bradley and an entertaining discussion of the German thinker Georg Simmel (TGE II 103–6); Joseph attributed his combativeness in part to a chivalrous desire to help those he thought in the minority, such as liberals in the Church.25 He was popular with students studying for exams because he would lay out a case in numbered arguments, but apparently took that trait into everyday life: he was once overheard saying to his wife at the end of a walk, ‘Now thirteenthly, my dear’.26 He was also stereotypically absent-minded, forgetting 24 26

Matheson, Life of Rashdall, p. 78. Blanshard, ‘Autobiography’, p. 29.

25

Matheson, Life of Rashdall, p. 253.

12 british ethical theorists from sidgwick to ewing appointments, inept at sports—he played field hockey with thirty-four other dons and said he was not even in the third eleven—and ignorant about machinery. Informed once that the back tyre of his bicycle was flat, he started to blow up the front one and, when alerted to his mistake, asked, ‘Don’t they communicate?’27 He was an enthusiastic cyclist, for whom ‘one of his ideas of heaven was riding down Boar’s Hill with his feet up and the view of Oxford before him’.28 After being passed over for an Oxford professorship in 1910 he began to spend more time on theological issues and in 1917 accepted a full-time ecclesiastical appointment as Dean of Carlisle. His health began to deteriorate in 1921 and he died in 1924. At birth John McTaggart Ellis McTaggart (1866–1925) had the last name ‘Ellis’, but his family added the second ‘McTaggart’ to secure an inheritance from his maternal uncle. He was precocious, becoming an atheist as a child and reading Kant at thirteen. After earning the only first in Moral Sciences at Cambridge in 1888 he was elected to a six-year Trinity Prize Fellowship in 1891; his dissertation for it was published as Studies in Hegelian Dialectic in 1896 and he became a lecturer at Trinity in 1897. Hegel was long an inspiration and also the subject of Studies in Hegelian Cosmology (1901), A Commentary on Hegel’s Logic (1910), and many articles. But McTaggart’s intellectual character, dedicated to clarity of expression and rigour in argument, was very different from Hegel’s, and Broad and Moore doubted that the interesting views he claimed to find in Hegel were really there. Broad said he ‘produced an extremely lively and fascinating rabbit from the Hegelian hat’, whereas other commentators produced ‘nothing but consumptive and gibbering chimeras’, and his achievement was all the more impressive since ‘the rabbit was, in all probability, never inside the hat, whilst the chimeras perhaps were’ (EHP 75; also Moore, ‘A’ 19). McTaggart found the other British Idealists aside from Bradley annoyingly high-minded, calling them ‘the sort of people who wanted to believe that they ate a good dinner only in order to strengthen themselves to appreciate Dante’ (EHP 79). He himself was an avid drinker and thought every undergraduate should get drunk at least once a year to prove to his tutor he was not a teetotaller. He presented his own metaphysical views, free of Hegelian trappings, in The Nature of Existence and is best known for his argument that time is unreal, first given in Mind in 190829 and distinguishing an A-series of times involving the concepts ‘past’, ‘present’, and ‘future’ and a B-series involving ‘earlier’ and ‘later.’ He thought reality is entirely spiritual, involving a plurality of individual minds each loving one or more others, and love was his highest value. It was reflected in his intense loyalty to institutions, especially his public school, Trinity, and England; he was a fierce nationalist in the First World War and, less admirably, helped secure Russell’s expulsion from Trinity in 1916 for his pacifism. He was gay, and read an influential paper to the Apostles defending homosexual love, under the title ‘Violets or Orange 27 29

Matheson, Life of Rashdall, pp. 61, 81. McTaggart, ‘Unreality of Time’.

28

Carritt, Fifty Years a Don, p. 30.

introduction

13

Blossom?’ When he surprised his friends by returning from a trip to New Zealand with a wife, he told them she shared his interests in metaphysical discussion and schoolboys. Moore called him a ‘very strange and fascinating personality’.30 He had a distinctive crab-like walk, moving sideways and often keeping a wall behind him; some thought he had developed this practice to avoid being kicked by bullies at school.31 He rode a custom-built tricycle around Cambridge, because he had become too fat to manage a bicycle.32 He would entertain students at his home for breakfast or lunch but sometimes became so lost in philosophical thought he forgot to serve any food. Though he was an atheist he strongly supported the Church of England and opposed its disestablishment; he loved ritual and insisted on strict attention to detail in Cambridge ceremonies. A conservative in national politics, he was a liberal inside the university and went beyond Sidgwick in supporting full membership for women. Moore and Broad both praised his writing, the latter calling him a master of English philosophical prose alongside Hobbes, Berkeley, and Hume (EHP 71).33 But his metaphysical system-building was already out of fashion by the first decade of the new century, when Broad said he had many admirers but almost no disciples (‘A’ 50).34 He reached retirement age in 1923 but was still lecturing and writing when he fell ill and died in 1925. At his request, the memorial brass for him in Trinity chapel, which is beside Sidgwick’s, contains no specifically Christian references. After excelling in mathematics and classics H.A. (Harold) Prichard (1871–1947) was elected an Oxford college fellow in 1895. He was a devoted tutor, often extending his students’ one hour per week to two, three, or more and discussing their essays with them line-by-line. The strain this involved contributed to a breakdown in his health that led him to resign his fellowship in 1924; in 1928 he was elected to a Professorship of Moral Philosophy that was less demanding. In ethics he published just two journal articles, ‘Mistake’ in 1912 and one on Aristotle in 1935, and two lectures; in epistemology just Kant’s Theory of Knowledge (1909), described as ‘a very good book about Prichard’s theory of knowledge, but not such a good one about Kant’s’,35 and a handful of articles. As a result the two posthumously published collections of his writings, Moral Obligation and Knowledge and Perception, contained many items that had not before appeared in print. In later years he worked on a book about ethics that Carritt thought would rank with The Methods of Ethics 36 but he never finished it; two partial drafts appear in his Moral Writings as ‘Manuscript on Morals’ and ‘Moral Obligation’. Despite his limited output he was an active and influential philosopher and often conducted philosophical discussions by correspondence. Many of his letters survive, partly because he would ask his

30 32 34 36

Moore, ‘Death of Dr. McTaggart’, p. 271. Geach, Truth, Love and Immortality, p. 12. Also Woolf, Sowing, p. 133. Blanshard, ‘Autobiography’, p. 61.

31 33 35

Geach, Truth, Love and Immortality, p. 10. Also Moore, ‘Death of Dr. McTaggart’, p. 271. Price, ‘Harold Arthur Prichard’, p. 343.

14

british ethical theorists from sidgwick to ewing

correspondents to return his original letter to him with their reply, so he could consult it when replying back to them. Ross said his letters ‘often exhibit the firmness and subtlety of his thought as forcibly as anything that he published’;37 they also show he often changed his mind. Having long held views like those in Ross’s The Right and the Good, he began to see flaws in them after that book was published and to develop an alternative. He was a deep thinker, often giving our school’s distinctive views their most forceful statement, but he could also tie himself in knots about what seem trivialities. Thus he denied that the ground of your obligation to do an act can be the fact that, if you did it, it would have certain properties, arguing that from a hypothetical premise only a hypothetical conclusion can follow; none of Ross, Broad, and Carritt could see the difficulty here.38 He was combative in philosophical debate and had a not undeserved reputation for dogmatism. In 1933 a committee he was on decided that none of the candidates who had written the examinations for the John Locke Prize, including the young Austin and Ayer, deserved to win it. Ayer attributed this decision largely to Prichard, saying, ‘As quite often happens, even with good philosophers, he could see no merit in views that differed radically from his own’. Ayer also called the general tone of Oxford philosophy in the period of Prichard’s dominance ‘surly and unadventurous’ and Prichard himself ‘philosophically gifted, but narrow and dogmatic’.39 But Austin, despite not sharing Prichard’s views, admired ‘the single-mindedness and tautness of his arguments, and the ferocity and the total lack of respect for great names with which Prichard rejected obscurity and lack of consistency in philosophy, ancient and modern’.40 H.H. Price’s obituary of him says, ‘The ruling passion of his life was the desire to discover the truth about ultimate questions’,41 and while one expression of that can be intellectual modesty and an openness to differing opinions, another can be passionate commitment to what you think you have discovered is true. Prichard spoke in short staccato sentences, and his favourite retorts included ‘I don’t know what is true, but whatever is, that isn’t’ and ‘Now you are wrapping it up in cotton-wool’.42 His prose was direct and free of jargon but often hard to read because of the density of its argumentation. He was short, wiry, and athletic and had played tennis for Oxford as an undergraduate. He was a lifelong and enthusiastic golfer, and his golf game was said to be like his philosophy: ‘his shots were sometimes short, but they were always straight’.43 He reached retirement age from his Professorship in 1937 and died ten years later. G.E. (George Edward, though he hated those names) Moore (1873–1958) arrived in Cambridge in 1892 and in his first year met Russell, who with McTaggart’s support 37 Ross, ‘Preface’ to Prichard, Knowledge and Perception; also Price, Critical Notice of Knowledge and Perception, pp. 103–4. 38 Prichard, letters to Ross of 10.5.40 and to Broad of 4.3.43; Carritt, Fifty Years a Don, pp. 27–8. 39 40 Ayer, Part of My Life, pp. 152, 77–8. Berlin, ‘Austin and the Early Beginnings’, p. 2. 41 Price, ‘Harold Arthur Prichard’, p. 332. 42 Price, ‘Harold Arthur Prichard’, p. 333; Carritt, Fifty Years a Don, p. 26. 43 Price, ‘Harold Arthur Prichard’, p. 332.

introduction

15

had him elected to the Apostles and persuaded him to switch from classics to philosophy. To Russell he ‘fulfilled my ideal of genius. He was in those days beautiful and slim, with a look almost of inspiration, and with an intellect as deeply passionate as Spinoza’s. He had a kind of exquisite purity.’44 He soon became the dominant figure among the Apostles, as Sidgwick had been before. After an undergraduate first he applied in 1897 and 1898 for the same Prize Fellowship McTaggart had had, submitting two versions of a dissertation on Kant’s ethics. It was sent to referees and their reports on it survive; those by Sidgwick and Bosanquet are especially interesting, since they found similar strengths and weaknesses but differed in which they emphasized. A little disappointed by the dissertation, Sidgwick said its merit ‘seems to lie in promise rather than performance: but I judge it to be very promising’. But Bosanquet said that while promise may suffice for an undergraduate first, a fellowship candidate should have begun to turn his promise into performance, which Moore had not. Bosanquet objected especially to a part of the dissertation arguing that the objects of thought are independent of our thought and unaffected by it; he thought that view had been refuted by Idealism. He concluded, ‘if [this piece] had been sent me for review by “Mind” . . . I should have treated it respectfully as a brilliant essay by a very able writer, but should have endeavoured to point out that its positive stand-point and consequently its treatment of the subject were hopelessly inadequate’. Moore nonetheless won the Fellowship and in a way got back at Bosanquet. He cut the offending part out of the dissertation and submitted it (of course) to Mind, where it was published in 1899 as ‘The Nature of Judgment’.45 His fellowship ran for six years, and though he later accused himself of laziness in this period (‘A’ 24–5), he published numerous papers and reviews, wrote the lectures later published as The Elements of Ethics, and then revised and expanded them into Principia. This book was rapturously received by the younger Apostles who later formed the Bloomsbury group. Lytton Strachey told Moore his book had ‘wrecked and shattered all writers on Ethics from Aristotle and Christ to Herbert Spencer and Mr Bradley . . . I date from Oct. 1903 the beginning of the Age of Reason’.46 Keynes called it ‘exciting, exhilarating, the beginning of a renaissance, the opening of a new heaven on a new earth’, and said Moore’s account of the good became the group’s ‘religion’. Looking back, he thought they had ignored Moore’s concern for the effects of acts on other people, instead pursuing the good just in their private lives. But through the Bloomsbury circle Principia influenced much of English art and culture.47 In 1904 Moore applied for a follow-up Research Fellowship at Trinity but was turned down, in part because of another negative report from Bosanquet. He then lived outside Cambridge for

44

Russell, Autobiography, vol. 1, p. 64. On the dissertation and referees’ reports see Regan, Bloomsbury’s Prophet, pp. 68–70, 99–103; the documents are in The Moore Papers, Trinity College Library, Cambridge. 46 Letter from Strachey to Moore of 11.10.1903, The Moore Papers, Cambridge University Library. 47 Keynes, ‘My Early Beliefs’, pp. 52–5. 45

16

british ethical theorists from sidgwick to ewing

seven years on a private income before returning as a lecturer in 1911; given his stature in twentieth-century philosophy, it is startling to learn that he did not have a permanent academic post until he was thirty-eight. Once in it, however, he was a dedicated lecturer, taking no sabbaticals in the next twenty-eight years and writing all his lectures afresh each year. In 1925 he became Professor of Philosophy and from 1921 to 1944 was the editor of Mind, a job he handled on his own; he confessed later that he sometimes accepted an article from a famous writer despite thinking its quality below what he would require from someone unknown (‘A’ 36). Both his philosophy and his writing changed from his earlier to his later years. Principia had great confidence in its judgements and, reflecting that, was brisk and forceful in its prose, but his later style, with its labourious repetitions and qualifications, was that of a man for whom deciding had become difficult. R.B. Braithwaite said that although Moore’s principal interest from around 1910 on was the philosophy of perception, he never published a book on it because ‘he saw the reasons against any view so clearly that he could never make up his mind which was on the whole the most defensible’.48 There are many testimonials to his character. Leonard Woolf thought him the only great man he had met or known, saying he ‘resembled Socrates in possessing a profound simplicity’,49 and he was often called innocent or childlike, for example by Wittgenstein.50 His modesty shines through his ‘Autobiography’ and is reflected in his travelling to Norway in 1914 to act in effect as secretary to Wittgenstein, whom he thought cleverer and more profound than himself (‘A’ 33), and in his taking and then publishing notes on Wittgenstein’s lectures of 1930–3.51 Many noted what Braithwaite called his ‘single-minded and passionate devotion to the search for truth’.52 He was a constant pipe-smoker, despite numerous efforts to quit, and entertained his friends by singing Schubert’s and other Lieder while accompanying himself on the piano. He reached the retirement age for his Chair in 1939 and spent 1940–4 in the United States, away from the dangers of war. When he returned to Cambridge he was too ill to attend philosophy papers or appear often in public; he died in 1958 just before his eighty-fifth birthday. E.F. (Edgar) Carritt (1876–1964) was a fellow of University College, Oxford from 1899 until after the Second World War; his 1960 typescript memoir Fifty Years a Don is in the college library. As an undergraduate at Hertford College he dined opposite a portrait of Hobbes, and ‘the study of his face convinced me that mischief, not vagueness, produced his famous ambiguity: “It is a Law of Nature that Men keep their covenants made” ’.53 He modestly attributed his first in Greats to the good luck of getting a question on Plato’s aesthetics, which he had been studying, and aesthetics remained a major philosophical interest. His first publication was an article on the

48 50 52

Braithwaite, ‘George Edward Moore’, p. 298. Malcolm, Wittgenstein: A Memoir, p. 80. Braithwaite, ‘George Edward Moore’, p. 305.

49 51 53

Woolf, Sowing, pp. 131, 137. Moore, ‘Wittgenstein’s Lectures in 1930–33’. Carritt, Fifty Years a Don, p. 6.

introduction

17

sublime,54 and he later wrote The Theory of Beauty (1914), What is Beauty? (1932), and An Introduction to Aesthetics (1949); he also gave what Bradley told him were the first lectures at Oxford on aesthetics not about Aristotle’s Poetics. His aesthetic views were influenced—some thought overly so—by Benedetto Croce’s and took a similar ‘expressive’ line. After taking up his fellowship he attended the Philosophical Tea, where ‘perhaps the most stimulating dialectics were those between Prichard, Joseph, and Professor J.A. Smith, the first a convert to Realism, the last an Idealist and Joseph a cross-bencher’. He ‘sometimes thought the controversies of Prichard and Joseph—for instance on the meaning of “whole” or “same”—would have made a popular music-hall turn’.55 He was known for an extremely compressed prose style; in a letter to Ross, Prichard said, ‘I expect I have carried brevity beyond even the extent to which Carritt carries it’.56 Brand Blanshard thought Carritt’s writing was influenced by the idea that in a perfect work of art there are no excrescences and everything contributes to the whole; he found the result at times elliptical and even crabbed.57 Carritt’s obituarist said his counterexamples were often devastating, ‘but many readers will fail to notice it because he does not waste unnecessary words hammering a nail home. For him it was enough to tap it once in the right place. If he could do so with light irony, so much the better.’58 Certainly his ethical writings often make novel points so briskly one can easily miss them. In politics he was a socialist, according to Blanshard ‘on moral grounds’, and the most left-wing member of the school. He was not a communist but his wife and five sons were; one became an editor of The Daily Worker and another died in the Spanish Civil War. Carritt participated in the ‘pink lunch’, a gathering of left-wing dons at his college that heard visiting speakers. He also gave the first Oxford lectures on dialectical materialism; an older philosopher asked him why he had given his lectures so queer a title, which no one had heard of. But his interest in this topic was characteristic. He thought that just as a true aesthetic theory should not change our particular judgements of beauty, so philosophy in general should not affect our conduct—and Marx’s metaphysical theory claimed to do so. He was a fine teacher and also physically energetic. As an undergraduate he rowed stroke for his college boat club and he continued to enjoy exercise throughout his life. When visiting with Blanshard for a year at the University of Michigan he ran around the block every morning before breakfast, and when the two came to a river on a walk would strip and jump in. To Blanshard he was ‘a new type of man. Comfort, clothes, and money meant little to him; beauty in art and nature meant much; and duty meant most of all.’ He continued writing well past his retirement and died in 1964. Alongside his work in ethics, W.D. (later Sir David) Ross (1877–1971) was also and even primarily a scholar of ancient philosophy, as distinguished as any in the 54 56 58

Carritt, ‘The Sublime’. Prichard, Letter to Ross of 11.12.36. Raphael, ‘Edgar Frederick Carritt’, p. 447.

55 57

Carritt, Fifty Years a Don, p. 26. Blanshard, ‘Autobiography’, p. 72.

18

british ethical theorists from sidgwick to ewing

twentieth century. Elected a fellow of Oriel College in 1902, he soon after became editor, at first jointly and then alone, of the Oxford translations of Aristotle, which appeared between 1908 and 1931. He himself contributed translations of the Metaphysics and Nicomachean Ethics, arguably the two most important Aristotelian works, and gave editorial guidance to the others. He prepared Oxford Classical Texts of six works of Aristotle and did complete editions, with introductions and commentaries, of five; his Aristotle of 1923 is a standard work, though more expository than critical, and Plato’s Theory of Ideas appeared in 1951. In a hugely successful academic career he was Provost of Oriel from 1929, president of the British Academy from 1936 to 1940, and vice-chancellor of Oxford University from 1941 to 1944. While he was Academy president many Jewish scholars fled central Europe; he did much to help them, both officially and unofficially, and welcomed several into his home. During the First World War he worked in the Ministry of Munitions and so impressed officials that he served on a succession of government boards over the next thirty years, often concerned with setting minimum or other wages. From 1947 to 1949 he chaired a Royal Commission on the Press, which recommended the creation of a Press Council that then governed the industry until the phone-hacking scandal of 2011. Largely because of his government work he was made OBE in 1918 and knighted in 1938. From early in his career he lectured on ethics, and when the Professor of Moral Philosophy fell ill in 1923 he was chosen to fill in as Deputy, which he did until 1928. It was during this period that he did his main work in ethics, publishing a series of journal articles that led to The Right and the Good. Raised a Scots Presbyterian, Ross was taciturn and even severe in character, disliking idle conversation and hating gossip; Ayer complained that he turned the Oriel Senior Common Room into ‘a stronghold of puritanism’.59 He did, however, enjoy playing charades. He was extremely conscientious. When the Moral Philosophy Chair fell open in 1928 he did not let his name stand; among his reasons were that a Deputy may be unfairly favoured by a selection committee and therefore should not in general stand, and that Prichard was a better moral philosopher than he and a better philosopher generally.60 In the event Prichard was selected. Ross’s many successes rested on talents for hard work, quick decision, and concentration on the task at hand, the latter allowing him to move easily from one task to another. A former student described visiting him at Oriel and finding him deep in writing the lectures that became Foundations of Ethics. The student was welcomed and his issues dealt with quickly and courteously, but as he left the room he could hear Ross’s pen racing across the paper again before he reached the door.61 In all his work, from his Aristotle commentaries to the government boards, Ross showed sound judgement and good sense. It was said of him, ‘He is not only an Aristotelian scholar, but he also has an Aristotelian frame of mind—moderate, critical, balanced, thorough, and above all, 59 61

Ayer, Part of My Life, p. 308. Blanshard, ‘Autobiography’, p. 60.

60

Clark, ‘Sir David Ross’, p. 534.

introduction

19

judicious’.62 A tall man, he held himself with natural dignity and played golf and tennis regularly. After enjoying robust good health he died in 1971, aged 94. As a Cambridge undergraduate C.D. (Charlie Dunbar) Broad (1887–1971) first studied natural science but in his last two years switched to philosophy. After an especially high first, he in 1911 won the six-year fellowship both McTaggart and Moore had had; his dissertation for it was published as Perception, Physics, and Reality in 1914. Before the fellowship was announced, however, he had accepted a teaching job at St Andrews. He decided to keep both positions, teaching most of the year in Scotland and spending a few months each summer at Trinity. This gave him two incomes, and with his Trinity funds ‘began that course of saving and investment which has been one of my main sources of interest and satisfaction in life’ (‘A’ 52). During the First World War he avoided military service, of which he was terrified, by doing chemistry research in Scotland for the Ministry of Munitions, and in 1922 returned to Trinity to replace McTaggart as lecturer in philosophy. He would write his lectures out in full in advance, and while lecturing read each sentence out twice; his lectures were the origin of all his later books, including Scientific Thought (1923), The Mind and Its Place in Nature (1925), Five Types (1930), and the twovolume Examination of McTaggart’s Philosophy (1933 and 1938). In 1933 he was appointed to the Professorship Sidgwick had held, but did not enjoy the position. It involved teaching graduate students, whom he liked less and the value of whose research he questioned; in addition, he ‘no longer believed in the importance of philosophy’ and ‘took little interest in its later developments’, thinking he had ‘shot [his] bolt’ (‘A’ 60–1). Yet he continued to produce incisive papers and reviews. From 1920 on he belonged to Sidgwick’s Society for Psychical Research, serving as its President in 1935–6. He did not have Sidgwick’s motive for involvement, hoping in fact that there is no afterlife, but thought the issue important enough to be tackled with the best, empirical, methods. He was utterly unathletic, unable to dance, swim, play any sport, or drive a car; his main outside interest was building elaborate model railways, which he did in some friends’ garden while a lecturer in Cambridge. He lived in rooms in Trinity—the same ones Newton had had—and ate in college. He was gay, though less flamboyantly so than McTaggart, and ended his ‘Autobiography’ by mentioning his ability ‘to make friends with the kind of young men whom I like and admire, despite great disparity in age’, saying he ‘derived more happiness from this than from any one other source’ (‘A’ 68). In politics he mixed distrust of democracy, Plato’s objections to which he thought conclusive, with distrust of a pure market economy, for destroying natural beauty and exploiting wage earners. By the 1950s, however, he thought the political balance had swung too far toward labour and was voting Conservative: ‘I cannot imagine myself at home in that collection of bone-heads unequally yoked to eggheads and decorated with a broad lunatic fringe,

62

Blanshard, ‘Autobiography, pp. 60–1. The quoted remark is from A.K. Stout.

20

british ethical theorists from sidgwick to ewing

which is the British Labour Party’. As for communism, ‘if nonsense imposed by violence attracted me, I would prefer the old vatted nonsense of the Roman Church to the thin pseudo-scientific vinegar provided by the Jesuits-without-Jesus of Moscow’ (‘A’ 47). As these remarks show, he was a brilliant writer, especially when expressing vitriol, and the chapters to follow will often quote him at length. (The index to Five Types must be the funniest in philosophy.) But he was also self-aware and highly self-critical. He did not overrate his philosophical contributions, and in his autobiography attributed to himself vices of laziness, physical and moral cowardice, and lack of resolution, concluding, ‘I am not the kind of person whom I like’ (‘A’ 66–8). In 1953 he gladly reached retirement age, wanting no longer to ‘occupy the ambiguous position of an unbelieving Pope, or the invidious one of the veteran who lags superfluous on the stage’ (‘A’ 65). In 1971 he wrote that experience on Cambridge committees had taught him ‘the desirability of retiring before one has become too “ga-ga” to realize just how “ga-ga” one is becoming’ (CE 15) and ceased to publish. He died the next year aged eighty-four. A.C. (Alfred) Ewing (1899–1973) was the youngest member of the school, born twelve years after Broad and a year before Sidgwick’s death. He earned a first in Greats at Oxford in 1920 and a DPhil, then not a common degree, in 1923, publishing his thesis for it as Kant’s Treatment of Causality. He won the John Locke Prize in 1921 and the Green Prize in 1926 but he did not win an Oxford college fellowship, and the reason may in part be what his obituarist G.R. Grice called ‘certain defects of personality’ that made him ‘not cut a good public figure’ and hampered him throughout his career.63 He spoke with a lisp and in a northern accent; the way he said ‘synthetic a priori’ could make a philosophical audience burst out laughing. He was socially awkward and dressed chaotically, often with a sweater over a waistcoat, and in old-fashioned clothes, for example as late as the 1950s in shoes that buttoned rather than laced up. A muscular problem gave him an unusual walk and meant he could not do much with his hands.64 After short stints teaching in Michigan, Newcastle, and Wales he moved to Cambridge in 1931 as a Lecturer in philosophy, becoming Reader in 1951 but not being elected to a college fellowship until 1962. Ayer, who despite holding very different views respected him as a philosopher, thought he was badly treated by Cambridge, largely because he never took up the new ideas foisted on the university by Wittgenstein.65 (I am also told his lectures were very dull.) He was a prolific writer. Alongside four books and many articles on ethics he published Idealism: A Critical Survey in 1934, A Short Commentary on Kant’s Critique of Pure Reason in 1938, The Fundamental Questions of Philosophy in 1951, and many papers on metaphysics, epistemology, and the philosophy of

Grice, ‘Alfred Cyril Ewing’, pp. 501, 510–11. This combination of traits makes one wonder whether he had an autism-spectrum disorder such as Asperger’s Syndrome. 65 Ayer, Part of My Life, p. 15. 63 64

introduction

21

religion. He was a trenchant early critic of logical positivism, being the first to notice that its verifiability criterion of meaning is itself neither analytic nor empirically verifiable and making telling objections to the claim that necessary truths are true by convention.66 He enjoyed travel, attending many international conferences and holding late in his career several visiting professorships in the US. He was modest and unworldly, quietly accepting his lesser philosophical standing; he never married but lived in Cambridge with his mother. Teasing him about his religious beliefs, Ayer once asked what he most looked forward to in the next world; he replied, ‘God will tell me whether there are synthetic a priori propositions’.67 Apart from philosophy his great interest was hill-walking, and he could out-walk people far younger than himself. He seemed oblivious to physical discomfort, not noticing even in pouring rain that his raincoat had slid halfway down his back. He once showed up at his Rambling Club with a new walking stick—a white one. He had no idea that a white cane meant anything special and spent the weekend trying to peel the paint off. Grice’s fine obituary of him is poignant, describing a man whose work was not appreciated at its true worth because of a change in philosophical fashion—and the arrogance of those who made the change—and irrelevant facts about his personality. It concludes by wondering ‘whether his work will have, at some time in the future, the attention which it should have had in the past’.68 Grice would be gratified to know, as would Ewing, that as parts of moral philosophy return to views like Ewing’s his contributions are becoming better known. Ewing retired from Cambridge in 1966 and moved to Manchester but continued to write; Value and Reality, giving a morally-based argument for the existence of God, appeared just after his death in 1973. Its publication came ninety-nine years after the first edition of The Methods of Ethics, supplying a little-noticed coda to a century of distinctive writing in ethics.

66 67

Ewing, ‘Meaninglessness’ (1937), ‘Linguistic Theory’ (1940). 68 Ayer, Part of My Life, p. 15. Grice, ‘Alfred Cyril Ewing’, p. 513.

1 Minimal Concepts The school agreed, first, on a view of the normative concepts that I will call conceptual minimalism. It says there are not a great many irreducible normative concepts, as some hold, but only a small number, in terms of which all normative judgements can be expressed. They also agreed on the leading candidates for this role, most centrally ‘good’ in the sense of ‘intrinsically good’ on the one side and ‘ought’, ‘right’, and ‘duty’ on the other; any other apparently normative concepts are either reducible to these few or not truly normative. This led them sometimes to analyse a non-basic normative concept in terms of a basic one and sometimes to argue that what may seem a normative concept is in fact merely descriptive. Because of this initial agreement, the school’s normative debates took place on common ground. However much they differed about what is in fact good or right, they expressed their views in the same language and using the same concepts. This made their disagreements mainly substantive: they were not debating which are the best concepts for expressing a normative view but which claims using a shared set of concepts are true. This chapter will examine their conceptual minimalism, contrast it with views that recognize a greater number of irreducible concepts, and distinguish the versions of it proposed by different members. It will also argue that minimalism makes for the clearest and most illuminating approach to normative ethics.

1.1 Thin Concepts The most radically minimalist views hold that there is just one basic normative concept. This was in different forms the view of Sidgwick, the early Moore, and, more tentatively, Broad and the middle Ewing. Sidgwick thought the one irreducible concept is that of what a person ought, has a duty, or, as some today would say, has a reason, to do. This concept was for him ‘too elementary to admit of any formal definition’ (ME 32) and implicit in all our normative judgements. Thus to call something good was, for Sidgwick, to say that people ought or have reason to desire and pursue it (ME 92n, 112, 381, 388; EEM 27; GSM 331). In Principia Moore likewise recognized just one basic concept, in his case that of what is intrinsically good. It too was ‘simple and indefinable’ (PE 15) and present in

minimal concepts

23

all other normative concepts; thus Moore defined the right as what will result in the most good possible (PE 24–6, 146–8, 167–8) and gave similarly reductive analyses of virtue, the appropriateness of emotions, and beauty (PE 171–2, 190, 201–2). Broad and Ewing proposed a different basic concept, that of the fitting or appropriate. In Five Types Broad said that to call an act right is to say it is appropriately or fittingly related to the agent’s situation (FT 164–5; CE 76),1 while Ewing gave a similar account of ‘ought’ (‘SN’ 4, 14; DG 132–3,135–7), and both proposed analysing the good as what it is fitting to desire or have a pro-attitude toward (FT 283; ‘SN’ 6–9, 13–14; DG 145–50, 152–3). Ewing discussed a further concept of moral obligation, connected to ideas about praise and blame, but in his middle works thought it too might be analysed in terms of fittingness, so you have a moral obligation to do an act if it is fitting for you to do it and not doing it makes you a fitting object of moral disapproval. Then all normative concepts would be reduced to the one concept of fittingness (‘SN’ 14; DG 168–70). Others in the school recognized two basic concepts. Shortly after Principia Moore abandoned its analysis of ‘right’ and came to treat this concept too as irreducible. (He still thought right acts maximize the good, but this was now a synthetic rather than analytic truth (E 73, 77).) Prichard and Ross too saw ‘right’ and ‘good’ as distinct, with Ross arguing against both a reduction of ‘right’ to ‘good’ and one of ‘good’ to ‘right’ (RG 7–11, 104–6, 111; FE 42, 278–83). The later Ewing likewise admitted two basic concepts. While still reducing ‘good’ to fittingness, he now took moral obligation too to be irreducible (ST 95–6). Despite their differences, these proposed basic concepts are all what are today called ‘thin’ normative concepts, and a first facet of the school’s minimalism was their belief that any ‘thick’ concepts can be reduced to thin ones. Since normative properties supervene on non-normative ones, anything right or good has non-normative properties that make it so, such as producing or involving pleasure. What makes ‘right’ and ‘good’ thin concepts is that they give no indication what these properties are; it is no part of the meaning of ‘x is right’, for example, that x will produce pleasure or has any other specific descriptive property. By contrast, thick concepts such as ‘courageous’, ‘generous’, and ‘malicious’ do have descriptive content: a courageous act must involve facing danger, a generous one be done from concern for others, and so on. Some philosophers argue that thick concepts, especially those connected to the virtues, are both irreducible to thin ones and central to sound ethical thought. We should ask not what is right or good, they say, but what would be courageous, generous, or malicious, concepts that blend normative and non-normative elements in an indissoluble whole.2

1 Broad denied that he was giving an ‘analytical definition’ of rightness, since rightness is just one kind of fittingness and its difference from other kinds is ‘specific and unanalysable’ (FT 165, CE 76). 2 McDowell, ‘Virtue and Reason’, and ‘Non-Cognitivism and Rule-Following’; Williams, Ethics and the Limits of Philosophy, pp. 130, 141–2; Dancy, ‘In Defense of Thick Concepts’.

24

british ethical theorists from sidgwick to ewing

Our school rarely addressed this view explicitly, more often just asserting that the basic concepts are thin. Thus Ewing said the two main ethical concepts are expressed by ‘good’ and ‘ought’ (E 5; also ‘SN’ 2), and Ross titled his book not The Courageous, Generous, Kindly, Dainty, and Dumpy but The Right and the Good.3 The closest engagement with the ‘thick’ view came in Sidgwick’s critique of what he called ‘aesthetic intuitionism’, which sees the virtues as excellences whose ‘nature does not admit of being stated in definite formulae’. He wrote: our notions of special virtues do not really become more independent by becoming more indefinite: they still contain, though perhaps more latently, the same reference to ‘Good’ or ‘Wellbeing’ as an ultimate standard. This appears clearly when we consider any virtue in relation to the cognate vice—or at least non-virtue—into which it tends to pass over when pushed to an extreme . . . Common Sense may seem to regard Liberality, Frugality, Courage, Placability, as intrinsically desirable: but when we consider their relation respectively to Profusion, Meanness, Foolhardiness, Weakness, we find that Common Sense draws the line in each case not by immediate intuition, but by reference either to some definite maxim or duty, or to the general notion of ‘Good’ or Wellbeing. (ME 392; also EEM 261)

Compare an act of courage such as the Spartans’ self-sacrifice at Thermopylae with the foolhardiness of refusing a robber’s demand for ‘A nickel or your life’. If thick concepts were irreducible, there could be no explanation of the moral difference between these acts; we would have to just ‘see’ that one is courageous and the other not. But as Sidgwick noted, there is an obvious explanation. The Spartans’ act was courageous because the factors independent of courage that tended to make it right outweighed those that involved harm to them and tended to make it wrong; more specifically, the good it achieved, by preserving Greek civilization, was greater than the good it cost, their lives. But resisting the robber is foolhardy because the good it hopes to preserve is trivial compared to what it risks, and something similar holds for other virtue-vice pairs, such as generosity and excessive self-abnegation: you are generous when what you give another is sufficiently valuable to justify the cost to you and self-abnegating when it is not. Because the ‘thick’ view rules out these explanations, it is less illuminating than our school’s minimalism and less true to everyday moral thought, which does not see the difference between virtue and vice as ineffable but explains it in roughly the way described above. Sidgwick characterized the virtues using thin concepts and a causal relation. For him they were dispositions that tend to result in right acts; since he thought right acts promote the good, that made the virtues dispositions that tend to promote the good. Moore and Carritt had a similar view (PE 171–2; TM 137), but others analysed the virtues using thin concepts and an intentional relation. Rashdall equated virtue with the ‘consciousness of a man who loves and wills the truly or essentially good’, and contrasted this with ‘mere capacities or potentialities of pleasure-production such as 3

See also Rashdall, TGE I 93, E 5; Ross, RG 123, FE 3, 10–11, 166; Broad, FT 257.

minimal concepts

25

might be supposed to reside in a bottle of old port’ (TGE I 65). Similar accounts were given by McTaggart, Broad, Ewing, and especially Ross. He initially equated virtue with action or a disposition to act from any of three motives, ‘the desire to do one’s duty, the desire to bring into being something that is good, and the desire to give pleasure or save pain to others’ (RG 134), where the third is directed at something that is in fact good but without thinking of it as such. Later he extended his definition to include desires for and feelings about these objects regardless of whether they issue in action (FE 290–1), but his basic idea remained the same: virtue involves an intentional relation to something good or right. I will defend this account in Chapter 10, but it also helps answer an influential argument against the reducibility of thick concepts. Due originally to John McDowell, this argument says that if the descriptive and normative parts of a thick concept could be separated, its extension would be determined entirely by the descriptive part, so we could tell what does and does not fall under it without knowing its evaluative point. That, the argument says, is impossible.4 But no plausible reductive analysis will take the descriptive part of a thick concept to completely specify the properties that make objects falling under it good or right; it will only indicate a general area in which those properties are found. It is part of the concept of distributive justice, for example, that whether a distribution is just depends on some property it has as a distribution and not, say, on the motives of those who brought it about. But that leaves open whether the relevant property is that of being an equal distribution, one proportioned to merit, or something different. That must be determined normatively, as McDowell says. And this response to his argument is strengthened if the virtues are understood in Rashdall’s and Ross’s way. If generosity involves wanting and doing what is good for another person, then what is generous depends crucially on what is good. If knowledge as well as pleasure is good, then giving someone slightly painful knowledge can be generous; if it is not, it cannot. An aspect of moral thought McDowell says a reductive analysis of thick concepts cannot capture in fact follows directly from its most plausible forms.5

1.2 Deontic Concepts Second, the school did not draw irreducible distinctions between uses of the thin concepts: there was just one basic ‘ought’ and one normatively relevant valueconcept, intrinsic goodness. This again marks a contrast with much present-day ethics. Many writers today distinguish between senses of ‘ought’, in particular moral ‘oughts’ on the one side and rational or prudential ones on the other, and discuss at length how they relate. Is moral action also rationally required? Or is it sometimes 4 5

McDowell, ‘Non-Cognitivism and Rule-Following’, p. 144. See further Elstein and Hurka, ‘From Thick to Thin’.

26

british ethical theorists from sidgwick to ewing

irrational? Similar distinctions are made on the side of value. Some talk not only of intrinsic goodness but also or even exclusively of attributive goodness, as in ‘good knife’ or ‘good person’, or of goodness-for, as in the ‘good for a person’. Others distinguish ‘moral’ from ‘non-moral’ goodness and may recognize further forms of value such as the aesthetic. Our school, by contrast, mostly treated the normative uses of ‘ought’ and ‘good’ as univocal. The use of a single undifferentiated ‘ought’ began with Sidgwick. His ‘methods of ethics’ included not only utilitarianism and deontology, which all would count as ‘moral’ views, but also egoistic hedonism, which many would not. (For him ‘moral’ and ‘ethical’ were synonyms (OHE 11).) He defined a ‘method of ethics’ as ‘any rational procedure by which we determine what individual human beings “ought”— or what it is “right” for them—to do’ (ME 1). Since egoism does this, it is on a par with and uses the same concepts as uncontentiously moral views such as utilitarianism. Sidgwick often described his methods in the language of rationality, as in the above reference to ‘rational’ procedures or in his claims that they issue ‘dictates of reason’ (ME 34) or prescribe ‘rational ultimate ends’ (ME 9, 23, 94, 201, 497–8). But this meant only that their prescriptions, unlike those of a singular imperative, are universal, telling everyone how to act in a generic situation (ME 33–4). On occasion he distinguished ‘moral’ judgements from ones about ‘what “is right” or “ought to be done” in view of the agent’s private interest or happiness’, which he called ‘prudential’ (ME 25–6; also EEM 43, PE 130–1). But even then he said prudential judgements are ‘“moral” in a wider sense’ and use the same ‘ought’ (ME 25); he also called your own happiness ‘a manifest obligation . . . prescribed by reason “categorically” ’ (ME 7) and took a duty to seek it to be part of common-sense morality (ME 327–8). His occasional talk of a prudential ‘ought’ therefore treated it not as a distinct concept but as the same concept applied to a specific subject matter, namely your own happiness. And that there is just one ‘ought’ was vital for his ‘dualism of the practical reason’, the for him irresolvable conflict between egoism and utilitarianism that threatened the coherence of practical thought (ME 506–9). This could not be a dualism of different ‘oughts’, which if different would not conflict; it had to involve competing claims using the same ‘ought’. Others had the same view. Ross, following as so often Prichard, said there is no moral duty to pursue your own pleasure. There are moral duties to pursue everyone’s knowledge and virtue, including your own, and to pursue others’ pleasure, but no such duty concerning your pleasure (RG 21; FE 72–5, 129–30, 272–4, 277, 284; Prichard, MW 10n, 135, 171, 204). But he did not add, as many today would, that there is some other duty or ‘ought’, say a rational or prudential one, to pursue your pleasure. He assumed that where there is no moral duty there is no duty at all, and this assumption was shared by his critics. Carritt, Broad, and Ewing argued against him that there is a duty to pursue your own pleasure but were happy to call it ‘moral’. Broad said we ‘condemn morally’ someone who acts imprudently (CE 274–5), while

minimal concepts

27

Ewing said that preferring a smaller present pleasure to a larger future one ‘must be in some slight degree morally wrong’ (DG 161; also Carritt, EPT 113–15). Or consider their view of the question, ‘Why should I be moral?’ If there are different ‘oughts’, this question is perfectly intelligible: it can ask whether we ought rationally or prudentially to do what we ought morally to do. But our school unanimously rejected the question. In ‘Mistake’ Prichard famously argued that the only reason to do your moral duty is that it is your moral duty; moral obligation is in that sense underivative and the demand to justify it on some other ground is ‘illegitimate’ (MW 19–20). A key assumption of his argument was that there is only one ‘ought’. As he later put it: to refer to a certain group of actions as right actions is to imply that we already know that they are right, and therefore, since ‘right’ is after all only a synonym for ‘ought to be done’, that we already know that we ought to do them. Yet to ask ‘why ought we to do them?’ is to imply that we have still to be convinced, and therefore a fortiori do not as yet know that we ought to do them. (MW 119)

Moore said the question ‘Why should I do my duty?’ is ‘puzzling’, since it reduces to ‘Why is duty duty?’ (EE 17–18); Carritt wrote, ‘If any one ask us, “Why ought I to do these acts you call my duty?” the only answer is “Because they are your duty”’ (TM 29); and Ewing said the explanation ‘why we should do what we ought . . . is that the ought itself is a reason so that there is no further room for question left’ (DG 33; also MP 216; ST 17, 56). Sidgwick may seem to have disagreed, since he said we ask ‘Why should I do what I see to be right?’ but do not ask ‘Why should I believe what I see to be true?’ (ME 5). But his explanation was that we are often uncertain about what is right and express our uncertainty by asking the ‘why?’ question (ME 6); this implies that if we knew what is right there would be nothing to ask. All the school therefore rejected the ‘Why be moral?’ question, and did so in part because they recognized just one basic ‘ought’. I believe they were right to do so. One and perhaps the central task of normative ethics is to determine what we ought all things considered to do, in an unqualified sense of ‘ought’, and if there are distinct moral and prudential ‘oughts’ they are of interest only insofar as they bear on judgements using this unqualified one. If you ought morally to do some act, that is of interest only if it supports the conclusion that you ought unqualifiedly to do it; likewise if you ought prudentially to do some act. But then the distinct ‘oughts’ are dispensable. Rather than say, ‘You ought morally to do x, therefore you ought unqualifiedly, other things equal, to do it’, we can say simply, ‘You ought unqualifiedly, other things equal, to do x’. That is the normatively important claim and the only one we need to make. Given the one basic ‘ought’ we can, if we want, distinguish moral and prudential uses of it by their content, as Sidgwick sometimes did. Thus we can say an unqualified ‘ought’ telling you to keep your promises or promote others’ happiness is, given its content, moral, whereas one concerning your own happiness is prudential. But then morality and prudence are

28

british ethical theorists from sidgwick to ewing

only what Ross called ‘departmental’ concepts, which apply a more generic concept to a specific subject matter or in a specific domain (RG 102). The generic concept is what does the normative work. Our school did not just use a single ‘ought’; they called it, more specifically, the ‘moral’ ought. And some who grant a single deontic concept may resist this language, saying the basic concept can be used to make more than just moral claims. But the issue here is really just terminological, concerning what the one ‘ought’ should be called, and has no substantive implications. (Sidgwick’s book would have the same content if its title were The Methods of Rationality.) And the school did have grounds for their use of ‘moral’. Many and even most of the ‘oughts’ they recognized, for example about keeping promises and benefiting others, are ones everyone calls ‘moral’. This was especially true for Prichard and Ross, who thought what is commonly called the prudential duty to promote your own happiness is not a duty at all. More importantly, the school shared Kant’s view that moral ‘oughts’ are distinctively categorical, prescribing acts not as means to some end you may desire but regardless of your desires. Since those among them who recognized ‘oughts’ concerning your happiness thought they too are categorical—recall Sidgwick’s remark about ‘a manifest obligation’—they had reason to call them ‘moral’. It may be objected that the issue was not just terminological for Prichard and Ross. Their claim that there is no duty to promote your own pleasure is plausible if ‘duty’ is read as moral duty, since morality arguably concerns only acts that affect others. But it is not plausible if it denies any duty or ‘ought’ concerning your pleasure. They were tacitly relying on a restricted sense of ‘duty’ in advancing what they presented as an unqualified normative claim. I doubt this is the correct explanation of their view. Both Prichard and Ross affirmed a moral duty to promote your own virtue (MW 2, 216–17; RG 21), and Ross added a similar duty to promote your knowledge (RG 21); he also emphatically denied that morality concerns only acts that affect others (RG 153). Both therefore took morality to have self-regarding parts. They may have been influenced here by Kant, who likewise denied a duty to promote your happiness,6 though Ross found his argument for that denial unpersuasive (FE 278). They may also have been affected by religious backgrounds that made them see morality as stern and hostile to pleasure. This is reflected in Ross’s treatment of the duty to promote others’ pleasure, since he thought their pleasure has infinitely less value than their virtue and much less value than their knowledge (RG 150–2). Whatever its origin, though, their claim that there is no duty to seek your pleasure seems to have been substantive rather than based on any confusion about concepts. That said, I find the claim in the way they made it implausible, and will argue below that room must be made for permissibly seeking your pleasure. As in the

6

Kant, Doctrine of Virtue, pp. 44–7.

minimal concepts

29

views of Carritt, Broad, and Ewing, however, this will be moral room, created using generic concepts that may as well be called moral rather than some other kind. If moral ‘oughts’ are categorical, what about hypothetical imperatives, which prescribe acts as means to an end you may desire, or prescribe what many call instrumental rationality? Do they involve an ‘ought’ separate from the moral one and potentially in conflict with it? Here many in the school took either of the two lines mentioned at the start of this chapter: to reduce the hypothetical ‘ought’ to the moral one or to deny that it is really normative. The central issue was clearly put by Ross. On the simplest or ‘narrow-scope’ reading of the hypothetical imperative, it is a conditional with an imperative consequent, so if the best way for you to get rich is to murder your wealthy uncle, it says ‘If you want above all to get rich, murder your uncle’. But then if you do want above all to get rich, simple logic yields ‘Murder your uncle’, a categorical and therefore moral imperative to murder, which Ross found unacceptable: ‘No one really thinks that the fact that a person desires a certain end makes it obligatory on him to will the means to it; if we think the end is a bad one (or that his desiring it is bad), we think that in spite of his desiring the end he ought not to take the means’. What the hypothetical ‘ought’ says, then, is only that ‘a man who desires certain ends can hope to get them only if he adopts certain means’, so its content is descriptive and ‘the categorical imperative is the only true imperative’ (FE 47). This was also Prichard’s view. In ‘Mistake’ he referred briefly to Kant’s ‘wrongly describing his so-called “hypothetical imperatives” as imperatives’ (MW 9); later he said that by the hypothetical ‘I ought to do so and so’ we really mean just ‘If I do not do so and so, my purpose will not be realized’ (MW 166; also 34, 43, 54–5, 126–8, 143–4, 188, 237). For him as for Ross there are two utterly different uses of ‘ought’, one normative and categorical and the other hypothetical and merely descriptive.7 The alternative line was to reduce the hypothetical imperative to the categorical, as on the ‘wide-scope’ reading of the former.8 It treats the hypothetical imperative not as a conditional with an imperative consequent but as an imperative to make a conditional true, as in ‘Make it the case that (if you want above all to get rich, you murder your uncle)’, which is equivalent to ‘Make it the case that (either you do not want above all to get rich or you murder your uncle)’. This reading does not have Ross’s unwanted implication. From the psychological fact that you want above all to get rich it does not follow that you ought to murder your uncle; if doing that is wrong, what follows is that you should stop wanting above all to get rich. What the imperative commands, on the best version of this view, is a kind of consistency between your desires and actions given your non-moral beliefs: if you believe that some act will best further a given end, you should not both above all want the end and 7

See also Laird, SMT 41–2; Carritt, TM 12. Hare, ‘Wanting: Some Pitfalls’; Hill, Jr, ‘Hypothetical Imperative’; Greenspan, ‘Conditional Oughts and Hypothetical Imperatives’; Broome, ‘Normative Requirements’. 8

30

british ethical theorists from sidgwick to ewing

not do the act.9 And the imperative commands this consistency categorically, since its directive is not conditional on your having any specific desires but says you should be consistent whatever your desires. Sidgwick hinted at the wide-scope reading when he said the hypothetical imperative ‘implies the unreasonableness of adopting an end and refusing to adopt the means indispensable to its attainment’ (ME 37). This was not consistently his view, since he elsewhere said a hypothetical imperative is addressed only to someone who desires the relevant end (ME 6–7), which is true only on the narrow-scope reading. But Broad made a clear wide-scope claim when he said the hypothetical imperative expresses an ‘obligation of practical consistency’, or says ‘A person who intends a certain end ought either to cease intending it or to take the most efficient means open to him to attain it. He ought not both to go on intending it and to do acts which make it impossible for him to attain it’ (CE 361, also 362–3; FT 162–3).10 The wide-scope reading is very different from Prichard’s and Ross’s non-normative one, which Broad thought ‘quite plainly untrue’.11 But it is like theirs in denying that the hypothetical imperative involves a separate ‘ought’ from the categorical one; there is again just one basic ‘ought’. As well as not distinguishing moral from other ‘oughts’, most in the school did not see important differences between the concepts ‘ought’, ‘right’, ‘duty’, and ‘obligation’. (The later Ewing is an exception, with his distinction between the ‘ought’ of fittingness and moral obligation.) Moore, Ross, Broad, and Ewing said ‘right’ differs from ‘ought’ and ‘duty’ in that, whereas several acts possible for you can be right, for example because they will produce equal amounts of good, at most one can be your duty (E 14; RG 3–4, FE 43–4; FT 164; DG 134). Echoing Kant, Sidgwick, Broad, and Ewing thought ‘ought’ and ‘duty’, unlike ‘right’, apply only when you have motives tempting you not to do what is right, so a perfect being would not have duties (ME 34–5, 217; EEM 269; FT 139, 164, 182; CE 353; VR 208);12 Prichard and Ross rejected this view (MW 55; RG 159; FE 48). But neither of these distinctions is philosophically important, since each involves just a departmental restriction of a more basic deontic concept. Some philosophers distinguish obligations from duties, or both from mere ‘oughts’, on the ground that the former arise from voluntary acts such as promises and are owed to specific people.13 But this difference concerns the ground of an ‘ought’ rather than what results, and is again merely departmental; as Sidgwick said, it defines a species of duty rather than duty as such (ME 218). Finally, some take ‘duty’ and ‘obligation’ to be stronger terms than ‘ought’, expressing a more stringent

Way, ‘Defending the Wide-Scope Approach’. In this passage Broad denied that the obligation of practical consistency is a ‘specifically moral’ one but did make it a categorical imperative. 11 Broad, Critical Notice of Prichard, Moral Obligation, p. 558. 12 Moore also said ‘duty’ is commonly used only for acts ‘which a considerable number of individuals are strongly tempted to omit’, but thought this a superficial fact that does not affect the meaning of the term (PE 168–9). 13 Whiteley, ‘On Duties’; Hart, ‘Legal and Moral Obligation’; Rawls, Theory of Justice, p. 113. 9

10

minimal concepts

31

or demanding requirement, so for example, duties are acts that are right and whose omission it is right for society to condemn or punish. Carritt and Ewing had views like this (TM 104–5; DG 125) but Ross rejected them, saying any ‘ought’ is a duty (RG 52–4). Again, however, the views involve no fundamental conceptual distinction. A more stringent ‘ought’ is still an ‘ought’, which remains the one deontic concept. What of the concept of a moral right? Theories of natural rights were not common in this period, with Sidgwick listing only Herbert Spencer as an important defender.14 But several in the school were sympathetic to the idea of natural rights, and in particular rejected Green’s claim that all rights depend for their existence on legislation or recognition by society (Prichard, MW 226–36, 243–9; Carritt, TM 95–100; Ross, RG 51–2). They did so, however, because they thought talk of rights is equivalent to talk of moral duties and thought those are independent of society. Prichard said the claim that a person has a moral right says either that it is not wrong for him to perform some act or that another person has an obligation to treat him in some way (MW 228). He here recognized both what are today called ‘liberty-rights’, which do not imply duties for others, and ‘claim-rights’, which do, but others in the school discussed only claim-rights. For Sidgwick ‘ “a right” is really an obligation regarded from a different point of view: i.e. regarded in relation to the person to whom the obligation is intended to be useful’ (EP 27), while Carritt and Ross thought rights always imply duties (TM 95–107; RG 48, 56). Ross wondered whether some duties, in particular of beneficence, may not imply rights in the people they concern, because rights are things we can in decency claim, and it is indecent to claim another’s beneficence (RG 53). But he rejected this view, largely because the term ‘a right’ is vague. Having first used it for a legal notion, he argued, we have extended it to non-legal contexts but in a way that leaves its content unclear. Since ‘we have found the meaning of “rights” more doubtful than that of “duties”’, we should ignore the former and frame all issues in terms of duties (RG 53–4). A final issue about the deontic concepts concerns the school’s preference for talk of ‘oughts’ and ‘duties’ rather than of the ‘reasons’ to do an act that figure so much in current moral philosophy. They did sometimes use the latter language as an alternative. Sidgwick asked ‘What then do we commonly regard as valid ultimate reasons for acting or abstaining?’ (ME 78), while Ewing discussed cases where ‘a prima facie reason against an act is outweighed by reasons for doing it’ (ST 126; also Prichard, MW 217). And the two terminologies are in key respects interchangeable. The claim that you have ‘a reason’ to do an act parallels the claim that you ‘ought other things equal’, or in Ross’s language have a ‘prima facie duty’, to do it; that you have ‘most’ or ‘decisive reason’ to do an act parallels the claim that you ‘ought all things considered’ or have a ‘duty simpliciter’ to do it. The issue is therefore essentially terminological. But our school tended to prefer the language of ‘ought’

14

Sidgwick, Critical Notice of Ritchie, Natural Rights, p. 386.

32

british ethical theorists from sidgwick to ewing

and ‘duty’, especially when stating their views formally or explicitly discussing the normative concepts. And there are important advantages to doing so. The main one concerns the notorious ambiguity of the word ‘reason’, which can refer either to normative reasons to do an act or to explanatory reasons why an event occurred, which when the event is an act are, more specifically, motivating reasons, or psychological states such as beliefs and desires that caused the act. These two types of reason are radically different; normative reasons concern what you ought to do, motivating reasons what you did or will do. But it is easy to conflate them, and some philosophers have done so, causing confusions and unsound arguments.15 The ambiguity of ‘reason’ has given false support to the view that what you ought to do depends on what you desire, as if normative reasons had to be based on motivating ones. It has also encouraged the fallacious argument that if normative reasons are independent of desires, they are as on ‘objective’ views, the explanation of action need not mention desires. But it can be true that you ought to do some act you do not now desire even though you will do it only if you do desire it; Ross had this combination of views. The term ‘reason’ is also silent on a key issue dividing the concepts ‘ought’ and ‘good’: whether it is a necessary condition for having a reason to do x that you be able to do x. (Some reasons-theorists assume yes, others no.) And talk of reasons can lead to underestimates of the commitments of normative realism. Some philosophers hold that, while belief in underivative truths about ‘good’ or ‘right’ commits one to metaphysically suspect non-natural properties, realism about reasons does not; it is philosophically innocuous. But though another person’s being in pain, which is a normative reason for you to help him, is a natural fact, its being a reason is not: it is every bit as ‘unscientific’ as facts about what is good or right. All these dangers are avoided if the primary deontic term is ‘ought’. Then the view that you ought to do what will most satisfy your desires must clearly be defended by substantive argument rather than assumed on the basis of wordplay, and normative facts are clearly distinct from natural ones. It is not even clear that ‘normative reason’ is an independent concept. Normative reasons are often defined as factors that ‘count in favour’ of an act,16 but there are different kinds of favouring. Something can favour an act causally, by tending to make it happen, or epistemically, by supplying evidence that it will occur. Since normative reasons do not favour acts in either of these ways, how do they do so? The only answer I can think of is that they favour an act by tending to make it one you ought all things considered to do, but then the concept of a normative reason is not basic but is defined in terms of ‘ought’.17

15

For elaboration, see Smith, Moral Problem, Chapter 4. Scanlon, What We Owe, p. 17; Dancy, Ethics Without Principles, pp. 6, 29; Crisp, Reasons and the Good, pp. 38, 43. 17 For similar views see Gert, Brute Rationality, Chapter 4; Broome, ‘Reasons’. 16

minimal concepts

33

It may be said that it is a mistake to try to define the concept of a normative reason; it is a primitive, which we understand in its own terms. This view is similar to one that treats ‘ought other things equal’ as a primitive, which I will argue in Chapter 3 is less persuasive than one that defines ‘ought other things equal’ in terms of ‘ought simpliciter’. But even if a view like this is defensible, what is gained by stating it using the term ‘reason’? The issue is, as I said, essentially terminological, and ‘ought’ both avoids the ambiguities of ‘reason’ and settles issues ‘reason’ does not. As terminology, it has clear advantages.

1.3 Value-Concepts The school took a similar view of the value-concepts. Here their one basic concept was that of intrinsic goodness or goodness simpliciter, which they assumed is had by states of affairs; other putative value concepts are either reducible to this one or not really evaluative. The first or reductive approach is illustrated by their account of ‘moral’ goodness, had by motives and traits of character and contrasted with the ‘non-moral’ goodness of states such as pleasure and knowledge. They saw this not as a distinct valueproperty but as the same property of intrinsic goodness when found in a specific type of object. In the first edition of The Methods Sidgwick said that ‘good’ as applied to acts has ‘the specific meaning of “morally excellent”’, but this is only ‘a special application of the fundamental notion of “Good” = “intrinsically preferable or desirable” ’ (ME1 93n; also GSM 330). McTaggart, Rashdall and Moore made similar claims (SDR 152n; TGE I 174–5; EE 139), while Ross defined ‘morally good’ as ‘good by being a certain sort of character or by being related in one of certain definite ways to a certain sort of character’ (RG 155; also 166, FE 290–1), so it connotes only goodness with a distinctive ground. There were minor differences about the exact extension of ‘morally good’. Prichard and Carritt applied it only to the motive of duty, calling other admirable motives such as compassion virtuous and therefore good but not morally so (MW, 15–16, 55–6, 61–2, 160, 216, 218; TM, 47–8; EPT, 26, 85, 125–6); Ross used ‘morally good’ for all good motives (‘NM’, 254; RG 134, 160–3). But they all treated moral goodness as ordinary goodness when had by a certain kind of object, and this view has the same merit as the unifying treatment of moral and prudential ‘oughts’. Even if moral goodness were a separate property, it would be of interest mainly for its bearing on more global value-judgements using an unqualified ‘good’, as in Plato’s question whether a virtuous but painful life is on balance better than a vicious but pleasant one. And then what matters is its contribution to that unqualified judgement rather than any distinctive properties it has as ‘moral’.18 18

Ewing is a partial exception, since he distinguished moral from ordinary intrinsic goodness on the ground that what it makes fitting is admiration rather than choice of an object for its own sake (‘SN’

34

british ethical theorists from sidgwick to ewing

The alternative approach is illustrated by Ross’s account of attributive uses of ‘good’, as in ‘good knife’ and ‘good burglar’. He thought these make a merely descriptive claim about effective means to a given end, so a good knife is one that can be successfully used for cutting and a good burglar one who successfully burgles, with no evaluative content in either claim (RG 65–7; FE 255–7). It is natural to connect this account with his view of the hypothetical imperative. In an attributive use the sortal ‘knife’ or ‘burglar’ indicates an end commonly sought through or by items falling under it, such as cutting or burgling, and a good knife or burglar is one that effectively promotes that end. This account of the attributive ‘good’ had been suggested to Ross by Prichard in a letter of 192819 and was shared by Carritt, Broad, and Ewing (TM 66–7; FT 57–8, 202; ‘SN’ 5; DG 46–7, 104–6). Rather than reduce this use to the predicative ‘good’, it treats it as not normative. The same divide is found in accounts of beauty. Moore again gave a reductive analysis, defining the beautiful as ‘that of which the admiring contemplation is good in itself ’ (PE 201). By contrast, Ross and Carritt said what is true in a judgement of beauty is only that an object has the power of producing a certain experience in minds, namely aesthetic enjoyment or thrill, so what is true is only something descriptive (RG 89, 127, 131; ‘MP’ 141–4; EPT 30).20 More complex issues are raised by the concepts of ‘my (or anyone’s) good’ and of ‘a good to’ or ‘the good for’ a person, where in present-day philosophy the latter are often associated with a concept of ‘welfare’ or ‘well-being’ that is contrasted with intrinsic goodness. In fact, one criticism of at least later members of the school is that they wrongly concentrated their evaluative inquiries on what is simply good rather than on the equally or more important concept of what is good for people. Moore, characteristically, gave a reductive account of both concepts. The only way something can be ‘my good’ or ‘good for me’, he wrote, is if it is simply good and a state of me: ‘when I talk of a thing as “my own good” all that I can mean is that something which will be exclusively mine, as my own pleasure is mine . . . , is also good absolutely’ (PE 99, also 170). But this was also, in his main discussion of ‘good’, Sidgwick’s view. He there defined ‘my good’ and the ‘ultimate good on the whole for

11–13; DG 116, 153–5, 164–5). But he thought both types of goodness involve the appropriateness of some pro-attitude, and would have had to say which of Plato’s two lives calls for more of a pro-attitude on balance. 19 Prichard said the meaning of ‘good knife’ ‘can be rendered so far as I can see without using the word good at all . . . A term for some kind of instrument e.g. “a knife” of course implies that we or some one else does or may want a certain thing—“a knife”, say, implies simply that we do or may want some body to become divided smoothly, & to say “My knife is a good knife” means that I, or some one, by using it can bring about some smooth severance which I want, to a considerable degree’ (Letter to Ross of 20.12.1928). 20 Ross and Carritt thought we normally take our aesthetic judgements to ascribe to objects a property independent of anyone’s feelings, but held that since there is no such property, our judgements are false (RG 128n; ‘MP’ 132, 141). Ross’s analysis was therefore not of a concept we do use but of one we should use if we want our aesthetic claims to be true. Ewing saw beauty as involving a distinct kind of non-natural fittingness, as in the harmonious fitting together of the elements of a work of art (DG 172–3; ST 94–5).

minimal concepts

35

me’ as ‘what I should practically desire if my desires were in harmony with reason, assuming my own existence alone to be considered’ (ME 112). The reference to ‘my own existence’ was added in response to an objection of Rashdall’s. In an earlier edition of The Methods Sidgwick had defined ‘my good’ just as what I ‘ought to desire’ (ME3 106–7, 108), but Rashdall pointed out that this dissolves the dualism of the practical reason. If my good is what I ought to desire, and what I ought to desire is the happiness of everyone, then in maximizing the happiness of everyone I also maximize my own good, which makes egoism and utilitarianism always agree (‘PS’ 218–19). Sidgwick therefore added the ‘my existence’ phrase (ME4 112), and the result was an account very close to Moore’s. ‘My good’ is whatever I ought to desire and is located in me, or is that part of the good, however ‘good’ is understood, that I see when I look just at my life.21 This was a common view of ‘my good’, but some in the school gave a different account of ‘a good to’ and ‘good for’, seeing them as merely descriptive. Prichard said, ‘we mean by “a good to us” something which pleases, i.e. excites pleasure in us’; the difference between this meaning and that of ‘good’ when it refers to an indefinable quality of goodness is therefore ‘radical’ (MW 174, also 8, letter to Ross of 20.12.28). Carritt held that while ‘good’ sometimes means ‘intrinsically good’, in such phrases as ‘one’s own good’ and ‘do good to others’ it means ‘conducive to somebody’s advantage or satisfaction’, where the satisfaction is of desires and the meaning is again descriptive (‘AG’ 51–3, 55, 60). He and Prichard may have been influenced here by Kant, who in a passage Prichard cited (MW 175) distinguished two ‘very different’ value-concepts under the labels ‘das Gute’ and ‘das Wohl’. ‘Das Gute’ applies to the morally good will and even more to the summum bonum combining it with the happiness it deserves; it is intrinsic goodness in something close to Moore’s sense. But ‘das Wohl’ indicates ‘only a relation to our condition of pleasantness or unpleasantness, or enjoyment or pain’; it is ‘not a concept of reason but an empirical concept of an object of sensation’, that is, one applied on descriptive rather than normative grounds.22 Prichard’s account of ‘good for’ can be broadened. To accommodate talk of the ‘good for’ non-human organisms, it can include conduciveness not just to pleasure or satisfaction but to any end an organism naturally seeks, as a plant seeks growth and flowering. To allow the ‘good for’ artefacts such as knives, it can include conduciveness to an end people commonly use a thing for. Prichard’s account can also be combined with Moore’s to yield a mixed view on which what is good for a person must be simply good, be a state of him, and meet some descriptive condition such as 21 Sidgwick was reverting here to an earlier view of his, since the first edition of ME had said a person’s good is what ‘it would be reasonable for each individual to seek for himself, if he considered himself alone’ (ME1 360; also 374). But this claim did not come in the chapter on ‘Good’, which in the first two editions did not define ‘my good’; when Sidgwick started to define it there, in ME3, he initially omitted the ‘himself alone’ phrase. 22 Kant, Critique of Practical Reason, pp. 168–9.

36

british ethical theorists from sidgwick to ewing

satisfying his desires. Russell gave accounts of this kind of two uses of ‘my good’, saying they refer to ‘any good that I desire’ and ‘my pleasure, or any state of mind in me which is good’ (‘EE’ 45–6). But whether wholly or partly descriptive, all these accounts support a similar conclusion to Moore’s: that there is no value-concept that is both ethically important and separate from intrinsic goodness. It may be objected that, whatever was true of later writers, this was not Sidgwick’s view. In discussing egoism he explicitly allowed a distinct ‘good for’ concept, saying that although an egoist can be refuted if he says his own pleasure is simply good, he is unassailable if he says only that his pleasure is good ‘for him’ (ME 420–1). Moore rejected this concept in the course of arguing against Sidgwick that egoism is incoherent (PE 96–102), but it was the concept of welfare or well-being and Sidgwick was right to see its importance.23 This objection rests on an anachronistic reading of Sidgwick, who no more than Moore had the present-day concept of welfare. That he sometimes used the phrase ‘good for’ proves little, since that phrase has many meanings. It can mean what is simply good and a state of a person, as Moore thought, or have a merely descriptive or mixed meaning, as Prichard and Russell did. It can also refer to what a person believes is good, and it can express the very different concept of agent-relative goodness. Here something is good ‘for’ a person if it is good from his point of view or relative to him, so he and perhaps only he ought or has reason to pursue it. Whereas agent-neutral goodness makes the same demands on everyone, the good from his point of view may give only him duties or reasons. Finally, ‘good for’ can express the present-day concept of welfare. Which meaning did Sidgwick intend in his discussion of egoism? It cannot have been the concept of ‘my good’ analysed earlier in The Methods, since that has no special connection to egoism; utilitarians too use this concept, saying we ought to maximize the good of or in everyone. But for the same reason it cannot be the concept of welfare, for that too has no special connection to egoism. Welfarist utilitarians say we ought to maximize the welfare of all, thereby using rather than rejecting the concept. So welfare too is irrelevant to Sidgwick’s argument. What is relevant, and what he must therefore have intended, is agent-relative goodness, or goodness from a given person’s point of view. If the egoist says that each person’s pleasure is agent-neutrally good he contradicts himself, since agent-neutral goods give duties to everyone. But if he says that each person’s pleasure is the only thing good from her point of view, it follows, perfectly consistently, that she should seek only her own pleasure, since it is the one good relevant to her. The agent-relative sense of ‘good for’ fits Sidgwick’s claims about egoism, and it also fits his official discussion of ‘good’ in Methods I 9. That discussion never mentions a distinct welfare-concept, but defines the good simply as what one ought to desire. And this definition allows

23

Baldwin, G.E. Moore, pp. 73–6, 116–19; Darwall, ‘Sidgwick, Concern, and the Good’, p. 291.

minimal concepts

37

agent-relativity, since what different people ought to desire can be different. In addition, the agent-relative sense fits Sidgwick’s language in his discussion of egoism, which speaks only once of what is ‘good for’ a person and more often of his ‘rational ultimate end’ (ME 420–1, 497–8). But a person’s rational end is what she ought to desire, and what different people ought to desire can again be different. Finally, the agent-relative sense fits Sidgwick’s view that impartial or agent-neutral good is the universalization of ‘good for’. This would be hard to explain if ‘good for’ expressed welfare: the pleasure of the group of all people cannot be ‘good for’ the group, because only individuals have a welfare.24 But if ‘good for’ expresses agent-relative goodness, we can define the impartial good as what is good from the point of view of all moral agents, so they all equally should desire and pursue it. In fact, Sidgwick’s oft-maligned talk of the good ‘from the point of view of the Universe’ (ME 382, 420) here receives a straightforward analysis, as what is good from all people’s point of view, or makes the same claim on them all.25 What is good from my point of view is what I and perhaps only I should seek; what is good universally is what everyone should seek. He himself gave this analysis, saying that, when used ‘absolutely’, ‘good’ means ‘desirable from a universal point of view’, or ‘what all rational beings, as such, ought to aim at realizing’ (EEM 27; also ME 112). Sidgwick therefore used two distinct concepts: ‘my good’, or what is good (however that is understood) and located in my life, and the ‘good from my point of view’, or agent-relative goodness. He did not clearly distinguish the two, because the only agent-relative view he considered was egoistic hedonism, which locates the good from my point of view entirely in me. But an agent-relative view can say the pleasure of my spouse or child is a greater good from my point of view than the equal pleasure of a stranger, or even that what is good from my point of view is only other people’s pleasure and not my own. (This was Ross’s final view (FE 282).) Nonetheless, neither of Sidgwick’s two concepts involves welfare in the present-day sense, and each is reducible to his basic concept ‘ought’, as connoting something one ought to desire. Nor is it evidence against this reading that Sidgwick sometimes used the term ‘wellbeing’,26 since he treated it as a mere verbal equivalent of ‘ultimate Good’ (ME 391, 392, 396–7, 453–4; OHE 1–3, 4–5, 11, 48–9; EEM 15), as did others such as Rashdall (TGE I 184, II 59, 107; E 65, 72). There is historical confirmation for this reading of Sidgwick. If Moore abandoned a welfare-concept crucial to Sidgwick’s ethics, as some have said, we would expect the reviewers of Principia to have noticed this fact. But they did not, most instead

24 In one paper Sidgwick wrote, ‘I shall assume that what is good on the whole for any individual agent is also good on the whole for human society, the world of living things, or the cosmos, whichever we take to be the larger whole of which the individual is a part’ (EEM 59). Since on most views the cosmos cannot have a welfare, Sidgwick must here have meant by ‘good for’ just ‘good located in’. 25 For the maligning, see, e.g., Williams, ‘Point of View of the Universe’; Gert, Morality, p. 132. 26 Crisp, ‘Pleasure and Hedonism’, p. 27n4.

38

british ethical theorists from sidgwick to ewing

commenting on Moore’s deep indebtedness to Sidgwick.27 Especially pertinent is the reaction of Sidgwick’s collaborator and defender, E.E. Constance Jones. Her ‘Mr. Moore on Hedonism’ began by noting key points of agreement between Moore and Sidgwick, including on a central concept of ‘what is intrinsically good, desirable in itself or as an end’ that is ‘indefinable’ and can only be applied ‘by intuition’;28 she saw no difference in their evaluative concepts. She did speak later of what is ‘good for’ or ‘a good to’ a person, but meant by that only what Sidgwick meant by ‘my good’, namely a good that is located in a person’s life and could not exist unless he did.29 Her critique of Moore concerned only his substantive claims about the good, where she defended Sidgwick’s hedonism against his attacks. An objector may say that if my reading of Sidgwick is correct, he was just as misguided as others in the school, failing to centre his ethics on the key concept of the good for a person. But again I think they were right to avoid this concept. Welfare as philosophers today understand it cannot be promoted just as such. It consists in other states of people, and there are familiar disputes about what these are: some say pleasure, others desire-satisfaction, yet others the items on an ‘objective list’. But imagine a moral view that accepts a hedonist view of welfare and then makes the following claims: we ought to maximize what is intrinsically good, what is good is everyone’s welfare, welfare consists in pleasure, so we ought to maximize pleasure. What does this view gain by its talk of welfare, and how in particular does it differ from one saying only that we ought to maximize what is good, the good is pleasure, and therefore we ought to maximize pleasure? (Those who reject the concept of intrinsic goodness can compare a view that says we ought to maximize welfare, which consists in pleasure, with one saying only that we ought to maximize pleasure.) The two views’ starting and end points are the same; they both say we ought to maximize what is good and ought to maximize pleasure. What of substance is added by the middle claims about welfare?30 They might add something if we had a clear grasp of the concept of welfare and a strong intuition that we ought to promote it that is independent of any judgements about what welfare consists in. But there is ground for doubt on both counts. Different writers on welfare have very different understandings of the concept. Some take it to be a normative concept, which implies or consists in claims about what we ought to pursue, with the ‘ought’ often called ‘prudential’.31 Others see it as a non-normative concept that needs to be supplemented by claims about ‘good

27 Anonymous, Review of PE, p. 847; Bosanquet, Critical Notice of PE, p. 255; Mackenzie, Review of PE, p. 379; Wilde, Review of PE, p. 582. The contemporary reaction to PE is helpfully surveyed in Welchman, ‘G.E. Moore and the Revolution’. 28 Jones, ‘Mr. Moore on Hedonism’, p. 429. 29 Jones, ‘Mr. Moore on Hedonism’, pp. 456, 458–9, 463. 30 Regan, ‘Why Am I My Brother’s Keeper?’, pp. 212–13. 31 Scanlon, What We Owe, p. 109; Parfit, On What Matters, vol. 1, pp. 101–2.

minimal concepts

39

simpliciter’ or ‘ought’ to yield directives about how we should act.32 Some view the concept’s relativity—the ‘for’ in ‘good for’—as tying it especially to subjective views, on which a person’s welfare is determined by her desires, feelings, or other subjective states; others read the ‘for’ objectively, as connoting a relation to capacities whose development benefits her whatever she desires.33 It is hard not to see this last, purportedly conceptual dispute as resting on competing substantive views about what is worth pursuing, and therefore hard to see the concept of welfare as having independent content. And without such content, what does it add to a normative view? A more radical objection rejects the school’s concept of intrinsic goodness entirely, as unintelligible or philosophically confused. It was first made in the 1950s by P.T. Geach, who argued that the only legitimate uses of ‘good’ are attributive, as in ‘good knife’ or ‘good human act’, and called the predicative ‘x is good’ a ‘peculiarly philosophical use of words’ not found in ordinary language and involving ‘words without knowledge’.34 Others have made similar claims, though often favouring ‘good for’ rather than Geach’s attributive ‘good’.35 But this objection’s claim about ordinary language is simply false. Predicative statements such as ‘pleasure is good’ are perfectly familiar English and are often made. Moore’s Principia, which centres on such statements, had a wide readership outside philosophy yet no one found its concepts problematic. Geach said the F in ‘a good F ’ must be a specific predicate rather than one as generic as ‘thing’, but Martha Stewart’s catchphrase is ‘It’s a good thing’, and the history parody 1066 and All That repeatedly describes events as ‘A Bad Thing’. The predicative ‘good’ also occurs in one of the best-known pieces of English prose, the biblical Book of Genesis. It describes how, after each stage in his creation of the world, God saw that what he had created was good. Thus, having created light, he ‘saw the light, that it was good’. God was not expressing relief that he had created good light rather than bad light, whatever that might be. Nor was he saying the light was good for some being; it was not good for him, since he needed nothing, nor was it good for anything else, since nothing else yet existed. He was saying the existence of light was better, in an unrelativized or predicative sense, than its non-existence. Yet none of the Bible’s millions of readers have found that claim unusual.36

32

Sumner, Welfare, Happiness, and Ethics, Chapter 7. Sumner, Welfare, Happiness, and Ethics, Chapters 1–2; Kraut, What is Good and Why, Chapter 2. 34 Geach, ‘Good and Evil’. Though Ross was one of the main ‘Objectivists’ his article targeted, Geach did not acknowledge that Ross recognized the distinction between attributive and predicative uses of ‘good’ and discussed it at length in both his books (RG 65–7; FE 255–7). He also sneeringly dismissed an article ‘by an Objectivist’ without identifying it as Carritt’s ‘An Ambiguity’ or letting readers examine Carritt’s arguments for themselves. 35 Williams, Morality, pp. 40–50; Foot, ‘Utilitarianism and the Virtues’; Thomson, ‘The Right and the Good’; Kraut, What is Good and Why. 36 For similar points see Butchvarov, Skepticism in Ethics, p. 17 (from where I borrow the Genesis example); Pigden, ‘Geach on Good’; Zimmerman, ‘In Defense of the Concept of Intrinsic Value’. 33

40

british ethical theorists from sidgwick to ewing

Can a concept found in ordinary language nonetheless be philosophically objectionable? It can, but it is hard to see why in this case. On views such as Sidgwick’s and Ewing’s the intrinsically good is what we ought, or what it is fitting, to desire for itself. Are there philosophical objections to ‘ought’ and ‘fitting’? Moore’s view treats ‘good’ as unanalysable, but ethics needs some unanalysable concepts and the critics’ views have their own. On Geach’s view, for example, ‘good’ must have a constant meaning in all its applications, from ‘good knife’ to ‘good burglar’ to ‘good act’, as ‘small’ does in ‘small mouse’ and ‘small elephant’. But he gave no indication what this meaning is and seemed to treat it as primitive. If it is primitive, his view is no freer of unanalysable concepts than Moore’s and no less problematic on that score. His rejection of ‘intrinsically good’ seems to have been pure prejudice.

1.4 Concepts and Claims As the above discussions show, the school’s minimalism involved some scepticism about everyday language, whose ethical terms they often found ambiguous. Prichard and Ross thought the ordinary-language ‘ought’ can refer either to a crucial ethical concept or to the merely descriptive one of what will further a given end; they, Carritt, and Ewing they also thought ‘good’ has different meanings when used predicatively, attributively, and in the phrase ‘a good to’. They greatly respected everyday judgements about what is substantively right and wrong, treating them as a touchstone for ethical theory. But they thought everyday ethical language is misleading and has often misled philosophers. Prichard thought many previous moral theories collapsed the moral into the non-moral ‘ought’ and were encouraged to do so by the word’s ambiguity (MW 43, 116, 142–5, 169, 188–93, 237), while Carritt said Plato and Aristotle alternated between two meanings of ‘good’ without realizing they were doing so (‘AG’ 52–6). A sound moral philosophy must use a more regimented language and attend carefully to distinctions ordinary language blurs. As Carritt said: in moral philosophy we learn to use language with much greater care and discrimination. In ordinary talk . . . we use terms like ‘right’, ‘moral’, ‘virtuous’, ‘good’, ‘meritorious’, ‘obligatory’, as if they all meant much the same; but when we have to speak accurately we find that actions and characters can be ethically commended on different grounds, and that we must either invent new words to indicate these differences or select which among the names we have hitherto used loosely is most appropriate to each. (EPT 5–6; also ‘AG’ 53)

Broad talked of ‘the disastrous results of blindly following the suggestions of language’ and called natural languages ‘defective and distorting media’ (‘SAP’ 105, 114). This view contrasts sharply with that of the school’s immediate philosophical successors, especially at Oxford, who so revered ordinary language that they made its careful study the main task of philosophy. Their view was famously expressed by Austin:

minimal concepts

41

our common stock of words embodies all the distinctions men have found worth drawing, and the connections they have found worth marking, in the lifetime of many generations: these surely are likely to be more numerous, more sound, since they have stood up to the long test of survival of the fittest, and more subtle . . . than any that you or I are likely to think up in our armchair of an afternoon.37

But our school thought ordinary language fails to draw crucial ethical distinctions and makes connections it should not make, and there is no reason to exclude this view a priori. Our language was developed in pre-philosophical times by people with non-philosophical interests. Why should the distinctions it draws always line up with those most needed in philosophy?38 It may be objected that the school’s view makes words like ‘ought’ and ‘good’ simply ambiguous, like ‘bank’ for a financial institution and for the side of a river, when they are not. (Geach made this objection to Carritt’s claims about ‘good’.) But the school could reply that the ethical ambiguities are subtler than those for ‘bank’ and easier to fall into. They could also point to certain commonalities, as Ross did for ‘good’. Quoting the OED definition of ‘good’ as ‘the most general adjective of commendation’, he said that what unites the different meanings of ‘good’ is that they are all used to commend, or express a favourable attitude to, some object (RG 90–1, FE 252–5, 284–5). He here anticipated Hare’s The Language of Morals, which began with the same dictionary definition and likewise said ‘good’ is used to commend.39 But there was an important difference between their views. Hare assumed that if ‘good’ is used to commend, that must be part of its meaning, so its linguistic function is to indicate the illocutionary force of commending or prescribing. But Ross recognized that we can perform a speech act by using a sentence whose meaning contains no specific reference to that force; thus I can commend your performance on a test not only by saying ‘I commend your performance’ but also by saying ‘You got the highest mark in the class!’ In the right tone, the last remark can be recognized as one of commending even though its literal content is descriptive. Following Alexius Meinong, Ross said an utterance can express an internal state of mind even though its semantic content concerns external objects, and he thought this regularly happens with statements containing ‘good’ (FE 254). They express approval of an object, but their literal content is that the object has a certain property, sometimes intrinsic goodness and sometimes that of conducing to an end, so the expression of approval unites what in their literal meanings are very different remarks. As well as simplifying the normative concepts, their minimalism allowed the school to address what they considered the most important questions in the most

Austin, ‘A Plea For Excuses’, p. 8. For a generalization of these points about ordinary language see Whiteley, ‘Mr. Warnock on Ordinary Language’. Broad said the if the ordinary-language philosophers charged his EMP with consisting ‘largely of difficiles nugae’, he would reply that their writings ‘consist largely of faciles nugae’ (‘A’ 60). 39 Hare, Language of Morals, p. 79. 37 38

42

british ethical theorists from sidgwick to ewing

straightforward way. They are known for their non-naturalist and intuitionistic metaethics, but their focus was often more on substantive questions about what is in fact good or right. Sidgwick’s Methods briefly discussed moral truth and knowledge but spent most of its time examining the competing normative theories of egoism, utilitarianism, and deontology. Moore’s Principia said more about metaethics, but only to clear the way for what it called the ‘primary ethical question’ of what is good in itself (PE 158, also 27, 138, 140, 184, 189, 222). Rashdall, McTaggart, Prichard, Carritt, and Ross too concentrated on normative questions, and though Broad and Ewing wrote more extensively on metaethics, this was mostly later in their careers and in response to criticism from non-cognitivists. At the start their main interest, like the others’, was in substantive moral questions. A great merit of conceptual minimalism is that it makes these questions transparently substantive. If Sidgwick’s conflict between egoism and utilitarianism is formulated using a single ‘ought’, the issue is plainly the substantive one of which view has the stronger claim to determine what we ought to do. But if it pits a prudential against a moral ‘ought’, the issue can seem to turn on conceptual issues about what prudence and morality in the abstract involve, which is misleading since substantive questions are never settled conceptually. There is an additional danger if the conflict is made one between ‘moral’ and ‘rational’ ‘oughts’. Since many take rationality to be the fundamental normative category, this can seem to give the so-called rational ‘ought’ priority, as if it must naturally outweigh any merely moral claim. But any specific rational ‘ought’, such as one telling us to do whatever will most satisfy our desires, makes a substantive claim about how we should act. It must therefore be accepted, if it is, on substantive grounds, and it can likewise be rejected on substantive grounds, as it was by our school: they thought it simply false that we ought to do what will satisfy an evil desire. Something similar holds for ‘good’ and ‘good for’. To decide whether a virtuous but painful life is better than a vicious and pleasant one, we need to weigh pleasure and virtue against each other. It does not help to call the one ‘good for’ us and the other simply ‘good’, and doing so is positively misleading if it suggests either that the issue turns on conceptual issues about ‘welfare’ or that the value of pleasure has some priority. The school’s conceptual parsimony did not mean their substantive views were always minimalist. This was true of Sidgwick: his preferred normative theories combined a single consequentialist principle of right with a single intrinsic good, pleasure. But Rashdall, McTaggart, and Moore, though also consequentialists, had more complex theories of the good. Moore used his one concept of intrinsic goodness in a pluralistic theory that includes a principle of organic unities whereby the value of a whole need not equal the sum of the values of its parts and recursive principles that generate higher-level goods out of lower-level ones. Prichard, Carritt, Ross, and Broad shared this pluralism about the good but also defended pluralistic theories of the right, replacing the consequentialists’ one duty with several distinct ones, such

minimal concepts

43

as to keep promises, pay reparations, and more. They made many claims using the one concept ‘good’ and many using the one concept ‘ought’. It is even arguable that the simplicity of their conceptual apparatus encouraged them to develop these richer substantive views. Because they were undistracted by conceptual issues, they were freer to notice and try to capture the many different elements in everyday moral thought and the complex ways they interact. Their procedure was in this respect a model, reflecting what I like to call a ‘salad dressing’ approach to moral theory. The key to a good salad dressing is to be a spendthrift with the oil and a miser with the vinegar. In moral theory the ideal is to be a miser with moral concepts and a spendthrift with moral claims.

2 ‘Ought’ and ‘Good’ The school recognized just two main ethical concepts: a deontic concept expressed by ‘ought’, ‘right,’ or ‘duty’ and a value-concept expressed by ‘good’. They differed about the relationship between the two—is one reducible to the other, and if so which one?—and on more specific issues concerning each. But first they had to distinguish them from each other.

2.1 ‘Ought’ vs. ‘Good’ Sidgwick opened his discussion of ‘good’ in The Methods by contrasting ‘right’ and ‘duty’ as ‘imperative’ concepts with ‘good’ as a merely ‘attractive’ one. If an act is right, he said, there is ‘an authoritative prescription to do it’, but a thing’s being good does not imply that we should prefer it to all other good things, since some of those may be better (ME 105–6). But though this distinction fits ‘right’ and ‘good’, it does not separate the deontic and value-concepts more generally. Like ‘right’, the valueconcept ‘best’ places its object at the top of an evaluative scale; if something is best, you should prefer it to all alternatives. And like one ‘good’, one ‘ought other things equal’ can be outweighed by another and so is not authoritatively prescriptive. In each of the categories there is a less prescriptive concept, ‘good’ or ‘ought other things equal’, and a more prescriptive one, ‘best’ or ‘ought all things considered’. And Sidgwick’s broader contrast between attractive and imperative concepts is puzzling coming from him, since he treated judgements about ‘good’ as imperative when he defined the good as what you ought to desire (ME 112–13) and said recognizing an end as ‘ultimately reasonable’ involves recognizing ‘an obligation to do such acts as most conduce to the end’ (ME 35).1 He was on firmer ground when he said ‘good or excellent actions are not implied to be in our power in the same strict sense as “right” actions—any more than any other good things’ (ME 113). Whereas a duty to do x presupposes that you can do x, ‘good’ and ‘evil’ apply also to ‘results out of the reach of human attainment—an abundant harvest next autumn or influenza in the winter’ (EEM, 59; also ME 110, GSM 330). Rashdall and Russell likewise thought the main difference between

1

See further White, ‘The Attractive and the Imperative’.

‘ought’ and ‘good’

45

deontic and value concepts is that the former require voluntary control (TGE I 135; E 12, 14; ‘EE’ 17); Prichard said the obligation to do x presupposes the physical ability to do it though not the psychological capacity given your character (MW 67) but did not apply that condition to ‘good’; and Ross endorsed ‘the Kantian principle . . . that “I ought” implies “I can”’ (RG 5) but nothing similar about ‘good’. This is indeed a principal difference between the two categories of concept. Unlike these last four, Sidgwick did not think all ‘oughts’ presuppose voluntariness. He accepted this for a ‘narrower’ sense of ‘ought’ but thought there is a ‘wider’ sense in which it can be true that I ought to know or feel something even though I cannot do so by an act of will; here the ‘ought’ implies only an ‘ideal or pattern’ that I ought in the narrower sense to imitate when I can (ME 33). Broad too recognized this wider ‘ought’ (FT 61, 163), but Ross rejected it. If we say someone ought to feel sorrow at a relation’s death when his character makes that impossible, he argued, we can mean only that his not feeling sorrow ‘is a bad thing, a manifestation of a bad character’ (FE 45). This seems right. If the wider ‘ought’ means only that something like compassionate sorrow is an ‘ideal’, it is hard to see how it differs from the claim that the sorrow is desirable or good. English usage certainly allows remarks such as ‘you ought to care more for others’, but the issue is where the philosophically important distinctions come. There is surely a vital difference between normative concepts that do and do not presuppose voluntariness; given that, we do best to confine ‘ought’, as Ross did, to the former. This initial difference implies a further one in the objects that fall under the two concepts. As Ross said, ‘right’ in its deontic use applies only to acts while ‘good’ applies to ‘things other than acts’, such as virtue, knowledge, and pleasure (FE 314). But this point should not be exaggerated. Like others in the school, Ross thought ‘good’ applies only to states of affairs and not, as on Kantian views, to entities such as persons. The bearers of value are what he called ‘objectives’, whose proper expression is a thatclause such as ‘that he is acting from a sense of duty’ or ‘that he has insight’. More specifically, they are objectives that exist and constitute ‘facts’; if someone could be happy, his happiness would be good if it existed but is not actually good until it does (RG 111–14). But voluntary acts can also be understood as objectives. If I have promised to do x, what I am obligated to bring about is ‘that I do x’, or the state of affairs in which I do x. The difference between items that are right and ones that are good is therefore not one of ultimate ontological kind. Both are states (in a broad sense), and differ only in that the former are under the agent’s control. This is how Rashdall and Russell saw the issue: for them ‘good’ was the wider term, applying to all items that can be evaluated, whereas ‘right’ in its deontic use governs only the subset that are voluntary acts.2

2 ‘Right’ also has non-deontic uses. Ross called satisfaction in another’s pleasure ‘right’ (FE 279), but in this usage the term is not governed by ‘ “ought” implies “can” ’ (FE 55). We can also speak of the right means to an end, or the right road or right tool, but those express hypothetical imperatives, which on Ross’s view are not normative.

46

british ethical theorists from sidgwick to ewing

A more important difference concerns the logical structure of the two concepts. Those who took ‘good’ to be unanalysable, such as Rashdall, Moore, and Ross, thought it ascribes a simple property to states of affairs, and the same view is possible of ‘right’. Then if I have promised to do x, my act of x-ing has a simple property of rightness, and something similar follows if ‘right’ is just a narrower concept than ‘good’. But if rightness is a simple property, my act’s being right makes the same normative demand on everyone, just as a state’s being agent-neutrally good does. And that is incompatible with the agent-relativity of the deontological duties in theories such as Prichard’s, Carritt’s, and Ross’s, which must be at least conceptually possible. On these theories my having promised to do x imposes an obligation mainly on me. If others can promote my x-ing, for example by reminding me of my promise, they may have no duty to do so, and even if they have one, it is far weaker than mine. The person mainly obligated by my promise is me, and that is not possible if obligation is a simple property had by acts on their own. The deontic concepts therefore require a different treatment, and at least two are possible. One takes ‘right’ to ascribe a relational property to the pair of me and an act or, since many individual acts would fulfil my duty, of me and an act-type. Another makes ‘right’ ascribe a property just to me, that of being obligated to perform an act of type x. (If others have a duty concerning my promise, it will involve the different act-type of helping me to x.) Prichard defended the second of these views, though on a different ground. As an extreme metaphysical realist, he thought that nothing that does not exist can have properties. And since acts are morally required before they are performed, we cannot follow Ross’s line on ‘good’ and say my x-ing would be obligatory if I did it; it must be obligatory now. That means the obligation must be a property of something that exists now, namely me: ‘when we make an assertion containing the term “ought” or “ought not”, that to which we are attributing a certain character is not a certain activity but a certain man’ (MW 99; also 167–9). This view was shared by Ross and Joseph (FE 155–6; SPE 61) and allows me to have an obligation concerning my promise that is different from and stronger than other people’s. These logical issues aside, Prichard and Ross thought that even when ‘good’ and ‘right’ apply to the same items, namely voluntary acts, they do so on different grounds. What makes an act good, and in particular virtuous or morally good, is only the motive from which it is done and not its conformity to external rules. But what makes it right is only its conformity to rules and not its motive. Sidgwick and Broad shared this view (ME 202–4; CE 18–19), but Prichard and Ross emphasized it more. In ‘Mistake’ Prichard said ‘the rightness of an action concerns an action not in the fuller sense of the term in which we include the motive in the action, but in the narrower and commoner sense in which we distinguish an action from its motive’. Its rightness ‘lies solely in the origination in which the act consists, whereas the intrinsic goodness of an action lies solely in its motive’ (MW 11–12, 14; also 3, 35–6, 59, 198–9). Ross proposed that we use the word ‘“act” of the thing done, the initiation of

‘ought’ and ‘good’

47

change, and “action” of the doing of it, the initiating of change, from a certain motive’. It will then follow that ‘the doing of a right act may be a morally bad action, and that the doing of a wrong act may be a morally good action; for “right” and “wrong” refer entirely to the thing done, “morally good” and “morally bad” entirely to the motive from which it is done’ (RG 7; also FE 122–3, Moore E 78–80). Prichard and Ross gave two arguments why rightness is independent of motives, one directed against the view that a right act must be done from the motive of duty, or the desire to do it because it is right. This motive presupposes the belief that the act is right, and in its simplest form this is the belief that the act is right in itself, or apart from its motive. But if this belief is true, it contradicts the proposed view about rightness, on which the act is right only if done from a certain motive. As Ross put it, the ‘whole expression’ of the view then ‘is in contradiction with a part of itself ’ (RG 5; FE 117; Prichard, MW 3, 11). And if the view tries to avoid the contradiction by making rightness require the more complex belief that the act is right only when done from the belief that it is right, the same issue arises about the embedded belief and generates an ‘infinite process’ (MW 157) or ‘infinite series of amendments’ (RG 5; also FE 117–18) that never yields a determinate belief. The proposed view could require only the simple belief that the act is right but allow that this belief is false. But that would make it a necessary condition for acting rightly that you have a mistaken belief about your act’s rightness, which is implausible. Prichard said this would involve ‘the paradoxical, and indeed untenable, idea that the existence of moral goodness depends on the existence of an illusion’ (MW 219; also 158), while Ross said ‘it can hardly be claimed that it is our duty to act from a mistaken thought’ (FE 118). So a view requiring the motive of duty for rightness either contradicts itself, generates an infinite regress, or makes rightness depend on a false belief.3 This argument concerns only the motive of duty, but Prichard and Ross had a more general argument appealing to the ‘ “ought” implies “can” ’ principle. To have a duty to act from a certain motive, they said, we must be able to have that motive, and in many cases we cannot: our desires are not under our control. Prichard said it ‘cannot be true that we ought to act on a certain desire; for at any moment its existence and degree is out of our power’ (MW 3); Ross said it ‘is not the case that I can by choice produce a certain motive . . . in myself at a moment’s notice, still less that I can at a moment’s notice make it effective in stimulating me to act’. Though I can now do what may develop that motive in the future, ‘[m]y present duty . . . cannot be to act here and now from it’ (RG 5; also FE 115–16). It is unclear how far Prichard and Ross intended these arguments to go. They said only that rightness is independent of the agent’s motives or desires, but they did not

3 In some passages Prichard and Ross wrote as if the view that right acts must be done from the motive of duty was Kant’s (MW 219, RG 5; also Broad, FT 119), but Ross later said Kant held that an act’s being done from the belief that it is your duty (one concept) gives it moral worth (a different concept) (FE 139). On this reading Kant’s view was close to theirs.

48

british ethical theorists from sidgwick to ewing

think desires the only mental states relevant to action. Prichard distinguished desiring from ‘willing’ and Ross distinguished it from ‘choice’ (MW 274; FE 192–4), each of which seems closer to the present-day concept of intention. Was their view only that an act’s rightness is independent of the agent’s desires or, more generally, that it is independent of all his mental states, including his intentions? Some philosophers make the second, stronger claim and use it to reject views that make intentions relevant to the rightness of acts. More specifically, they use it to reject the doctrine of double effect, which says an act that causes a bad effect can be wrong if the effect is intended as an end or means but permitted if the effect is only foreseen. Thus, a military attack that kills civilians can be forbidden if their deaths are intended as means to the end of victory, but not if they are foreseen but unintended side effects of an attack on a military target. These philosophers say the doctrine of double effect mistakenly uses facts about an agent’s mental states, which bear only on the evaluation of his character, to assess the rightness of his acts.4 Did Prichard and Ross intend a general claim like this or only a narrower one about desires? We cannot tell by examining their views about double effect, since they never discussed it. But we can ask what force their arguments have against double effect. The argument about the motive of duty is not relevant, and the one about ‘“ought” implies “can”’, though relevant, does not support as strong a conclusion as they drew. Consider a duty of the kind Prichard and Ross were most concerned to deny: a positive duty to benefit others from a benevolent desire for their happiness. This duty requires us to make a conjunction true: that we both benefit others and do so from a benevolent motive. Since we cannot make a conjunction true unless we can make both of its conjuncts true, and cannot in this case make the second one true, the ‘ “ought” implies “can” ’ argument successfully rules this duty out, and in fact rules out all positive duties to act from a mental state. Ewing tried to resist this conclusion, arguing that if you have two motives favouring a given act, one good and one bad, you can concentrate your attention on the facts that prompt the good motive and away from those that prompt the bad one; in this way you can have a duty ‘to act in a certain state of mind’ (DG 138–40).5 As he acknowledged, however, your attention is something you can control to some extent (DG 139). Prichard and Ross could therefore accept his initial premise, affirming a duty to perform whatever mental acts of attending may strengthen good motives but no positive duty to act from those motives as such. The double effect duty, however, is not positive but negative, forbidding rather than requiring action from a certain mental state. It therefore requires us to make a conjunction false: that we both cause a bad effect and do so intending it as an end or means. And we can do this so long as we can make one conjunct false, which we can do by making the one about action false. Even if we cannot control our mental states, we can satisfy double effect by not causing the bad effect, and this is in fact how the doctrine’s proponents imagine it being applied. We ask whether in doing an act that will Thomson, ‘Self-Defense’, pp. 292–4; Scanlon, Moral Dimensions, Chapter 2. Scanlon makes the ‘ “ought” implies “can” ’ argument (pp. 58–60, 88). 5 See also Robinson, Review of RG, p. 349. 4

‘ought’ and ‘good’

49

cause a bad effect we would intend that effect as an end or means; if the answer is yes, we choose not to do the act. We avoid the forbidden conjunction by not causing the bad effect, thereby satisfying double effect even though we cannot control our intentions. The same point holds for negative duties about desire. Consider a view that forbids inflicting pain, even when that would otherwise be justified, if it is done from a vicious desire, as in an example of Sidgwick’s where a prosecutor prosecutes someone he believes is guilty but does so from malice (ME 202–3). Here again the duty, if merely negative, does not violate ‘“ought” implies “can” ’. The prosecutor can ask whether, if he did prosecute, he would do so from malice, and, if he would, he can decide not to prosecute; he too can make a conjunction false by making its conjunct about action false. It is striking that critics of the view that motives are irrelevant to rightness usually cite negative duties as counterexamples. Ewing said ‘it still seems paradoxical to say that I have done what I ought if I . . . give . . . a candidate who deserves it a third-class mark not because I think he deserves it but because I dislike him’ (DG 137–8). Steven Sverdlik says it can be wrong to marry or put a child up for adoption from a desire for money, wrong to have an abortion from a trifling motive such as convenience, and wrong to refuse to sell your house to a black or Jew from racism.6 All these examples involve duties not to act from certain mental states, and therefore do not violate ‘ “ought” implies “can”’ as a positive duty to act from a motive or intention would. Sidgwick thought that in his example the prosecutor would not be wrong to prosecute, and that may be the correct all-things-considered conclusion. Even so, the prosecutor may have a duty other things equal not to act from malice, and this duty, though outweighed by a stronger one here, may not be outweighed in other cases where the reasons favouring a malicious act are less weighty. Sidgwick also worried that if you have two motives for the same act, one virtuous and one vicious, you cannot determine which will cause the act if you do it. But this only means that if you do the act you risk violating your duty not to act from a vicious motive, and, like other risks, this one must be weighed against whatever positive reasons there are for running it. If the act’s benefits are very great and your potential motive only slightly bad, the risk of acting viciously may be worth taking; if the benefits are small and your motive horrible, it may not. Either way, you can fulfil your duty other things equal not to act from a vicious motive by not doing the act. Sidgwick, Prichard, and Ross may have held the strong view that an act’s rightness is independent of any accompanying mental states, intentions or desires, and that view may be true. But it does not follow from conceptual truths such as that ‘ought’ implies ‘can’. They exclude positive duties to act from good mental states but not negative duties to avoid acting from bad ones; the latter exclusion must be defended on other grounds.

6

Sverdlik, ‘Motive and Rightness’, pp. 339–41; also his Motive and Rightness.

50 british ethical theorists from sidgwick to ewing

2.2 Reducing ‘Ought’ to ‘Good’ Though Prichard, Carritt, and Ross thought ‘right’ and ‘good’ are distinct concepts, others tried reducing one to the other. The first possibility is to define ‘right’ in terms of ‘good’. This was, famously, Moore’s view in Principia. Calling ‘good’ ‘the notion upon which all Ethics depends’ (PE 142), he said he would confine his use of ‘right’ to that in which it denotes ‘what is good as a means, whether or not it is also good as an end’ (PE 18; also 21). Nor did he see this as an arbitrary stipulation: since ‘ “right” does and can mean nothing but “cause of a good result,” . . . the assertion “I am morally bound to perform this action” is identical with the assertion “this action will produce the greatest possible amount of good in the universe”’ (PE 147; also 24–6, 146, 148, 167–9, 180–1). It was also, if a little less clearly, Rashdall’s view. He said ‘good’ rather than ‘right’ is ‘logically the primary conception’ (TGE I 135), so ‘“right” is meaningless except in reference to the good’ (TGE II 42) and ‘[t]hat action is right which tends to bring about the good’ (TGE I 135; also I 266, II 42, 97, E 14, 61). He therefore treated Sidgwick’s axiom of benevolence, which says we should promote the good of everyone impartially, as a tautology, which ‘simply assert[s] that more good is always more valuable than less good’ (TGE I 147). He sometimes called ‘good’ and ‘right’ ‘correlative terms’, related in the same way as ‘concave’ and ‘convex’ (TGE I 138; E 14; ICE 46), which suggests equal importance rather than primacy for either one. But his more usual view, like Moore’s, reduced ‘right’ to ‘good’ and the idea of promotion. He and Moore sometimes explicated the good as what ‘ought to exist’ or ‘ought to be’ (TGE I 135; E 14, 25, 75; PE viii, 17, 115, 118, 121). They did not intend this as a reductive analysis, and it cannot be one. Since, as Prichard and Ross later argued, states of affairs cannot have obligations (MW 9–10, 167–8, 223–5; RG 105), the ‘ought’ in ‘ought to be’ can only be Sidgwick’s wider one, which connotes an ideal or standard. And that ‘ought’ is not clearly distinguishable from ‘good’. While Rashdall merely asserted that ‘right’ reduces to ‘good’, Moore argued for the view. When we call an action our duty, he said, we mean that ‘the performance of that action at that time is unique in respect of value’. But no dutiful act can be unique in value in the sense that it is better than anything else in the world; then every right act would be better than anything else, which is contradictory. It can only be unique in the sense that the world will be better if it is performed than if any alternative to it is, so ‘duty’ must mean ‘what maximizes the good’ (PE 147–8). But this argument is unpersuasive. If its initial claim about the ‘unique value’ of a dutiful act is not to beg the question, it must use ‘value’ in a broad sense, so it says the act is most supported by whatever factors make acts right. But then it does not support the specifically consequentialist conclusion that right acts maximize the good, since some factors may be deontological. Only if it understands ‘value’ more narrowly, as

‘ought’ and ‘good’

51

involving intrinsic goodness, does it yield that conclusion; and then it assumes what it is meant to prove. That issue aside, the reductive analysis is open to a decisive objection first made by Russell. Moore’s open-question argument says ‘good’ cannot mean the same as ‘pleasant’, since then ‘Everything pleasant is good’ would be equivalent to the tautologous ‘Everything pleasant is pleasant’. Russell pointed out that the same argument can be made against Moore’s definition of ‘right’. If someone asks why he should do what will have the best results, he said, we can only answer, ‘Because you ought to do what will have the best results’. And this reply ‘distinctly adds something. The same arguments by which good was shown to be indefinable can be repeated here, mutatis mutandis, to show the indefinability of ought.’7 Ross likewise said that, like ‘the good is the pleasant’, ‘the right is what produces the most good’ is synthetic rather than analytic, since philosophers have disagreed about it. Even if ‘a true statement about what is right’, it cannot be ‘what we mean by right’ (RG 8–9; also FE 42). Though neither Russell nor Ross said so, similar objections can be made to others of Moore’s analyses, for example of beauty as that the admiring contemplation of which is good. It reduces ‘it is good to admire beauty’ to ‘it is good to admire the things it is good to admire’, and though the latter claim is not completely tautologous, since it implies that there are things it is good to admire, it is surely less informative than the one it is meant to analyse. To Moore’s credit, he accepted Russell’s objection almost immediately. In his 1942 ‘Reply to My Critics’ he reported that he was ‘inclined to agree’ with Russell’s point when he read it (‘R’ 558), and he explicitly rejected the Principia analysis in his Ethics of 1912. There he called it ‘quite plain’ that the words ‘duty’ and ‘expediency’ have different meanings, since otherwise ‘it would be a mere tautology to say that it is always our duty to do what will have the best possible consequences’ (E 73; also 25–6, 76–7). He still thought right acts do maximize the good, but this consequentialist view was now only substantively self-evident and its denial not self-contradictory (E 76–7).8 In his ‘Reply’ he said it ‘follows from the very nature’ of intrinsic value that what has it ought to be promoted, but the connection was again synthetic rather than analytic, since a claim about intrinsic value neither includes nor is identical to one using ‘ought’ (‘R’ 574–7).

7

Russell, Review of PE, p. 330. The Independent Review was a politically liberal monthly launched in October 1903 by a group containing several of Moore’s and Russell’s Apostle friends. For a time it planned to publish a book-length ‘manifesto’ with opening philosophical essays by Moore and Russell, but the two disagreed about who should write on what and the idea came to nothing; see Levy, Moore, pp. 252–9. 8 This had also been his view before PE. In lectures given just before the book, he said the principle ‘you ought always to do that which is a means to what ought to be’ seems to be self-evident but cannot be proved and does not follow from the claim that something ought to be or is good (EE 118). He therefore only held the PE definition for a very short time.

52

british ethical theorists from sidgwick to ewing

Consequentialism was the dominant view in late nineteenth- and early twentiethcentury ethics, in either hedonistic or ideal versions. Sidgwick saw clearly that its core principle is substantive and must be defended as such, and that was also the prevailing view later in the period. In 1926 Laird wrote, ‘When we say . . . that duty is the adoption of the best, we are not saying simply that duty is duty, or that the best is the best. We are asserting a necessary and fundamental connection between value and obligation. . . . this connection is a synthetic connection, a union of things significantly different’ (SMT 21–2; also 25). In the middle of the period Rashdall and Moore claimed more ambitiously that consequentialism is conceptually true, but that view is not sustainable; whatever its substantive relation to ‘good’, ‘ought’ is not definable in its terms.

2.3 Reducing ‘Good’ to ‘Ought’: The Fitting-Attitudes Analysis The contrary view that reduces ‘good’ to ‘ought’ or some other deontic concept was also defended, first by Sidgwick and Brentano and later by Broad and especially Ewing. It is more plausible than Moore’s reduction and has re-emerged in presentday philosophy under the heading of the ‘buck-passing’ or ‘fitting-attitudes’ analysis of value.9 Sidgwick’s treatment of ‘good’ is often misunderstood. He began his main discussion of the concept by considering Hobbes’s view that a person’s good is whatever he desires as an end or for itself. Several objections led him to replace it with the view that a person’s good is ‘what he would now desire and seek on the whole if all the consequences of all the different lines of conduct open to him were accurately foreseen and adequately realised in imagination at the present point of time’ (ME 111–12), an analysis in terms of hypothetical rather than actual desires. Many have taken this to be his final view, so he accepted an ‘informed-desire’ account of ‘good’.10 But this was not his final view. He said this account involves an ‘ideal element’, since the things it identifies as good are not always actually desired, ‘but the ideal element is entirely interpretable in terms of fact, actual or hypothetical, and does not introduce any judgment of value . . . ;—still less any “dictate of reason”’ (ME 112). The analysis is, in other words, entirely descriptive, and Sidgwick thought that a mistake: ‘It seems to me, however, more in accordance with common sense to recognise—as Butler does— that the calm desire for “my good on the whole” is authoritative; and therefore carries with it implicitly a rational dictate to aim at this end’. An adequate analysis of ‘good’

9

See, e.g. Scanlon, What We Owe, pp. 95–100. Rawls, Theory of Justice, pp. 416–17, 421; Brandt, Theory of the Good and the Right, pp. 59, 155; Griffin, Well-Being, p. 39. That Sidgwick rejected this account is noted in Schneewind, Sidgwick’s Ethics, pp. 224–6; Parfit, Reasons and Persons, p. 500; Shaver, ‘Sidgwick’s False Friends’. 10

‘ought’ and ‘good’

53

must therefore be normative, but we can keep its normative aspect ‘merely implicit and latent’ if we define my good as ‘what I should practically desire if my desires were in harmony with reason, assuming my own existence alone to be considered’ (ME 112). Since reason is a faculty that issues dictates using ‘ought’, this in effect equates my good with what I ought to desire, and Sidgwick elsewhere made that equation explicit. The fourth edition of The Methods explained what it means for my desires to be ‘in harmony with my reason’ by saying ‘my good on the whole is what I “ought” to desire’ (ME4 110; also 112), and there were similar remarks in the final edition, such as that ‘we define “Good” as that at which it is ultimately reasonable to aim’, that ‘we may define “good” as “what one ought to aim at”’, and that the desirable or good ‘is what each person ought to desire’ (ME 92n, 381, 388; also EEM 27; GSM 145, 331). As Broad saw (FT 176), Sidgwick’s final view of ‘good’ was not an informed-desire but a normative one that reduces ‘good’ to ‘ought’. Any such reduction must respect the distinctive features of ‘good’, in particular the fact that it is not governed by ‘ “ought” implies “can” ’ and therefore does not apply only to products of voluntary choice. Sidgwick did this to some extent by making the good what you ought to desire rather than what you ought actively to pursue. Since you can often want things you cannot bring about, ‘ought to desire’ extends more widely than ‘ought to produce’. But he still faced a difficulty. It can only be true that you ought to desire something if you can desire it, and this is not always so: as Prichard and Ross argued, your desires are not under your control. Sidgwick addressed this difficulty in the fourth edition of The Methods. Acknowledging that ‘irrational desires cannot always be dismissed at once by voluntary effort’, he said the analysis of ‘good’ cannot use ‘ought’ in ‘the strictly ethical sense’ but only in ‘the wider sense in which it merely connotes an ideal or standard’ (ME4 110–11; also Broad, FT 163). As we saw above, however, there is reason to question the independence of this wider ‘ought’. In particular, if the claim that you ought to desire some end is just the claim that this desire is an ideal, how does it differ from the claim that the desire is good, and how then is Sidgwick’s analysis not circular, defining a good state of affairs in terms of a good desire? One could try to avoid the difficulty by defining the good as what you ought to desire if you can. But Sidgwick did not take this line and was right not to. If ‘good’ has normative force, it surely has it for all people, not just those with a certain psychology, and it must in particular have it for all people if it is to yield the right conclusions about action. That you ought to desire x implies, in conjunction with a hypothetical imperative requiring the most effective means to your ends, that you ought, if able, to do what will most promote x. We want to say, with Prichard, that everyone who physically can promote a good end ought to do so, whatever his desires. And that will follow only if everyone ought to desire the end. Later versions of the analysis followed Sidgwick in using a concept not governed by ‘“ought” implies “can”’. Brentano defined the good as that the love of which is correct, where ‘correct’ is a broad concept that applies also to beliefs, which are

54

british ethical theorists from sidgwick to ewing

correct when they are true.11 In Foundations Ross said that while virtue and knowledge have an unanalysable property of goodness, pleasure is good only in a secondary sense in which the good is what it is right to feel satisfaction in (FE 275–6, 279), and where ‘right’ does not require voluntariness (FE 55). But the most sustained effort of this kind was Broad’s and Ewing’s proposal to define the good as what it is ‘fitting’ to desire or have a pro-attitude toward. Broad made this proposal with his characteristic tentativeness, saying ‘it might be held . . . that “X is good” means that it is fitting for every rational being to desire X’ (FT 278; also 283, CE 109, 114, EMP II 659), but Ewing made it the centre of ‘A Suggested Non-Naturalistic Analysis of Good’ and The Definition of Good and defended versions of it through the rest of his career. Moore once explicated ‘good’ by saying a good experience is one ‘worth having for its own sake’ (‘GQ’ 122–3), and though Moore did not intend this as a definition Ewing thought it is one, giving just what the ‘man in the street’ means by ‘good experience’ (‘SN’ 7; DG 147–8). He then generalized the definition beyond experiences by saying the intrinsically good is what ‘ought to be chosen for its own sake’ (‘SN’ 7) or is ‘worth choosing or producing for its own sake’ (DG 148). The ‘ought’ here is not that of moral obligation, which he understood in a distinctive way, but that of fittingness as understood by Broad, so the good is what it is suitable or fitting to choose for its own sake (‘SN’ 8, 14; DG 150–1). Ewing thought responses other than choice can also be fitting. The good is what we ‘ought to welcome, rejoice in if it exists, seek to produce if it does not exist . . . approve its attainment, count its loss a deprivation, hope for and not dread its coming if this is likely, avoid what hinders its production, etc.’ (DG 149; also ‘SN’ 8). These are all what Ross called pro-attitudes, and Ewing’s final definition of ‘good’ was as ‘a suitable object of pro-attitudes’ (‘SN’ 9) or a ‘fitting object of a pro-attitude’ (DG 152). But he did not think all pro-attitudes are fitting to something good, nor always the same ones. Sidgwick and Broad defined ‘good’ in terms of desire, and Ewing sometimes included desire as a relevant pro-attitude (‘SN’ 8). But at other times he argued that it is not usually a suitable response to something good. One argument was that desire is an ‘uneasy’ or unpleasant emotion, which the less we feel, the better (‘SN’ 9; DG 153). But desire is not always unpleasant—mild hunger can feel positively good— and even when it is, it can be at the same time bad insofar as it is unpleasant and fitting given its relation to its object. A second argument was that, if desire is more than an uneasy emotion, it becomes a striving to bring about its object, and then a definition in terms of desire merges into one in terms of choice and pursuit, which he preferred (‘SN’ 9; DG 153). But desire extends more widely than active pursuit and is explanatory of it: the explanation why you pursue an object is that you desire it and believe you can produce it. Later he said he had been wrong to define the good in

11

Brentano, Origin of Our Knowledge, p. 18.

‘ought’ and ‘good’

55

terms of feelings, since it is false that you should proportion your feelings to the degrees of value in their objects: ‘There is no obligation to work oneself into a terrific rage at the crimes of Hitler and Stalin, or to try to feel a million times as miserable because a million people have met a horrible death as we should on seeing one man thus die’ (ST 88). This is a stronger argument, but surely there is some merit in proportioning your feelings: if one of two people similarly related to you now feels intense pain and the other mild pain, you should feel worse about the first person’s pain. There can also be agent-relativities about feeling, so it is more fitting to feel sympathy for your child’s pain than for a stranger’s somewhat more intense one. And other factors can be relevant: it can be less fitting to feel strongly about events in the distant past or that were less unlikely.12 Given enough such factors the attitudes relevant to goodness can include not just conations but also desires and feelings. More importantly, Ewing thought different attitudes can be appropriate to different objects. The fitting response to pleasure or knowledge is to choose or welcome it for its own sake, but other things call, differently, for admiration. This is true of beauty and non-moral cleverness but especially of morally virtuous actions such as ones of self-sacrifice, though here what is fitting is a specifically moral form of admiration (‘SN’ 11–13; DG 153–6, 166–7). He was unsure whether moral admiration is irreducibly distinct from other forms of admiration or differs only in its object, which is an action with volitional qualities such as devotion to a good end. But he was clear that admiration differs from choice and welcoming and is appropriate only to some objects, for example self-sacrifice but not pleasure. He also denied that it is fitting to choose or welcome self-sacrifice for its own sake (‘SN’ 12; DG 154), but his reason was unpersuasive: a self-sacrificing act can be regrettable and worth avoiding given its costs for the agent but worth choosing for itself given its motive. That aside, he thought an attitude that is appropriate to one kind of object may not be appropriate to another. This introduced an element of pluralism into his view. Some fitting-attitudes analyses, such as Sidgwick’s, Brentano’s, and Broad’s, are doubly monistic. They think all goods are in the same ontological category, for example, are states of affairs, and think the same response is appropriate to them all, either a specific attitude such as desire or a generic one such as love or any pro-attitude; though there may be different pro-attitudes, what matters about them is just that they are pro-attitudes. But some present-day versions of the analysis are pluralistic at both points. They recognize goods in different categories, for example states of affairs and also entities such as persons. And they think different attitudes are fitting to different objects, sometimes because they are in different categories—what suits a state of affairs may not suit a person—and sometimes even if they are not. The result is a pluralistic account of ‘good’, with different kinds of value defined by different fitting attitudes.13 12 13

I discuss several such factors in Virtue, Vice, and Value, pp. 118–22. Anderson, Value in Ethics and Economics, pp. 8–16; Scanlon, What We Owe, pp. 87–100.

56

british ethical theorists from sidgwick to ewing

Ewing’s view was not pluralistic in the first way, since like others in the school he thought all goods are states of affairs. But it was pluralistic in the second, since he thought that what is fitting is sometimes to choose an object for its own sake, sometimes to admire it non-morally, and sometimes to admire it morally. If moral admiration is a sui generis attitude, that makes for three kinds of value; even if it is not, there are two kinds, and Ewing’s framework allows in principle for as many as there are distinct fitting attitudes. It therefore departs to some degree from the school’s conceptual minimalism, since it recognizes if not completely distinct value-concepts, then different species of a generic concept. Ewing thought his analysis has several advantages over a non-reductive view like Moore’s, the first of which is to reduce the number of unanalysable normative concepts. It reduces ‘good’ to the same concept of fittingness that appears in judgements about how you ought to act; if Ewing’s other deontic concept of moral obligation could likewise be analysed in terms of fittingness, as in his middle works he thought it might, fittingness would be the one basic normative concept (‘SN’ 14; DG 169–70). Relatedly, he thought the view allows a better defence of metaethical nonnaturalism. By positing only one unanalysable normative concept, it generates a ‘minimum’ non-naturalist theory that admits ‘the least in the way of non-natural concepts’ (‘SN’ 14). And he considered it better for this concept to be ‘fittingness’ than ‘good’. Like some other philosophers, he denied that he was aware of an indefinable property of goodness, but he thought it impossible to deny that there is an indefinable relation of fittingness (‘SN’ 14; DG 178; E 103–4; ST 81–4). He also thought the fact that this is a relation helps explain why naturalism has been popular, since philosophers tend not to notice the importance of relations (DG 179). And a naturalist account of a person’s good as what he in fact desires involves less of an error than if ‘good’ were indefinable, since it retains half the correct analysis, its reference to a psychological attitude, and omits only the other half about fittingness (DG 178–9). Ewing also thought the analysis dissolves the dispute between ideal consequentialism and deontological views like Ross’s. It makes the claim that something is good equivalent to the claim that you ought to have pro-attitudes to it, which means in part that you ought other things equal to choose it, or have a prima facie duty to produce it. But then if acts can themselves be intrinsically good, as ideal consequentialism in principle allows, Ross’s claim that you have a prima facie duty to keep promises can be expressed as the claim that promise-keeping is intrinsically good, and ‘to give a list of the different kinds of intrinsic goods is just to give a list of our prima facie obligations’ (‘SN’ 9–11; also DG 186–96). What are usually thought of as the great alternatives of consequentialism and deontology are in fact terminological variants of the same moral view. This particular argument was too hasty, because it ignored the agent-relativity of Ross’s deontological duties, or the fact that my duty to keep my promise is stronger

‘ought’ and ‘good’

57

than anyone else’s duty to help me keep it. This feature of deontology is not captured if promise-keeping is simply or agent-neutrally good, as Ewing seemed to assume, for then everyone has the same duty to ensure that I keep my promise. It requires instead that there be agent-relative values, so my keeping my promise is a greater good from my point of view than it is from other people’s. (It also requires time-relativities, since my duty to keep a promise now is stronger than my duty to ensure that I will keep my promises in the future.) But the fitting-attitudes analysis allows relative values, as Ewing sometimes recognized. The point had emerged earlier, in the debate between Sidgwick and Moore about egoism. They both thought any duty must be to promote the good, but whereas Sidgwick thought this duty can be egoistic, so each person’s duty is just to promote his own good, Moore thought egoism is self-contradictory, because the idea that something is just ‘my good’ or only good ‘for me’ is self-contradictory (PE 97–102). This last view seems right if goodness is, as Moore thought, an unanalysable property. It is hard to see how an object can have this kind of property ‘for me’ but not ‘for you’, or ‘from my point of view’ but not ‘from yours’. Surely it either has the property or not. Compare being square. An object cannot be square for one person but not for another; it either is square or not. (It can look square to one person but not to another, but looking square is different from being square.) The same is true of goodness if that is unanalysable, which means the only duty to promote the good must be agent-neutral. But Sidgwick’s analysis of ‘good’ allows relativities, since what each person ought to desire can be different, for example it can be true that each person ought to desire just his own pleasure. This was precisely Sidgwick’s line in defending egoism; the egoist cannot be refuted if he holds only that his pleasure is good ‘for him’ rather than agent-neutrally (ME 420–1, 497–8). The same point emerged in Ross’s discussion of a duty concerning your own pleasure. In The Right and the Good he argued that since pleasure is good and there is a duty to promote anything good, you must have a duty to promote your pleasure (RG 24–6). At the same time, he thought the intuitive view is that there is no duty to promote your pleasure (RG 24–6) and he omitted that duty from his official list of duties (RG 21). In Foundations he resolved this tension by distinguishing two senses of ‘good’. In the first and ‘most proper’ sense, goodness is an unanalysable property had by virtue and knowledge, and there are agent-neutral duties to promote these states in everyone. But in a secondary sense ‘good’ means a ‘worthy object of satisfaction’ or what it is ‘right to feel satisfaction in’. Pleasure is good only in this secondary sense, and what is good relative to you is only other people’s pleasure and not your own; you should therefore feel satisfaction only in their pleasure, and your only duty concerning pleasure is to promote theirs. If goodness is unanalysable, Ross held, it must be agent-neutral, but a fitting-attitudes analysis allows agent-relativity (FE 271–83). Though he did not connect it to his argument about consequentialism and deontology, Ewing also made this point. In his first discussion he noted that, given

58

british ethical theorists from sidgwick to ewing

his analysis of ‘good’, ‘something, A, might be better than something else, B, relatively to one agent but not to another’. This would not mean ‘that it possesses one quality for me in itself and a contradictory quality for you, but simply that I ought to choose A rather than B and you B rather than A’ (‘SN’ 19–20). He was not entirely comfortable with this idea, however, and in The Definition of Good he rejected it, giving as one reason not to define ‘good’ as what you ought to desire that this would allow agent-relativity (DG 163, 192–3). But later he reverted to his earlier view, saying ‘there is no contradiction in admitting that what is really the greatest good from the point of view of A is not the greatest good from the point of view of B . . . It is only because there is still a lurking idea of good as an objective quality distinct from what we ought to prefer . . . that there seems to be a contradiction’ (ST 107–8). That, as Sidgwick and Ross saw, is correct: whereas the nonreductive view of ‘good’ precludes agent-relativity, the fitting-attitudes analysis allows what is good from one person’s point of view to differ from what is good from another’s, and therefore allows my keeping my promise to be a greater good for me than it is for you. In doing so it allows ideal consequentialism and deontology to coincide. Some more recent philosophers have proposed reducing all moral views to versions of consequentialism, using agent- and time-relative claims about value.14 Ewing gave an early version of this proposal, and though he did not include in it the required relativities, he did show how a reductive analysis of ‘good’ makes them possible.

2.4 The Fitting-Attitudes Analysis: Objections In a conversation reported by O.K. Bouwsma, Wittgenstein dismissed Ewing’s fitting-attitudes analysis, which he took to equate ‘good’ with ‘what it is right to admire’, as follows: The definition sheds no light. There are three concepts, all of them vague. Imagine three solid pieces of stone. You pick them up, fit them together, and you now get a ball. What you’ve now got tells you something about the three shapes. Now consider you have three balls of or lumps of soft mud or putty—formless. Now put the three together and mold out of them a ball. Ewing makes a soft ball out of three pieces of mud.15

These remarks do not show Wittgenstein at his best. He misrepresented Ewing’s view, which does not put three concepts together but puts two together and says they are equivalent to a third. Nor did he substantiate the charge that Ewing’s concepts are vague. What exactly is ‘muddy’ about ‘right’, ‘admire’, or, what Ewing actually used,

14 Dreier, ‘Structure of Normative Theories’; Smith, ‘Neutral and Relative Value’; Portmore, Commonsense Consequentialism. 15 Bouwsma, Wittgenstein: Conversations, pp. 41–2.

‘ought’ and ‘good’

59

‘pro-attitude’? Wittgenstein did not condescend to say. His main theme in the surrounding discussion was that ‘good’ has many uses, but Ewing recognized that and distinguished ten in The Definition of Good (112–17). Wittgenstein’s critique did not engage with his position at all. There are, however, more serious objections, some to Ewing’s specific version of the analysis and others to it as a whole. One concerns the claim that the intrinsically good is what it is fitting to have a proattitude toward ‘for its own sake’. This is meant to distinguish intrinsic from, for example, instrumental goods, which it is fitting to choose or welcome for their effects. But what is it to choose something ‘for its own sake’? If Ewing had shared Moore’s view that intrinsic goodness can depend only on intrinsic and not relational properties, he could have said it is to choose a thing for its intrinsic properties. But he rejected that view, saying something can be intrinsically good in some contexts and not others (DG 114). He instead defined intrinsic goodness negatively, as what is not good just for causing something else that is good, and this suggests a similarly negative account of ‘for its own sake’: you choose something for its own sake if you do not choose it just as a means. But this is insufficient if there are other forms of non-intrinsic goodness such as goodness as a sign, where it is fitting to welcome a thing as evidence of something else that is intrinsically good, as a smile is evidence of pleasure. Ewing could extend his negative account to exclude these other attitudes, but given the importance of intrinsic goodness there ought surely to be a positive characterization of the response it makes fitting. Ewing did not give one, and it is not clear how he could. A second issue concerns Ewing’s version of pluralism about the good. He thought some things should be chosen and welcomed for their own sakes while others call for admiration and some more specifically for moral admiration. This requires these attitudes to be distinguished from each other, and the question is how this is done. Ewing may have thought admiration differs from welcoming introspectively, or in how it feels, but that is not plausible. (It is even less plausible that moral admiration differs introspectively from admiration more generally.) And the attitudes cannot be distinguished by the type of good that is their object, since the point of the fittingattitudes approach is to analyse values in terms of attitudes. So exactly how do the different attitudes differ? This difficulty is highlighted by an argument of Ross’s. He thought knowledge and virtue should be admired but that it would not be right to admire another’s pleasure; that calls only for satisfaction. But he thought the difference between admiration and satisfaction is that the former involves the thought that its object is good in the unanalysable sense: ‘admiration is not a mere emotion; it is an emotion accompanied by the thought that that which is admired is good’, whereas ‘[w]e often take satisfaction in things that we do not think good, but only pleasant’ (FE 278–9, 282). This is a plausible claim, since admiration does seem to involve a judgement of value, but it is not one Ewing could make, since it assumes a prior understanding of

60

british ethical theorists from sidgwick to ewing

goodness. Nor could he say, as some today might, that admiration involves the thought of its object as admirable. Since the thought that something is admirable is the thought that it ought to be admired rather than, say, feared, it assumes that we have already distinguished those emotions and cannot be used to make the distinction. This difficulty does not arise for all pluralistic versions of the analysis, and in particular not for ones that distinguish attitudes by how they relate to their object. Some hold that while the fitting attitude to good states of affairs is to want to bring them into existence, that to persons is respect, which involves a desire not to harm or destroy them but no desire to create them. A similar view can be had about states of affairs, so some call for a desire to produce them but others just for a desire not to destroy them. But Ewing’s pluralism did not take this form; he seemed to distinguish admiration from welcoming in some other way. And it is not clear what, without circularity or a dubious appeal to introspection, that could be. A more general objection is that fitting-attitudes analyses rule out an attractive explanation. They treat it as a brute fact, with no deeper explanation, that you ought or that it is fitting to desire, say, pleasure for its own sake. But the objection says there is an explanation: the reason you ought to desire pleasure is that pleasure is good, and likewise for anything else you should desire. It is natural to say ‘you ought to desire x because x is good’ but not ‘x is good because you ought to desire it’, though the opposite should be true given the fitting-attitudes view. On both counts ‘good’ cannot mean ‘what ought to be desired’. In Foundations Ross made this objection for objects it is fitting to admire, saying the reason we ought to admire them is that they are good (FE 278–9, 282), but he did not make it more generally since he accepted a fitting-attitudes account of the goodness of others’ pleasure.16 The objection was made more generally by Blanshard. In a critique of Ewing he said, ‘the conviction persists that if it is fitting to favour an object, this is because the object is good already . . . It still seems to me truer to say that we favour the object because it is good than to say the object is good because we may fittingly favour it.’17 This not a completely decisive objection. It may be that though a fitting-attitudes analysis of ‘good’ is correct, many of us do not realize this and so make explanatory claims that are false. But usually when we are given a correct analysis of a concept, we see that it is correct and adjust our claims to accommodate it. This does not seem to happen with ‘good’: even when we have an analysis like Ewing’s, ‘you ought to desire x because it is good’ seems explanatory while its converse does not.

16 Ross, surprisingly, did not make the objection in his lengthy discussion of ‘good’ in RG, Chapter 4, where he argued that its only ethically relevant sense is unanalysable. But he did not really consider fittingattitude views there. He discussed the view that good is ‘what ought to be acknowledged’ (RG 111) but seems to have meant by this ‘ought to be acknowledged to exist’. 17 Blanshard, Reason and Goodness, pp. 284–6; also Zimmerman, ‘Good and the Right’, p. 347. Blanshard’s account of ‘good’ was naturalistic; Ewing criticized it effectively in ‘BVG’.

‘ought’ and ‘good’

61

Ewing addressed this objection and said the proposed explanation is not needed. A sufficient account of why you ought to desire something is given by ‘the concrete, factual characteristics of what we pronounce good’, for example by the pleasantness of a pleasant experience (DG 172; also 157, 176, ST 99). This reply may work for the claim that you ought to desire some particular object, but not for the general claim that you ought to desire, say, (any) pleasure. There Ewing’s view can say only that you ought to desire pleasure because it is pleasure, which is just to repeat that you ought to desire it. Yet it seems natural to say you ought to desire pleasure because it is good. An argument about explanation has also been made by supporters of the analysis. Imagine that some state of pleasure is one you ought to desire and choose. The fact that it has the natural property of pleasantness, they say, gives a ‘complete explanation’ of why you ought to desire it. There is no ‘further work’ that could be done by the fact that it is good; in particular, its goodness does not ‘add a further reason’ to choose it over and above its being pleasant.18 But the claim about a ‘further reason’ misrepresents the non-reductive view. It will say a state’s being pleasant makes it intrinsically good, which in turn makes it one you ought to desire and choose; given the transitivity of explanation, this means the state’s being pleasant makes it one you ought to desire and choose. The explanation in terms of goodness is not a different or additional explanation to the one in terms of pleasantness; it is a later part of the same explanation.19 But what does goodness add to this explanation, or how does it make the explanation better? The claim that pleasure is good might seem to imply that it would not be right to desire pleasure if it were not good, but since normative truths hold necessarily the antecedent of this conditional cannot be true. What goodness does add is explanatory unification. Imagine that you ought to desire pleasure, knowledge, and virtue. On the fitting-attitudes view these ‘oughts’ are irreducibly plural, with no common ground. But there is a common ground if all three states have an unanalysable property of goodness: then the explanation why you ought to desire each is that it is good. The resulting unification goes only so far, since the facts that pleasure, knowledge, and virtue are good remain plural. But the reference to goodness does unify claims about what you ought to desire, and, as we have seen, ‘you ought to desire x because x is good’ does seem explanatory. In ‘A Suggested Non-Naturalistic Analysis’ Ewing acknowledged that the nonreductive view allows a kind of unification his own view does not, but said this is no great advantage if the ‘oughts’ in his theory, though distinct, are such that fulfilling any one tends to promote the fulfilment of others (‘SN’ 15–18). He was here appealing to a distinctive claim of his moral epistemology: that it counts in favour of a moral view if its duties are mutually supporting in this way. But not only is that 18 19

Scanlon, What We Owe, p. 97; Dancy, ‘Should We Pass the Buck?’, p. 164. Zimmerman, ‘Good and the Right’, pp. 340, 344.

62 british ethical theorists from sidgwick to ewing epistemological claim contestable, it is doubtful that his own set of ‘oughts’, which was similar to Ross’s, satisfies it; after all, Ross emphasized that his duties conflict. And it is hard to see how a conceptual issue such as whether ‘good’ is definable can turn on partly empirical questions about how far, say, paying reparation promotes truthtelling or beneficence. In Second Thoughts Ewing conceded that ‘there does seem to be something plausible’ in the idea that the obligation to adopt a pro-attitude to a thing must be based on the thing’s goodness, but offered the following explanation: sometimes we ought to have one pro-attitude to an object because we ought to have another, so, for example, the reason we ought to pursue something is that we ought to welcome it once we attain it (ST 100). But this reply presupposes that the objection cannot be made about the second attitude, and it can: ‘you ought to welcome it because it is good’ seems just as plausible as ‘you ought to pursue it because it is good’. The objection about explanation raises delicate issues, and its final strength is hard to assess. But it seems to count somewhat in favour of a non-reductive view of ‘good’ that it allows explanations the attitudes view excludes. A further objection that has emerged in the recent literature is the ‘wrong-kind-ofreasons’ objection. On some present-day versions of the attitudes analysis a thing is intrinsically good if you have reason to desire it for its own sake. But imagine that an evil demon will torture you unless you want to eat a saucer of mud for its own sake. This gives you a reason to want to eat the mud, but surely does not make doing so intrinsically good.20 This objection was not discussed in Ewing’s day; it involves the kind of artificial counterexample that became common in philosophy in the second half of the twentieth century but was not much used in the first. But he could meet it by emphasizing specific features of the concept of fittingness. For even if the demon’s threat gives you a reason to want to eat the mud, that desire surely does not fit the mud. Can we explain why? The concept of fittingness connotes a kind of match between a response and its object, where the one somehow parallels or complements the other. This will be so here if the content of your attitude matches the properties of its object that are the ultimate explanation of why you ought to have the attitude. Consider another’s pleasure. The property that makes it something you ought to desire is just its pleasantness, and your desire ‘fits’ the pleasure if it is directed to it as pleasant rather than as having some other property. Then what you desire it for matches what at bottom makes it worth desiring. Or consider an act that will cause another pleasure. The reason you ought to desire it is that it will cause pleasure, and your desire fits the act if it is for it as having that instrumental property. So a fitting desire is for the property in its object that makes it desirable. But this condition is not satisfied by your desire in the demon example. There you want to eat the mud for its own sake, or

20

Rabinowicz and Ronnow-Rasmussen, ‘Strike of the Demon’.

‘ought’ and ‘good’

63

for the taste and other sensory qualities the experience will involve, but the explanation of why you should want to do that act is its having the relational property that wanting it will prevent you from being tortured. The content of your desire does not match the ultimate explanation of why you ought to have it, and the desire is not fitting.21 As so understood, the concept of fittingness allows an answer to the wrong-kindof-reasons objection, but it does so at considerable cost. For it makes fittingness not a basic or irreducible concept, as Broad and Ewing thought, but a departmental restriction of the more basic concept ‘ought’. A desire is fitting if it both is one you ought to have and meets a further condition about matching content to ground. This takes us back to the issue Sidgwick wrestled with, about whether the ‘ought’ in ‘ought to desire’ satisfies ‘“ought” implies “can”’. This issue arises for all versions of the analysis. Ross thought another’s pleasure is good in the sense that it is ‘right’ to feel satisfaction in it, where this does not require voluntary control. But he had objected that Sidgwick’s and Broad’s wider ‘ought’ is indistinguishable from ‘good’, and it is hard to see how the same objection does not apply to his own view: if a right satisfaction need not be one you can produce, how does it differ from a good satisfaction, and how is Ross’s analysis not circular? The issue is obscured in present-day versions of the analysis that use the concept of a reason but leave it unclear whether having a reason to do x requires being able to do x. This creates a dilemma. If reasons do require voluntariness, the resulting analysis does not cover all the cases it should; if they do not, it seems to explain ‘good’ in terms of ‘good’. Ewing was certainly aware of this issue. In his initial discussions he said that although the concept of moral obligation is governed by ‘“ought” implies “can” ’, that is not true of the ‘ought’ of fittingness, so that even if we cannot alter our emotions ‘at a moment’s notice’, we can ‘speak of the emotions a man ought to have in the sense of “the suitable emotions for him to have” ’ (‘SN’ 13–14; also DG 133, 150–1, 165; ST 92–4). This move avoids the first horn of the dilemma but lands Ewing on the second: if what is fitting need not be under our control, what separates the claim that an emotion is fitting from the claim that it is good, beyond a departmental restriction about content and explanation? His final discussion was somewhat different. There he took our inability to control our emotions as a reason to define ‘good’ in terms not of them but of conative attitudes that, while ‘not entirely in our power, . . . are so to a much greater extent and more directly than our feelings’. Then the deontic concept

This proposal is close to the ‘dual-role’ one suggested by Rabinowicz and Ronnow-Rasmussen at the end of their article, but with two differences. It is tied to the concept of fittingness and requires the content of the attitude to match the ultimate explanation of why you ought to have it, not just some explanation. It is therefore not open to a counterexample they raise against their proposal, where the demon will torture you unless you admire him for his determination to torture you if you do not (p. 419). Though in this example the demon’s determination explains why you should admire him, it is not the ultimate explanation, which is that his determination will make him cause you pain if you do not admire him for that property. 21

64

british ethical theorists from sidgwick to ewing

in the definition could be, not fittingness or, as he then called it, reasonableness, but moral obligation, and the good can be what we have a moral duty to seek (ST 92–3). But this took him back to the other horn of the dilemma, since ‘good’ as so defined does not apply when conative attitudes are either not possible or, as he allowed is sometimes the case, not under our control. The fitting-attitudes analysis aims to reduce ‘good’ to a deontic concept, but this concept cannot be the strict moral ‘ought’ governed by ‘ “ought” implies “can”’, since the attitudes the analysis calls for are not always under our control. And if the concept is not governed by ‘ “ought” implies “can”’, it is unclear how it differs more than departmentally from ‘good’. Broad’s and the earlier Ewing’s appeal to fittingness can be seen as an attempt to negotiate these twin demands by using a concept that is distinct both from ‘ought’, because it does not require voluntariness, and from ‘good’, so it occupies a space between the two. But does that space really exist? And do we really have an independent grasp of the concept that occupies it, rather than understanding it just as one intermediate between two more basic concepts? Our school held that ‘ought’ and ‘good’ are the primary ethical concepts, but differed about their relations. Some took both concepts to be unanalysable, so ‘you ought to promote the good’ is a synthetic rather than analytic truth, but others proposed reducing one to the other. The more plausible reduction analyses ‘good’ in terms of ‘ought’ or some other deontic concept, and though that view’s merits were claimed in the period, its shortcomings were also pointed out.

3 Kinds of Goodness and Duty Whatever the relations between ‘good’ and ‘ought’, ‘duty,’ and ‘right’, there are questions about the more specific forms these concepts can take. One concerns how intrinsic goodness, which is the central kind of goodness, differs from instrumental and other kinds. Another concerns Ross’s contrast between ‘prima facie duties’ and ‘duties proper’ or ‘actual duties’, while a third concerns objective versus subjective duties, the former depending on the truth about the situation in which you act and the latter on your beliefs about that situation.

3.1 Intrinsic Goodness The best-known account of intrinsic goodness is Moore’s, which says a state’s intrinsic goodness can depend only on its intrinsic properties and not on its relations to other states. This account was not stated explicitly in Principia but guided the book at two points. One was its method of testing for intrinsic value by asking whether a universe containing only a state such as pleasure and nothing else would be good (PE 91, 93–5, 157, 187–8, 197, 208); the point of this ‘method of isolation’ is precisely to prevent judgements of intrinsic value from being affected by relations. The other was Moore’s specific interpretation of his principle of organic unities. Stated generically, this principle says the value of a whole need not equal the sum of the values its parts would have if they existed alone, so if a and b would each have five units of value alone, the value that results when they are combined by relation R into aRb need not be ten but can be, say, fifteen.1 On what I will call a ‘variability’ interpretation of this principle, the value of a part can change when it enters a whole, so a’s value can go from five to ten when it is related by R to b. Moore rejected this possibility, saying ‘The part of a valuable whole retains exactly the same value when it is, as when it is not, a part of that whole’ (PE 30). Instead, he assumed a ‘holistic’ interpretation, on which if a whole’s value differs from the sum of the values of its parts, that is because alongside the unchanged values in those parts there is a further value in the whole ‘as a whole’ that must be added to the values in its parts to determine its value ‘on the whole’ (PE 213–22). In the example above, though a and b still have their five units of

1

Ewing gave a generic statement of the principle like this in VR 215.

66 british ethical theorists from sidgwick to ewing value, there can be a further five units of value in the whole aRb considered as a separate entity and with R now internal to it, or, in what this really amounts to, the obtaining of R between a and b. This holistic treatment of organic unities is required if a state’s intrinsic value can depend only on its intrinsic properties, and Principia duly assumed it. In his later Ethics Moore used the method of isolation to define intrinsic goodness: ‘By saying that a thing is intrinsically good [our theory] means that it would be a good thing that the thing in question should exist, even if it existed quite alone’ (E 27). And ‘The Conception of Intrinsic Value’ extracted the abstract idea underlying the method: ‘To say that a kind of value is “intrinsic” means merely that the question whether a thing possesses it, and in what degree it possesses it, depends solely on the intrinsic nature of the thing in question’ (‘CIV’ 260).2 Moore’s account seems to have been an innovation. In the first edition of The Methods Sidgwick spoke often of what is ‘intrinsically’ good or desirable; from the second edition on he dropped that language in favour of ‘ultimate good’, though it is unclear why.3 But he seems to have allowed that ultimate goodness can in principle be affected by relations. At one point he considered the view that the virtues, which are ‘largely valued as means to ulterior good’, are on that basis also ultimately good. He did not dismiss this view as self-contradictory, saying only that it is ‘difficult to conceive’ anything as ‘both means and end’ in respect of the same property (ME 396). A similar admission occurred in his discussion of goods such as knowledge. He had previously argued that the good must consist in states of consciousness (ME 113–15), but he recognized that this does not imply that only pleasure is good. Our good can in principle depend on ‘objective relations’ between our minds and external objects, as when our beliefs about those objects are true. He rejected this idea, but he did not say it contradicts his claim about states of consciousness or makes the bearers of value wholes combining, say, a belief, an external fact, and a correspondence relation between them. Instead, he treated it as making the value of the belief depend on its standing in that relation (ME 398–405). Rashdall had a similar view. He often spoke of ‘intrinsic value’, but like Sidgwick allowed it to be affected by relations, saying pleasure is good only when accompanied by virtue (TGE I 72) and knowledge and virtue good only given pleasure in them (TGE I 153, II 37–8; also TGE I 75-6, 78, II 21, 33, 39–40). Later opinion was divided. Russell and Ross followed Moore, with Ross saying ‘by calling a thing intrinsically good we mean that it would be good even if nothing else existed’ (RG 75, also 74, 114–15; Russell, ‘EE’ 37–8, 55). But others took the broader view. In a passage 2 In EE Moore had spoken of what is good ‘as an end’ or ‘in itself ’ rather than ‘intrinsically’, and had given no positive account of it, taking it to be the only goodness there is (EE 47–8, 115–20, 121–3). But even there he said that if something ‘is once good, then it is always good, and nothing can make a change in this respect’ (EE 119; also 115, 121). 3 Jones, while defending Sidgwick against Moore, was happy to talk of ‘intrinsic’ value as a concept shared by the two (‘Mr. Moore on Hedonism’, p. 429 and passim).

kinds of goodness and duty

67

directed at Moore, McTaggart said, ‘if A. loves B., what is good is not the relation between them, but the state of A. in being one of the terms of that relation’ (‘IV’ 443). Prichard accepted as ‘by no means odd’ Kant’s view that happiness is good only conditionally, or if its subject is worthy to be happy (MW 50), and Ewing said, ‘By the expression, “good-in-itself,” . . . I do not mean that anything good-in-itself would necessarily remain good if other things were different’ (MP 14n.; also 164–6); later he explained ‘intrinsically good’ by saying ‘something might really have the characteristic goodness in some contexts and yet not have it in others, or have it only in a lower degree’ (DG 114; also ST 98, 131–2; ‘U’ 107; VR 215; Laird SMT 45–6; Broad ‘RSE’ 100B). Moore’s view defines intrinsic goodness positively, in terms of its own nature, but many who took the broader view defined it only as what is not instrumentally good. In the first edition of The Methods Sidgwick said ‘intrinsic or ultimate Good’ is what is ‘Good per se, and not as a means to any further end’ (ME1 367), and he took a similar line in the final edition (ME 109); Ewing said that by ‘good-in-itself ’ he meant ‘simply “good itself,” in opposition to good as a means’ (DG 114; also MP 14n, 165–6; ‘U’ 107; McTaggart, NE II 398). These negative definitions face the difficulty, raised in the preceding chapter for Ewing’s fitting-attitudes view, that there may be other forms of non-intrinsic goodness than the instrumental. One noted by several writers is goodness as a sign, where one thing is evidence for another that is intrinsically good,4 while another is generated by the principle of organic unities. Imagine that a has no value on its own, b has five units on its own, and the whole aRb has ten units. Here a has a kind of derivative goodness, because it is necessary for the additional goodness in aRb, but this is not instrumental goodness, since a does not cause aRb. Moore distinguished this form of goodness from the instrumental (E 107), and Ross called it ‘contributive goodness’ (RG 72). A negative definition can be extended to exclude these additional forms, so the intrinsically good is what is not instrumentally, evidentially, or contributively good. But it will then draw what seems an arbitrary distinction between those relations that make for non-intrinsic goodness and those, like being the pleasure of a virtuous person, that affect intrinsic goodness. Should there not be an explanation of these differences? A negative definition also seems to get the explanatory relations backward. Surely we do not understand instrumental goodness first and define intrinsic goodness as what is not that; intrinsic goodness is primary, and we understand instrumental goodness as causing it. This requires a positive account of intrinsic goodness. The broader view can meet this demand by understanding intrinsic goodness as that portion of the overall goodness of the world that is located in a given state, or attributable to it rather than to some other. If a is a causal means to b, is evidence for b,

4

Feldman, Doing the Best We Can, p. 26; Thomson, ‘On Some Ways’, pp. 100–1.

68

british ethical theorists from sidgwick to ewing

or contributes non-causally to b, none of what makes the world good is located in a; it is all, as far as these two are concerned, in b. But if a is the happiness of a virtuous person, then even though her happiness would not be good if she were vicious, it is still the happiness that has value. The point can also be made in terms of appropriate attitudes. If a is a means to or evidence for b, we should desire and welcome a but only because we should desire and welcome b; the fittingness of the former attitudes derives from the fittingness of the latter. But this need not be so if a is good conditionally on its standing in a relation to b. Then it can be fitting to desire and welcome a even though it is not fitting to desire and welcome b, say, because b has zero value. This account of intrinsic goodness makes judgements about the values of worlds primary and ones about the values of particular states derivative, but Moore’s method of isolation likewise reaches a judgement about a particular state such as pleasure from a judgement about a universe containing only pleasure. And we do seem to have intuitive judgements about where values are located. This issue aside, the broader view of intrinsic value has the advantage of greater flexibility. Moore’s view requires all organic unities to be interpreted holistically, so the values of the parts remain unchanged and any additional value is located in the whole as a whole, but the broader view also allows the variability interpretation and even a mixture of the two. It does not reach different substantive conclusions, since the same final assessment of a complex state can always be reached in any of the three ways. If a and b would each have five units of value on their own, the whole aRb can have fifteen units either because there is an additional five units in R’s holding between them, because one of them increases its value to ten, or because one increases its value to eight, and there is an additional two units in R. But sometimes one interpretation better fits an organic value than the others. Retributive desert, for example, which makes it good if a vicious person suffers pain, is best understood holistically. We do not want to say that either the vice or the pain becomes less bad by its association with the other; each remains as evil as before. What is good is only the relational fact that they occur in the same life. But claims about the values of knowledge and achievement are better put in variability terms. If a justified belief of yours is true and therefore constitutes knowledge, the additional value this involves seems to be located in the belief, and therefore in your mind and life, rather than in a whole combining the belief, a fact, and the relation between them. Similarly, if you successfully achieve a goal, that increases the value of your activity in pursuing it. Or consider a diminishing-marginal-value claim, such as that the more whooping cranes there are in the world, the less extra value there would be if there was an additional whooping crane.5 It would be artificial to say that each crane adds the same value to the world but there is an increasingly negative value in the population of cranes as a

5

Hurka, ‘Value and Population Size’; Carlson, ‘Organic Unities, Non-Trade-Off, and Additivity’.

kinds of goodness and duty

69

whole. Better to say the value a crane adds to the world gets less the more other cranes there are, that is, to make a variability rather than a holistic claim. Moore’s narrow account of intrinsic goodness therefore brings no concrete benefits, since it makes no difference to the final values of states of affairs, while sometimes excluding the most plausible formulation of an organic-unities view; this makes the broader view preferable. Some philosophers use ‘intrinsically good’ in Moore’s way and ‘good as an end’ for the broader concept that can be affected by relations,6 but I see little point in this distinction. The two forms of goodness play the same normative role: each is to be desired and pursued for its own sake, and each contributes in the same way to the overall goodness of the world. They are therefore both best called ‘intrinsic goodness’.7 Given a positive account of intrinsic goodness, the secondary forms can be defined as what causes intrinsic good, is evidence for it, and so on. But some in the school denied that these are forms of goodness at all. In The Elements of Ethics Moore said ‘“Good as a means” means nothing but “a means to good”. . . .What is only “good as a means” is not good at all’ (EE 116). He abandoned this view in Principia, where he spoke often of what is good as a means, but it was shared by Prichard and McTaggart (MW 213–14;8 NE II 398), as well as by Ross, who called instrumental goodness ‘not goodness at all, but the property of producing something good’ (RG 133; also FE 140). This view is understandable if goodness is, as they all thought, an unanalysable property, since then what causes or is evidence for this property does not itself have it. But it is less well motivated given a fitting-attitudes analysis, since it surely is fitting to desire and welcome a cause of intrinsic goodness as causing it, evidence for it as evidence, and so on. And even if ‘good’ is unanalysable there is no serious harm in talking of instrumental and other secondary forms of goodness, as these philosophers all sometimes did.

3.2 Prima Facie Duty In The Right and the Good Ross distinguished two forms of the deontic concepts ‘duty’ and ‘ought’. The first he called ‘prima facie’ or ‘conditional’ duty, which many today would call duty ‘other things equal’ or ‘ceteris paribus’. The other was ‘duty proper’, ‘duty sans phrase’, ‘actual duty’, or ‘absolute duty’, which many would call duty ‘all things considered’ or ‘simpliciter’ (RG 19–20, 28). These phrases all point to the same distinction, which is vital for pluralist moral theories such as Ross’s. They contain multiple principles of duty, each identifying a property of acts that counts in 6 Korsgaard, ‘Two Distinctions in Goodness’. Korsgaard calls goodness as an end that depends on relational properties ‘extrinsic’ goodness, but this is misleading if it suggests that this goodness comes entirely from outside the good thing. The value of a virtuous person’s happiness may depend in part on its occurring in the same life as his virtue, but it also depends on its intrinsic qualities as happiness; only happiness and not misery is good when had by someone virtuous. 7 8 See further my ‘Two Kinds of Organic Unity’. See also his letter to Ross of 20.12.1928.

70

british ethical theorists from sidgwick to ewing

favour of their being right or, for negative duties, wrong. But an act can fall under several principles, and to determine whether it is right or wrong on balance we need to weigh the relevant duties against each other or assess the act in light of everything true about it. The initial competing duties are prima facie duties; what results from the weighing is a duty proper. Consider an example of Ross’s where you have promised to meet a friend but on the way encounter accident victims needing help (RG 18). You have one prima facie duty to keep your promise and another to help the victims, that is, you ought, other things equal, to do each of these things. But you cannot do both and must therefore weigh the duties against each other. If the prima facie duty to help the victims is stronger, your duty proper, or what you ought all things considered to do, is to help them; if the duty to keep the promise is stronger, you should do that. Ross therefore called prima facie duty a ‘parti-resultant’ attribute and duty proper a ‘toti-resultant’ one (RG 28), because the former rests on just one of an act’s properties while the latter takes account of them all. Like much in Ross’s moral philosophy, the distinction between these forms of duty derived from Prichard. It was implicit in his positive normative theory in ‘Mistake’ (MW 12–14) and was developed further in a lecture of the late 1920s that has only recently been published (MW 77–83).9 But Ross emphasized the distinction more than Prichard and elaborated it more fully, so it is rightly associated primarily with him. The distinction was also, for all its familiarity today, a major innovation. It had not been drawn by the earlier consequentialists, who accepted just one moral duty, to maximize the good, and expressed it using one unqualified ‘ought’. In any situation, they held, what Ross would call your duty proper is to produce the most good you can. They could have said that when an outcome is good but not the best you have a duty other things equal to produce it, but as Sidgwick’s discussion of ‘imperative’ concepts showed, they did not. The distinction also does not figure in deontological theories modelled on Kant’s, where the negative duties constraining pursuit of the good are absolute and cannot be outweighed. If it is never permissible to lie, even to a would-be murderer,10 the duty not to lie can be expressed using the simple ‘ought’ or ‘duty proper’. What Ross’s distinction allows is precisely a non-absolute or moderate deontology, where there are constraints on how you may promote the good but they can in some cases be outweighed. This deontology does not permit you to lie if this will have just slightly better consequences, but if lying will prevent a major disaster, such as the deaths of thousands of people, it can be permitted and even required; then the prima facie duty not to lie is outweighed by a stronger duty to prevent harm. The novelty of Ross’s distinction is shown by how often earlier philosophers wrote as if the only alternatives are consequentialism and absolute deontology. Sidgwick 9 10

H.H. Price noted Ross’s debt to Prichard on this point I; see his Critical Notice of RG, pp. 343, 354. Kant, ‘On a Supposed Right to Lie’.

kinds of goodness and duty

71

recognized that a deontological morality will contain multiple principles that, until they have exception-clauses added, must be weighed against each other. But as we will see in Chapter 7, his detailed critique of these principles often ignored the possibility of a moderate deontology, for example by moving directly from the premise that common sense takes some account of consequences to the conclusion that it considers only consequences (ME 281, 316, 348). The same slide is apparent in Rashdall and Moore. Having identified a deontological view as one that pronounces acts right or wrong ‘without reference to their consequences’ (TGE I 80; also 91–2), Rashdall stated its rule of veracity as ‘Do not lie under any circumstances whatever’ (TGE I 92; also 193–4) and took ‘objections to that sort of Intuitionism which declares that certain rules of action are to be followed irrespectively of consequences’ to show that we must ‘accept the utilitarian formula in so far as it asserts that conduct is good or bad only in proportion as it tends to promote the Well-being of human society’ (TGE I 91; also 83–4; E 57–60; ICE 41).11 Moore similarly contrasted his consequentialism only with ‘the strictly Intuitionistic view that certain ways of acting [are] right and others wrong, whatever their results might be’ (PE 106; also 148; EE 171–3; E 74–7), again ignoring the moderate view intermediate between those two. But that view is easy to ignore if one lacks the concept of prima facie duty. Even some deontologists lacked this concept. In a book published just before The Right and the Good Carritt denied that there can be conflicts of duties, saying ‘these duties which cannot be done if I fulfil my paramount duty are really only hypothetical duties, which I ought to have fulfilled in other circumstances’, and are called duties only ‘by an empirical generalization because generally . . . we have judged it our duty to keep our promise or obey our parents or the law’ (TM 106, also 114). As Ross emphasized, however, a key feature of a prima facie duty is that it is still present when it is outweighed. It remains ‘an objective fact involved in the nature of the situation’ (RG 20), and its doing so has consequences, or leaves traces. It makes it appropriate to feel, if not guilt, then ‘compunction’ about not fulfilling the outweighed duty, and can ground a further duty to compensate the person to whom the duty was owed (RG 28). Even Ross, in his earlier book Aristotle, seems not yet to have had the concept. He there imagined Aristotle’s virtuous agent deciding how to act by consulting a ‘right rule’ that tells him what to do now (A 195, 215, 221). But if the rule were only one of prima facie duty it could not do that, since it would have to be weighed against other rules.12 Ewing was by no means exaggerating when he said Ross ‘made one of the most important discoveries of the century in moral philosophy in recognizing the fundamental character of these prima facie duties’ (ST 126). As Ross recognized, ‘prima facie’ is not the ideal term for this concept. It suggests that its referent is a kind of duty proper rather than something related to duty proper, 11

Rashdall recognized the possibility of non-absolute deontology in his later E, p. 46. For another deontologist of the period who seems to have assumed his principles must be absolute see Stocks, ‘Morality and Purpose’, pp. 65–7. 12

72

british ethical theorists from sidgwick to ewing

and also suggests that it is merely an appearance of duty, which Ross insisted it is not (RG 20). Prichard had used the term ‘claim’ (MW 79), but Ross thought that describes matters from the point of view of the person affected by an act rather than from the agent’s point of view, and does not fit self-regarding duties such as to seek knowledge and virtue in oneself (RG 20; FE 85). In The Right and the Good he stuck with ‘prima facie duty’, though in Foundations he preferred Carritt’s suggestion of ‘responsibility’ (FE 85). The latter term has not caught on, however, and today this form of duty is mostly called ‘prima facie’ or sometimes, in more accurate Latin, ‘pro tanto’. The question, however, is how the concept of prima facie duty is to be understood, and in particular how it relates to duty proper. Here Ross gave two different accounts. In The Right and the Good he defined prima facie duty in terms of duty proper, which he took to be the primary deontic concept. He first said ‘prima facie duty’ refers to ‘the characteristic . . . which an act has, in virtue of being of a certain kind (e.g. the keeping of a promise), of being an act which would be a duty proper if it were not at the same time of another kind which is morally significant’ (RG 19). As a formal definition this would be circular, since it says a prima facie duty would be a duty proper if there were no other, competing prima facie duties.13 And it is really a consequence of a more basic idea Ross introduced later in this book, of prima facie duty as a tendency to be duty proper: ‘We have to distinguish from the characteristic of being our duty that of tending to be our duty. Any act that we do contains various elements in virtue of which it falls under various categories. In virtue of being the breaking of a promise, for instance, it tends to be wrong; in virtue of being an instance of relieving distress, it tends to be right’ (RG 28). By ‘tendency’ Ross meant something quasi-causal. That an act is an instance of promise-breaking counts in favour of its being wrong, in the sense of tending to make it wrong or potentially explaining why it is wrong; that it is an instance of relieving distress tends to make it right. So prima facie duty is defined as helping to determine duty proper, and it follows that any act that falls under just one prima facie duty must be your duty proper, as his first definition said, since then its right-making tendency is unopposed. Ross illustrated this tendency idea by an analogy to physical forces: ‘Qua subject to the force of gravitation towards some other body, each body tends to move in a particular direction with a particular velocity; but its actual movement depends on all the forces to which it is subject’ (RG 28–9; also FE 86). That a force pushes a body eastward gives the body a tendency to move east, in the sense that the force will make it do that if no other forces intervene; if they do intervene, its movement will depend on the directions and strengths of them all. The force is a real presence, with real causal power even when it is outweighed by contrary forces, and a prima facie duty is likewise a real presence even when it is outweighed. It is still there ‘pushing’ an act to be right even when that is not the final result.

13

Dancy, Moral Reasons, p. 97.

kinds of goodness and duty

73

This tendency account of ‘prima facie duty’ fits some of Prichard’s remarks about the concept. In the lecture that talked of ‘claims’ he said that when an act is right, some aspects of its circumstances ‘render us bound to do’ it or ‘render the particular action a duty’(MW 77). The account also figured in Ross’s first mentions of prima facie duty, in two articles of the late 1920s. Each equated prima facie duty or rightness with a ‘tendency to be our duty’ or ‘tendency to be right’, and explicated that with the analogy to physical forces (‘BO’ 126–7; ‘IME’ 95–6). Broad too spoke of ‘tendencies to be right’ or ‘right-tending’ characteristics and compared them to forces (FT 222; CE 229–30). In one respect this comparison is misleading. If a body is subject to a strong force pushing it north and a weak one pushing it east, the result is movement in a direction slightly east of north, so the weak force affects the final outcome; this is possible because there are infinitely many directions a body can move. But there can only be two results of a conflict of prima facie duties: that an act is simply right or that it is simply wrong. If the duty to keep a promise is outweighed by the duty to help accident victims, the result is that helping the victims is simply right, just as it would have been with no promise in play. But the disanalogy is not decisive, since the two cases are similar in other respects. And Ross thought an outweighed prima facie duty does make a difference, if not to the rightness on balance of an act then by making it appropriate to feel compunction and sometimes to pay compensation. Despite Ross’s clarity, his view of prima facie duties has been misunderstood. In a 1949 article P.F. Strawson said that to say ‘promise-keeping tends to be right’ is to say more acts of promise-keeping are right than are wrong; he then argued that this statistical claim cannot be a necessary truth as Ross and other intuitionists held.14 But Ross did not mean anything remotely statistical by ‘prima facie duty’. He meant that every act of promise-keeping has a property that tends to make it right, and this is compatible with most such acts’ being wrong; this will be so if this right-making property is usually accompanied by a stronger wrong-making one. (Compare an eastward force that is usually accompanied by a stronger westward one.) Ewing made this reply to Strawson (ST 109–11). John Searle took Ross’s ‘official’ view to be that prima facie duties merely seem to be duties, and proposed as an improvement the idea that they are truths about what you ought to do ‘other things being equal’ that contribute to conclusions about what you ought to do ‘all things considered’.15 But what Searle called Ross’s ‘official’ view is one Ross explicitly rejected in both his books (RG 20; FE 85), while the supposed ‘improvement’ is the view The Right and the Good clearly affirmed. If Ross initially defined prima facie duty in terms of duty proper, he took a different line in Foundations, treating prima facie duty as a primitive concept and

14

Strawson, ‘Ethical Intuitionism’, pp. 29–30.

15

Searle, ‘Prima Facie Obligations’.

74

british ethical theorists from sidgwick to ewing

suggesting that duty proper is derivative. Though he did not note the change, it was substantial. The new view was borrowed from Broad’s Five Types, published in the same year as The Right and the Good and presenting similar ideas in different language. Broad first said that to call something ‘right’ is to think of it ‘as a factor in a certain wider total situation’ and to mean ‘that it is “appropriately” or “fittingly” related to the rest of this situation’ (FT 164). Though he did not intend this as a definition of rightness, it makes rightness on balance a matter of fittingness. But later in the book he said an act’s rightness depends on two factors, its ‘fittingness or unfittingness to the total course of events as modified by it’ and its ‘utility or disutility’, or the goodness or badness of its consequences; here fittingness is just one of two factors relevant to rightness (FT 218–20). But he thought even this narrower fittingness can be the result of several component fittingnesses. If you are an elector to an office and one candidate has previously done you a favour, preferring him to a better-qualified candidate would fit one aspect of the situation, since it would show him gratitude, but not others, since it would be an act of bad faith to the institution that hired you and unjust to the other candidates (FT 219). Broad seems to have thought judgements about these individual fittingnesses are primary, since an act’s fittingness on balance is a ‘function’ of its individual fittingnesses (FT 220); we understand fittingness first with respect to specific aspects of a situation and only derivatively to the situation as a whole. In Foundations Ross used these ideas to explain the concept of prima facie duty. Now he said you have a prima facie duty to do an act of a given type if that would be fitting to an aspect of your situation, for example that you promised to do an act of that type. You may have a prima facie duty to do a different act if that would fit another aspect of your situation, and the act that is right on balance is the one that best fits your situation as a whole, or weighing all the different fittingnesses against each other (FE 51–3, 79–82). He rejected Broad’s later distinction between fittingness and utility, saying an act’s having good consequences must be a kind of fittingness if rightness in general is a kind of fittingness (FE 81). But by using Broad’s ideas he gave an independent characterization of prima facie duty, one not defining it in terms of duty proper, and the concept so characterized again seems primary. We do not first understand fittingness to a whole situation and then understand fittingness to an aspect of it as contributing to that; we first understand fittingness to an aspect. If this is right, Ross’s account in Foundations runs in the opposite direction from that in The Right and the Good: from prima facie duty to duty proper rather than vice versa.16 This raises questions about what, in Foundations, the role of duty proper is. There are, unfortunately, two different things ‘duty proper’ can mean: duty all things considered and duty simpliciter. The first is, like prima facie duty, a conditional 16

This may not be entirely true of FE, given the passage on p. 86 that again explains prima facie duty in terms of a tendency toward absolute duty, with an analogy to natural forces.

kinds of goodness and duty

75

concept; it says you ought to do an act given certain facts about it, though these are now all the facts rather than just one or a few. The second concept is unconditional, saying simply that you ought to do the act. But both concepts, unlike that of prima facie duty, do not admit of degrees. A prima facie duty can be stronger or weaker, but an act either is your duty all things considered or simpliciter or it is not.17 Ross did not distinguish these meanings of ‘duty proper’, and it is unclear clear which he intended in The Right and the Good. His calling duty proper a toti- rather than parti-resultant attribute may suggest duty all things considered, but his contrasting it as ‘duty sans phrase’ and ‘absolute duty’ with ‘conditional duty’ suggests duty simpliciter. And the latter seems on balance the better reading, so The Right and the Good treated duty simpliciter as the primary deontic concept and defined prima facie duty as helping to determine that. A prima facie duty does this by first contributing to a judgement about duty all things considered, which means Ross’s view requires the additional principle that you ought simpliciter to do what you ought all things considered to do. But if prima facie duty is defined as contributing to unconditional duty, this principle is analytic. If all an act’s morally relevant properties taken together tend to make it right, it follows that it is right, just as if the net effect of the forces acting on a body is to push it a little east of north, it necessarily moves east of north. But the situation in Foundations is different. Now prima facie duty is the primary concept, and duty all things considered can be defined as resulting from that. But what of duty simpliciter? How is it to be understood, and is it needed? Prichard raised this issue in his correspondence. In a 1938 letter to H.J. Paton he said he used to accept a view like Ross’s, with duties proper determined by prima facie duties, but came to doubt it after Ross published his book. That view, he now said, requires two distinct senses of ‘ought’ and ‘duty’, the prima facie and the proper, and he found that ‘very difficult’; it also requires a principle linking them, which he thought cannot be given.18 These difficulties do not in fact arise for the view in The Right and the Good, which defines prima facie duty in terms of duty proper, making the two not really distinct. But Prichard often treated the concept of prima facie duty or claim as primitive, saying to one correspondent that he could not explain what it is.19 That had also become Ross’s view in Foundations, and it does raise Prichard’s difficulties. Though ‘duty all things considered’ can be defined in terms of prima facie duty, duty simpliciter cannot. What then is its role? Prichard’s response was to abandon the concept of duty proper. As early as 1931 he wrote to Ross saying, ‘At present . . . I am wholly out for the view that what there is

17

Of two acts that are all things considered wrong, one can be more seriously wrong; thus murder is more seriously wrong than breaking a promise. But in a central sense of ‘wrong’ an act either just is wrong or just is not. 18 Prichard, letter to Paton of 31.1.1938; also letter to Laird of 8.10.1938. 19 Prichard, letter to Collingwood of 10.2.1933.

76

british ethical theorists from sidgwick to ewing

is not the act which I ought to do but acts which I ought to do (i.e. your prima facie duties not your duty sans phrase).’20 Later he said he was ‘disposed to think that your prima facie duties are the only duties there are’ and that the most we can say about an act is that it is your strongest prima facie duty.21 The same view appears in the one letter of his that has been published, to Ross in 1932: the more I consider it the less I can make sense out of ‘the act which I am bound to do’—as distinct from ‘an act which I am bound to do’—and the more I get to think that the only fact corresponding to the phrase is ‘the act which I ought to do more than I ought to do any other’, and that your ‘a prima facie’ duty is really a duty, your ‘my duty sans phrase’ is, really that of a man’s duties which he most ought to do (MW 287).

He was here rejecting the concept of duty simpliciter, and may also have been rejecting duty all things considered. Though it is conditional, the latter concept is not comparative; it says that given all the facts about an act, you ought noncomparatively to do it. But Prichard wanted to use only comparative concepts such as that of a stronger or your strongest prima facie duty; this resulted in a view structurally similar to the ‘scalar consequentialism’ defended by some recent writers. Consequentialist views start by ranking acts by the amount of good they will produce: this one will produce more, this one less, and this one the most possible. They then usually make a cut on this ranking and say only acts above the cut are permitted or right. Maximizing views make their cut at the top of the scale, so an act is right only if no alternative has better consequences; satisficing ones count acts as right so long as they come reasonably far up the scale. But some philosophers worry that any such cut is arbitrary and impossible to justify; they propose abandoning judgements of rightness in favour of a purely comparative or scalar consequentialism that can say one act has better consequences than another or even the best consequences possible, but not that any act is non-comparatively right or wrong.22 Prichard’s later view is similar: a ‘scalar deontology’ that talks only of degrees of deontological duty but never of non-comparative duty. One act can be more oblitatory than another but never simply obligatory. A common objection to scalar consequentialism is that it is not sufficiently actionguiding, and the same difficulty arises for Prichard’s view. If all we can say is that one act is more obligatory than another, we have not decisively recommended the first act over the second; each remains morally permitted. But surely a moral view should be able to make the stronger recommendation, and that requires the non-comparative claim that the first act is simply right. In fact, if judgements about degrees of value or obligatoriness are not tied to judgements of non-comparative rightness, it is unclear that they guide action at all. If the claim that one act is more obligatory than another

20 21 22

Prichard, letter to Ross of 9.9.1931. Prichard, letter to Ross of 28.1.1938; also letter to Ross of 14.7.1938. Slote, Common-Sense Morality, Chapter 5; Norcross, ‘Scalar Approach to Utilitarianism’.

kinds of goodness and duty

77

does nothing to support the conclusion that you ought simply to do the first act, how does it direct your action at all? If we do make non-comparative deontic judgements, there are two options, represented in Ross’s two books: to treat duty simpliciter as the primary concept and define prima facie duty and duty all things considered in its terms, or to take prima facie duty as primary and, perhaps dispensing with duty simpliciter, see duty all things considered as derived from prima facie duties. Both views are determinately action-guiding, but there are several reasons to prefer the former. There are, first, objections to the specific version of the second view in Foundations. If its account of prima facie duty in terms of fittingness is to be informative, the concept of fittingness must have independent content: that doing x fits an aspect of your situation cannot mean just that in your situation you ought other things equal to do x. Fittingness has some distinctive content, suggesting a kind of complementarity whereby one thing matches another or fills a normative gap the other created, and this complementarity is present in some prima facie duties. Fulfilling a promise fits the fact that you made it, by satisfying a specific demand the promise created; likewise for the duty of gratitude, which matches a benefit in one direction with a benefit in the other, and for the duty to compensate for harm. But it is harder to see independent content in the claim that an act fits the fact that in your situation it will produce the most good. Here there seems to be no distinctive complementarity, or nothing beyond the fact that you ought other things equal to maximize value. Broad sensed this in his later distinction between fittingness and utility, and it echoes our discussion in Chapter 2 of fittingness as used to analyse ‘good’. There I said fittingness is not a primitive concept but involves an ‘ought’ with a distinctive ground, involving matching, and the same conclusion seems warranted here. If the prima facie duty to promote the good does not involve any complementarity, there must be a generic ‘ought’ present in it and all other prima facie duties. Some duties, for example about promise-keeping and gratitude, do involve fittingness, but only because they involve a prima facie ‘ought’ with the right basis. But fittingness does not explain the prima facie ‘ought’; it presupposes it. There is a further objection if the fittingness used to explain prima facie duty is the same as that used to analyse ‘good’, as Broad and Ewing proposed. Since our desires and pro-attitudes are not under our control, the fittingness relevant to ‘good’ cannot be governed by ‘“ought” implies “can” ’. But just as you cannot have a duty proper to do x unless you can do x, so you cannot have a prima facie duty to do it unless you can. So does the concept of fittingness require voluntariness or not? It may be said that while the concept itself does not require voluntariness, it is true that you ought other things equal to do x only if doing x is fitting to an aspect of your situation and you can do it. But this makes the ‘“ought” implies “can” ’ condition an external addition to the normative content of ‘ought other things equal’ when it seems to be internal. It seems part of the core of any such ‘ought’ that you can act on it, and this is not true if the prima facie ‘ought’ involves the same fittingness applied to desires.

78

british ethical theorists from sidgwick to ewing

The second view can avoid these objections by not explaining prima facie duty in terms of fittingness, and the simplest option here is to treat it as a pure primitive, which cannot be explained at all. This was at times Prichard’s view, and it makes for a direct opposite to the first view, on which duty simplicter is a pure primitive. But of the two concepts, duty simpliciter seems to me the better candidate for this role. If we reject scalar views, the concept of prima facie duty is a contributory concept, in that judgements about it contribute to conclusions about duty all things considered or simpliciter. And it seems odd to treat a contributory concept as primary. We do not treat the concept of force as a primitive and understand the movement of bodies just as what results from the interaction of forces; we understand movement independently and forces as contributing to that. We likewise understand intrinsic goodness first and the forms that promote it, such as instrumental goodness, derivatively. Likewise with duty it is natural to treat duty proper, and in particular duty simpliciter, as the primary concept and define prima facie duty as what helps to determine that. The result of a contribution seems conceptually prior to what contributes, and we acknowledge this if we define prima facie duty in terms of duty simpliciter. That was Ross’s view in The Right and the Good, and though he was later influenced by Broad to take a different line his first thoughts were in my view his best. His introduction of the concept of a prima facie duty was a huge contribution to philosophical ethics, giving precise expression to an idea implicit in everyday moral thought and showing clearly how, as earlier philosophers had not seen, non-absolutist deontology is possible. But the best understanding of this concept does not treat it as a primitive, instead defining it by its contribution to judgements about duty simpliciter that are the primary deontic ones.

3.3 Objective vs. Subjective Duty Your belief about what is right in a situation depends on others of your beliefs, both non-moral and moral. It depends, first, on your beliefs about empirical questions such as what consequences the different acts available to you will have and, if you are a deontologist, whether a possible act will do or allow harm. If you derive your belief about what is right from general moral principles, it also depends on what you think the true principles are. Since any of these beliefs can be false, two types of deontic judgement are possible. An objective judgement assesses acts in light of the actual facts, either non-moral or moral; a subjective judgement assesses them in light of your beliefs or perhaps your evidence about them. If you have the justified belief that a pill will cure a friend’s illness when in fact it will kill her, giving her the pill is objectively wrong but can be subjectively right. Is one of these the true judgement of rightness? Or are they just different judgements, each appropriate to its own context? Sidgwick drew one distinction between objective and subjective rightness. He defined utilitarianism as the view that ‘the conduct which, under any given

kinds of goodness and duty

79

circumstances, is objectively right, is that which will produce the greatest amount of happiness on the whole’ (ME 411; also EEM 3). But he recognized that many nonutilitarians say it can never be right to do what you believe is wrong; he therefore called such action ‘subjectively wrong’ even when it is ‘objectively right’ (ME 207). This distinction is not relevant to first-person deliberation, since you cannot there distinguish what is right from what you believe is right. But Sidgwick thought it can make a difference in a third-person case where you can influence someone not to do an act he thinks is right but you think is wrong. He thought common sense usually condemns this influence unless the objectively wrong act is very seriously wrong,23 but said the subjective view is ‘too simple to admit of systematic development’ and concluded that ‘our investigation must relate mainly to “objective” rightness’ (ME 207–8). Later he added that subjective rightness in his sense presupposes objective rightness: when it says you ought to do what you believe you ought, the second ‘ought’ must involve an ‘objective standard’ (ME 394). Rashdall sometimes made objective-sounding remarks such as ‘actions are right or wrong according as they tend to promote or to diminish universal Well-being’ (TGE I 100; also I 91, 93, 101, 184, 217, 219, II 38; E 60, 71), but at others said the right act is ‘that which (so far as the agent has the means of knowing) will produce the greatest amount of good’ (TGE I 184; also 81, 87–8, 185; E 60; ICE 41). Moore, however, recognized only objective rightness, in Principia defining the right as what will actually produce the most good and in Ethics defending this view against subjective alternatives. If you do what in fact has bad consequences because you sincerely and justifiably think it will have good ones, he wrote, you are not to blame and we may be tempted to say you acted rightly. But ‘wrong’ and ‘blameworthy’ are different concepts, and the truth is that what you did was wrong (E 80–2).24 This is not a decisive argument, since there can be more than one kind of wrongness and the features that make for blameworthiness can also make for wrongness in a subjective sense. Russell took this line, in ‘The Elements of Ethics’ recognizing several senses of ‘right’, from a fully objective one in which the right act is that which actually has the best consequences to a fully subjective one in which it is whichever act you believe is right. He thought the most plausible of the more objective views equates the right act with the one that, given all the available evidence, has the highest probability of producing good consequences, while on the best subjective view it is the one that, given your beliefs about moral principles, you would think objectively right if you considered the question properly. But his main point was that there are different senses of ‘right’ that can yield different verdicts on the same act (‘EE’ 25–36). 23 Broad was more favourable to acts that prevent another from doing what he falsely believes is right (CE 153–4). 24 In EE Moore argued that the view that it is right to do what you think is good leads to ‘Antinomianism’, or the view that murder can be right if done from conscience (161–2). Though a good reason not to take this subjective rightness to be the only kind of rightness, this does not show that it is not one kind of rightness among others.

80

british ethical theorists from sidgwick to ewing

In The Right and the Good Ross recognized only objective rightness and, echoing Moore, said the idea that it is right to do what your evidence suggests is right mixes the concept of rightness with ones about praiseworthiness and moral goodness (RG 32). That was also Prichard’s initial view (MW 82), but in ‘Duty and Ignorance of Fact’ he reversed his position and argued that the only true rightness is subjective. Ross was persuaded by Prichard’s arguments, and with frequent references to them now also defended a purely subjective view (FE 64, 159–68). Prichard began by noting that there are arguments for each of the competing views. It counts in favour of the objective view that it makes our duty depend on the facts about our situation and not, as seems less plausible, on our beliefs about it (MW 87). The objective view also fits several aspects of our everyday practice. Often we are uncertain of our duty because we are uncertain of relevant non-moral facts, and to resolve our uncertainty try to discover them (MW 87; Ross, FE 147). We also often try to change another’s mind about his duty by changing his mind about the nonmoral facts, for example about what a candidate for political office will do if elected. This would be pointless if the subjective view were true, since then an act based on false beliefs could still be right. So our advising practices seem to assume the objective view (MW 92–3; Ross, FE 151). On the other side, the objective view implies that we can never know which act is our duty, since that would require knowing, among other things, what consequences different possible acts will have, which we can never do (MW 87–9; Ross, FE 149–51). This objection assumes a conception of knowledge as requiring certainty and allows that you can have a ‘strong opinion’ or even a justified belief about your duty (MW 88–9; Ross, FE 150). But it is close to an important present-day argument for subjectivism: that objective rightness cannot always be used in deliberation. If you are uncertain what effects the acts possible for you will have, you cannot follow the rule ‘Do what will actually have the best effects’. What you can follow is only a rule relativized to your beliefs, in particular your beliefs about the probabilities that acts will have certain effects. If you can know what your beliefs are, you can know what is right relative to them, or subjectively. Prichard also thought the objective view has counterintuitive implications. Consider whether, when driving, you ought to slow down before entering a main road. On the objective view the answer depends on whether there actually is traffic on the road. If there is none, you do nothing wrong if you enter at full speed, and can congratulate yourself afterward for acting rightly. But Prichard thought that if you cannot tell in advance whether there is traffic, you have ‘definitely a duty to slow down’ (MW 93; Ross, FE 152). This objection too anticipates a present-day one, which likewise says the objective view gives the wrong verdict in cases involving probabilities, though the leading case now involves three options, where the intuitively right choice among them is certain to be objectively wrong.25

25

Jackson, ‘Decision-theoretic Consequentialism’, pp. 462–3.

kinds of goodness and duty

81

Prichard initially concluded that there is an ‘impasse’, with each view having some strengths and some weaknesses (MW 95). He could then have said there are two concepts of rightness, each appropriate in a different context and not in conflict with the other; perhaps subjective rightness best fits first-person deliberation and objective rightness external assessment. But in 1928 he had said, ‘There are not two senses of the term duty, and plainly, a given action is either a duty or not’ (MW 82), and in ‘Duty and Ignorance of Fact’ he continued to ask which concept of duty is the true one. Ross likewise said ‘the question remains, which of the characteristics—objective or subjective rightness—is ethically the more important, which of the two acts is that which we ought to do’ (FE 147). Prichard then gave an argument for the subjective view, but despite Ross’s calling it ‘conclusive’ (FE 148) it was strangely indirect. His argument started with a claim on a different topic: the content of our duties, whether objective or subjective. A natural thought is that a duty is always to bring about an effect in the world, and this had been Ross’s view in The Right and the Good; if you have promised to return a book to your friend, your duty is to see to it that he actually gets the book (RG 42–3). But Prichard argued that producing effects of this kind is not something you can strictly do; if you put the book in the mail, it will reach him only if the postal service works, he has not moved, and so on. What you strictly cause must be something ‘wholly due’ to you (MW 85; Ross, FE 148), and that can only be your ‘willing’ or ‘setting yourself ’ to get him the book, or, more precisely, your setting yourself to perform the bodily movements that will, in the right external circumstances, result in his getting the book. So the true content of your duty is to perform an inner mental act of willing or setting yourself to act (MW 84–7, 95–8; Ross, FE 148–9, 153–4). Especially when it says an effect must be ‘wholly due’ to you, this argument assumes the ‘contributory consequences’ or ‘share-of-the-total’ view of causal responsibility. On this view, if two events, each necessary for an effect but neither sufficient, combine to produce it, each is responsible for only part of the effect, for example, for five of its ten units of value if that is its total value. This is a mistake. If the other necessary conditions are already in place, for example if the postal service will work, your putting your friend’s book in the mail makes all the difference between his getting it and not and is therefore responsible for all that effect. Given your choice to mail the book, the postal service’s working also makes all the difference between his getting it and not, and therefore also is responsible for all that effect. But there is nothing contradictory in two causes’ each being fully responsible for an effect, and therefore nothing problematic in saying your duty is to ensure that your friend gets his book.26 This will in fact be your duty only if the

26 If each of two individually sufficient but not necessary events cause an effect, for example, if two people each cast a vote when only one vote is needed, each makes no difference to the effect’s coming about and has no responsibility for it; this again is not contradictory. See Regan, Utilitarianism and Co-operation, pp. 13–17; Parfit, Reasons and Persons, pp. 67–70.

82

british ethical theorists from sidgwick to ewing

postal service will work, and you may not know whether this is so, but an objective view allows that you can have duties you do not know you have. So Prichard’s initial conclusion, that your duty is only to set yourself to do an act, seems to rest on a false view about responsibility. Even if it is granted, however, this conclusion is, as Prichard recognized, neutral between the objective and subjective views. In his election example, the objective view will say your advisee should set himself to vote for the candidate who is in fact best, the subjective view that he should set himself to vote for the candidate he believes is best. But Prichard thought the conclusion nonetheless favours the subjective view: once it has become common ground that the kind of activity which an obligation is an obligation to perform is one which may bring about nothing at all, viz. setting ourselves to bring about something, we are less inclined to think that, for there to be an obligation to perform some particular activity, it must have a certain indirect effect. To this extent the modification diminishes the force of the objective view without in any way impairing that of its rival. (MW 98)

It is hard to see his point here. Why should the fact that what you have a duty to do is of a type that can have no effects prevent us from believing that, in order to be your duty, it must have certain effects? The argument seems a non sequitur. Even if it were sound, the argument would not address what Prichard called the ‘outstanding difficulty’ of the subjective view: that it makes your duty depend not on the properties an act would have if you did it but on your beliefs about those properties. He tried to meet this difficulty by appealing to his view that duty is a property not acts but of agents, so to say you ought to do an act is to assert a fact about you. If duty is a fact about you, he argued, there is nothing to prevent its depending on your having certain beliefs, and in fact it must depend on some property of you. So the apparent difficulty facing the subjective view is ‘the result of a mistake’ (MW 99–100). This argument too is puzzling. Even if duty is a property of you, why can your being required to do a given act not depend on that act’s having some property? And even if it can depend only on a property of you, why must this be a purely internal property such as your having certain beliefs, rather than a relational one such as your being able to produce certain effects, or being such that if you set yourself to produce these effects you will? It is again hard to see how Prichard’s argument does anything to favour the subjective view. In Foundations Ross amended Prichard’s view at a key point. Prichard had made subjective rightness relative only to your non-moral beliefs and not to your conclusions about what is right in your situation; he positively rejected the latter view (MW 99; also Broad, CE 25–6). But Ross thought rightness should be relative to an agent’s ‘complete state of opinion’, both non-moral and moral, so what is right is ‘the selfexertion which he thinks to be morally most suitable in the circumstances as he takes them to be’, as on Sidgwick’s ‘subjective’ view (FE 161–2). He recognized that this

kinds of goodness and duty

83

view invites the objection that it lets your believing an act is right, however unreasonably, make the act in fact right, and he agreed that nothing can have a property by being thought to have it. But he argued that this is not the case here: your thinking an act suitable to your beliefs about what your circumstances are makes it suitable to your belief about which act is suitable to your circumstances, which is a different belief (FE 162). But this complicated answer succeeds only if you accept the subjective view of rightness and relativize your initial judgement of suitability to your beliefs about the non-moral facts; it does not apply if your initial judgement is that an act is right simply or objectively. And it is hard to see how the answer applies to beliefs about moral principles. Imagine that you believe the right act is always the one that will produce the most pleasure, and use that principle to arrive at specific moral conclusions. Is your belief in that principle not self-validating on Ross’s view? There is a view of subjective rightness that is like Ross’s in relativizing rightness to both non-moral and moral considerations but not open to this objection. Recall Russell’s preferred subjective view, which looks not to your actual beliefs about the non-moral facts but to your evidence about them, or to what you would believe about them if you reasoned from your evidence in the best way. A similar view can be taken of your beliefs about moral principles, so what is subjectively right depends on whatever beliefs about principles are justified by your moral evidence, or are those you would have if you reasoned properly about principles. These beliefs need not be true, if your moral evidence is misleading, nor the ones you actually have, if you reason badly. But on the resulting doubly subjective view the right act is the one identified as right by the combination of beliefs about non-moral facts and moral principles you would have if you reasoned correctly from both your non-moral and moral evidence.27 This is an attractive view of subjective rightness, especially if subjective rightness is connected, as many philosophers believe, to praise and blame. If it is appropriate to blame you for acting on false non-moral beliefs when you could and should have had true ones, the same should hold for false beliefs about principles. But the view is not open to the same objection as Ross’s, since it does not make any moral beliefs selfvalidating. At the same time, however, it gives no support to the claim that subjective rightness is the only true rightness. The doubly subjective form may again be just one form of rightness alongside other, more objective ones. A final discussion was Ewing’s. Like Russell, he was most concerned to distinguish senses of ‘ought’ and ‘right’, and in The Definition of Good identified three. The first is a fully objective sense in which what you ought to do depends on all the facts, both moral and non-moral, of your situation. Next is a fully subjective sense, connected to ideas about blame and in which you ought to do whatever you believe you ought to do. A third sense is tied to your evidence, so the act you ought to do is the one it

27

Smith, ‘Consequentialism’.

84

british ethical theorists from sidgwick to ewing

seems preferable to choose in the light of the data, for example about its consequences, that you have or could get without more trouble than it is worth (DG 118–32).28 Ewing thought the first, fully objective sense is rarely applied to particular acts except by philosophers like Moore (DG 118–19; E 68), while the third is the most common (DG 121, 128). But he did not think there is any one true sense; there are just different ‘oughts’, and what is important is not to confuse them with each other (E 67–8; also Carritt, EPT 21–2). He thought the first or Moorean sense can be used to state general prima facie duties (DG 119–20) and is needed to explain the importance of the third sense, since the reason you ought to do acts that are right in the evidence-based sense is that you will then more likely approximate to what is right in the objective sense (DG 128–9). He also argued, like Sidgwick, that the fully subjective sense presupposes some other sense. The claim that you ought to do what you believe you ought to, even when you are wrong about what you ought to, would be ‘absurd’ or involve ‘self-contradiction’ if its second ‘ought’ did not have a different sense from its first (‘SN’ 3; DG 120–1). Ewing thought the required sense is the third, evidence-based one, but it could equally well be the first, fully objective one. Either way, the content of the belief about duty which the fully subjective sense says you ought to act on must involve duty in an at least partly objective sense. At one point Ewing said his senses of ‘ought’ do not involve distinct ultimate concepts, just the same concept applied in different relations (ST 104). But there and elsewhere he drew a distinction. He said his first and third senses involve Broad’s concept of fittingness, saying an act is fitting either to your situation or to your evidence about it (‘SN’ 4; DG 132–3), but he thought the second sense involves a different concept of moral obligation, tied to concepts of sin and blameworthiness. (He sometimes wrote as if you are to blame only for doing what you believe is unfitting (‘SN’ 2–3; DG 120–1, 132–3, 135, 168, 189–90).) This raises the question of how the two concepts relate, and in particular whether one can be reduced to the other. He denied that fittingness can be analysed in terms of moral obligation, since what you are obligated to do is what you believe you ought in the fittingness sense to do (DG 168). But in his middle works he contemplated the converse reduction, so you have a moral obligation to do an act if and only if it is fitting (in some sense) for you to do it and, if you do not do it, you are a fitting object of moral disapproval (‘SN’ 14; DG 168–9).29 He did not insist on this analysis, saying only that if it was correct and 28 Ewing noted that, in place of his third sense, Carritt, Ross, and Broad used one tied to your actual beliefs about the non-moral facts, but he thought this sense less common than the one based on evidence and set it aside (ST 102–3). His earlier MP had recognized only the first and third senses, saying that if a person has false beliefs about what is right in both these senses he cannot act rightly at all (MP 3n., 9). ‘SN’ seemed to dismiss the first, fully objective sense and to recognize only the second and third (‘SN’ 2–3). 29 In ‘SN’ the ‘ought’ used in this analysis was the third, evidence-based sense (‘SN’ 14), though it was there called the first sense, and that was probably also the view of DG. But the analysis could equally well

kinds of goodness and duty

85

could be combined with a fitting-attitudes analysis of ‘good’, fittingness would be the one basic normative concept (‘SN’ 14; DG 169). There are several objections to this sweeping reductive view. It is ambiguous in its use of fittingness, which must satisfy ‘“ought” implies “can” ’ when it is applied to acts but cannot do so when it is used to define ‘good’. Nor, as I have argued, is fittingness really an ultimate concept rather than a generic ‘ought’ with a distinctive type of ground. And in Second Thoughts Ewing rejected the analysis of moral obligation he had earlier proposed, though for a different reason. He now argued that in order to be subjectively under a moral obligation, or under an obligation because you believe you are, you must believe not only that it would be fitting for you to do an act but also that, independently of that belief, you have an obligation to do it (ST 91).30 Later he added the argument that a judgement of obligation makes an authoritative demand, as a mere statement of the fittingness of an act does not (VR 190). But whether or not these are cogent arguments, they led him to hold that fittingness and moral obligation are distinct concepts, so his three senses of ‘ought’ do not apply the same concept differently but involve two different ones. Nonetheless, he continued to hold, with Russell and against Moore, Prichard, and Ross, that there are multiple senses of ‘ought’, some perhaps more common or useful but none the one true sense. That, like many of Ewing’s views, seems sound and sensible, though it departs somewhat from the school’s conceptual minimalism. There are different forms of rightness, some looking more to the actual facts of your situation and some more to your beliefs about those facts. But none is the one most central form; each is appropriate to its own context and its own type of deontic judgement.

use the fully objective sense. Note that the first condition in the analysis, that your act actually be fitting, contradicts Ewing’s more frequent claim that you can be obligated to do an act even when your belief that it is fitting is false. And the second condition leaves open the possibility, which he elsewhere closed, that there are other grounds for blame than not doing what you think fitting. 30

Ewing attributed this argument to R.B. Brandt (ST 91n2).

4 Non-Naturalism Our school may be best known for their metaethics, which combined nonnaturalist moral realism with intuitionism about how moral truths are known. For most of them metaethics was not their main interest; they wrote more on normative topics. But they shared a distinctive view of the subject. They held, first, that moral judgements claim to be true, in the same sense of ‘true’ as scientific and mathematical judgements. They were cognitivists, treating judgements such as ‘pleasure is good’ as, in the present-day terminology, truth-apt. They also thought some such judgements are true. They were moral realists, and thought more specifically that there are properties of goodness and duty we can truly ascribe. Finally, they thought moral truths are sui generis, with a subject matter distinct from that of all other truths. Goodness and duty are neither identical to nor reducible to any of the natural properties studied by science, nor to any non-moral properties posited by theology or metaphysics, and no conclusion about them can be inferred from non-moral premises. They here affirmed the ‘autonomy of ethics’, and combining that with a belief in moral truth yields non-naturalist moral realism. This kind of realism has epistemological implications. If moral truths are sui generis, we cannot discover them by the empirical methods of science or whatever methods lead to religious or metaphysical knowledge. We must instead know them by direct insight or intuition. We just see that some moral propositions are true, and our developed moral knowledge is based on this seeing. This chapter will examine the school’s non-naturalism and the next their intuitionist moral epistemology.

4.1 Moral Realism Sometimes the school took their moral realism for granted. In the Preface to the first edition of The Methods Sidgwick said casually that he would assume there is something it is right and reasonable to do and ‘that this may be known’; later he said he assumed ‘there is such a thing as moral truth and error’ (ME v; ME1 23). The first sentence of Moore’s Principia said ethics is concerned with the ‘truth’ of judgements about ‘good’ and ‘ought’, with no awareness that this might be questioned (PE 1; also Ewing MP 6). At other times they affirmed their realism emphatically. In a letter of 1880 Sidgwick wrote, ‘Duty is as real a thing to me as the physical

non-naturalism

87

world’ (M 347), while Rashdall said moral judgements ‘express facts about the ultimate nature of Reality’ as much as the empirical sciences do (E 8; also 14–15; ICE 38–40, 53–4) and Ross wrote, ‘The moral order expressed in these propositions is just as much part of the fundamental nature of the universe . . . as is the spatial or numerical structure expressed in the axioms of geometry or arithmetic’ (RG 29–30, also 15, 20n; FE 3–4, 77; KET 26, 85). These claims assume that moral judgements can be true, and early members of the school again just assumed that cognitivist view. They were doubtless influenced by the grammatical form of judgements like ‘pleasure is good’, which mirrors that of judgements like ‘the table is round’. But before the 1930s they did not consider the rival non-cognitivist view that moral judgements express attitudes or issue commands and so cannot be true. One might think they had encountered this view in Hume, but they read him too as a cognitivist. Their further claim that some moral judgements are true reflected a broader philosophical realism they shared. Moore and Russell are known for mounting a realist challenge to Idealism around 1900, but Sidgwick was a Cambridge realist well before them and Prichard, Carritt, and Ross belonged to an Oxford realist school founded, again before Moore, by Cook Wilson. Their realism concerned not just the external world but also abstract entities such as numbers. Moore’s first anti-Idealist paper defended the independent reality of propositions, understood as abstract objects of thought,1 while Russell said he and Moore believed in ‘a pluralistic timeless world of Platonic ideas’.2 It is not a large step from this Platonism to non-naturalism about values, and the school often defended their belief in moral truth by analogy with truth in mathematics. Hence Ross’s comparing the ‘moral order’ to the ‘numerical structure . . . of arithmetic’ and Moore’s likening the ontological status of ‘good’ to that of numbers (PE 110–11). The school gave few positive arguments for moral realism, again tending just to assume it, but they did address objections to it. One was that of a critic who denies that he has the concepts of goodness and duty; he does not, he says, think in those terms. Sidgwick, Rashdall, Moore, and Ross thought this objection cannot be decisively refuted but nonetheless found it unpersuasive. Whatever the critic says, he will sometimes use moral concepts, especially when he feels his interests are threatened (ME 35, 211; E 14–15; EE 17; FE 28). A different objection concedes that we have moral concepts but says all judgements containing them are false, because the properties of goodness and duty they ascribe are not real; accepting cognitivism but not realism, it posits an ‘error theory’ of ethics of the kind associated with J.L. Mackie.3 This was an especially relevant objection for Carritt and Ross, who accepted an error theory of aesthetic judgement. Carritt said our judgements that things are beautiful ‘generally mean to attribute to 1 3

2 Moore, ‘Nature of Judgment’. Russell, ‘My Mental Development’, p. 12. Mackie, Ethics, Chapter 1; Mackie originally defended the theory in ‘Refutation of Morals’ (1946).

88

british ethical theorists from sidgwick to ewing

the thing a quality independent of anybody’s thoughts or feelings. But so far as they do assert this, there are reasons for thinking that perhaps none of them are true in the sense in which they are thus meant’ (‘MP’ 132; also 141; TB 106; EPT 30, 180–1 186; IA 23–4, 124). Ross shared this view (RG 128n), but they both thought the moral and aesthetic cases are different, because the way we apprehend some moral judgements shows them to be self-evidently true. Carritt said, ‘The truth of some judgments and the existence of some duties are self-evident’ (TM 28; also ‘MP’ 145, 147; EPT 2–3, 43), while for Ross ‘It seems as self-evident as anything could be, that to make a promise, for instance, is to create a moral claim on us in someone else’ (RG 20–1n, also 29–30, 82). Ewing too considered an error theory of ethics but rejected it, saying it is ‘impossible to deny that there are some ethical propositions which we have a right to hold true’ (DG 38, also 30–1, 76–7; MP 6, 7–8; E 1, 109–10; ST 29; VR 246). These claims are unlikely to persuade an error theorist, and one reviewer of The Right and the Good, Richard Robinson, found its only serious defect in its not squarely facing the question ‘Do what we mean by “right” and “good” really exist?’ Ross, he charged, moved too quickly from premises about our moral concepts to conclusions about moral facts.4 In his review Robinson thought his question can be answered affirmatively, primarily by an argument Ross gave briefly and attributed to Cook Wilson. It says an error theorist must explain where our moral concepts come from and could do so if ‘good’ and ‘right’ were complex, since then we could have formed them by combining other concepts. But if they are simple concepts, we can only have acquired them by apprehending a real property, since there cannot be a ‘fictitious “simple idea”’ (RG 82). This argument has no force against Mackie’s error theory, which finds the origin of the moral concepts in our feelings, or in ‘the projection . . . of moral attitudes’,5 and Robinson himself came later to accept this theory.6 Against it Ross and his fellow non-naturalists had just the assertion that some moral propositions are self-evident. This assertion assumes that our intuitions reliably indicate moral truth, and this can be questioned on Darwinian grounds. Like other things about us, it can be argued, our moral beliefs are the products of natural selection, present in us because earlier humans who had them were better able to survive and reproduce. Why should we trust beliefs with that origin? Since we would have them whether or not they were true, their being true would be a massive accident.7 Sidgwick addressed one objection of this kind when he dismissed issues about ‘the Origin of the Moral Faculty’ as irrelevant to ethics. All our cognitive faculties, he argued, including for mathematical knowledge, developed out of more primitive

4

5 Robinson, Review of RG, p. 346–7. Mackie, Ethics, p. 42. Robinson, ‘Emotive Theory’, which cited Mackie’s 1946 article, noted the contrast between Ross’s treatments of beauty and of moral concepts (86–9), and rejected the Cook Wilsonian argument Robinson had earlier endorsed (99–100). 7 Street, ‘Darwinian Dilemma’. 6

non-naturalism

89

ones, and the fact that a faculty had one rather than another origin is no reason to distrust its deliverances now (ME v–vi, 212–13; also Rashdall, E 37–40, ICE 61–76, 82–4; Moore, E 48–9; Ross, RG 12–15, FE 15–17; Ewing, DG 23–5, 55–6). The more important Darwinian objection concerns specific moral judgements and says that if a particular belief about right or wrong resulted from evolutionary processes, our finding it even highly intuitive is not good evidence for its truth. Sidgwick also addressed this objection. While granting it force against some moral beliefs (ME 213), he argued that it has none against his own principles of prudence and rational benevolence: ‘No psychogonical theory has ever been put forward professing to discredit the propositions that I regard as really axiomatic, by showing that the causes which produced them were such as had a tendency to make them false’ (ME 383). This is a plausible claim, at least about benevolence. Since this principle requires us to care equally for all people, our belief in it cannot be explained by the mechanism of kin-selection, which favours behaviours that benefit our close genetic relatives, nor as a case of reciprocal altruism, where we benefit those who may benefit us in turn. On the contrary, by requiring us to sacrifice our good for genetically distant people we may never meet again, it seems positively counteradaptive, and it is even more so if it is extended, as it was by Sidgwick, to require concern for non-human species. A principle of impartial benevolence is central to many moral traditions, but it is hard to see how it could be directly favoured by natural selection.8 Nor is this true only of benevolence. This principle is hard to explain in Darwinian because it is impartial, but there are equally impartial deontological principles telling us not to kill or lie to anyone, no matter how genetically distant or whether we will see him again. Sidgwick thought biological arguments are superfluous against these principles if there are philosophical objections to them (ME 383; also Broad, CE 179–80), but their form makes them equally unlikely products of natural selection. A standard objection to non-naturalism is that it is metaphysically extravagant, positing properties and relations unlike those used to explain the natural world: its belief in ‘simple, indefinable properties that are of a peculiar non-natural or normative sort’ is ‘hard to defend’.9 Our school would not have been moved by this objection, since they thought there are many non-natural properties besides the moral ones. Nor would they have been moved by the specific version of this objection in Mackie’s ‘argument from queerness’.10 He assumed that moral judgements are intrinsically motivating, so to judge an act right is necessarily to have some desire to do it. What he found ‘queer’ about moral properties is that they would combine this motivating force with objective reality, or have an ‘objective prescriptivity’ whereby to know something has one is automatically to desire it. As we will see, our school said little about how moral belief relates to action, and if anything thought moral 8 9

See also de Lazari-Radek and Singer, ‘Objectivity of Ethics’. 10 Frankena, Ethics, p. 103. Mackie, Ethics, pp. 38–42.

90

british ethical theorists from sidgwick to ewing

belief is not itself motivating but influences behaviour only in conjunction with a desire to seek what is good or do what is right. For them moral properties, though objective, lack the prescriptivity that makes for Mackie’s specific queerness. Some philosophers distinguish two versions of non-naturalism, one open to objections about metaphysical extravagance and the other not. The first posits non-natural properties of goodness or duty that make moral judgements true; the second, which they think is more defensible, says only that some judgements about we people ought or have reason to do are true but does not talk of non-natural properties or give them that grounding role.11 Some commentators read this distinction back into our period. Sidgwick, they say, affirmed only the modest version of non-naturalism, but Moore and those who followed added claims about real moral properties that invited avoidable objections and led to non-naturalism’s demise later in the century.12 The interpretive contrast here is overdrawn. Sidgwick did not discuss Moore’s non-natural property of goodness because he analysed ‘good’ in terms of ‘ought’, and his claim that duty is ‘as real a thing as the physical world’ hardly seems metaphysically modest. The initial reviewers of Principia did not see it as altering Sidgwick’s ontology but emphasized its continuity with Sidgwick, while Moore himself said little about his non-natural property and even tried to limit his metaphysical claims, saying that though goodness is an object and therefore ‘is somehow’, it does not ‘exist’, and in particular does not exist in a ‘supersensible reality’. There is no such reality, and metaphysicians are mistaken if they assume that what does not exist in space and time must exist somewhere else if it is to be at all (PE 110–12, 123–5, 140–1). Moore can even be read here as holding a version of the modest view defended by Derek Parfit, which says normative properties ‘exist’ but only in a ‘non-ontological’ sense that has no metaphysical implications;13 note also his later claim that value-predicates like ‘good’ are ‘of a quite different kind’ from natural ones, since they seem not to ‘describe’ the intrinsic nature of what has them (‘CIV’ 274).14 I doubt Moore unambiguously intended a view like Parfit’s, since he never clearly explained his contrast between being and existing or his denial that ‘good’ describes. But he did try to make his view less metaphysical. Ewing’s views are illustrative here. In his later Second Thoughts he did defend a modest non-naturalism, denying that there are objective moral properties in part because they are metaphysically ‘mysterious’ (ST 35–7, 43–7, 50–3, 77). He was here

11 The second version is defended in Nagel, View From Nowhere, p. 144; Scanlon, What We Owe, pp. 62–4. 12 Mackie, ‘Sidgwick’s Pessimism’, p. 323; Schneewind, Sidgwick’s Ethics, pp. 205, 222, 303–4; Shaver, ‘Sidgwick’s Minimal Metaethics’; Crisp, ‘Sidgwick and the Boundaries’, p. 59; Schultz, Henry Sidgwick, pp. 188, 198–9; Phillips, Sidgwickian Ethics, pp. 29–30. 13 Parfit, On What Matters, vol. 2, pp. 464–87. 14 This claim can also be read as consistent with non-cognitivism.

non-naturalism

91

rejecting a view he took Moore, Ross, and himself to have previously held (ST 50), which suggests that for him that view was ontological. He did not now say, like some present-day non-naturalists, that moral judgements can be true without there being real properties that make them so. On the contrary, he accepted the ‘correspondence’ theory that truth is always ‘dependent on reality’. He just thought moral truths need not be made true directly, by specifically moral properties; they can be made true indirectly, by natural properties such as that an act will maximize pleasure, just as counterfactuals are made true indirectly by categorical facts (ST 43–7). As he later acknowledged (VR 195), he could not explain exactly how his indirect moral truthmaking was to work, and the idea seems problematic.15 But his correspondence assumption implies that, without this idea, non-naturalism must posit real moral properties. Prominent though it now is, modest non-naturalism seems not to have been a possibility our school, Moore perhaps excepted, considered. They thought judgements about goodness and duty can be true, and, the later Ewing aside, that there are therefore properties of goodness and duty. That our talk of these might not be ontological seems not to have occurred to them. These interpretive questions aside, the idea that non-naturalism can be modest is controversial, denied by its critics as well as by some adherents.16 In Parfit’s version it assumes we can say goodness ‘exists’ without intending anything ontological; many will dispute that. In his and other versions it assumes moral judgements can be true without any properties’ making them true; many will also dispute that. The metaphysical issues here are beyond the scope of this book; thus I will not decide whether there can or cannot be moral truth without moral truth-makers. If there cannot, as I suspect, then non-naturalism requires robust moral properties of the kind Rashdall, Ross, and the early Ewing, if not so clearly Moore, affirmed. A different objection says non-naturalism cannot explain the supervenience of moral properties: that anything good or right is made so by a non-moral property such as its being or causing pleasure. The supervenience relation is necessary, since any two items with the same non-moral properties must have the same moral ones. But non-naturalism cannot explain the necessity as analytic, because it denies that moral conclusions follow from non-moral premises, and so cannot explain it at all.17 Our school were certainly aware of supervenience. It is implied in Sidgwick’s axiom of justice (ME 208–9, 379–80), Moore assumed it when he said a thing’s intrinsic goodness ‘depends . . . on’ its intrinsic properties, so anything exactly like a

15 A discussion of Second Thoughts that presses this difficulty is Olson and Timmons, ‘Ewing’s First and Second Thoughts’. 16 For two of the latter see Svavarsdottir, ‘Objective Values’, p. 157; Enoch, Taking Morality Seriously, pp. 122–7. 17 Blackburn, ‘Moral Realism’; Blackburn, Spreading the Word, pp. 182–90. He takes the argument to support non-cognitivism.

92

british ethical theorists from sidgwick to ewing

good thing must be exactly as good (‘CIV’ 260–1; also ‘R’ 588),18 and Ross did the same when he called value a ‘dependent or consequential’ quality (RG 120–3, also 79, 88; FE 168, 280; Prichard, MW 173; Broad, FT 177–8, CE 118, 137–8, 201–2; Ewing, MP 166, DG 53, ST 70, VR 109). But they did not see anything problematic in this relation. Where the objection assumes that all a priori necessities are analytic, they thought there are many synthetic a priori truths, such as that any figure with three sides has three interior angles or that every event has a cause. And Moore, Ross, and Broad explicitly called the necessity in moral supervenience synthetic (‘CIV’ 271–2; RG 120; KET 42; CE 206–11, also 202–3), involving a synthetic a priori connection between the properties of, say, pleasantness and goodness, such that, necessarily, anything pleasant is good. Their view will not appeal to those who reject such connections, but then the issue is not supervenience as such but the general possibility of synthetic a priori truths.19 The school also addressed Mackie’s ‘argument from relativity’, which holds it against non-naturalism that people and cultures disagree morally, with no accepted method of resolving their differences. If there were objective moral truths knowable by intuition, would we not agree on what they are?20 Some denied that there is as much disagreement as the objection contends. Moore and Russell thought most people agree about what is ultimately good (PE 188; ‘EE’ 54), while Ewing said that in civilized communities there is near-universal acceptance of the principles of prima facie duty (ST 40; also ‘RSE’ 85). Some said different acts are in fact right in different societies; thus Ewing thought slavery, though wrong today, may have been right in ancient Egypt, as then involving the least bad consequences (DG 19; E 128–9; also Ross, FE 18). And some moral disagreements can rest on disagreements in non-moral belief; thus two cultures can share the goal of promoting their members’ happiness but differ about what will best do so (Sidgwick, GSM 226; Ross, FE 18–19). Another response denied that moral disagreement always counts against nonnaturalism. Even if some moral truths are self-evident, some argued, we can intuit them only if we meet certain conditions, so agreement from people who do not meet the conditions should not be expected. Prichard attributed much disagreement to a lack of ‘thoughtfulness’ in reaching moral conclusions (MW 6, 14n), while Ewing said people sometimes disagree because they have had different experiences or because some concentrate on just one aspect of a complex situation; thus militarists see the unselfish heroism war brings out but underestimate its evils. Ewing also thought our moral judgements can be affected by ‘desire and prejudice’, as we avoid

In Contemporary Moral Philosophy Warnock said that on Moore’s view goodness is ‘quite unconnected with’ a thing’s other properties, and can alight ‘inexplicably and at random’ on anything (p. 14). As Cox pointed out (‘Warnock on Moore’), this charge ignores Moore’s explicit remarks about supervenience. 19 See Klagge, ‘An Alleged Difficulty’; Enoch, Taking Morality Seriously, pp. 142–50. 20 Mackie, Ethics, pp. 36–8. 18

non-naturalism

93

reaching a correct moral conclusion because it would require us to do something we prefer not to (DG 20–1; E 130–1; also Ross, FE 20). Prichard said awareness of an obligation is possible only for ‘a developed moral being’, where ‘different degrees of development are possible’ (MW 14n), and to some the development was partly cultural. Rashdall asked why, if we know more today about mathematics than earlier humans did, we cannot also have developed in moral understanding (TGE I 84–5; also E 53–5, 88–9). And Ross said, ‘the nature of the self-evident is not to be evident to every mind however undeveloped, but to be apprehended directly by minds which have reached a certain degree of maturity, and for minds to reach the necessary degree of maturity the development that takes place from generation to generation is as much needed as that which takes place from infancy to adult life’ (RG 12, also 29; Carritt, EPT 139). These claims may not fully answer the argument from relativity, as similar more recent claims may not.21 More generally, the school’s brand of moral realism is a controversial view many will reject. But after being scorned for many decades nonnaturalism has had prominent recent defences,22 and though some say its current versions are more defensible than earlier ones, the alleged differences are often illusory. In Sidgwick, Moore, Ross and their fellows we find at least a classic presentation of a still-living metaethical view.

4.2 The Autonomy of Ethics: The Open-Question Argument The realist belief that some moral judgements are true is shared by naturalist views, such as ones that equate ‘x is right’ with ‘x will maximize pleasure’ or ‘x will maximize biological fitness’. What distinguishes non-naturalism is its belief that moral truths are sui generis, so ethics has an autonomous subject matter. The best-known defence of this belief is in Moore’s Principia, which focussed on ‘good’ and famously said it is indefinable: ‘If I am asked “What is good?” my answer is that good is good, and that is the end of the matter. Or if I am asked “How is good to be defined?” my answer is that it cannot be defined, and that is all I have to say about it’ (PE 6; also 8–9, 10, 15, 17). For Moore ‘good’ was a ‘simple notion’, like ‘yellow’ (PE 7–8); it cannot be explained in other terms but must be grasped immediately or in itself, and any attempt to define it involves what he called the ‘naturalistic fallacy’ (PE 10). This label is misleading, since Moore rejected not only ‘natural’ definitions of ‘good’ but also ones in metaphysical or theological terms. He also found the fallacy 21 Nagel, View From Nowhere, pp. 147–8; Shafer-Landau, Moral Realism, Chapter 9; Enoch, Taking Morality Seriously, Chapter 8; Parfit, On What Matters, Vol. 2, pp. 543–8. 22 Nagel, View From Nowhere; Scanlon, What We Owe; Shafer-Landau, Moral Realism; Huemer, Ethical Intuitionism; Parfit, On What Matters; Enoch, Taking Morality Seriously.

94

british ethical theorists from sidgwick to ewing

not just in explicit definitions of ‘good’ but in any attempt to infer a conclusion containing ‘good’ from premises that do not (PE 37, 49, 57, 114). But this inferential fallacy rests on the one about definition, since non-ethical premises can entail an ethical conclusion only if there is an analytic connection between the two of the kind a definition provides; thus ‘x is pleasant’ can entail ‘x is good’ only if ‘good’ means ‘pleasant’. Moore assumed this point in his critique of John Stuart Mill’s argument that pleasure is the only thing desirable or good because it is the only thing desired. By equivocating on the meaning of ‘desirable’, he argued, Mill in effect defined ‘good’ as ‘desired’ and so committed the naturalistic fallacy (PE 66–73). By a ‘definition’ Moore could have meant an account of three things: the meaning of the word ‘good’, the content of the concept ‘good’, or the nature of the property of goodness. He said he was not interested in the meaning of the word ‘good’, as in a description of how it is used (PE 6, 8), and is best read as asking whether the concept ‘good’ admits of analysis, where a correct analysis of a concept can reveal more than people consciously have in mind when they use it. He described analysis somewhat awkwardly as identifying the ‘parts’ of a concept (PE 8), but this can be understood as follows. The concept ‘bachelor’ combines the concepts ‘unmarried’ and ‘man’; someone falls under the first just in case he falls under the other two. They are therefore ‘parts’ of ‘bachelor’, which the analysis shows is compounded from them. What Moore denied is that ‘good’ is ‘composed of any parts, which we can substitute for it in our minds’ (PE 8), that is, that it is a compound. He seems to have assumed, controversially as we will see, that whatever is true of a concept is true of its associated property, and therefore often drew conclusions about the ‘quality’ or ‘property’ of goodness. (He sometimes spoke of the ‘object or idea’ or ‘object or notion’ that ‘good’ stands for (PE 6, 7), where the first term in each pair suggests the property and the second the concept.) But we do best to interpret his central metaethical claim as concerning, at least initially, the concept ‘good’ and denying that it is composed of simpler concepts. As he later realized, this claim failed in two ways to capture his real intention. Since he thought the natural concept ‘pleasant’ too is indefinable (PE 12–13), a naturalist equation of ‘good’ with ‘pleasant’ would leave ‘good’ indefinable; yet he wanted very much to reject that equation (‘P’ 13–14, 18). And his central claim would not be undermined if ‘good’ could be analysed using other ethical concepts such as ‘ought’. Though he always resisted such analyses (‘R’ 598–9), it would leave his core antinaturalism intact (‘P’ 5, 13–14). Taking these two points together, we can restate his central claim as that no moral concept is identical to or reducible to non-moral ones; as a group, the moral concepts are distinct. His argument for this claim has come to be known as the ‘open-question argument’, but it had two different forms. The first asks us to consider a naturalist definition of ‘good’ as, say, ‘pleasant’. If this definition were correct, the argument says, the apparently substantive thesis ‘pleasure is good’ would be equivalent to the tautology ‘pleasure is pleasure’, which it clearly is not (PE 12, 14–15, 16); it would also be

non-naturalism

95

equivalent to ‘what is good is good’, which it is not (PE 16). If these equivalences fail, however, ‘good’ and ‘pleasant’ must be distinct concepts. This argument may seem to beg the question against the naturalist, for whom ‘pleasure is good’ is a tautology. But Moore was appealing to our everyday understanding of ‘good’ and arguing that naturalism does not fit it, or asking us to reflect on that understanding and see, through his examples, what ‘good’ cannot be.23 At one point he suggested a deeper explanation for his argument. Naturalists want to persuade us that what they call good is what we ‘really ought to do’, he wrote, and in so far as they do this, their teaching is ‘truly ethical’. But it cannot be ethical if it says only that pleasure is pleasure (PE 12). The suggestion here is that naturalist views leave out the connection of ‘good’ to directives containing ‘ought’, or to what many would call its ‘normativity’. This explanation is surprising from Moore, since he believed that ‘good’ is the primary ethical concept and ‘ought’ derivative, but some take it to be what makes his argument work: it is because ‘good’ is normative that no naturalist analysis of it succeeds.24 The second form of the argument, from which it takes its name, says naturalist definitions make certain coherent thoughts impossible, or close what seem to be open questions. Moore’s initial example was the proposal, apparently due to Russell, that ‘good’ means ‘that which we desire to desire’.25 If this definition were correct, he said, then once we knew that we desire to desire some x, it would be settled that x is good. But it is not settled: it is perfectly possible to ask whether it is good to desire to desire x, or whether, even though we do desire to desire x, x really is good. The same is true of other naturalistic definitions: we can know that something is pleasant or desired or approved and still ask whether it is good (PE 15–17, 20–1). This second form of the argument is really the flipside of the first. If the first says naturalism turns ‘pleasure is good’ into a tautology, the second says it turns that claim’s negation—‘x is pleasant but not good’—into a contradiction. That too violates our everyday understanding of ‘good’. Moore had a special reason for making his argument. He was a pluralist about the good, believing there are several ultimate values, and he thought naturalism rules this out: ‘If we start with the conviction that a definition of good can be found, we start with the conviction that good can mean nothing else than some one property of things’ (PE 20). But nothing in the concept ‘good’, he countered, implies that there is just one good thing. It may be replied that ‘good’ can be defined disjunctively, say as ‘pleasure or knowledge or virtue’, yielding a pluralistic version of naturalism. But Moore did not distinguish concepts from properties and may have assumed, as

23

Frankena said the open-question argument contributes nothing to the case against naturalism, which rests on a direct appeal to inspection of our concepts (‘Ewing’s Case’). But the argument is best read as an aid to that inspection rather than as an alternative to it. 24 Frankena, ‘Obligation and Value’; Darwall, ‘Moore, Normativity, and Intrinsic Value.’ 25 Russell, ‘Is Ethics a Branch of Empirical Psychology?’ p. 78; see also Pigden, ‘Desiring to Desire’.

96

british ethical theorists from sidgwick to ewing

others have, that disjunctive properties are not real.26 If that is so, then to make goodness a natural property is indeed to rule out plural values. A more general argument may be implicit here. Whereas a naturalist definition commits one to a substantive ethical view, for example, that only pleasure is good, non-naturalism does not: ‘if we recognise that, so far as the meaning of good goes, anything whatever may be good, we start with a much more open mind’ (PE 20). ‘Good’, Moore may have been arguing, is a critical concept, which far from being tied to a substantive ethical view can be used to challenge any such view, no matter how initially appealing or widely held. That naturalism precludes this critical role is then a further objection to it.27 Though the open-question argument says simply that naturalist analyses of ‘good’ have counterintuitive implications, it may rest on assumptions about the concept’s normative and critical character that explain why the implications are false and are its true basis. Moore presented Principia’s anti-naturalism as revolutionary, saying only Sidgwick before him had stated the view clearly (PE 17). But this claim was exaggerated. Sidgwick had not only called the basic ethical concept indefinable (ME 32–3) but had used the first form of the open-question argument: he said ‘good’ cannot mean ‘pleasant’ for ‘any persons who affirm—as a significant proposition and not as a mere tautology’ that pleasure is good (ME 109; also GSM 145), and argued similarly against Bentham’s suggestion that ‘right’ means ‘conducive to the general happiness’ (ME 26n). He did not think Bentham can have really meant this, given the objection it invites, and was therefore more charitable than Moore, who found the naturalistic fallacy ‘in almost every book on Ethics’ (PE 14). But he agreed about what would be a mistake if it was made. In 1885 Rashdall praised Sidgwick for seeing that ‘no experience of what is, no predictions of what will be, can possibly prove what ought to be. He neither dismisses the “ought” as a figment (with Bentham), nor involves the whole discussion in inextricable confusion (with J.S. Mill) by failing to distinguish between the desirable and the desired’ (‘PS’ 214; also TGE I 53).28 McTaggart said ‘no amount of “is” can produce the slightest “ought” ’ (SHC 266; also 96), while the reviewers of Principia often endorsed its criticisms of Mill, Spencer, and Green but said they were ‘not altogether original’, had for the most part ‘already been brought out by other critics’, and even were ‘the standard criticisms’.29 To Moore’s claim that only Sidgwick had seen that ‘good’ is indefinable Rashdall replied that this ‘was taught with sufficient distinctness by Plato, . . . Aristotle, and a host of modern writers who have studied in their school’ (TGE I 135–6n).30 26

Armstrong, Universals and Scientific Realism, vol. 2, pp. 19–23. Adams, Finite and Infinite Goods, pp. 77–8. 28 Sidgwick too had noted the ambiguity in Mill’s use of ‘desirable’ (EEM 31, ME 388n). 29 Bosanquet, Critical Notice of PE, p. 261; Mackenzie, Review of PE, p. 378; Wilde, Review of PE, p. 582. 30 A.N. Prior cited Cudworth, Shaftesbury, Hutcheson, Price, Whately, and Dugald Stewart as anticipating the first version of the open-question argument (Logic and the Basis of Ethics, Chapter 9), and 27

non-naturalism

97

The autonomy thesis was therefore perfectly familiar in 1903, but Moore emphasized it more than others and pressed the case for it more forcefully; in addition, the second form of the open-question argument was original to him. The thesis is therefore not unreasonably associated primarily with him, and partly under his influence was shared by later writers. In The Theory of Good and Evil Rashdall said ‘good’ is unanalysable (TGE I 47–8, 102, 135–6, II 106) and used the first form of the open-question argument (TGE I 47–8; E 16, 22), while McTaggart said ‘it is impossible to define good and evil in terms of anything else’ (NE II 398). Prichard called the concept of obligation ‘indefinable, . . . what Locke would have described as a simple idea’ (MW 115, also 12–13, 116–17, 169, 173) and said an ‘ought’ can only follow from another ‘ought’ (MW 9–10, 47–8); Ross made similar claims (RG 6–12, 78, 92–4; FE 27–8, 183–4), as did Broad and Ewing (FT 111, 165–6, 173–4, 281–2; CE 359; MP 6, 182, 216; ‘SN’ 1–2, 4, 6; DG 41–2, 57, 67–8, 109–10, 122–3; E 1, 11–12, 55, 100–3, 113–14).31 Before the 1950s the school did not use what later became a common argument for the autonomy thesis, one appealing to the ‘judgement-internalist’ view that moral judgements are intrinsically motivating. On this view it is a necessary condition for sincere assent to a moral claim that you have some motivation to act as it directs. This explains, it is argued, why no ‘ought’ can be derived from an ‘is’, since no conclusion with motivating force can follow from premises that lack it. The school’s early members either did not accept this argument’s internalist premise or did not think it important. In Principia Moore said casually, ‘we hardly ever think a thing good, and never very decidedly, without at the same time having a special attitude of feeling or will towards it; though it is certainly not the case that this is true universally’ (PE 131; also 135–6; Carritt; EPT 118–20). He here rejected internalism as a conceptual truth, accepting only a weaker empirical generalization,32 but he did not connect that generalization to his argument that ‘good’ is indefinable, which was already complete. Sidgwick said ‘when I speak of the cognition or judgment that “X ought to be done” . . . as a “dictate” or “precept” of reason . . . , I imply that in rational beings as such this cognition gives an impulse or motive to action’ (ME 34, also 77). This statement sounds internalist, but Sidgwick did not use it in his argument that ‘ought’ is indefinable, which again was complete, nor is the internalism clear. Since Sidgwick thought motivation always involves desire (ME 60, 202–3, 362–3), any internalism he accepted would have to say the belief that an act is J.N. Keynes and T.H. Huxley as late-Victorian believers in the autonomy of ethics (pp. 43–5): ‘Moore’s achievement has not been to work a revolution in Moral Philosophy, but simply to help keep alive, in our own age, the eighteenth century tradition of sanity and rigour which Sidgwick . . . kept alive in his’ (p. 107). In Second Thoughts Ewing said his new ‘indirect truth-making’ view allowed some derivations of ethical conclusions from non-ethical premises, though they were not ‘formal deductions’ (ST 48). This was another departure from his earlier, more standard non-naturalism. 32 In EE he had denied the generalization, saying ‘We often, as we say, know or feel that a thing is good, and yet feel no attraction toward it’ (153). 31

98

british ethical theorists from sidgwick to ewing

right causes a desire to perform it, which in turn causes action. While that may have been his view, he may instead have held that moral beliefs cause action only in combination with a desire to do what is right, which would make his view ‘externalist’; and he did often speak of a desire to do what is right (ME 5, 39, 52, 101, 204, 206, 447). He never decided clearly between the two views,33 and the same is true of Rashdall, some of whose remarks favour internalism and some its opposite (TGE I 104–6, 121, 140–1; E 32, 41; ICE 112–15). This pattern continued with later members. In ‘Mistake’ Prichard said the sense of obligation that motivates moral action is not a desire, because it does not have a ‘purpose’ (MW 14–15); later he said the belief that an act is right ‘involves of its very nature emotion’, that the belief and emotion are ‘inseparable’, and that the combination of the two can move us to act (MW 73). These claims all sound internalist, but in ‘Duty and Interest’ he took the opposite line, saying all action involves desire for a ‘purpose’ and in moral action the desire is to do what is right (MW 37–8). Later he said the belief that an act is right arouses a desire to do it in the same way as the belief that something will be enjoyable does (MW 129–30); since beliefs about enjoyment motivate only given a desire for enjoyment, this too suggests externalism. Ross was more consistently externalist, saying we have a general desire to do our duty just as we have a general desire for food, so the desire to do an act we think right ‘arises in us at a particular time just as does the particular desire to eat’ (RG 156–8; also ‘NM’ 260–1; FE 205–6, 237; KET 27, 38; but contrast FE 226–8). Broad made similarly externalist remarks (FT 107–8, 282–3; CE 140, 152, 218, 223, 353, 355), as did Ewing, who, like Ross, compared the desire for what is good with the desire for a particular type of food (E 89, also 143, 147; DG 51–2; ST 13, 17; ‘BET’ 87–8; but contrast ST 74–5). Whether externalist or not, none of these philosophers used the claim that moral judgements are motivating to argue for the autonomy thesis; they thought it evident on other grounds that moral concepts are distinct. Though some may think internalism gives the best explanation of this fact, it is not one our school gave.

4.3 The Open-Question Argument: Objections The open-question argument has persuaded many philosophers that ethics is autonomous, but it has also met with objections. One says that since the argument can only consider naturalist definitions one by one it does not rule out the possibility that some analysis it has not tested is sound. But its proponents have examined the most plausible such analyses, those fitting the ethical views people actually hold, and it is hard to see some alternative doing better. 33 Bradley complained of Sidgwick’s unclarity on this point (‘Mr. Sidgwick’s Hedonism’, pp. 76–8), while Broad read him as having the second, externalist view (FT 146); see further Shaver, ‘Sidgwick on Moral Motivation’.

non-naturalism

99

Moreover, the failure of a definition like ‘“good” means “pleasant” ’ turns less on its specific content than on general features of ethical concepts such as their being normative and critical. If so, its failure generalizes. A more important objection invokes the ‘paradox of analysis’ identified by C.H. Langford in his contribution to The Philosophy of G.E. Moore.34 An analysis of the kind Moore said cannot be given of ‘good’ is supposed to be informative, telling us things about a concept we did not already know. But then the truth it states, though analytic, must be non-obviously so, or one not everyone could see in advance is analytic. That means Moore’s argument fails, since even if ‘x is pleasant, but is x good?’ seems an open question, it may not really be one. The concepts it contains may be non-obviously equivalent, and the naturalism asserting their equivalence may despite our resistance be true. The difficulty here was implicit in Principia. When Moore said he was not seeking an account of what we mean by the word ‘good’, he implied that a successful analysis need not reflect what is consciously in our minds. How then could he appeal to what is in our minds to refute naturalist claims? Ross and Ewing recognized this difficulty and had a response. Ross allowed that we can be unsure whether a given concept is analysable, since we can use a term for a complex ‘while yet the complex is not explicitly present to our minds’. We must therefore look closely at any proposed definition of ‘good’: If it is the correct definition, what should happen is that after a certain amount of attention to it we should be able to say, ‘yes, that is what I meant by “good” all along, though I was not clearly conscious till now that it was what I meant.’ If on the other hand the result is that we feel clear that ‘that was not what I meant by good’, the proposed definition must be rejected’ (RG 92–4; also FE 259–60; Ewing, DG 41–4).35

If the concept X is analyzable as A + B + C, so the latter are the criteria for its application, and we apply X correctly, we must already have some grasp, even if only subconscious, of A, B, and C. But then we should be able, after reflection, to see the analysis as correct when we are given it, since it only makes explicit what we implicitly know. Moore is sometimes said to have assumed our concepts are transparent to us, so we always know from the start what they contain.36 Ross made only the weaker and more defensible assumption that our concepts can become transparent once we have an analysis of them. And it is even more plausible that, if we are given an incorrect analysis, we can see its elements are not ones we were implicitly

34 Langford, ‘Moore’s Notion of Analysis’, p. 323; Baldwin, G.E. Moore, pp. 208–14. Something close to the objection had been made earlier in Field, Moral Theory, pp. 54–5. 35 Sidgwick too thought that even when an analysis tells us something new we will recognize it as correct when given it (ME 353); Broad also made points like Ross’s though with more scepticism about how decisively they meet the difficulty (FT 173–4; CE 110–15). 36 Baldwin, G.E. Moore, pp. 88, 210–11; Darwall, Gibbard, and Railton, ‘Towards Fin de Siecle Ethics’, p. 115.

100

british ethical theorists from sidgwick to ewing

using. If on reflection most of us think ‘good’ is not equivalent to ‘pleasant’, that is strong evidence that it is not.37 A more recent objection targets Moore’s slide between talk of concepts and of properties, or his assumption that if the concept ‘good’ is distinct from any nonmoral concept, the property of goodness must be distinct from any non-moral property. Against this critics have said there are non-analytic property-identities, such as between water and H2O. The concepts ‘water’ and ‘H2O’ are not identical, nor do the words ‘water’ and ‘H2O’ mean the same. Nonetheless, scientists have discovered that water is H2O, so the property of being water is identical to that of being H2O. The concept ‘good’ may likewise be distinct from any natural concept while the property of goodness is identical to a natural property, so all we refer to when we talk of ‘good’ is, say, pleasure.38 This objection assumes that what is true of a natural-kind property like water can also be true of moral properties like goodness and rightness. But natural-kind properties have special features that allow non-analytic identities and are not had by moral ones. As Parfit has argued, the property of being water is internally complex. It is that property, whichever it is, that has a different, second-order property. More specifically, it is that property, whichever it is, that has the different property of being what explains the behaviour of the stuff in lakes, rivers, and streams in our world. When scientists discovered that the atomic structure H2O explains the behaviour of that stuff, they showed that being water is identical to being H2O, but that is because the property of being water invites an identity of that kind. Its structure contains an ‘explicit gap’ that being H2O fills,39 and a similar gap is present in the concept ‘water’ and in the meaning of the word ‘water’ when it is used for the natural kind. In this use, ‘water’ refers to whatever has the property that explains the behaviour of the stuff in our lakes, even if it does not behave like our water. Applied on a Twin Earth where the stuff in lakes has the atomic structure XYZ and H2O behaves differently, our word ‘water’ refers to H2O and not to XYZ, because it refers to whatever property plays a certain role for us. If this is why water is identical to H2O, a similar identity will be possible for goodness or rightness only if it too is whichever property has a different, secondorder property. But this is not plausible. To say something is good is not to say it has another property that plays some role; it is just to say it is good or ought to be desired. Consider the proposal that to be good is to have that property or cluster of properties that have caused speakers in the past to call things ‘good’.40 It is open to an openquestion objection, since we can know something has the properties that have led us

38 See also Pigden, ‘Desiring to Desire’, pp. 256–9. See e.g. Brink, Moral Realism, Chapter 6. Parfit, On What Matters vol. 2, pp. 298–303, 329–38; also Gampel, ‘A Defense of the Autonomy’; Stratton-Lake, ‘Introduction’, pp. 9–11. 40 Boyd, ‘How to Be a Moral Realist’. 37 39

non-naturalism

101

to call things ‘good’ but wonder whether it really is good. Nor can it be defended as stating a non-obvious analyticity, since it fails Ross’s test. When we are presented with it we do not recognize it as what we always had in mind; on the contrary, we see it is not what we meant. Though goodness could be a natural property if it was structured like the property of being water, it is not. There is a further objection to the analogy. Our use of ‘water’ is indexed to the actual world, so it refers to what explains the behaviour of the stuff in lakes here; only so can it refer to H2O on Twin Earth rather than to XYZ. A Twin-Earth use of ‘water’ is similarly indexed, so it refers to XYZ in our world. It follows that if we say something is ‘water’ and people on Twin Earth say it is not, we are not disagreeing any more than people in Toronto and London are when they each call their location ‘here’. But this cannot be true of the moral concepts: if we say something is good and people on a Moral Twin Earth say it is not, we are disagreeing. Goodness is again not like water.41 In later years Moore called his presentation of the open-question argument in Principia a ‘mass of confusions’ and ‘certainly fallacious’ (‘P’ 3; ‘GQ’ 127). But he still thought the argument’s main point is sound (‘P’ 2), and that seems right; the argument at least shows that moral concepts are distinct from non-moral ones and play a very different role.

4.4 Responses to Subjectivism Non-naturalism opposes subjectivist views that tie moral judgements to people’s desires or feelings. At first our school thought the only possible such view is subjective naturalism, which equates a moral judgement with the assertion that some persons have or would have some feeling. They criticized this view at length. Sidgwick initiated the critique with a brief refutation of the view that to say an act is right is to report that you, the speaker, approve of or are satisfied by it. He argued that on this view there can be no moral disagreement, since if someone else says the same act is wrong the two of you report perfectly consistent psychological facts (ME 27; also EEM 60). Moore gave a lengthier critique in his Ethics, first addressing three subjectivist views about ‘right’: to say an act is right is to say that you have a certain feeling toward it, that most people in your society have that feeling, or that someone has that feeling. He objected, first, that on these views the same act can be both right and wrong; this will be so on the first view if one person says the act is right, another says it is wrong, and both report their feelings accurately (E 37–42, 45–6). But as Stevenson later pointed out, this objection fails. The subjectivist views make moral judgements relativized; on the first, your claim that an act is right says only that it is Horgan and Timmons, ‘New Wave Moral Realism’; more specific criticism of Boyd is in Rubin, ‘Is Goodness a Homeostatic Property Cluster?’ 41

102

british ethical theorists from sidgwick to ewing

approved of by you. But in deriving his contradiction Moore used unrelativized concepts; when presented with Stevenson’s critique, he said his objection had been ‘a sheer mistake’.42 He also objected, as Sidgwick had, that the first view makes moral disagreement impossible (E 42–3), but he did not press this objection against the second and third. Though the second view rules out disagreement between societies, it allows it among members of the same society, and neither it nor the third involves the ‘absurdity’ that moral disagreement never happens (E 46). But surely it is problematic if societies cannot disagree morally. Moore had a third objection to both these views and to a fourth, which says that to call an act right is to say the majority of all humans have a feeling toward it. Applying the open-question argument, this objection says you can think an act is right even though you do not think most people in your society, someone or other, or the majority of humans have any feeling; conversely, you can know most people have the required feeling yet wonder whether the act is right (E 46–7, 60; also Sidgwick, ME 28). He did not press this objection against the first view but could have: you can feel positively about an act yet wonder whether it is right. If so, his final objection applies to all subjective naturalist accounts of ‘right’, and he later applied both it and the others to the parallel views about ‘good’ (E 66–9). In his critique Ross cited Moore’s ‘convincing refutation’ of subjective naturalism about ‘right’ (RG 11; also FE 22–6) and added further objections of his own. One said the feeling of approval the view cites cannot just be approval in a generic sense; it must be the specifically moral approval we feel toward right acts but not toward, say, works of art. This approval is not just a feeling that arises in us; it is an intellectual emotion that presupposes the judgement that what we approve of is right, a judgement that cannot be reduced to one about feeling (FE 23; also Sidgwick, ME 27). This objection echoes his critique of fitting-attitudes analyses of ‘good’ (FE 278–9, 282) and is telling against subjective views that often say little about what distinguishes the feeling they say moral judgements report. His lengthier critiques of subjectivism about ‘good’ (RG 75–104; FE 258–62) again supplemented Moore’s points with others, such as that this view makes it impossible to desire something because it is good and conflicts with our belief that many things that are desired are bad (RG 97–8, 99–103). Broad ascribed subjective naturalism to Hume and made objections to it like these last two of Ross’s (FT 86). He did not take them to be decisive but added that on Hume’s view the way to settle moral disputes is to collect statistics about how people feel; he found this ‘simply incredible’ (FT 114–15, 281; also Moore, E 54–5, ‘NMP’ 330; Ewing DG 40–1, E 100). Ewing too rejected subjective naturalism (DG 4–10, 42 Stevenson, ‘Moore’s Arguments’, pp. 73–7; Moore, ‘R’ 547–51. Ewing too said Moore’s objection fails if speakers use ‘right’ in different senses, each relativized to its user, but he could not believe ‘right’ always changes meaning in this way (DG 9).

non-naturalism

103

40–2, 53–70), and taken together the school’s objections to it are surely decisive. But the more credible subjective views are not naturalist but non-cognitivist. The autonomy thesis that separates non-naturalism from naturalism is shared by non-cognitivist views, which say moral judgements express emotions or issue prescriptions. Though subjective, they avoid the objections to subjective naturalism, for example by allowing moral disagreement. If one person says an act is right and another says it is wrong, their statements do not literally contradict each other but nonetheless conflict, as contrary commands or emotions do. Non-cognitivism takes moral judgements to be distinct not because they assert a distinct kind of truth but because they do something other than assert, and moral conclusions cannot be derived from non-moral premises because they have a kind of meaning such premises lack. Though ignored early in our period, non-cognitivism emerged forcefully in the 1930s in writings by Ayer and Stevenson, the former of whose Language, Truth, and Logic was the main focus of the school’s initial response.43 Ayer defended the logical positivists’ verifiability criterion of meaning, which says a meaningful sentence must be either empirically verifiable or analytic. Since moral sentences are neither, he argued, they require some other treatment, and he proposed that they have no factual significance but instead express feelings, in particular of moral approval or disapproval. This was metaethical emotivism as part of a broader positivist view. Given their belief in synthetic a priori truths, our school strongly resisted Ayer’s verifiability criterion,44 but they objected especially to his use of it to ground his emotive theory. Carritt said this theory’s motivation was ‘not any consideration of [moral judgements’] own nature, but the desire to support a peculiar view of truth (‘MP’ 132–3), and Ross and Ewing concurred (FE 35; ‘BET’ 91). If a criterion of meaning does not fit the natural understanding of our moral language, they argued, that is not a reason to adopt a counterintuitive theory of ethics but instead one to abandon the criterion. Broad wrote that too many writers on ethics: seem to have little direct acquaintance with or interest in moral problems. They have accepted a general theory of judgment and significance which was devised without reference to moral phenomena, and into which such phenomena fit very badly, and they now devote themselves with infinite ingenuity to stretching the facts to fit the theory.45

The school also had more specific objections. Carritt argued that Ayer’s attempt to distinguish his emotivism from subjective naturalism, which Ayer agreed Moore had Ayer, Language, Truth and Logic, Chapter 6; Stevenson, ‘Emotive Meaning’. The earliest defence of non-cognitivism was in Ogden and Richards, Meaning of Meaning (1923), which influenced Ayer and Stevenson. There were also oral defences of it in the early 1930s by Austen Duncan-Jones (Broad, CE 107–10). 44 Broad, ‘SAP’; Ewing, ‘Meaninglessness’ and ‘Linguistic Theory’; Ross, FE 35–6. 45 Broad, Critical Notice of Prichard, Moral Obligation, p. 566; also CE 190. By contrast, Prichard ‘was really at home in the facts about which he was philosophizing’ and had a ‘healthy naivety and commonsense which made him immune to clever talk and ingenious fictions’. 43

104 british ethical theorists from sidgwick to ewing refuted, fails. A moral judgement is not an involuntary symptom of feeling, as sweating is of pain, but a voluntary attempt, using conventions of language, to convince your audience of something; if its subject is, as Ayer held, your feelings, it is indistinguishable from a report of those feelings (‘MP’ 133–4, EPT 32–3; also Ewing, DG 10–12). Though this objection is still made,46 it is not persuasive. It has no force against versions of non-cognitivism that assimilate moral judgements to imperatives, which are expressed conventionally, and there are also conventional expressions of feeling, such as ‘boo!’ and ‘hurrah!’ Carritt also argued that, if emotivism is true, speakers who make moral utterances are deluded about what they are doing. Ayer conceded that most speakers think that in making a moral utterance they assert a moral proposition; if there are no such propositions, they must be mistaken about their action (‘MP’ 134–5; EPT 33–4; also Ewing, DG 12–14, ‘BET’ 90). This is a much better objection and has recently been restated in terms of speechact theory. If the speech-act you perform depends on your intentions, how can someone who intends to assert a moral proposition be performing the different speech-act of expressing a feeling or issuing a prescription?47 Ross added a characteristic objection. To Ayer’s claim that moral judgements express a ‘special sort of moral disapproval’ he replied that ‘whatever be true of dislike, it is impossible to disapprove without thinking that what you disapprove is worthy of disapproval’ (FE 34; also Broad, CE 242–3). He also gave an early version of the ‘Frege-Geach objection’. It says that while non-cognitivism may give a plausible account of the use of moral terms in assertions such as ‘You ought to do x’, it does not fit their use in non-assertoric contexts such as questions (‘Ought I to do x?’) and the antecedents of conditionals (‘If he ought to do x, he will do x’): there no attitude is being expressed nor is any prescribing going on.48 Ross wrote: The only moral judgments of which it could with any plausibility be maintained that they are commands are those in which one person says to another ‘you ought to do so-and-so’ . . . But if we are to do justice to the meaning of ‘right’ or ‘ought’, we must take account also of such modes of speech as ‘he ought to do so-and-so’, ‘you ought to have done so-and-so’, ‘if this and that were the case, you ought to do so-and-so’, ‘I ought to do so-and-so’. Where the judgement of obligation has reference either to a third person, not the person addressed, or to the past, or to an unfulfilled past condition, or to a future considered as merely possible, or to the speaker himself, there is no plausibility in describing the judgement as a command. But it is easy to see that ‘ought’ means the same in all these cases. (FE 33–4)

Some of Ross’s counterexamples, such as third-person judgements, tell only against the view that moral judgements express singular imperatives and not against the more sophisticated view, defended later by Hare, that they address general prescriptions to everyone; nor do they apply to the view that moral judgements express

46 48

47 Jackson and Pettit, ‘Problem for Expressivism’. Cuneo, ‘Saying What We Mean’. Despite its earlier use by Ross the objection is usually credited to Geach, ‘Assertion’.

non-naturalism

105

attitudes. But his example of ‘ought’ in a conditional is central to the Frege-Geach objection, as is his claim that an account of ‘ought’ must fit all its uses. Though he may not have grasped the objection’s full import, he made its central point. If these initial responses to non-cognitivism were hostile, some later ones were more sympathetic. In his contribution to The Philosophy of G.E. Moore Stevenson argued that ‘x is right’ means ‘I approve of x; do so as well’. Moore replied that this analysis would be better without its first, subjective naturalist part, so it says only ‘Approve of x!’ (‘R’ 542). He then said, of the resulting non-cognitivist view: I certainly think that [it] may be true: that is to say, I certainly think that I don’t know that it is not true. But this is not all. I certainly have some inclination to think that it is true, and that therefore my own former view is false. . . . And, if you ask me to which of these incompatible views I have the stronger inclination, I can only answer that I simply do not know whether I am any more strongly inclined to take the one than to take the other. (‘R’ 544–5, also 546–7, 554)

His indecision did not last long. After Moore’s death Ewing reported that he ‘completely retracted this statement in the later years of his life. . . . Moore told me orally that he still held to his old view, and further that he could not imagine whatever in the world had induced him to say that he was almost equally inclined to hold the other view’.49 Even if briefly, however, he saw non-cognitivism as an equal rival to non-naturalism. This also came to be Broad’s view. Having early on found non-cognitivism ‘plausible enough to deserve very serious consideration’ (CE 110), he turned against it when Ayer connected it to logical positivism. But in the late 1950s he reverted to his earlier view, saying the ‘non-predicativists’ Stevenson and Hare had made ‘valuable contributions’ and helped make ethics ‘one of the liveliest branches of philosophy’ (CE 307). The cause of his reversal was a new argument for noncognitivism starting from judgement-internalism. If moral judgements are intrinsically motivating, it says, they must express mental states that involve motivation; since attitudes do this and beliefs do not, the judgements must express attitudes. Having earlier made externalist claims, Broad now said the state of mind expressed by saying an act is wrong ‘always tends to evoke a reaction against doing the act in question’; if this psychological claim is analytic, as ‘might fairly be alleged’, this would be ‘a point in favour’ of non-predicativism (CE 306–7, also 367–8). This argument, which goes back to Hume, was not made by Ayer and made only briefly by Stevenson,50 but it was central to later writers such as Hare and is closely associated

49 Ewing, ‘G.E. Moore’, p. 251. Blanshard likewise said Moore’s sympathetic belief ‘was only temporary. He held it in the early forties, when for six months he was my guest in the USA. When I asked him later (in 1950, as I remember) whether he still felt equally drawn to the two positions, he answered with an emphatic No; he was now more inclined to think his earlier view the right one. In 1955, I put the same question again, and received the same answer, still more emphatically given’ (Reason and Goodness, p. 269n.) 50 That Ayer and Stevenson did not use the internalist argument is argued in Mahon, ‘Emotivism and Internalism’. The twentieth-century origin of the argument seems to be Field, Moral Theory, pp. 46–64.

106

british ethical theorists from sidgwick to ewing

with non-cognitivism today.51 It is also, unlike Ayer’s appeal to verifiability, based squarely on a claim about moral judgements as such and made Broad see noncognitivism in a new light. The change in Ewing’s view was even greater. Having rejected non-cognitivism in The Definition of Good (DG 10–16), he had come by Ethics of 1953 to think, like Broad, that there is something important in the thought that moral judgements have an emotive and practical function (E 115–16, 120–1). And in Second Thoughts he embraced the internalist view that a moral judgement must express ‘some inclination in favour of the act or aversion from it’ (ST 74), and, based partly on that view, defended a ‘middle way’ in metaethics that combines non-cognitivist and nonnaturalist elements. On this mixed view, to say an act is right is both to express an attitude in favour of it and to say there are good reasons for this attitude, so the attitude is ‘justified and indeed imperatively required by the facts’, where this last claim expresses a proposition and can be true (ST 3, 61, 77; VR 96, 102, 193–4). It was in connection with this claim that he denied non-natural properties and posited indirect truth-making. But asserting truth was not all a moral judgement did on his later, mixed view; it also expressed an attitude. These later sympathies should not be exaggerated. Moore and Broad never fully endorsed non-cognitivism, seeing it at best as an equal rival to non-naturalism, while Ewing’s view retained a non-naturalist element in its claim about justification. But there are no similar instances of sympathy for metaethical naturalism, which the school always resisted.52 I think this is because the autonomy-of-ethics thesis, which non-naturalism and non-cognitivism both affirm, was more important to them than their moral realism. While the former was crucial for their approach to normative ethics, the latter was not. Embracing naturalism would have required a major change in their style of moral argument, making it empirical rather than a priori. But accepting non-cognitivism would have made little difference to it, giving just a different interpretation to arguments that remained essentially the same. Though the contrary is sometimes alleged,53 non-cognitivism does not make right and wrong depend on anyone’s attitudes. It says that to judge lying wrong is to express a negative attitude to lying, but this attitude is not conditional on any other attitudes. You do not say, ‘Lying is wrong when done by someone with a negative attitude to lying’; you say simply ‘Lying is wrong’, and your attitude therefore applies to all acts of lying, whatever the liar’s attitude to his own act. If you see someone who does not have a negative attitude to lying lie, your attitude to his act is negative. 51 Meyerhoff, ‘Emotive and Existentialist Theories’, p. 777; Hare, Language of Morals, pp. 20, 29, 171; Urmson, Emotive Theory, pp. 18–19, 20–1; Nagel, Possibility of Altruism, pp. 7–8. 52 Some accepted naturalist analyses of some uses of ‘good’, but none accepted a naturalist account of the moral ‘ought’. 53 Dworkin, ‘Objectivity and Truth,’ p. 112; Enoch, Taking Morality Seriously, pp. 36–7. Of our school only the very late Ewing made this objection to non-cognitivism (VR 97–8); perhaps the others saw that it addresses a straw man.

non-naturalism

107

Similarly, if you imagine a possible world where you yourself lack a negative attitude to lying and lie, your attitude, from the actual world, to your lie in that world is negative. You can therefore say ‘Lying would be wrong even if I did not have a negative attitude to it’ and even ‘Lying would be wrong even if no one had a negative attitude to it’. Non-cognitivism takes moral judgements to express attitudes, but if the judgements are categorical rather than hypothetical in form the attitudes too must be categorical. But then the normative methodology of non-naturalists and non-cognitivists can be essentially the same. Each can start from moral judgements that are intuitive, in the sense that they are immediate rather than based on inference. Each can then reason from those judgements toward a more comprehensive moral view, perhaps seeking what is today called ‘reflective equilibrium’.54 The two views will interpret the initial intuitions differently, one seeing them as beliefs that claim to be true, the other as expressing attitudes. But this difference will not affect how reasoning from them proceeds: on both views it can involve efforts to harmonize judgements about different subjects and at different levels of generality. And to someone observing the reasoning it can be unclear and unimportant whether the reasoner interprets his activity in a non-naturalist or non-cognitivist way; either way the procedure will be the same. Sidgwick, Rashdall, and Moore thought their consequentialist principle is objectively true, but their defence of it would not have differed significantly if they took their acceptance of it to express an attitude. Prichard, Carritt, and Ross could likewise have made the same objections to consequentialism if they interpreted their deontology non-cognitively.55 The overlap between the two metaethical views was noted in a hostile way by Alasdair MacIntyre. He said non-cognitivism only makes explicit what was implicit in Moore’s theory: that instead of rational argument on moral questions there can only be unsupported assertions of the speaker’s preferences.56 MacIntyre’s own ‘rational’ argument involved a relativist appeal to the traditions of one’s society that both non-naturalists and non-cognitivists will reject as insufficiently critical; they will also deny that ethics can proceed without intuitions. But on the larger metaethical point MacIntyre was right. As interpretations of moral thought, nonnaturalism and non-cognitivism are sufficiently close that in reading our school’s normative writings we can largely set their realism aside.

54

Rawls, Theory of Justice, pp. 20–1. Their arguments are also consistent with a version of error theory that says our moral judgements express categorical, universal attitudes, though we do not take them to do just that, and says we can continue making them so long as we understand them for what they are; Robinson’s version in ‘Emotive Theory’ was of this type. 56 MacIntyre, After Virtue, pp. 16–17. 55

5 Intuitionism Alongside their non-naturalism, the school shared a distinctive epistemological view. They thought we can know some moral truths by direct insight or intuition, and were therefore moral intuitionists. Intuitionism is a near-corollary of non-naturalism. If moral truths are distinct from natural ones, it would seem they cannot be known by the empirical or scientific methods that reveal natural truths. This conclusion does not strictly follow, since it is in principle possible for the same method to yield knowledge of different kinds. But the natural truths are often defined epistemically, as those that can be known by the methods of science; this was at one time Moore’s view (‘P’ 13; also PE 40).1 Given this definition, non-natural truths must be known in some other way. Moreover, if moral truths are necessary—if they hold in all possible worlds—our method of discovering them must be in general capable of yielding necessary knowledge, as empirical methods are not. Some moral truths can be known by inference from others, but this cannot on pain of regress be the case for all moral truths; some must be known to be true or probable non-inferentially or by themselves, that is, by intuition. Our school did not often make this argument for intuitionism explicitly. There is a passage where Sidgwick hints at it (ME 97–8) and a few where Ewing does (MP 187; ‘RI’ 38–9; E 136), but they seem for the most part just to have assumed that moral knowledge is intuitive.2 I have said their style of moral argument was consistent with non-cognitivism, but their further descriptions of it always placed it in a nonnaturalist context and my discussion will do the same.

5.1 Intuition and Self-Evidence Sometimes the school characterized intuition merely negatively. Moore said that by calling certain propositions ‘intuitive’ he meant ‘merely to assert that they are incapable of proof ’ and implied ‘nothing whatever as to the manner or origin of our cognition of them’ (PE x, also viii; EE 29, 67, 163). In this he echoed Sidgwick, who

1 In PE Moore sometimes identified the natural as what exists in time whereas the non-natural does not (PE 40–1, 100–11), but surely if an experience can be pleasant at one time and not another, it can be good at one time and not another. 2 Sturgeon, ‘Ethical Intuitionism’, pp. 190–1; Phillips, Sidgwickian Ethics, pp. 54–8.

intuitionism

109

said intuitive truths are known ‘not as the result of reasoning’ (ME 211, also 97) and ‘may be properly cognisable without being viewed in connexion with other propositions’ (EEM 29). These particular claims do not distinguish intuitive knowledge from all other kinds, since perceptual beliefs too are arrived at without reasoning. The better negative claim was Ewing’s, who said ‘intuition’ in ethics means only ‘cognition otherwise than by inference or by mere observation’ (MP 10, also 14, 186; VR 43; Sidgwick, ME 34, n1). Other characterizations were more positive, for example that intuitive knowledge is ‘immediate’ or involves ‘direct apprehension’ or ‘direct inspection’ of a proposition, or that truths knowable by intuition are ‘self-evident’. Moore explained this last term by saying a self-evident proposition is ‘evident or true, by itself alone’ (PE 143); Sidgwick expressed the same idea more fully when he said self-evident truths are such that ‘when their terms are properly understood, the perception of their absolute truth is immediate and irresistible’ (ME 229, also 379); and Broad called a proposition selfevident when it is ‘such that a rational being of sufficient insight and intelligence could see it to be true by merely inspecting it and reflecting on its terms and their mode of combination’ (‘SAP’ 102–3). These last claims come close to present-day definitions of self-evidence on which if you adequately understand a self-evident proposition you are justified in believing it to be true, and, if you do believe it, know it.3 They also successfully distinguish intuition from sense-perception, since you cannot know an object in front of you is red just by understanding that proposition; you have to look. So a positive account of intuition says some moral truths are such that merely understanding them suffices for knowing them. In giving this account the school assumed that we can intuit moral truths, or have a faculty for doing so.4 They tended to call this faculty ‘reason’ and took it to have a wider field of operation than just morality; they thought it is by reason that we grasp self-evident logical principles such as modus ponens, mathematical truths such as 2 + 2 = 4, and metaphysical truths such as that every event has a cause. Some explicitly contrasted this view of intuition as rational with the idea that we make moral judgements through a ‘moral sense’. To Sidgwick the term ‘sense’ suggested ‘a capacity for feelings which may vary from A to B without either being in error’ (ME 34), but though this may be true of taste it is not true of vision, which can be accurate or not about the external world. He made a better point when he said even particular moral judgements are implicitly universal, since they assume a particular right act has properties that would make any similar act right (ME 34; also OHE 229; Broad, FT 177–8); this is relevant because universals are usually thought to be known by reason, as more specifically are necessary truths. And whereas ‘moral sense’ suggests

3

Audi, The Good in the Right, pp. 48–9; Parfit, On What Matters, Vol. 2, pp. 490, 508–9. Ewing actually denied that there is a faculty of intuition, at least in any sense that goes beyond our having specific cases of intuitive knowing (‘RI’ 45). 4

110

british ethical theorists from sidgwick to ewing

a single-purpose faculty with a specifically moral subject matter, ‘reason’ suggests a more widely applicable one. The claim that we can intuit moral truths is controversial and generates an epistemological version of Mackie’s argument from queerness: just as nonnaturalism posits properties unlike any found by science, so intuitionism posits a mysterious faculty of knowledge. But for the reason just given, one version of this objection misfires. Mackie denied that we know moral truths by ‘a special sort of intuition’, while Sharon Street has objected to a ‘highly specialized, sophisticated capacity, one specifically attuned to the evaluative truths in question’.5 But our school assimilated moral knowledge to knowledge of other a priori truths and often said we have as much reason to trust our moral intuitions as to trust our beliefs in the axioms of logic and mathematics. Sidgwick said his moral axioms ‘do present themselves as self-evident; as much (e.g.) as the mathematical axiom that “if equals be added to equals the wholes are equal”’ (ME 383, also 507), while Ross said ‘In our confidence that these propositions [about prima facie duty] are true there is involved the same trust in our reason that is involved in our confidence in mathematics’ (RG 30, also 32–3, 82; KET 42, 85; Rashdall, TGE 147–8, E 14–15, 40, ICE 38–9; Prichard, MW 18–20; Carritt, EPT 2–3, 43; Ewing, MP 10–11). There may be special objections to moral intuition that do not apply to its mathematical form. One argues from judgement-internalism to the conclusion that moral judgements, unlike mathematical ones, are not truth-apt. Another says we can explain why we make the moral judgements we do without presupposing that they are true but cannot do the same for our mathematical judgements; there we need to posit truth. And there are general objections to all claims to synthetic a priori knowledge. These last must give an alternative account of our logical and mathematical knowledge, and though several have been proposed it is an open question whether they succeed. But these larger issues are beyond the scope of this book, and it is in any case not essential to resolve them if the school’s normative theorizing can be separated from their metaethics. The important point here is that they did not posit a sui generis faculty of moral intuition but thought we know moral truths in the same way we know other a priori truths. If you intuit a self-evident proposition you at least apprehend its truth, but Sidgwick and Ross sometimes suggested that you also apprehend its self-evidence (ME 383; EEM 25; RG 33, 39–40). Robert Audi has argued that this is a mistake: a proposition’s being self-evident is a fact about its epistemic status, or about how it can be known, which you need not grasp when you intuit that it is true.6 But the mistake here is not a large one. If you believe a proposition just by understanding it you will normally be aware

5 Mackie, Ethics, p. 39; Street, ‘Darwinian Dilemma’, p. 143. Mackie acknowledged that the objection is weakened if non-naturalism can find ‘companions in guilt’, that is, other a priori truths known by intuition. 6 Audi, The Good in the Right, pp. 41–3.

intuitionism

111

that this is what you are doing, and if you take your belief to be justified you must believe the proposition can be known in that way. In a passage quoted above Sidgwick called the perception of self-evident truth ‘irresistible’ (ME 229). Some members emphasized how self-evident propositions force themselves on belief, and took this to be special evidence of their truth; Rashdall said ‘we have no reason for believing anything except the fact that we cannot help believing it’ (ICE 39; also McTaggart, ‘IV’ 434; Ewing, DG 32). But many propositions compel belief, including perceptual ones, and they can do so even when they are illusory; the stick in the water irresistibly looks bent. The main intuitionist claim should be just that if you understand a self-evident proposition, you are justified in believing it. This claim does not imply that self-evident propositions are obvious or will be known by everyone; the school repeatedly denied this. Prichard said, ‘I don’t think the apprehension of the self-evident easy to reach’,7 and this will be true if understanding a self-evident proposition, though sufficient for knowing it, is itself difficult or requires preparation. Right before defining self-evident truths as ones we can know by understanding them, Sidgwick said ‘their certainty cannot be seen except by minds carefully prepared’ (ME 229); elsewhere he said that for a truth to be selfevident to you there is often ‘required some rational process connecting it with propositions previously accepted’ (EEM 29). This point was especially emphasized by Ewing, who said ‘the best and most reliable intuition comes after reasoning and not before’, since an immediate ethical judgement ‘may be the fruit of long experience and thought about similar situations, without being itself a logical deduction or induction’ (E 141–2; also MP 193–4; ‘RI’ 43, 60–1, 63–4; VR 123). Rashdall, Carritt, and Ross thought you can only intuit successfully if your society and indeed humanity as a whole have developed intellectually (TGE I 84–5; EPT 139; RG 12), and a similar development is required in you. Ross said a principle of prima facie duty is self-evident ‘not in the sense that it is evident from the beginning of our lives, or as soon as we attend to the proposition for the first time, but in the sense that when we have reached sufficient mental maturity and have given sufficient attention to the proposition, it is evident without any need of proof ’ (RG 29). The only reliable moral opinions are therefore those of ‘thoughtful and well-educated people’ (RG 41), and something like that had been Sidgwick’s view. He preferred ‘the moral thought of the reflective few to that of the unreflective many’ (EEM 24), and to test his principles for consensus considered mainly the opinions of philosophers (ME 384–6). Broad and Ewing sometimes suggested that intuition has emotional preconditions: only if you feel certain emotions, which for Broad were ones of moral approval and disapproval, can your reason recognize acts as right or wrong (FT 108–10, 178–9, 268–70, 281–2; MP 194, 215). Neither explained exactly how emotions contribute to

7

Prichard, letter to Laird of 7.7.1938.

112

british ethical theorists from sidgwick to ewing

moral knowledge, and it is unclear how long Broad held this view, since he later made the Ross-like argument that approving an act presupposes judging that it is right (CE 242–3). But he and Ewing at one time thought moral intuition also requires emotion. Partly because they recognized these preconditions, many in the school emphasized the difficulty of achieving moral knowledge and the consequent fallibility of our intuitions. But these points emerged most clearly in Sidgwick’s account of the conditions for self-evident knowledge, which is the fullest the school gave.

5.2 Sidgwick’s Conditions Sidgwick gave this account at the start of his review of common-sense morality, and used it first to argue that deontological principles are not self-evident and then to argue that his own consequentialist principles are. The account lays down four conditions that distinguish ‘self-evident truths . . . from mere opinions’. The first condition says ‘the terms of the proposition must be clear and precise’; it is commonly called the clarity condition. The second says, ‘the self-evidence of the proposition must be ascertained by careful reflection’, while the third says ‘the propositions accepted as self-evident must be mutually consistent’ since ‘any collision between two intuitions is a proof that there is error in one or the other, or in both’. The fourth requires general agreement on the proposition, or a consensus about it: ‘since it is implied in the very notion of Truth that it is essentially the same for all minds, the denial by another of a proposition that I have affirmed has a tendency to impair my confidence in its validity’ (ME 338–42). There is an initial puzzle about this account. Its four conditions are supposed to be jointly necessary for self-evidence, but the second tells us to ascertain ‘the selfevidence of the proposition’. If this condition itself establishes self-evidence, what work is left for the others to do? Robert Shaver has proposed resolving this puzzle by taking the four conditions to establish something more than self-evidence. That a proposition is self-evident, he argues, follows from its satisfying the second condition, or that plus the first or clarity condition. But its satisfying the third and fourth conditions, which concern its coherence with other similar propositions and with other people’s beliefs, is necessary for it to have ‘the highest degree of certainty attainable’ (ME 338). What the tests together concern, then, is an epistemic status of ‘highest certainty’ beyond mere self-evidence.8 But this interpretation does not fit Sidgwick’s texts. He introduced the conditions as distinguishing just ‘self-evident truths’ (ME 338), and when defending his consequentialist principles did not claim for them any more than self-evidence (ME 379–89). Moreover, his remark about ‘the highest degree of certainty’ did not

8

Shaver, Rational Egoism, p. 64; Schultz, Henry Sidgwick, pp. 190, 200–4.

intuitionism

113

contrast that with self-evidence but only with a proposition’s being ‘apparently selfevident’ (ME 338), and a contrast between apparent and real self-evidence appears often in his work (ME 34n2, 211, 383, 508; EEM 25, 30, 31–2, 33–4, 166, 170). I think the best reading takes him to have been a little careless in stating his second condition, which in its best version asks us to establish the apparent self-evidence of a proposition, or establish that it really is apparently self-evident. The picture is this.9 You start by being in what seems to you to be a mental state of apprehending a proposition as true just by understanding it. Your actually apprehending a proposition in this way does not guarantee that it is true, any more than your seeing a stick as bent guarantees that it is bent. But you need to confirm that you really are in that mental state, and that is what the second condition requires. Sidgwick’s justification for it was precisely that we are often mistaken about our mental states; we tend to confuse genuine intuitions with ‘mere impressions or impulses which to careful observation do not present themselves as claiming to be dictates of Reason’ or with ‘mere opinions’ that repetition has given ‘a false appearance of self-evidence’. Since ‘any strong sentiment, however subjective, is apt to transform itself into the semblance of an intuition’ (ME 339, also 211–12, 340–1; EEM 25, 31), we need to detect such semblances. This requires ‘careful reflection’ (ME 339), which for British philosophers of Sidgwick’s day meant introspection or self-examination more generally.10 Thus whether you really are having an intuition ‘can only be decided by each person by direct introspection or reflection’ (ME 211, also 212, 362, 383; EEM 43), or by that plus a survey of possible distorting causes of your belief, such as common acceptance in your community or a strong desire that the belief be true (ME 339). The second condition therefore requires a self-survey. You are in what seems to you to be a mental state of apprehending a proposition as true just by understanding it, or of apparent self-evidence. You then apply introspective and other tests to determine whether it really is a mental state of that kind. If it is, it involves a real case of apparent self-evidence, and you can then apply the other conditions to see if the proposition is simply, and not just apparently, self-evident. Of these other conditions, the first or clarity condition is closely tied to the second and perhaps inseparable from it. An unclear proposition can lead you to believe falsely that you are having an apparent intuition; more strongly, if a proposition is unclear, it is hard to see how you can have the understanding of it that intuition requires. But the third and fourth conditions involve additional tests. You can apprehend each of two propositions as true and find that they contradict each other, or apprehend one as true and find that other people reject it. The situation when a belief that passes the first and second tests fails the third or fourth can be 9

Also Phillips, Sidgwickian Ethics, p. 60. An early and influential use of ‘reflection’ to mean introspection is in Locke, Essay Concerning Human Understanding, Bk. II, Chapter 1, para. 4. 10

114

british ethical theorists from sidgwick to ewing

described in two ways. We can say your initial mental state was not in fact one of intuiting the proposition, so the third and fourth conditions address the same question as the second but by non-introspective means. Or we can say you did intuit the proposition, but it turned out to be false. Sidgwick did not distinguish these possibilities, but my formulations have favoured the second, which allows an intuition to be false, because it seems closer to his texts; it was also preferred by Ewing. After saying it does not ‘matter very much’ which description we use, Ewing said he preferred the one where intuition can err because the alternative ‘suggests that there is some specific recognizable psychological state, that of having intuitions, that has the proud privilege of being infallible’, which he did not think is true (E 139; also VR 43–4). Sidgwick’s official statement of his third condition requires an apparently selfevident proposition to be consistent only with other apparently self-evident propositions (ME 341; also EEM 166), and with this restriction the condition is compelling. If two equally well-grounded beliefs contradict each other, you have no basis for preferring one to the other and should abandon both. But sometimes he spoke as if a self-evident belief must harmonize with all your beliefs about its subject matter, whether apparently self-evident or not (EEM 128, 167, 169, 170; ME 400), and this stronger condition is less plausible. If an apparently self-evident belief clashes with one that does not meet that condition, should the latter not yield? His fourth condition put him on one side of a present-day epistemological debate about ‘peer disagreement’, and in fact he gave a classic statement of the view that others’ disagreement should reduce your confidence in a belief.11 We can understand his reasoning for the condition as follows. You have a capacity for intuiting selfevident truths that is generally reliable but sometimes leads you astray. If someone with an equally reliable capacity does not share your intuition, one of the two capacities must be malfunctioning, and you have no reason to believe it is not yours (EEM 32). Unfortunately this reasoning conflicts with his own description of intuition as involving ‘immediate’ knowledge, for it implies that even in favourable cases you make an inference from the general reliability of a type of mental process to its reliability now. And there is already conflict with that description in the list of four conditions, since together they make the process of determining self-evidence anything but ‘immediate’. Given an initial seemingly intuitive belief, you engage in selfexamination, including a survey of possible unreliable causes of your belief, test its consistency with other beliefs, and check what other people believe. While the conditions as a whole try to secure against error, they also make the determination of self-evidence far from simply a matter of ‘immediate and irresistible’ perception (ME 229).

For a more recent statement see Elga, ‘Reflection and Disagreement;’ the contrary view is defended in Kelly, ‘Epistemic Significance’. 11

intuitionism

115

At the same time, the conditions make the intuition of moral truth in several ways fallible; certainly your initial sense that a moral judgement seems true is no guarantee that it is, since your belief can fail any of the second, third, and fourth conditions. Some commentators have contrasted Sidgwick’s fallibilism with the allegedly more dogmatic views of later writers such as Moore, Prichard, and Ross; this is meant to be another case where insights of Sidgwick’s were lost.12 But the contrast is again overdrawn, since many of those later writers likewise insisted that intuition is fallible. After calling the fundamental ethical truths ‘intuitions’ Moore denied ‘that any proposition whatever is true, because we cognise it in a particular way . . . in every way in which it is possible to cognise a true proposition, it is also possible to cognise a false one’ (PE x). Rashdall said repeatedly that we can be mistaken about moral questions (TGE I 85, 145, 211–12; E 35), as did Ewing (MP 185, 191; ‘RI’ 52–4, 56, 59; DG 28; E 139–40; VR 43–4). And they did so because, though not stating them as systematically, they shared many of Sidgwick’s conditions. Moore applied the first or clarity condition when he said the only way to guard against error in intuition is ‘by taking care, that, when we try to answer a question of this kind, we have before our minds that question only, and not some other or others’ (PE viii; also 223). Just as Sidgwick said those who deny that pleasure is the only good confuse instrumental with intrinsic evils (ME 402–6), so Moore said Sidgwick confused the question whether pleasure is necessary for value with the question whether it is the only value (PE 92–5). More generally, he thought Bentham, Mill, and Sidgwick ‘never even asked themselves the question which they professed to answer’ but confused it with another question; he also thought ‘the whole world would agree with us, if they could once clearly understand the question upon which we want their votes’ (PE 145). Only Ewing at the end of his career thought there can be intuitions of propositions that are not entirely clear (VR 46, 127). The second condition’s concern about mistaking your mental states was also shared. Rashdall spoke of ‘the real difficulty of distinguishing mere feelings or aversions which may be only prejudices due to inheritance or environment or superstition from real judgements of value’ (TGE I 211–12, also II 407, 411, 425–6). Thus a ‘strictly educated Scotchman’ should lose his horror of Sunday music when he learns the history of the traditions about Sabbath-observance; more seriously, belief in retributive punishment should weaken when we understand its origins in ‘the instinct for vengeance’ (TGE I 213, II 404, I 291, 305). Moore said we need to introspect to determine how firmly we really hold a belief (EE 162–3, 166), while Ewing worried that people ‘will be content to accept a passing fancy at the first moment instead of taking pains to make as sure as possible that their supposed intuition is a genuine and true one’ (MP 185, also 191), since ‘a genuine intuition may

12

Schultz, ‘Introduction’, pp. 28–9, 59n84; Henry Sidgwick, p. 197.

116

british ethical theorists from sidgwick to ewing

well be mixed up with false beliefs accepted on authority or derived from mistaken inference’ (E 139; also ‘RI’ 41, 63–5; Carritt, TM 44). The third condition was also widely accepted, since all recognized that conflicting intuitions cannot both be true (Rashdall, TGE I 89–90; Moore, EE 162, 165–6), while Ewing in particular emphasized a coherence test for moral truth (MP 207–9; ‘RI’ 47, 50, 56–7, 59, 62–3; DG 30, 83–93, 207–11; E 3–4, 23, 132–4; ST 72–3, 132–4). Many also accepted a consensus condition. Moore said others’ dissent will reduce your confidence in a proposition and therefore your justification for believing it (EE 163, 167–8) and called moral disagreement an ‘unsatisfactory state of things’ as our inability to prove a moral judgement is not (PE 75). He also claimed consensus about his views on the good (PE 188–9) and tried to explain away disagreements with them as resting on unclarities (PE 92–6). Ross said the only check on a belief about goodness ‘is that of inquiring whether others have an opposite opinion, or no opinion’ (‘BO’ 119), while Broad said that if most people have a certain moral opinion, a wise person will ‘attach very great weight to this fact’ (CE 154; also Ewing, VR 116). But the importance of consensus appears most clearly in the weight later members placed on agreement with common-sense morality. When Ross said that, unlike natural science, ethics must start from ‘what we really think’ (RG 39–40; FE 1–3; also Prichard, MW 2, 29–30, 121–2), he was appealing partly to his readers’ own intuitive judgements but also to ‘the main moral convictions of the plain man’ and ‘the opinions of the many or those of the wise’ (RG 21n; FE 3). For Carritt ‘men’s reflective conclusions on simple moral questions’ are ‘the philosopher’s only data’ (EPT 6),13 while Ewing said a philosophy must ‘be in accord with what we cannot possibly help believing in ordinary life’ (DG 32; also Moore, EE 195). These writers tested their theories against everyday beliefs because they assumed the generally accepted morality of a developed society will be broadly correct; in doing so they implicitly valued a kind of consensus. This aspect of their view leads to an objection contrary to that of dogmatism: that their methodology was inherently conservative, so instead of innovating ethically they merely reproduced the conventional morality, and in particular the middle-class morality, of their day. There certainly is something to this objection, since a procedure that seeks broad consensus can only go so far beyond common opinion. But the limit is not absolute. A moral reformer can detect contradictions in existing morality and try to remove them; she can also intuit a proposition no one has thought of before and, once it has been presented to them, others can intuit it too. The spread of this second kind of reform may take time, especially if the rest of society have to go through a process of moral development as the reformer did before them. But several members emphasized that such reform has happened. Rashdall said, ‘The higher moral ideals have in the first instance been recognized by exceptional minds’, on Likewise, ‘The only data for an aesthetic theory are aesthetic experiences, either first-hand or at second-hand from presumably sensitive minds’ (IA 7; also 19–20). 13

intuitionism

117

whose insights ‘the average man is largely dependent’ (E 88–9), while Ross praised the ‘creative morality’ of ‘a man who, going beyond the routine of the duties commonly recognized by those round him, becomes convinced of some new duty and devotes his life to the discharge of it, as for example Wilberforce did when he devoted his life to the abolition of the slave-trade and of slavery’ (FE 189–90; also RG 12); Carritt made a similar claim about Socrates (EPT 139). Our school may not have innovated much themselves; they were not moral reformers on the order of Bentham and Mill. But extreme conservatism was no more inherent in their intuitionism than extreme dogmatism was, and much of the conservatism there was reflected a laudably fallibilist concern for consensus with other minds.

5.3 Certainty and Inference If there are several conditions for self-evidence, there are several ways intuition can fail, and if you have not applied all the conditions to a moral belief you should not hold it with full confidence. But what if you have applied them all and not found that it fails any? How confident should you be of it then? Prichard and Ross thought you can often be certain. They had a conception of knowledge on which it involves ‘certainty or complete absence of doubt’ and on which any belief held with less than full conviction is at best an instance of ‘right opinion’, which is not in the same genus as knowledge (RG 146–7; also Prichard, MW 24). Their claim that we ‘know’ certain moral propositions therefore implied that we can be certain of their truth, and Ross said, ‘we seem to be able to judge without any possibility of doubt that certain simple types of action are right’ (‘BO’ 123) and contrasted our merely probable opinions about duty in concrete situations with ‘the certainty that attaches to our recognition of the general principles of duty’ (RG 30; also 36).14 But this approval of moral certainty was questionable even by his and Prichard’s lights. If apprehending the self-evident requires, as Prichard said, both ‘a developed moral being’ and a high degree of ‘thoughtfulness’ (MW 14n7; also Ross, RG 29), how can you be sure you satisfy those conditions and so can do the apprehending? Sidgwick, too, often spoke as if once you have applied all his conditions you can be certain of a belief. He said judgements about particular acts do not present themselves as ‘indubitable and irrefragable’ (ME 100), in contrast with judgements about principles, and sought principles ‘more absolutely and undeniably true and evident’ (ME 102) or whose ‘absolute truth’ we can immediately perceive (ME 229).15 He called his fourth condition ‘an indispensable negative condition of the certainty of

14 Audi is therefore wrong when he says there is ‘no equivalent suggestion’ in Ross to Sidgwick’s claim that moral knowledge can be certain (The Good in the Right, p. 31). 15 The last of these remarks came in his description of claims dogmatic intuitionists make, but since he was taking them to argue that their principles satisfy his conditions, they are evidence for this reading.

118

british ethical theorists from sidgwick to ewing

our beliefs’, contrasted the situation when it is not satisfied with ‘scientific certitude’ (ME 342), and claimed to have ‘clear and certain ethical intuitions’ (ME 387, also 391; EEM 25, 43). These statements are again at odds with his own epistemology. Even if you have introspected carefully, you may falsely believe you are intuiting when you are not; a proposition may be consistent with all those that seem selfevident to you but conflict with others you have not thought of; and though everyone you know accepts a proposition, some people you have not surveyed may not. He was being truer to his own position when he said, ‘I disclaim the pretension of establishing absolute truth or absolute error’ (EEM 169) or said he had only given a proposition ‘as high a degree of certainty as I can hope to attain under the existing conditions of human thought’ (EEM 32). For the most part, however, his view was not notably more fallibilist than Prichard’s and Ross’s, since he too held that we can be certain of some moral beliefs. Nor was his argumentative practice more fallibilist. Moore, Prichard, and Ross often made more confident moral claims than they should have, backed by overly hasty assertions of consensus. But as we will see in Chapter 7, Sidgwick did the same with his moral axioms, which he did not subject to nearly as careful scrutiny as he did deontological principles and for which he too claimed more agreement than is plausible. Ewing more consistently emphasized the fallibility of moral intuition, and he denied that it need involve certainty. Early on he said we can believe our intuitions often ‘have a very good chance of being approximately true’ (MP 186, also 191), but insisted they do not always give us certainty as against ‘a probable opinion’ (‘PKE’ 42). In Second Thoughts he opposed confining the term ‘intuition’ to cases involving certainty and added that a directly cognized proposition may present itself to us only ‘as having an inherent plausibility’ (ST 66, 72; also ‘RI’ 52–3, 56–7, 59–60; ‘RSE’ 81–2; VR 44). If this means we sometimes initially apprehend self-evident propositions only as probable, it suggests a more thoroughly fallibilist view. Just by understanding a proposition we can sometimes see it as probable, and its passing further tests can make us judge it even more probable, but at no point can we have certainty about it.16 A related question is whether a proposition we can know as self-evident can also be known in some other way, in particular by inference. Some members denied this. Moore said self-evident propositions ‘are incapable of proof or disproof ’ and that in their case ‘no relevant evidence whatever can be adduced: from no other truth, except themselves alone, can it be inferred that they are either true or false’ (PE x, viii, also 143–4, 148, Rashdall TGE I 100). Prichard too denied ‘the possibility of proving what can only be apprehended directly by an act of moral thinking’ (MW 19), while Ross spoke of propositions ‘that cannot be proved, but that just as certainly need no Audi defends a fallibilist view like this, saying it involves only ‘soft’ rather than ‘hard’ self-evidence (The Good in the Right, pp. 53–4; also Phillips, Sidgwickian Ethics, pp. 59, 86n9). 16

intuitionism

119

proof ’ (RG 30; also Carritt, TM 28).17 They did not always hold to this view, since they sometimes argued for abstract moral claims by appealing to intuitions about particular cases. And the view itself seems to rest on a confusion between explanatory and epistemic issues. If a proposition known as self-evident is axiomatic, there can be no explanation of why it is true, but it does not follow that there cannot be inferences to its truth from some of the judgements it explains. The law of gravity is not explained by facts about particular falling bodies, but those facts are evidence for the law. Moore confused evidentiary and explanatory senses of ‘reason’ when he argued that inferring a proposition from others would supply ‘a reason why it is true’ but a proposition’s being self-evident means ‘it has absolutely no reason’ (PE 143).18 Sidgwick did better when he noted ‘Aristotle’s distinction between logical or natural priority in cognition and priority in the knowledge of any particular mind’ (EEM 29; also ME 98–9). The logically prior is the more explanatory and is what an ideal mind would know first, but our minds often first know truths we will later explain, and reason from them to what supplies their ground. So in ethics we can first intuit particular moral truths and then reason from them to explanatory principles. Ross held that self-evident truths not only cannot be proved but ‘need no proof ’ (RG 30), and this makes sense if intuition yields indubitable knowledge. If you already are certain of a proposition, what could an inferential justification add? But if the result of even careful intuition is only a high-probability belief, there is room for inference to increase that probability, as Ewing saw: ‘We must not suppose that, because an intuition is not proved true by reasoning, therefore it cannot be supported by reasoning’ (E 133). When a belief seems only plausible, ‘confirmation by other ethical beliefs’, involving ‘tests which can be brought under the general heading of “coherence”’, becomes important (ST 72; also ‘RI’ 47, 50, 56–7, 59, 62–3; ‘BET’ 78–9; VR 45). Ewing was thinking here of more than just Sidgwick’s negative test of noncontradiction. It is not just that a proposition’s contradicting other self-evident ones counts against its being true; its positively cohering with other propositions, or forming a unified system with them, counts in its favour. And there is room for this favouring if your initial grasp of the proposition, though immediate, is of it only as probable rather than as certainly true.

5.4 Levels of Intuition Moral intuitionists can differ about what the primary objects of intuition are. Are they abstract principles, particular moral verdicts, or something in between? In the earlier parts of The Methods Sidgwick wrote as if the primary intuitions concern very abstract principles. He said particular moral judgements do not present 17 In an earlier article Ross had taken the opposite view, saying ‘the fact that something can be inferred does not prove that it cannot be seen intuitively’ (‘BO’ 121). 18 Audi makes this critique of Moore’s argument in The Good in the Right, pp. 13–14.

120

british ethical theorists from sidgwick to ewing

themselves as self-evident, so they do not satisfy his second condition (ME 100), and deontological principles, though somewhat general, do not satisfy the conditions as a whole. The only genuinely self-evident principles are certain ones ‘of too abstract a nature, and too universal in their scope, to enable us to ascertain by immediate application of them what we ought to do in any particular case’ (ME 379), that is, his axioms of justice, prudence, and benevolence. In Book III he argued that these axioms, intuited on their own, establish the formal elements of utilitarianism as what results ‘when the demand for really self-evident first principles is rigorously pressed’ (ME 388). It is not clear whether he thought his hedonist account of the good, which he defended later and which gives utilitarianism its content, is also selfevident or is known only in some lesser way. But at least here he wrote as if the only true intuitions are of abstract principles (also ME 214; EEM 24). His defence of utilitarianism also involved the negative claim that no other principles are self-evident, which he defended in his critique of deontology. But this negative claim is an essential part of testing by his conditions. The third of these requires a proposition to be consistent with all others that are apparently self-evident or meet the other conditions. If deontological principles did this, they would contradict the utilitarian ones and show them not to be self-evident. He therefore had to show that they fail some other conditions. But Sidgwick did not give only this abstract argument. Having apparently established utilitarianism in Book III on the ground that its core principles meet his four conditions, he argued in Book IV that it coheres with common-sense morality, in particular by giving its various principles a unifying rationale and enabling it to clarify them when they are vague or conflict.19 This later argument seems to give weight to particular intuitions. What has aptly been called ‘the puzzle in Sidgwick’s moral epistemology’20 concerns how the coherentist argument appealing to common-sense morality in Book IV relates to the argument about self-evidence in Book III. One interpretation says the later argument is needed to complete the earlier one. Sidgwick’s conditions require consistency with other propositions and with other people’s beliefs, and in showing that utilitarianism fits with common-sense morality he was showing that it passes these tests.21 But the conditions require consistency only with other apparently self-evident propositions and with others’ beliefs about them, and Sidgwick denied that common-sense beliefs about particular cases or principles are of that sort. So common-sense beliefs do not seem relevant to his consistency and consensus tests. Moreover, there was no suggestion in Book III that its argument for utilitarianism, which already included a survey of others’ opinions

19

Schneewind, Sidgwick’s Ethics, pp. 279–85, 331–6. Phillips, Sidgwickian Ethics, p. 65; also Brink, ‘Common Sense and First Principles’. Shaver, Rational Egoism, pp. 70–1; Crisp, ‘Sidgwick and the Boundaries’, pp. 68–9, 72; Schultz, Henry Sidgwick, pp. 190, 200–4. 20 21

intuitionism

121

(ME 384–6), was incomplete. On the contrary, its concluding affirmations of utilitarianism were firm and emphatic (ME 387–9, 406–7, also xix, xxi; EEM 24). These affirmations did have an unstated qualification. At the end of Book IV Sidgwick worried that his axiom of benevolence and a version of his axiom of prudence that supports egoism contradict each other and are therefore not self-evident (ME 507–8). But he did not there take this worry, which he left unstated in Book III, to suggest the need for a different kind of argument. On the contrary, he assumed that if his axioms did not contradict each other, they would be firmly established. A better resolution of the puzzle exploits Sidgwick’s distinction between ‘logical priority’ and ‘priority in the knowledge of any particular mind’. An ideal mind could apprehend the axioms as self-evident, and some people, perhaps including Sidgwick, can do that now. But others cannot, often because they have moral beliefs that contradict the axioms. They can nonetheless be led to accept the axioms by an argument that starts from beliefs they now have and shows that some other propositions lie behind and explain them, so the latter and not the former are fundamental. Unlike a direct appeal to self-evidence, this second argument is in a sense ad hominem, since it is directed to a particular person with a particular set of beliefs. It therefore lacks the universal applicability and full justificatory force of the direct argument for self-evidence, which Sidgwick hoped it will eventually lead the person to accept (EEM 29–30). But it does seem to be what Sidgwick gave in Book IV. After noting that first principles cannot in the strict sense be ‘proved’, he said, ‘if Utilitarianism is to be proved to a man who already holds some other moral principles’, the argument must start with premises he accepts and establish ‘a conclusion actually superior in validity’ to them (ME 419). Unless his principles are taken as at least provisionally valid, ‘the so-called proof does not seem to be addressed’ to him; that requires ‘a line of argument which on the one hand allows the validity, to a certain extent, of the maxims already accepted, and on the other hand shows them to be not absolutely valid, but needing to be controlled and completed by some more comprehensive principle’ (ME 420). What seems to be envisioned here is an inferential justification starting from beliefs another thinks are axiomatic and showing that they in fact depend on one that is more fundamental. That fundamental principle is not, at least initially, intuited directly but is reasoned to from beliefs it will eventually supplant. The Book IV argument is therefore a secondary argument directed at those who cannot yet grasp the axioms’ self-evidence. Given its starting-point in beliefs that, read strictly, are false, the argument is not as cogent as a direct appeal to the four conditions. But it can lead someone to accept the axioms and eventually to see their self-evidence.22 22 For similar interpretations of the argument of Book IV, see Singer, ‘Sidgwick and Reflective Equilibrium’; Skelton, ‘Sidgwick’s Moral Epistemology’. A related but different reading is in Phillips, Sidgwickian Ethics, pp. 62–84.

122

british ethical theorists from sidgwick to ewing

Sidgwick continued to think, then, that the most reliable intuitions are of abstract principles, and that was also Moore’s official view. In Principia he took the propositions for which ‘no relevant evidence whatever can be adduced’ to be claims that answer the question ‘What kind of things ought to exist for their own sakes?’ (PE viii). The generalization ‘Pleasure is the only good’ has the right form to be an axiom (though it is false), but ‘this experience is good’ does not, because its truth follows from the facts that the experience is pleasant and that all pleasure is good. Moore sometimes followed his official view, as in his bald assertions that ‘by far the most valuable things . . . are . . . the pleasures of human intercourse and the enjoyment of beautiful objects’ (PE 188) and that it is ‘self-evident that knowingly to do an action which would make the world, on the whole, really and truly worse than if we had acted differently, must always be wrong’ (E 77).23 But at other times he defended abstract claims about the good by appeal to particular intuitions, and this was in fact required by his method of isolation, which determines whether something like beauty or pleasure is intrinsically good by asking whether a world containing only that state is good. His applications of this method often involved imagining quite specific worlds, for example, a beautiful one with ‘mountains, rivers, the sea; trees and sunsets, stars and moon’ (PE 83) or, in his critique of hedonism, one with ‘no contemplation of beauty, no personal affections—but in which the greatest possible pleasure would be obtained by a perpetual indulgence in bestiality’ (PE 95, also 197). Here general conclusions about the value of beauty or pleasure were derived from judgements about particular possible worlds. The contrary view to Sidgwick’s and Moore’s official one was taken in their earlier writings by Carritt and Ewing, who thought the most reliable intuitions concern particular cases. Carritt rejected the assumption ‘that general definitions of right conduct are more certainly recognized to be true than particular judgements of what we ought to do. . . . particular judgements about the right act in some particular situation, real or imaginary, are our only data for defining right acts’ (TM 31, also 70–1, 84–5, 114–15, 138–9). Ewing said that if an intuition concerns general principles, this ‘greatly reduces its authority and value’, (MP 119n); an ethical conclusion ‘cannot be deduced from universal laws’ but is ‘the cognition that an individual act in an individual situation is right or an individual complex whole better than another such whole . . . the only really adequate premise is the individual situation in all its individuality’, because in ethics ‘we see the particular before the general’ (MP 187–8, 201, also 160–1, 174–5, 179, 183–4, 193–4, 202). They did not deny that general rules

Baldwin objects that Moore’s ‘direct proclamations’ of his values ‘will not do’, since what is needed are ‘interlocking conceptions of the self and the community which show how ethical values have a place in human life’ (G.E. Moore, pp. 129, 117; also 118–19). Moore would reply that any attempt to make Baldwin’s vague proposals concrete will either rely tacitly on the intuitions it is meant to avoid or not yield substantive results. 23

intuitionism

123

can be formulated and are sometimes useful in moral thinking, but insisted that our evidence for them is always particular intuitive judgements. Prichard and Ross had an intermediate view. They denied that we can intuit abstract moral principles just as such, as Sidgwick and Moore thought; what prompts intuition is always a particular act or state of affairs. But what we intuit in a particular situation is implicitly general, for example, that an act’s having a certain property tends to make it right and will therefore tend to make any other act with that property right. Prichard said we only come to see the truth of a moral principle ‘in connexion with a particular instance’. I see ‘that I ought to pay some particular debt and then recognize that this obligation does not depend on the action being one of paying, say, £5 to another who would like it . . . but simply on its being one of paying to another something which I owe him . . . ; I then see that similar obligations will hold elsewhere’ (MW 5). Or ‘we recognize . . . that this performance of a service to X, who has done us a service, just in virtue of its being the performance of a service to one who has rendered a service to the would-be agent, ought to be done by us’ (MW 13, also 20, 77, 118). Prichard is sometimes read as a radical particularist; Ewing took him to hold ‘that there are no general ethical principles, but we know immediately what is right in each particular case, and leave it at that’ (MP 157). But though Prichard said, ‘what we directly recognize is always a particular obligation’, he added that in recognizing it ‘we imply the truth of a principle’; when we ‘recognize that we ought to do that act in virtue of its being the repayment of something promised . . . we thereby imply the principle that we, and indeed anyone, ought to repay what we have, or he has, promised to repay’ (MW 63). Ross had a similar view. In The Right and the Good he said, ‘we see the prima facie rightness of an act which would be the fulfilment of a particular promise, and of another which would be the fulfilment of another promise, and when we have reached sufficient maturity to think in general terms, we apprehend prima facie rightness to belong to the nature of any fulfilment of promise’ (RG 33, also 36). In Foundations he described a situation where he saw a blind man crossing a street through traffic and instinctively helped him. He denied that his belief that he should do this was arrived at by inference from principles; on the contrary, we all first saw rightness ‘as belonging to particular acts’, but because we saw it as belonging to them ‘in virtue of a particular character they possessed . . . the general principle was later recognized by intuitive induction as being implied in the judgements already passed on particular acts’ (FE 168–71, also 84, 84, 315; Broad FT 145–6, 177–8, 271–2; Ewing ‘BET’ 80).24 The primary intuitions concern particular acts but see certain properties 24 Prichard took this view not only about moral knowledge but about a priori knowledge more generally, including of space: ‘Though we come to apprehend a priori the nature of space in general, the apprehension is not prior but posterior in time to the apprehension of individual spatial relations . . . it is reflection on the general nature of space, the apprehension of which is involved in our apprehension of individual spaces or rather of bodies in space, which gives rise to the apprehension of the totality of spaces’ (Kant’s Theory of Knowledge, pp. 42, 46).

124

british ethical theorists from sidgwick to ewing

as making the acts at least prima facie right, so the intuitions are implicitly general and lead by a ‘short and inevitable step’ (FE 320) to a principle. To contrast this view with Carritt’s and Ewing’s consider a question about the latter: can you know a particular act is right without knowing which non-moral properties make it right? In the abstract this seems possible. Mental states supervene on states of the brain, but you can know you are feeling a certain pain without having any idea what brain-states underlie it. But the parallel seems unlikely in the moral case, especially since, as Carritt and Ewing recognized (TM 72; MP 183–4), you cannot intuit that an act is right without knowing certain non-moral facts about it, such as that it will repay a debt, give another pleasure, and so on. Carritt and Ewing are therefore best read as holding that in intuiting an act’s rightness you are guided by beliefs about its right-making properties but not consciously; the beliefs are implicit, like your grasp of the criteria for applying a concept you cannot explicitly analyse. One difference between the Prichard-Ross and Carritt-Ewing views is therefore that the former takes the primary intuitions to involve explicit, and the latter only implicit, knowledge of right-making properties. Carritt’s and Ewing’s early view may seem closer to everyday moral thought. One lesson of recent discussions of, say, the trolley problem is that we often make particular moral judgements without being aware of the criteria that guide us in making them. Thus, we can judge it permissible to turn a trolley away from five people toward one but wrong to throw one person in front of the trolley, but not be able to say which properties of the two acts explain these differing verdicts. But Prichard and Ross were not trying to characterize everyday moral thought. They were describing what they took to be intuitions of self-evident truth, and it is arguable that these more plausibly involve the conscious application of criteria. If a moral judgement rests on merely tacit beliefs about grounds, how can you be sure those beliefs really are intuitive in the sense of Sidgwick’s second condition, or really cohere with similar beliefs and are widely shared? Though Ross thought we can be certain of the principles of prima facie duty, he denied that we can ever know a particular act is right. For that we would have to know not only which prima facie duties apply but also how they weigh against each other, and he thought we can never do that weighing. When two or more prima facie duties conflict, as he thought in real cases some always do, you can only ‘form the considered opinion (it is never more) that in the circumstances one of them is more incumbent than any other’ (RG 19). Your judgement has ‘none of the certainty that attaches to our recognition of the general principles of duty’ and allows you only to ‘believe something not self-evident at all, but an object of probable opinion, viz. that this particular act is (not prima facie but) actually right’ (RG 30, 33, also 31; ‘BO’ 123, 127; FE 186–9, 190–1, 315; Broad, FT 281–2). It may be thought that these claims depend on his demanding conception of knowledge, on which it involves certainty, and allow that you can have justified true beliefs, or what others would call knowledge, about particular cases. But when he emphasized the difficulty of weighing

intuitionism

125

duties and the wide disagreements about weights among those who agree what the duties are (FE 190, 187–8), he seems also to have been denying knowledge in a more ordinary sense. This denial marks a further difference between his view and that of Carritt and Ewing, who thought you can know an act is right.25 But his claims about particular moral knowledge were surely exaggerated. As Price asked, ‘May I not say that I know that I ought to save a drowning man rather than keep an appointment to play shovehalfpenny?’26 Or, since a trivial promise may be tacitly conditional on there not being some great evil you could prevent,27 can you not be as certain that it is permissible to cause one person mild pain if that is necessary to save a million people’s lives as you are that there is a general prima facie duty of gratitude? Ross may eventually have backed off his extreme view. In Foundations he first said we can never know our duty in a particular situation, but added, ‘where the one good is much the greater or the one obligation much the more stringent, we seem to be able to grasp these facts with certainty’ (FE 190–1). Despite denying that we can be certain of particular moral judgements, Ross sometimes appealed to them when defending his principles. One of his objections to utilitarianism was that it tells you to break a promise if that will result in a little more good (RG 34–5, 38), and like Moore he defended general claims about the value of virtue and pleasure by imagining specific possible worlds containing them (RG 134–5, 138–9). But his and Prichard’s view about the primary objects of intuition often made them emphasize a different kind of argument. The claim that an act’s being a keeping of a promise tends to make it right is partly explanatory. It says that if the act is right, its being a promise-keeping helps explain why it is right, and in many cases is the primary explanation. And Prichard and Ross often appealed to intuitions about explanation in defending their normative views. While Ross sometimes argued that utilitarianism yields incorrect particular judgements, he more often said that, even when it yields correct ones, it gives them the wrong explanation: ‘When a plain man fulfils a promise because he thinks he ought to do so, it seems that he does so with no thought of its total consequences . . . What makes him think it right to act in a certain way is the fact that he has promised to do so’ (RG 17, also 19, 24, 36–9). This was an appeal to ‘what we really think’, not about what is right, but about the explanatory question of why it is right. He did the same when he said the main test of utilitarianism is ‘do we really come to the conclusion that such an act as promise-keeping owes its rightness to its tendency to produce maximum good, or to its being an act of promise-keeping?’ (FE 69, also 65–8, 113,

25

There is no clear indication that Prichard shared this view of Ross’s; on the contrary, he sometimes wrote as if what you apprehend immediately is that, in virtue of having a certain property, a particular act is right all things considered (MW 5, 13, 20, 63, 77, 118). 26 Price, Critical Notice of RG, p. 344. 27 Ross argued that promises often have tacit conditions in FE 94–9.

126

british ethical theorists from sidgwick to ewing

187). Prichard likewise said, ‘Suppose we ask ourselves whether our sense that we ought to pay our debts or to tell the truth arises from our recognition that in doing so we should be originating something good . . . We at once and without hesitation answer “No”’ (MW 10). Or: ‘in thinking of our keeping our promise to X as a duty, we are thinking of the action as rendered a duty by its being the keeping of a promise. . . . . if we were to maintain that conduciveness to the agent’s advantage is what renders an action right, we should have to allow that any of our ordinary moral convictions . . . is simply a mistake (MW 29–30). For both Prichard and Ross our everyday judgements include ones about why acts are right, and a theory that contradicts those judgements must be rejected. Robert Nozick has said utilitarians’ attempts to derive the usual precepts of justice from their consequentialist principle ‘do not yield the particular result desired, and they produce the wrong reasons for the sort of result they try to get’.28 In emphasizing the second of these objections Prichard and Ross were responding in part to the tendency of utilitarians to point, plausibly or not, to remote or indirect effects of, say, promise-keeping that they say brings their theory closer to common sense. But they were also influenced by a view about the primary objects of intuition that makes them largely explanatory, so they concern, to quote Ross’s chapter title, ‘what makes right acts right’. They were not alone in emphasizing these intuitions; others such as Rashdall did too (TGE I 73, 99, 192, 203; E 66–7). But their distinctive moral epistemology made intuitions about explanation especially salient. There were, then, three views about the primary objects of moral intuition: abstract principles, particular acts, and, given such acts, truths about the properties that make them right. But few in the school adhered just to one view. Sidgwick, Moore, and Ross often appealed to particular moral judgements, Carritt raised issues about explanation (TM 69–70), and Ewing affirmed the general principle that ‘we ought always to produce the greatest good in our power’ (‘SN’ 10–11; also DG 188; E 82; ST 104–5). And it is consistent with their general moral intuitionism to imagine a methodology that gives some initial credence to judgements at all these levels and then tries to harmonize them, sometimes adjusting the more abstract to the more particular and sometimes doing the reverse, as in the method of reflective equilibrium. Though not explicitly endorsed by its members, this coherentist method is often a reasonable fit with the school’s practice and is for many present-day philosophers the best one available. Ewing came closest to endorsing this method, saying ‘when intuitive views differ, use may be made of inference to support one or other of the clashing views, especially by showing that it fits well into a coherent ethical system’ (DG 30, also 211). But ‘coherence’ can mean different things. Two judgements cohere minimally if they do not contradict each other; more positively, they can reinforce each other, as when a

28

Nozick, Anarchy, State, and Utopia, p. 202.

intuitionism

127

general principle gives a particular judgement a satisfying rationale, thereby increasing our confidence in that judgement, or when a particular judgement that is attractive in itself would be true if a given principle were true. But Ewing had a further, distinctive conception of coherence. He thought a moral view like Ross’s should not posit just a ‘chaos of prima facie duties’ (‘SN’ 15; ‘BET’ 78); its duties should form a ‘system’, but to do this they need not all follow from one principle. They can also form a system if ‘to fulfil any one, on principle and in general, harmonise[s] with and forward[s] the fulfilment of the others’, as he thought is true of Ross’s principles; thus he thought keeping promises tends to promote the good, involves truth-telling, and more (‘SN’ 16). He even made it a test of moral truth that different prima facie duties mostly favour the same acts (DG 207–10) or that violating one commonly involves violating others (E 133; also ‘RI’ 57; ‘BET’ 79; ST 132–4). Whereas the more common view is that two judgements cohere if one’s being true increases the probability that the other is true, Ewing held, differently, that two duties cohere if one’s being fulfilled increases the probability of the other’s being fulfilled. He took this to be evidence for the duties because ‘we must expect good on principle and per se to produce good rather than evil and vice versa’ (DG 208), and this assumption can be questioned. Why should fundamental duties not often clash or point in different directions? Whether or not the assumption is sound, however, Ewing used it to ground a distinctive coherentist argument, one arguing from a kind of extensional overlap between principles to their more probable truth. With or without this argument, a reflective equilibrium methodology is fully consistent with the school’s intuitionism; to embrace it they had only to give up the view that intuitive judgements can be certain, as they in any case had reason to do. It is worth repeating, however, that whether it takes this coherentist line or treats some moral judgements as fixed, their practice of appealing to moral intuitions is as compatible with non-cognitivism as with non-naturalism. For non-naturalists the intuitions will involve partial or attempted insights into sui generis moral truth; for non-cognitivists they will express immediate or underived moral attitudes. On each view some judgements may be privileged over others; for example, judgements about abstract principles may be seen as most reliable or may express the attitudes that most shape our other attitudes. But on both views the details of moral argument can be essentially the same.

6 Moral Truths: Underivative and Derived The autonomy-of-ethics thesis says we cannot infer moral truths from non-moral ones but also makes an explanatory claim: that moral truths are not grounded in non-moral ones. The most fundamental such truths are underivative, in the sense that there is no explanation why they are true. If we ask ‘why ought I to promote the good of others?’ there may be no answer; that may be just something we ought to do. The same may hold for other duties such as to keep promises and for ultimate truths about the good. There may be no explanation why pleasure or knowledge is good; each just is. Our school all accepted this explanatory claim, and it shaped their views in important ways.

6.1 False Derivations Sidgwick, for example, assumed the claim when he called his fundamental principles ‘axioms’. Rashdall and the Moore of Principia thought truths about duty follow from ones about the good but ultimate truths about the good are underivative; later Moore took the same view of his consequentialist principle of right. Prichard, Carritt, Ross, and Broad thought the basic duties, which for them included deontological ones, have no deeper ground, as likewise for basic claims about the good. Prichard put the point epistemically when he said our sense ‘of the rightness of an action of a particular kind is absolutely underivative or immediate’ (MW 12), but his reason was in part that the duty’s truth is underivative: because it has no further explanation, it must be apprehended in itself. The school therefore rejected all attempts to ground morality in non-moral truths, such as the religious view that for an act to be right is for it to be commanded by God. The classic refutation of this view comes in Plato’s Euthyphro and poses a dilemma: either God commands what he does because it is right, in which case morality is independent of him, or he does not, in which case his commands are arbitrary and we have no reason to obey them.1 Sidgwick echoed this refutation when he said God ‘is

1

Plato, Euthyphro 10a–11b.

moral truths: underivative and derived

129

Himself conceived as . . . prescribing what is right, and designing what is good’, where ‘the notions “right” and “good” are used absolutely, without any reference to a superior lawgiver’ (ME 505, also 31, 79–80; EEM 62; GSM 156, 322–4). Rashdall, Moore, Laird, Carritt, Ross, and Ewing gave similar arguments (E 79, 92–3; PE 128; E 65; SMT 17–18; TM 5–6; KET 25; DG 106–10; E 111–12; VR 130–1, 186–7), and some went further, using their moral beliefs to determine what is true of God. Rashdall, though a priest, said, ‘as in the days of Plato it is a paramount duty of Moral Philosophy to lay down Canons for Theology’ and argued that, since theories of retributive justice are false, we need ‘a considerable amendment of popular ideas about . . . divine punishment’ (TGE I 311–12). Ewing said that, whatever may seem to be in the Bible, true Christian doctrine involves ultimate salvation for everyone: It raises me to white-hot indignation to think of all the cruel misery that must have been inflicted on many millions by what . . . I should have called the criminal stupidity that grafted on the religion of a loving God the belief that this loving being was going to have the majority of the human race tortured forever as punishment for their sins. (VR 229, also 188, 232–3)

He also denied the doctrine of the incarnation, in part because if Christ were God he would not have presented his teachings so unclearly, leading to the multiplication of sects and ‘wars of religion, horrible persecutions and great mental anguish’ that followed (VR 275–6). The school likewise rejected attempts to ground morality in biology. Sidgwick said no truths about the good follow from Darwin’s theory of evolution (ME 79, 83; GSM 1–2, 144–6); Rashdall applauded his analysis, hoping it would end ‘the present craze for extracting ethical theory from a study of the habits of mollusca and crustacea’ (‘PS’ 203–4) and arguing at length for ethics’ independence from biology (TGE II 356–413). To Moore the fact that nature has so far tended in a certain direction was no reason to think that direction a good one (PE 46–9, 54–8; also Russell, ‘EE’ 23–4; Laird SMT 17–18; Carritt, TM 32–4; Ross, FE 12–14), while Ewing and Broad responded to two biologists’ defences of evolutionary ethics2 by reaffirming the gap between ‘is’ and ‘ought’ (‘RSE’ 68–75; ‘RSE’ 100C–100F; CE 178–87). For Broad the class of ‘competent persons’ on moral subjects did not include ‘the eminent natural scientists who from time to time take a holiday from their professional labours to instruct us in ethical theory’ (CE 194). They also rejected metaphysical groundings of morality, in particular the Idealist ones of Bradley and Green. Though a Hegelian, McTaggart said truths about ultimate reality ‘would not give even the first step’ toward proving that something is good (SHC 96, also 266–7). Others addressed the claim that metaphysics identifies a ‘true self ’ whose full realization therefore constitutes our good; the conclusion only follows, they said, if the concept of the ‘true self ’ is already normative, identifying some human capacities as most worth developing (Sidgwick, ME 79; Rashdall, TGE II 2

Waddington, ‘Relations Between Science and Ethics’; Huxley, Evolutionary Ethics.

130

british ethical theorists from sidgwick to ewing

62–70; Moore, PE 113–15; Carritt, TM 50; Ewing, DG 110–11). Kantian efforts to ground morality in the concept of a ‘rational will’ likewise fail because we can only understand such a will as one that wills what is independently good or right (Moore, PE 126–39; Broad, FT 125–8), and an ideal observer is one who is ‘not influenced in his approvals and disapprovals by circumstances other than those relevant to the real goodness or badness, rightness or wrongness, of what he approves or disapproves’ (Ewing, E 94; also ST 20; VR 101; Broad, FT 262–3). On one metaphysical topic there was disagreement. Sidgwick argued that most moral claims do not presuppose libertarian free will but are compatible with determinism, so in ethics we can set issues about free will aside (ME 66–70, 72–6, 285; OHE 260–6). But he thought claims about retributive desert as well as our everyday practices of praising and blaming do require libertarian freedom and must be abandoned if determinism is true (ME 71–2, 284, 291, 349). McTaggart, Rashdall, Moore, Russell, Ross, and the earlier Ewing agreed that moral duty as such is compatible with determinism (SDR 140–85; TGE II 302–55; E 84–95; ‘EE’ 36–44; FE 192–205; MP 35–6, E 153–62, ST 156–86), though Carritt, Joseph, Broad, and Ewing in his last book thought it requires libertarian freedom (TM 125–35; EPT 129–38; SPE 1–15; CE 105; MP viii–ix; VR 246–54).3 Carritt, Joseph, and Ewing thought we therefore have such freedom, but Broad found that ‘self-evidently impossible’ and concluded that obligation is probably a ‘delusive notion’. (He said this in his inaugural lecture as Knightbridge Professor at Cambridge. At the dinner afterward a colleague from another discipline said that if Broad really believed that about his subject, then, if he had any duties, the first would be to resign his chair (EHP, xi)). But no other members shared Sidgwick’s belief in a special connection between retributive desert or blaming and libertarian freedom. While accepting determinism, Ross thought it good for happiness to be distributed in accordance with moral merit and saw no inconsistency in these claims: a reward for a virtuous action can be fitting even if the action was necessitated (FE 246–51). Broad too thought there can be an organic unity between wrongdoing and pain that makes their combination less bad than either on its own even if the wrongdoing was determined (FT 198–205; also McTaggart, SDR 163; Moore, EE 136–8; Ewing, MP 35–6, E 167, ST 162–82). While some thought morality as a whole requires libertarian freedom, only Sidgwick thought metaphysical claims about the will count against specific moral views.4 There is another point where Sidgwick may have been an outlier. In several passages he seems to have argued to a moral conclusion from a metaphysical premise

McTaggart and Rashdall both made the argument, later given by Hobart in ‘Free Will as Involving Determination’, that moral duty requires determinism, since we can only be responsible for our acts if they follow causally from our characters (SDR 178–85; TGE II 335–8). 4 McTaggart had said determinism undermines retributive desert in SHC 140 but abandoned that view in his later SDR. 3

moral truths: underivative and derived

131

about the self. Considering the egoist’s demand for equal concern for all the times in one’s life but no parallel concern for others, he wondered why, given a Humean theory of the self, we cannot ask why ‘one part of the series of feelings into which the Ego is resolved [should] be concerned with another part of the same series, any more than with any other series’ (ME 417–18). And when explaining why a consistent egoist cannot be refuted he said, ‘It would be contrary to common sense to deny that the distinction between any one individual and any other is real and fundamental’ (ME 498; also EEM 44), as if the metaphysical separateness of persons supports normative egoism.5 These arguments are hard to fit into his overall position. If he thought premises about the self directly entail a moral conclusion, that would violate his insistence that no ‘ought’ can be derived from an ‘is’. They therefore require a suppressed moral premise, such as perhaps John Rawls’s claim that ‘the correct regulative principle for anything depends on the nature of that thing’.6 But then what is the status of this premise? Is it self-evident, as it presumably must be if it is to help establish a fundamental moral principle? If so, can its self-evidence be shown? Sidgwick never tried to show it or to give any defence of connecting moral views to claims about the self. Rashdall and Ross may seem to have argued from a metaphysical premise when they cited, as one point favouring a theory of the good valuing knowledge, pleasure, and virtue, that it harmonizes with the traditional division of our mental faculties into cognition, feeling, and conation by positing one good for each (TGE I 75–6, RG 140; also Ewing, E 72). But neither placed much weight on this consideration; each argued first for the three goods directly and gave the argument about faculties only as an addendum. Neither therefore departed seriously from the core belief that the ultimate moral truths are underivative.

6.2 Duty and Self-Interest Another type of grounding relates all duties, including any concerning other people, to the agent’s own good, so you ought to do an act when and only when it will benefit you. It has several versions, based on two cross-cutting distinctions. Psychological versions of this grounding claim that everyone in fact desires only their own good and draw normative conclusions from that claim. Normative versions say that, whether or not they do so, everyone ought to desire only their good. Low-minded versions equate a person’s good with something like her pleasure or desire-satisfaction, so her duties must promote that; high-minded versions, found in ancient philosophers like Aristotle, locate the good in a state of flourishing that

5

I discuss this last argument further in Chapter 11.

6

Rawls, Theory of Justice, p. 29.

132

british ethical theorists from sidgwick to ewing

typically has the exercise of the moral virtue as a constituent. Again, however, your ultimate duty is always to promote that good. Sidgwick discussed two low-minded psychological groundings, starting with Bentham’s and Mill’s argument that since each person desires only her own pleasure, everyone ought to seek the pleasure of all. Though he said he accepted this argument when young, he later rejected it on two grounds: from the premise that each desires her own pleasure it cannot follow that she ought to pursue everyone’s—the two goals conflict—and, as Butler showed, we do not we desire only our pleasure (ME xv–xix). He also considered the argument that since each person desires only her own pleasure, she ought to seek only that. He replied that if the argument’s premise is that she always desires her greatest possible pleasure through time, its conclusion is not really normative, since its ‘dictate’ is one she cannot fail to fulfil. But if it makes only the weaker claim that she always wants some pleasure, nothing strictly follows about how she should choose among pleasures; that could be on some non-hedonic ground. And, again, we do not desire only our pleasure (ME 40–56). Sidgwick took a low-minded normative grounding of moral duty more seriously, because he took egoistic hedonism seriously. More specifically, he thought that if fulfilling your utilitarian duty always maximized your pleasure, that would dissolve the dualism of the practical reason and fully justify the duty. But he did not think that, theological assumptions aside, it does do this. Obeying common-sense moral rules often costs you pleasure (ME 162–75), and obeying utilitarianism does so even more given that view’s stricter duty of benevolence (ME 497–503). As for highminded groundings like Aristotle’s, Sidgwick thought they merely restate the duties in an unhelpful way (ME 375–8, 392; OHE 57–8, 63–4; GSM 90–7). Prichard too discussed egoistic groundings; they were the main object of his argument that it is a ‘mistake’ to try to justify moral duty or ask why we should be moral. In Chapter 1 I said the school rejected the ‘why be moral?’ question because they thought there is no ‘ought’ other than the moral one in terms of which the question can be asked. This was indeed Prichard’s view, but it yields only a verbal victory against the view that you ought, say, to keep a promise only when doing so will promote your pleasure. That view can just redescribe its claim that you ought always to maximize your pleasure as a ‘moral’ one. The result will not be a justification of moral ‘oughts’ as a whole but of specific duties such as to keep promises in terms of your good. Prichard had two more-than-verbal arguments against this view, the first an epistemological argument that parallelled Moore’s reply to scepticism about the external world. We start out, Moore said, believing we have direct awareness and therefore knowledge of external objects. A sceptic like Hume may propose a theory of knowledge according to which we are directly aware only of our experiences and form beliefs about external objects by inference from our experiences. But since all such inferences fail—from premises about experiences only conclusions about

moral truths: underivative and derived

133

experiences can follow—our beliefs about external objects are unjustified. Moore replied that we have less reason to believe whatever theory underwrites this sceptical view than we do for our initial belief that we know external things. The latter belief is the more secure, and far from being refuted by Hume’s theory, it shows that theory to be false.7 Prichard made a similar argument in ethics. We initially believe we have a duty, say, to keep promises. A sceptic like Glaucon in the Republic may ask whether we really do have that duty, given that fulfilling it will often cost us happiness. In doing so he assumes the egoistic theory that we have reason to do something only when it will promote our happiness, and therefore have no reason to keep a promise when it will not. Prichard replied that we have less reason to accept this theory than to retain our initial belief that we ought to keep promises; since the latter is better grounded, it shows Glaucon’s theory to be false. Prichard himself made the analogy with nonmoral scepticism. Though we initially believe we have mathematical knowledge, a sceptic like Descartes can make us doubt that. But the correct response is to reflect on our original mathematical convictions and see that they were in fact knowledge, so whatever theory generated the scepticism is false (MW 18–19, also 8). So in ethics the sceptic’s demand for a proof that we have a duty to keep promises involves: the mistake of supposing the possibility of proving what can only be apprehended directly by an act of moral thinking. Nevertheless the demand, though illegitimate, is inevitable until we have carried through the process of reflection far enough to apprehend the self-evidence of our obligations, i.e. the immediacy of our apprehension of them. (MW 19–20)

Just as not everyone is persuaded by Moore’s reply to external-world scepticism, so some may be unmoved by Prichard’s defence of moral duty, especially since it rests on an appeal to self-evidence. But it is strengthened by his conceptual minimalism. Some who say we ought always to promote our happiness call this a demand of ‘rationality’, as against merely ‘moral’ requirements to keep promises or benefit others. And they think rationality has priority, so moral requirements must be shown to be rational before they can be accepted as binding. But if there is only one ‘ought’ there can be no such priority. That you ought always to promote your own happiness is a substantive claim on a par with ones about promise-keeping and beneficence and must be weighed substantively against them. I think many will agree with Prichard that in this weighing the latter claims win out. Prichard’s other argument was explanatory; it said an egoistic justification of moral duties gives them the wrong explanation and so distorts the moral phenomena. In ‘Mistake’ he made the point very briefly. To the claim that acting on moral rules is justified because it gives us what we want, he said this is ‘not an answer, for it fails to convince us that we ought to keep our engagements; even if successful it only makes

7

Moore, Some Main Problems, Chapter 6.

134

british ethical theorists from sidgwick to ewing

us want to keep them. And Kant was really only pointing out this fact when he distinguished hypothetical and categorical imperatives’ (MW 9). Prichard was here addressing a low-minded psychological grounding, and continued to discuss that in later writings. He thought many philosophers—Plato, Aristotle, Green, even Butler and Sidgwick (!)—were psychological egoists who tried to justify moral duties as a means to an end we already desire. But their view ‘resolves the moral “ought” into the non-moral “ought” ’ and so abolishes it. Since the hypothetical imperative makes only the descriptive claim that an act will further some end, their approach represents ‘our being morally bound to do some action as if it were the same as the action’s being one which we must do if our purpose is to become realized’, so ‘strictly speaking the theory is not a theory of obligation, or duty at all, but, if anything, is a theory that what are called our obligations or duties are really something else’ (MW 188, also 43, 169). This second argument does not apply to normative egoistic groundings. If their premise is that we ought always to pursue our good, they do not deny the moral ‘ought’ but derive one categorical imperative from another. But Prichard thought these versions too distort the phenomena. They say the reason you ought to keep a promise is that this will promote your good, so your ultimate duty in the situation concerns you. This is not, he said, what we believe; we think you ought to keep the promise just because you made it and regardless of whether that will benefit you (MW 26–7, 29–30, 122–3, 188). Though he did not do so, he could have made a similar argument about promoting another’s happiness. An egoistic grounding says the ultimate explanation why you ought to do this is that it will make your life happier or more flourishing, and that is again not the right explanation. The right explanation is that promoting the other’s happiness will make her life better, so your duty is not just superficially but at bottom about her. Ross made this point when he said, of benefiting others, ‘[i]t is for their sake that we feel bound to act so, not for our own’ (FE 276; also ‘IME’ 97–8; RG 16). All these claims reflect his and Prichard’s belief that the primary moral intuitions are explanatory, about what makes acts right. While normative egoistic justifications do not distort morality as a whole by grounding it in something non-moral, they distort particular duties by giving them a mistaken rationale. The broadest autonomy-of-ethics thesis says normative truths cannot be derived from non-normative ones. A more specific thesis says moral ‘oughts’ cannot be derived from non-moral ones, often because there are none of the latter. And the most specific theses say particular moral truths such as ‘you ought other things equal to keep your promises’ have no further explanation even in moral terms; they are true just because they are. Our school all accepted the first two claims and some version of the third. The clarity and explicitness with which they did so is one of their distinguishing marks in the history of ethics.

moral truths: underivative and derived

135

6.3 Theorizing Morality: Two Reasons Though the school thought some moral truths are underivative, they thought others can be derived. Particular moral judgements follow from principles, and more specific principles from more general ones. The task of moral theory is then to identify the most general and explanatory principles. For Sidgwick the impulse to moral theory arose from flaws in our everyday judgements. They do not always yield clear verdicts on particular cases, and even when they do, the judgements we make at one time often conflict with our judgements at others as well as with those of other people. Because they are ‘loose, shifting, and mutually contradictory’ (ME 338, also 6, 12, 100), they demand systematization. These complaints are not surprising from someone who thought the most reliable intuitions are of abstract principles, but they were shared by others. Ross, for example, took the verdicts of the ‘best people’ to be his ‘foundation’ but said we must ‘compare them with one another and eliminate any contradictions they may contain’ (RG 41); since they are not all true, consistent, or clear, we must ‘clear up, so far as we can, ambiguities that lurk in them’ (FE 1; also Rashdall, TGE I 81–3, 217, II 424–7; McTaggart, SHC 97; Broad, FT 80–1; Ewing, MP 119, E 1–4). Sidgwick sought a ‘scientific’ treatment of ethics, but, like a theory in the natural sciences, a moral theory can do two things. It can help us discover truths we otherwise could not, here yielding moral judgements about cases everyday thought cannot decide. It can also explain truths we already know, telling us why an act we are confident is right is in fact so. But different members of the school emphasized a different one of these two functions. Sidgwick mentioned the explanatory function when he said philosophy seeks ‘unity of principle’ (ME 6) and complained that a deontology with multiple duties ‘seems an accidental aggregate of precepts, which stands in need of some rational synthesis’, so without denying its claims about what is right, ‘we may yet require some deeper explanation why it is so’ (ME 102). But several facets of his thought reflected a greater interest in discovery. One was his frequent objection that everyday morality and its deontological systematization are incomplete, not making determinate judgements in cases where a more unified theory like hedonistic utilitarianism does. This need not be a defect from an explanatory point of view. If there are areas of moral indeterminacy, they can be explained as ones where two moral considerations that cannot be weighed precisely conflict. But it is a flaw if we want a theory to yield new moral truths. Another was his understanding of a ‘method of ethics’. He defined this as ‘any rational procedure by which we determine what individual human beings “ought” . . . to do’ (ME 1) and distinguished it from an ultimate reason or moral principle, saying ‘almost any method may be connected with almost any ultimate reason by means of some—often plausible—assumption’ (ME 83). If you accept an end as paramount, your ‘method’ is whatever enables you to identify the acts that

136

british ethical theorists from sidgwick to ewing

promote this end (ME 8), and it does not matter why you accept the end. You can pursue only your own pleasure because you think you ought or do so without any normative thoughts whatever. Either way, you use the same method (ME 77n), and to Sidgwick ‘difference of Method’ was ‘paramount’ (ME 83). This was reflected in his treatment of perfectionism, which he introduced as one of the two main types of consequentialism (ME 9) but quickly set aside. He said most perfectionists equate perfection mainly with moral virtue and that with obeying rules like those in a deontological morality; he also thought they never tell you to sacrifice your own virtue for others’. It follows that ‘any method which takes Perfection or Excellence of human nature as ultimate End will prima facie coincide to a great extent with that based on what I called the Intuitional view’, and so is ‘a special form of this latter’ (ME 11; also 83). From an explanatory point of view, there is a large difference between a theory that says your ultimate reason for acting in a certain way is to maximize your perfection and one that says doing so is simply required, but Sidgwick treated them as equivalent. In a letter he said some minds are interested in the question ‘What is right’ and others in ‘Why is it right’; his book was written mainly for the first minds (M 337–8). Equally telling was his attitude to the dualism of the practical reason. He thought the contradiction this involves would be resolved if there was a God who rewards virtue and punishes vice in the afterlife, making morally right acts maximize the agent’s happiness. He was not sure there is a God and doubted whether the fact that his existence would reconcile two principles is sufficient reason to believe in him (ME 503–9). But he thought God’s existence would have that effect and spent much of his life seeking empirical evidence for it. Moore, Prichard, and Broad found his view puzzling: the egoistic and utilitarian principles make contradictory claims about which property makes acts right, and the conflict between them does not disappear if acts with the one property always have the other. Moore thought Sidgwick here committed ‘the characteristic fallacy of Empiricism—the fallacy of thinking that an alteration in facts could make a contradiction cease to be a contradiction’ (PE 103), while Broad said ‘No God, however powerful and however benevolent, can alter the fact that these two principles are logically incompatible’. Their coincidence would only ‘make it a matter of practical indifference whether I acted in accordance with one or other of two principles, one of which must be false’ (FT 253, also 159–60, EHP 67; Prichard MW 151, 197). These are powerful objections if the dualism involves competing explanatory principles, but for precisely that reason Sidgwick seems not to have seen it that way. He seems to have been interested in his principles only as yielding concrete judgements and so as equivalent when their verdicts coincide. Other members emphasized the explanatory function, sometimes because they thought competing moral theories usually agree in their concrete verdicts. This had sometimes been the Idealists’ view. In his critique of Sidgwick, Bradley said pleasure ‘is an inseparable element in the human end’, so aiming at the greatest pleasure of all will usually lead you aright, but ‘what we hold to against every possible modification

moral truths: underivative and derived

137

of Hedonism is that the standard and test is in higher and lower function, not in more or less pleasure’.8 Green too thought it often makes ‘little practical difference’ whether we equate the good with virtue or with pleasure;9 even so, the philosophical truth was given by his ‘Theory of the Good as Human Perfection’. Some ideal consequentialists had a similar view. McTaggart defended a view on which right acts promote a ‘supreme good’ that is ‘absolutely spiritual, absolutely timeless, absolutely perfect’ and involves individual souls in complete harmony with each other (SHC 100, 95). But since we cannot apply this theory in practice we should use hedonistic utilitarianism as our everyday ‘moral criterion’; though it will not always guide us correctly, it will do so better than any alternative (SHC 99, 118–28).10 Moore said we should usually follow those rules obeying which will preserve society and that they should be the same given any plausible theory of the good (PE 157–8, also 62, 90; EE 87–8; E 19). But he thought the primary ethical question is ‘What is intrinsically good?’, in part because its answer explains why we should follow the rules. Of the ideal consequentialists only Rashdall thought the true theory of the good makes a large difference to which acts are right, so it plays a discovery as well as an explanatory role. He argued that a consequentialism that values knowledge intrinsically supports a stricter prohibition of lying than one that values only pleasure (TGE I 191–6), and that valuing virtue yields better accounts of temperance, sexual purity, humility, and suicide (TGE I 197–213). In his review of The Theory of Good and Evil Moore complained that Rashdall had not proved there are cases where what produces the most perfection does not also produce the most pleasure,11 but surely it is Rashdall’s that is the default assumption and Moore’s belief in general extensional equivalence that needs positive defence. Prichard, Carrritt, Ross, and Ewing too thought a true moral theory will both yield correct judgements and explain them, but they tended to emphasize the second function. Carritt denied that we need moral theory to tell us what our duties are; if we are thinking correctly, we already know what is right, either in particular situations (his earlier view) or as our prima facie duty (his later one). At best philosophy can save us from false generalizations, such as that the right act always produces the most happiness (TM 71; EPT 1–6).12 At the same time, it can identify the grounds of a given duty (TM 69–70; EPT 12) and organize our different judgements about those grounds. Like McTaggart, some in this group distinguished 8 Bradley, ‘Mr Sidgwick’s Hedonism’, p. 97; also 82, 98, 115–16; Ethical Studies, pp. 135–6, 138–9 (but contrast top of 138); Appearance and Reality, p. 358. 9 Green, Prolegomena, sec. 332; also secs. 351, 355–6, 361, 363. 10 McTaggart’s later NE said similarly that our judgements about the relative amounts of good and evil in the universe will be the same given different views about the ultimate good-making qualities (NE II 412–13). 11 Moore, Review of TGE, p. 449. Later Moore changed his view, calling it ‘extremely unlikely’ that what maximizes pleasure always also maximizes the good overall (E 101–2). 12 Also Carritt, ‘Ethics in Philosophical Education’. He likewise denied that philosophical aesthetics gives any guidance in producing or appreciating art (TB 28–30; IA 18–20).

138

british ethical theorists from sidgwick to ewing

a mere ‘criterion’ of right action from an explanatory principle and sought the latter. Thus Prichard said of Kant’s categorical imperative, ‘No one could suppose that the reason why an act ought to be done consists in the fact that everyone could do it. Even Kant could not have supposed this. The difficulty escaped him because it didn’t occur to him that his criterion of moral rules must express what, on his view, is their reason’ (MW 59; also Moore, PE 91, E 19–21, 101–2). And Ross wanted to know ‘what makes right acts right’ (RG 16; also FE 68–9). For most members of the school after Sidgwick a moral theory must above all explain, so it gives the ultimate reason why particular judgements we may already know are true are so.

6.4 Theory vs. Anti-Theory Whereas Aristotle said the goal of his Nicomachean Ethics was ‘not knowledge but action’,13 as if reading it will make us better people, Sidgwick said, ‘my immediate object—to invert Aristotle’s phrase—is not Practice but Knowledge. I have thought that the predominance in the minds of moralists of a desire to edify has impeded the real progress of ethical science’ (ME vi, but contrast 215). Moore, Russell, Carritt, and Ross too made the goal of ethics knowledge rather than action (EE 28; PE 20; ‘EE’ 13–14; TM 70–1; FE 314), and Broad said pungently: A healthy appetite for righteousness, kept in due control by good manners, is an excellent thing; but to ‘hunger and thirst after’ it is often merely a sign of spiritual diabetes. And a white-heat of moral enthusiasm is not perhaps the most favourable condition in which to conduct the analysis of ethical concepts or the criticism of ethical theories. (FT xxiv–xxv, also 284–5; CE 136)

Why should philosophical ethics aim at knowledge? Insofar as it includes Broad’s ‘analysis of ethical concepts’ or metaethics more generally, it is a theoretical subject, and a similar conclusion follows from two other theses. If, as Carritt thought, we already know what is right and wrong, moral theory has only an explanatory function; and if judgement-externalism is true, moral beliefs do not affect action directly but only given a separate desire to do what is right. That philosophy cannot supply this desire may be why some in the school left the practical task to preachers and editorial writers (Moore, EE 12; Broad, CE 136; Ewing, ST 123). In The Methods Sidgwick intended something different. He wanted to put aside our interest in identifying the true moral theory and consider dispassionately what conclusions follow from different ethical premises (ME vi, 12–14, 338); his ethics would then not be practical because it would not endorse specific moral claims. In fact his book defended utilitarianism against deontology and, as we will see, was not always fair in doing so. And elsewhere he took a different line, saying the essays in his Practical Ethics aim at ‘not knowledge but action’ because they discuss principles that can be applied in everyday life as against the more abstract ones studied in 13

Aristotle, Nicomachean Ethics, 1095a4–5.

moral truths: underivative and derived

139

philosophy (PE 5). This was close to a suggestion of Rashdall’s and Ewing’s: that moral philosophy is not practical because it does not know the empirical facts needed to make concrete judgements (‘LC’ 474–5; TGE II 435–8; E 10; ST 123–4). This does not make ethics entirely impractical if sound moral judgement requires knowledge both of principles and of facts, but it does abstract from some practical issues. However exactly they understood its aim, the school all pursued some version of the project of systematizing morality. But some Idealists, especially Bradley, rejected moral theory. Bradley expressed this view first in Ethical Studies and then more vehemently in The Principles of Logic: Just as Logic has been perverted into the art of reasoning, so Ethics has been perverted into the art of morality. They are twin delusions we shall consign, if we are wise, to a common grave. But I would not grudge Casuistry a Christian burial. I should be glad to see it dead and done with on any terms; . . . it is not only the Catholic priest, but it is also our Utilitarian moralist, who embraces the delusion which has borne such a progeny. If you believe, as our Utilitarian believes, that the philosopher should know the reason why each action is to be judged moral or immoral; . . . you have wedded the mistake from which this offensive offspring has issued.14

Bradley here rejected moral theorizing as ‘casuistry’, and Bosanquet did the same when he objected that Moore in Principia ‘follows Sidgwick implicitly’, for example ‘in the crucial point of the acceptance of Casuistry’.15 It is unclear how Bradley could reject theory given his acceptance of a consequentialist moral structure and rejection only of hedonism about the good. (‘For morality I am sure no law is absolutely imperative except the law, Realize the good’.16) Perhaps he thought that while right acts all maximize the good, the good cannot be theorized or measured.17 He continued: A code of morality is . . . impossible. To anticipate the conclusion in each special case you would have to anticipate all possible cases; for the particular condition which makes this conduct right here and wrong elsewhere, will fall outside the abstractions of the code.18

A moral theory tries to identify one or a small number of properties that make all right acts right. But there are, he argued, no such properties. Any right act is made so

14

Bradley, Principles of Logic, p. 269. Bosanquet, Critical Notice of PE, p. 255. The editor of Mind, Moore’s Cambridge teacher G.F. Stout, wanted to have an eminent philosopher from outside Britain review the book, such as Brentano or C.S. Peirce, but there were difficulties about each and then Bosanquet volunteered himself. Knowing the latter’s hostility to Moore, Stout wrote Moore saying, ‘I presume this proposal does not commend itself to you’, but in the end published Bosanquet’s review (Levy, Moore, p. 235; Regan, Bloomsbury’s Prophet, p. 174). 16 Bradley, ‘Mr Sidgwick’s Hedonism’, p. 122. 17 Even this reading has to be reconciled with Bradley’s claim that the good consists in ‘homogeneity and specification’, or, in present-day language, organic unity or unity-in-diversity (Ethical Studies, pp. 74–81, 247–50). Perhaps he thought this characterization of the good, though correct, cannot be applied in a rule-like manner to particular states of affairs. 18 Bradley, Principles of Logic, p. 270. 15

140

british ethical theorists from sidgwick to ewing

by the totality of its properties, and we can only see its rightness if we consider them all together. Sidgwick responded briefly to Bradley’s claims (EEM 24), and Moore was implicitly doing the same when he called casuistry ‘the goal of ethical investigation’. He took casuistry to consider more specific types of act than philosophical ethics and to require empirical knowledge, so it can be attempted only at the end of ethical study (PE 4–5; also EE 10–11). Laird and Ewing too endorsed ‘casuistry’ (SMT 62–3; MP 156), while Rashdall disputed Bradley at length in a lecture presented at Sidgwick’s invitation in Cambridge, then published as a journal article and, with revisions, in The Theory of Good and Evil. He noted that Bradley himself used a principle about maximizing the good when discussing issues such as punishment and infanticide (‘LC’ 480n; TGE II 426–7), and argued that an anti-theory view is inconsistent with the fact that we cannot say one act is right and another wrong without being able to point to some other difference between them (‘LC’ 462–3; TGE II 425). This argument would not have impressed Bradley, who thought an act’s rightness depends on all its properties, but Rashdall could reply that some properties have no moral relevance: killing someone is no more or less wrong if he has forty-seven rather than forty-six eyelashes. If it can be shown that most properties are in this way irrelevant, the remainder can be theorized. A further response to the anti-theory view was Moore’s principle of organic unities. It was a central tenet of Idealism that a whole is prior to its parts both ontologically and evaluatively. If a and b are joined by relation R to form aRb, that changes both the nature of a and b and their value; a human arm within a human body is a different entity than if severed from the body and has a different value. The Idealists therefore rejected the device of assessing a thing’s value by imagining it existing on its own, which Bradley found in Sidgwick and Bosanquet in Moore: this separates the thing from relations that affect its value. And a whole’s value cannot equal the sum of the values of its parts, since outside it the parts are different things.19 Moore’s principle met the Idealists partway. He agreed that the value of a whole can differ from the sum of the values of its parts, but insisted that it is still a sum. The value ‘on the whole’ of aRb is just the value of a plus the value of b plus the value of aRb ‘as a whole’, or plus the value of R’s obtaining between them (PE 214–16). And as a sum this value on the whole can be calculated, so recognizing organic unities requires no abandonment of moral theory; it just requires a theory to be more complex. It has been argued that Moore could not accept organic unities without also accepting Bradley’s attack on casuistry,20 but there is no basis for this claim. That there can be value in the relations between states of affairs does not for a moment imply that there cannot be rules governing that value. 19 Bradley, Ethical Studies, pp. 126–7, ‘Mr Sidgwick’s Hedonism’, p. 95; Bosanquet, Critical Notice of PE, p. 255. 20 Baldwin, G.E. Moore, pp. 127–8.

moral truths: underivative and derived

141

Moore denied that a part changes ontologically when it enters into a whole (PE 33–4)21 and he also held, as his conception of intrinsic value requires, that a part has the same value in the whole as outside it (PE 30). Bosanquet objected that while ‘the doctrine of organic membership is always the better for being overhauled, . . . Mr. Moore’s criticisms do not seem to me to make any impression on it, resting as they do on the non-modifiability of subjects by relations’.22 But as I said in Chapter 3, Moore’s principle need not be stated in his particular holistic way. A variability interpretation says a part’s entering into a whole can change its value, so a has more value in aRb than outside it. This seems to be how Rashdall understood the principle (TGE I 68–9, II 40), and it still allows for summation. There can be principles that explain how the values of states are changed by their relations, so recognizing organic unities again implies nothing anti-theoretical. Moore was less clear about the epistemology of organic unities. His main claims suggest that we know a thing’s value ‘on the whole’ by first knowing its component values and then summing them, but in one passage he defined a thing’s value ‘as a whole’ as ‘the difference between the value of the whole thing and the sum of the value of its parts’ (PE 215). Read epistemically, this suggests that we first know a thing’s value on the whole and infer its value as a whole by subtraction. This second view is closer to the Idealists’ and something like it was held by the early Ewing, who thought we can be more confident in our judgement of the total value of a complex than in our judgements of the values of its elements (MP 173–4, 179–81). But the view still has no anti-theoretical implications. Using Sidgwick’s distinction between logical priority and priority in knowledge, we can say that, even if we cannot know a thing’s value as a whole first, it comes first in explanation, since together with the values in the parts it explains the thing’s value on the whole. The principle of organic unities can be interpreted differently and combined with different epistemologies, but many are consistent with a theorizing approach to ethics. A major contribution of Moore’s was to show that we can reject atomistic views like hedonistic utilitarianism while still pursuing what Bradley rejected as the ‘casuistical’ search for explanatory moral principles.

6.5 Degrees of Theory Despite sharing the general project of theorizing morality, our school differed on how far it should be pursued. There was a sequence of views here from earlier to later members. Sidgwick was the most thorough systematizer. His preferred normative theories were doubly monistic, with a single principle of right, either ‘maximize your own good’ or ‘maximize the good of all’, and a single good, pleasure. For him this monism

21

See also his ‘External and Internal Relations’.

22

Bosanquet, Critical Notice of PE, p. 261.

142

british ethical theorists from sidgwick to ewing

was methodologically required: pluralistic theories have to weigh competing duties or goods and, since they cannot always do so determinately, are not fully scientific. The ideal consequentialists retained Sidgwick’s one principle of right but recognized a plurality of intrinsic goods, for example Rashdall’s pleasure, knowledge, and virtue or Moore’s aesthetic contemplation, love, and rewarding of desert. And in affirming them they rejected Sidgwick’s methodological demand. Rashdall said, ‘there is nothing really scientific in seeking to make a problem more simple and less complicated than it really is’ (‘LC’ 468), while Moore said, ‘We have no title whatever to assume that the truth on any subject-matter will . . . possess any particular form of “unity”. To search for “unity” and “system”, at the expense of truth, is not, I take it, the proper business of philosophy’; instead ‘things good and bad are many and various’ (PE 222–3, also 185–6, 202; E 106; Russell, ‘EE’ 53–4). These writers did not think our ultimate duties are many, but not for Sidgwick’s methodological reason. They just found the one consequentialist principle self-evident. Later members, in particular Prichard, Carritt, Ross, and Broad, did think there are many principles of right and held doubly pluralistic views, with several basic duties and several intrinsic values. Ross, for example, listed seven prima facie duties, of which three involved promoting the good but four were deontological, and four ultimate goods. And in defending their views they echoed Rashdall and Moore against Sidgwick. Prichard and Carritt denied that there must be a single ground of duty (MW 1, 5, 14, 62, 77; TM 75, 140; EPT 11–13, 69); Ross said, ‘loyalty to the facts is worth more than a symmetrical architectonic or a hastily reached simplicity’ (RG 23, also 19, 78–9; FE 79, 82–3, 189, 206, 319) and denied that a theory’s being easy to apply counts toward its truth (FE 90–1); and Broad called hedonistic utilitarianism ‘far too simple . . . to do justice to all the facts’ (FT 239, also 223, 283–4; CE 128–30, 228–33). There was, then, a rough progression from the double monism of Sidgwick’s utilitarianism through the mixed view of the ideal consequentialists to the double pluralism of the deontologists. But there was also a partial backlash. Joseph thought the duties in a view like Prichard’s were ‘an unconnected heap’ and found this ‘disconcerting to philosophy, which attempts to bring a diversity of facts under some unity of principle’ (SPE 67). Ewing agreed (E 81–2; ‘BET’ 73–4), and both thought pluralist deontologies also violate the principle, which like Moore they found self-evident, that it can never be wrong to do what will result in the most good (SPE 26; ‘SN’ 11; DG 188; E 66–7, 82; ‘BET’ 69).23 They proposed versions of ideal consequentialism in which the goods to be promoted include not just states such as pleasure and knowledge but also act-types such as promise-keeping and gratitude or, more plausibly, in which acts such as promise-breaking and ingratitude are intrinsically evil (SPE 88–103; MP 5, 170–1, 200, 202–3, 217; ‘SN’ 11; DG 187,

23

See also Laird, SMT 21; Pickard-Cambridge, ‘Two Problems (III.)’, p. 333.

moral truths: underivative and derived

143

190–3, 207; E 82–3; ‘U’ 109–10; ‘BET’ 74, 81). They hoped thereby to accommodate the intuitions that motivate deontology, such as that it can be wrong to break a promise even when that will result in more happiness, in a consequentialist framework. We will discuss this later ideal consequentialism in Chapter 7, but in two ways it did not depart that far from Prichard and Ross. Ewing combined it with a fittingattitudes analysis of ‘good’ on which to call promise-keeping good is to say one ought other things equal to choose it. As he recognized, this made his view similar to Ross’s, differing only in that it satisfies the principle about always promoting the good and gives all its duties a unifying rationale. And even without that analysis the view is not really less pluralistic. It reduces the number of basic duties from Ross’s list to one but does so only by adding an intrinsic good or evil for each duty it eliminates; the result is the same number of underivative moral factors. Theories that are alike in being pluralistic can be so to differing degrees. Some contain a large number of underivative duties or goods, perhaps dozens or more of each, while others recognize just a few ultimate factors and are therefore more unified. Prichard had little interest in unity. He held that every moral judgement implies a principle (MW 63), but never tried to enumerate the basic principles and even tended to multiply them, saying your duty to promote your own virtue is different from your duty to promote your child’s virtue, which is different from your duty to promote a stranger’s (MW 216–17).24 Ross was more systematic. He recognized just five prima facie duties when his three instances of promoting the good are combined, and though he called his list ‘not necessarily complete’ (RG 23, also 20), does not seem to have envisaged many additions to it. Instead he tried to derive other duties from it, saying for example that the duty not to lie follows from the duty to keep promises (RG 22, also 27–8; FE 76). He likewise listed four basic goods and said he could not discover any good that ‘is not either one of these or a combination of two or more of them’, so aesthetic enjoyment and mutual love combine pleasure and knowledge in the one case and virtue, knowledge, and pleasure in the other (RG 140–1). The result was an only modestly pluralistic theory, with only a small number of basic duties and goods and efforts to reduce other ones to these. If one aim of systematization is to reduce the number of ultimate moral factors, another, emphasized by Sidgwick, is to increase the number of acts we can judge. Pluralism works against this aim, since it requires us to weigh duties or values against each other, which is not always easy. The pluralists often said there are no rules for weighing conflicting factors, which must be done in each case by direct intuition (Rashdall, TGE II 433; Prichard, MW, 6, 79; Ross, ‘BO’ 127, RG 19, 23, 31, 33, 41–2; Broad, CE 81, 128–30; Ewing MP 183–4); here, they said, theory runs out. 24

In early writings Prichard recognized a plurality of intrinsic goods, including knowledge and pleasure (MW 10), but later held that only virtue is good (MW 173; letter to Ross of 20.12.1928).

144

british ethical theorists from sidgwick to ewing

I said Ross’s denial that we can know how duties compare is exaggerated, and the denial that there are any rules for weighing is also exaggerated since comparative judgements must at least satisfy formal conditions such as transitivity. If the duty to save two lives outweighs the duty to save one life, and the duty to save one life outweighs the duty to keep a trivial promise, then the duty to save two people’s lives must outweigh the duty to keep the promise. Even if there is no independently graspable rule for comparing a duty about lives and one about promises, our judgements about them must generate a coherent ranking, which then expresses an at least partial measure of their strengths. If we cannot resolve all conflicts, the ranking will not yield determinate verdicts in all cases, but it will yield them in some. Ross on particular duties aside, this was for the most part our pluralists’ view, most explicitly about the measurement of value. In their critique of theory, the Idealists, surprisingly, gave pleasure as their main example of a good that cannot be measured. Starting with Sidgwick (ME 131, 140–50), our school agreed that we cannot in practice measure it on a complete cardinal scale or say that one pleasure is 2.753 times as good as another, but said we can make the ordinal judgement that one pleasure is more intense and therefore better than another, and can sometimes make rough cardinal judgements. We can say one pleasure is much or only a little better than another, and can sometimes assign rough numbers to the difference. Rashdall said ‘if it be a matter of indifference to me whether I enjoy one minute of one pleasure or two minutes of another, I may reasonably be said to regard the one pleasure as twice as pleasant as the other’ (TGE II 23, also 26–8).25 McTaggart thought he could never say that two experiences are exactly equally pleasant, but was sure he got more than twice as much pleasure from a plate of turtle soup as from a plate of pea soup (SHC 117, quoted in Rashdall, TGE II 23n; also NE II 242–3). And these rough measures can be arrived at using the method of ‘supervaluations’ or ‘intersection sets’.26 We can certainly state precise cardinal measures of pleasure, in the sense of writing them down, and though we cannot choose one of these as uniquely correct we can reject some as unacceptable. McTaggart would reject a measure that makes the turtle soup only one-and-a-half times as pleasant as the pea, and might also reject one that makes it ten times as pleasant. This would leave him with a set of measures all of which are not unacceptable, and he could accept as true simpliciter any verdicts endorsed by all the measures in this set. If they all say the turtle soup is at least twice as pleasant, it is at least twice as pleasant. The same method can be used for goods such as knowledge and virtue, where it again yields rough cardinal judgements when all acceptable measures agree. Moore made a judgement of this kind when he called aesthetic contemplation and personal love ‘by far the most valuable things’ we can know (PE 188), as did Ross when he said 25 This argument assumes that the value of a pleasure is a linear function of its intensity, so given a fixed duration, a pleasure that is twice as intense is always twice as good. This assumption can be denied. 26 Sen, Collective Choice, Chapters 7, 7*; Hurka, Perfectionism, pp. 86–7.

moral truths: underivative and derived

145

‘an action of sublime heroism is many times as good a thing as some faint and passing shade of pleasure’ (RG 144) or that we can sometimes say ‘B is worth not less than m times and not more than n times as much as A’ (FE 183; also RG 143). Ewing compared weighing different goods to marking examination papers, where we balance qualities such as originality, common sense, and grasp of the subject; though there is no algorithm for this, we can often be certain that one candidate is ‘much better’ than another (MP 168, also 172; and McTaggart, NE II 447–8). Similar judgements can be made about competing duties. The school’s deontologists did not discuss cardinal measures of duty, but they are needed if relieving someone’s pain is not certain to result in your breaking a promise but has, say, a 0.6 probability of doing so. Then you need to multiply a cardinal measure of the strength of your duty to keep the promise by the 0.6 probability, and can get at least a rough one by supervaluing over acceptable precise measures of the duty’s strength.27 Though our school did not know this device, it allows claims like those they made. There were some partial exceptions to this view of measurement. In The Theory of Morals Carritt was sceptical about measuring, saying ‘there is not a certain amount of the same indistinguishable stuff ’ present in, say, the pleasure of work and that of smell (TM 25; also TB 62n). But in his later Ethical and Political Thinking he allowed, like other members, that pleasures are at least partly quantifiable (EPT 61–2). Broad rejected McTaggart’s views on the measurement of value as ‘stuff and nonsense’ (EMP 670) but had himself counted an act’s utility, or the total value of its consequences, as one right-making feature (FT 218–22), and it is not clear what his later objection was. Sometimes he thought McTaggart had too simple-minded a view of how the value in a whole results from the values in its parts, assuming too often that it must be a sum (EMP 676–7, 690). At other times he argued that, though value can be measured, what the measures define is only a ‘constructed’ and not an ‘intrinsic’ or real magnitude (EMP 671–3). But he gave no argument why a constructed magnitude like the average happiness cannot be morally important, and in general did not object to quantification as such; on the contrary, his discussion of the subject is the most technical the school produced.28 In affirming rough but only rough comparisons our pluralists took a line intermediate between two extreme views. One says an acceptable moral theory can and must yield a complete set of moral verdicts and be easy and even mechanical to apply. Sidgwick took this view, and it is shared by some present-day philosophers and economists whom it leads to favour monistic theories such as preference27 In a letter to Ross, Prichard said he could not see how to weigh the probability of an act’s fulfilling a duty against the strength of the duty. To him, finding an act A, which is twice as likely to produce half as much benefit as B, morally equivalent to B is ‘nonsense’ (letter to Ross of 5.10.1937). But his talk of ‘weighing’ suggests the mistaken assumption that the two factors must be added; if they are multiplied, the difficulty vanishes. Then whatever scales we use to measure the factors, doubling the probability that an act will fulfil a duty doubles the strength of the reason to do it, as does doubling the duty’s strength. 28 For a critique of Broad on this aspect of McTaggart see McKerlie, ‘McTaggart on Love’, pp. 80–2.

146

british ethical theorists from sidgwick to ewing

utilitarianism. The other extreme is the anti-theory view of Bradley, Bosanquet, and those neo-Wittgensteinians who reject principles and rest all moral judgement on particularized perception.29 Despite their differences these two views both assume that if morality is to be theorized it must be theorized completely, with no indeterminacies and no reliance on case-by-case intuition. The first view thinks this condition can be satisfied and proceeds to theorize monistically; the second thinks it cannot and rejects theory altogether. But our pluralists denied the assumption, and were right to do so. We can theorize morality partly but not completely, or formulate measures that are roughly but not precisely cardinal; though this reduces the need for case-by-case intuition, it does not eliminate it. A final question concerns how our pluralists understood their intermediate view. On one reading it is merely epistemic, assuming the moral truth is completely precise and determinate and denying only that we can know it all. On a metaphysical reading it says the truth itself is partly indeterminate: if a duty of beneficence conflicts with a duty to keep a promise, it may be true neither that the first outweighs the second, nor that the second outweighs the first, nor that they are exactly equal in strength.30 Though the school rarely addressed this issue directly, the epistemic reading seems closer to most of their views. Sidgwick assumed that every pleasure has a precise intensity and therefore value (ME 124, 146), and a similar view is suggested by analogies other members made. Rashdall compared measuring value with counting the number of sheep in a flock or assessing the strength of different lights (TGE II 25, 49–50), both of which are determinate magnitudes, while Ross’s suggestion that moral truths are as much part of the ‘fundamental nature of the universe’ as mathematical ones (RG 29–30) suggests that they are equally precise. Ross explicitly said truths about value are determinate when he assumed ‘that good is objectively quantitative . . . , but not that we can assign an exact quantitative measure to it’ (RG 34n, also 142–4; FE 183), and he implied a similar view in his account of conflicts of duty. If he had taken the metaphysical view he would have said that when we cannot decide whether the duty to help another or to keep a promise is stronger, there may be no truth of the matter. But he said that when good people make different judgements about duty ‘they cannot all be right’, even though ‘it is often impossible to say which is right’ (FE 189); he also called a right act a ‘fortunate act’, since if we cannot be certain that what we are doing is right, it is only luck if it is (RG 31; also ‘BO’ 123). These remarks make the characteristically realist assumption that the truth, in this case about particular duties, extends further than our power of apprehending it. That we cannot always know what is best or right does not mean there is not a fully determinate moral truth that we can try to grasp as far as we can. The limitations on moral theory inhere not in its subject matter but only in our ability to pursue it. 29 30

See e.g. McDowell, ‘Virtue and Reason’. Parfit affirms metaphysical indeterminacy in On What Matters, vol. 1, pp 32–3.

moral truths: underivative and derived

147

6.6 Inherent Explanations A moral theory explains some moral claims in terms of others, but there are two ways it can do so. The difference between them is not sharp; there is a continuum. But there is a clear contrast between the continuum’s two ends. At one end are what I call ‘external’ explanations. They explain a moral claim by deriving it from another that uses different concepts and concerns a different, supposedly more fundamental topic. A grounding of duty in non-moral claims about biology or self-interest fits this description, but so do some internal to morality. Utilitarian justifications of duties to keep promises or care more for your children in terms of the effects on general happiness are external, since the explaining topic is so different from the explained one. So is a derivation of a retributive theory of punishment from a claim about fairness, on the ground that there is an unfairness that needs correcting if some have restrained their impulse to break the law while others have not.31 Equally external is the view that the duty to keep a promise follows from the duty not to cause harm, since in promising you create an expectation in your promisee that it will harm her to disappoint.32 But the grand exemplar of this approach is Rawls’s attempt to justify an egalitarian principle of distributive justice by showing that it would be chosen by rational agents trying to maximize their share of primary goods in an ‘original position’. The justification is not non-moral, since its specification of the original position is guided by moral considerations, but it is external in my sense, since claims about choice behind a veil of ignorance are at some distance from thoughts about economic distribution. By contrast, an ‘inherent’ explanation derives a moral claim from another that is more abstract but concerns the same subject and uses similar concepts, so the explaining claim is continuous in content with the one it explains. The explaining claim must have intuitive appeal in itself if the explanation is to succeed, but if it does, connecting the explained one to it increases our warrant for believing the latter. But the derivation does not ground what it explains in something radically different; it just extracts its core idea and states it more abstractly.33 Our school sometimes gave external explanations. Sidgwick’s justifications of common-sense rules were external, as was Ross’s derivation of the duty not to lie from the duty to keep promises. But they more often gave inherent explanations, and many of their greatest contributions fit this pattern. One was Sidgwick’s grounding of utilitarianism in two axioms requiring equal concern for the goods of all people and for goods at all times (ME 381–2), which express the ideal of impartiality found in common-sense ideas like the Golden Rule more abstractly. Another was the analysis of virtue given by Rashdall and shared by

31 32 33

Morris, ‘Persons and Punishment’; Sher, Desert, Chapter 5 A recent defence of this view is in Scanlon, What We Owe, Chapter 7. I introduced the external/inherent distinction using different terminology in ‘Normative Ethics’.

148

british ethical theorists from sidgwick to ewing

McTaggart, Moore, and Ross. It equates virtue with loving for itself, for example desiring for itself, what is independently good, such as another’s happiness, and hating for itself what is evil; it can also include wanting to do what is independently right. This analysis is continuous with everyday thoughts about benevolence, malice, and conscientiousness but again states their core idea more abstractly. Moore’s principle of organic unities allows many inherent explanations, including his of ‘vindictive punishment’ as taking the combination of vice and pain in the same life to be good as a whole, and sufficiently so that inflicting the pain makes the situation on balance better (PE 215). This explanation does not ground retributivism in some non-retributive idea such as fairness but clarifies its logic. Ross’s concept of prima facie duty does not give his plural duties an independent rationale but illuminates their structure and interactions. It too is inherent, as is Broad’s description of ‘selfreferential altruism’ as giving you stronger duties to benefit people who are specially related to you, so you can be seen as ‘a centre of a number of concentric circles’, with those closest to you in an innermost circle and those to whom you have the weakest duties farthest out.34 This description does not derive the duties’ agent-relativity from something that is not agent-relative, as an external explanation might; it assumes agent-relativity, because the relationships are always to you. External moral explanations are more ambitious and, if successful, more satisfying. If the topic of their explaining claim really is more fundamental, they give deeper explanations, and if beliefs about that topic really are more secure, they do more to justify belief in the explained one. But inherent explanations also explain, with the abstractness of their explaining claims allowing them to unify diverse phenomena; they also clarify the views they explain and raise new questions about them. Sidgwick’s axioms bring out what is distinctive in utilitarian beneficence and show its formal similarity to a well-known view about self-interest; they therefore deepen our understanding of utilitarianism. Rashdall’s analysis of virtue relates it to other values, as a higher-level good involving fitting attitudes to them, and unifies the virtues and vices as common sense cannot. Moore’s version of retributivism has the important implication that deserved pain, though good as deserved, is evil as pain and to be regretted for that reason; it also connects retributivism to other organic-unities claims. Broad’s description unifies our duties to family, friends, and co-nationals and raises the question of what determines closeness in the relevant sense. Much recent moral philosophy has sought external explanations, either non-moral or moral; this has been true of Kantian, Aristotelian, contractualist, and many other approaches. Philosophers have been ambitious, wanting the most fundamental explanations they can find. But many have been sceptical of merely inherent ones, thinking they do not do what an explanation should. One critic of Moore on punishment said his organic-unities claim ‘merely restates what the retributivist needs to

34

Broad, ‘Certain Features in Moore’s Ethical Doctrines’, p. 55; also CE 279–80.

moral truths: underivative and derived

149

explain’ and so is just a ‘prologue’ that should lead to ‘independent justifications of our beliefs about desert’.35 Others have found the concentric-circles account of selfreferentiality insufficient; it needs to be supplemented by an argument deriving special obligations from, say, the goods special relationships make possible in our lives.36 This scepticism has in turn reflected doubts about whether these moral claims can be underivatively true; many have felt they need further explanation. Certainly Rawls did not think he could present his egalitarian theory as true just in itself; he had to ground it in something external like an argument about choice in an original position. By contrast, our school saw no difficulty in principle in underivative moral truths, and its later members affirmed several such truths, making underived claims about, for example, desert, virtue, or self-referentiality and using them in inherent explanations. And they were often sceptical about external explanations, objecting that they either do not yield the right results or give them the wrong rationale. Prichard, for example, took this double line against the expectation-based account of promisekeeping mentioned above. You can create an expectation by saying ‘I have resolved not to do x’, he said, but doing that is not the same as promising and does not put you under the same obligation. And the reason why you ought to keep a promise is not that you have created an expectation but that you have ‘bound’ yourself ‘directly’ (MW 258–9; also Sidgwick, ME 304).37 Objections like these are always possible against an external explanation, given the gap between its explaining and explained claims. But they arise less often for inherent ones, where the two are closer. Our school’s acceptance of underivative moral truths led them to reject non-moral groundings of morality, but it also shaped their positive views, leading them away from external explanations and toward inherent ones. These explanations do not explain as much as an external one would and may therefore disappoint some. But often the extra explaining cannot be done, and the inherent approach illuminates the details of our moral views as ones with grander ambitions often do not.

36 Sher, Desert, pp. 72, 19. Scheffler, ‘Relationships and Responsibilities’. For similar objections to expectation theories of promising see Owens, ‘Simple Theory of Promising’, pp. 53–67; Shiffrin, ‘Promising, Intimate Relationships, and Conventionalism’, pp. 486–92. Prichard thought the duty to keep promises needs an explanation, but found it in an initial general promise he took us all to have made to keep our future promises (MW 259–65). Its strangeness aside, this explanation does not ground the duty of promise-keeping in something unrelated to promises and so still treats some such duty as underivative. 35 37

7 Consequentialism vs. Deontology Despite their shared methodology, our school often differed on normative questions. An important divide was between those earlier members who were consequentialists and the later ones who defended deontology. In 1902 Rashdall said there is ‘a general consensus . . . that Ethics must be “teleological”’, (‘CAV’ 148; also TGE I 217, II 41), and in the decades around 1900 consequentialism was indeed the dominant moral view. It was, however, defended in different ways. Rashdall and the Moore of Principia took the consequentialist principle to be analytically true, while the later Moore, Laird, Joseph, and Ewing held that it is synthetic but declared simply that it is self-evident. Though this is one way of defending consequentialism, it does not offer much in the way of argument to non-believers. Sidgwick tried to do more, arguing in The Methods that no non-consequentialist principles satisfy all four of his conditions for self-evidence but the principles underlying consequentialism do. His arguments deserve extended examination.

7.1 Sidgwick: Against Deontology The target of Sidgwick’s negative argument was ‘Dogmatic Intuitionism’, a theory that tries to systematize common-sense morality while retaining its deontological character, so it contains both principles about promoting the good and ones constraining how we may do so. His model for it was William Whewell’s Elements of Morality, from which he said his initial antipathy to this view derived (ME xv–xvi). He sometimes denied that deontological principles satisfy his second condition: they do not even ‘present themselves as self-evident’, and any appearance that they do is dissolved by ‘reflection’ (ME 383; also EEM 25). But he usually gave a more complex argument. If these principles are left as ‘somewhat vague generalities’, such as ‘you ought to keep your promises’, they do not satisfy the clarity condition, but they do seem self-evident and assent to them is ‘approximately universal’. Once we try to give them ‘the definiteness which science requires’, however, by adding qualifications that make their application determinate and resolve conflicts between them, the consensus disappears and their self-evidence ‘becomes dubious or vanishes altogether’. In one formulation, in other words, the principles seem to pass the second test of apparent self-evidence and the fourth about consensus but are neither clear nor mutually consistent. In another they pass the first and third tests but are too

consequentialism vs. deontology

151

complicated to pass the second or fourth. But in no formulation do they pass all four tests (ME 342–3). Sidgwick applied this argument to a series of principles: about benevolence, gratitude, distributive and retributive justice, law-abidingness and promise-keeping, truth-telling, and more, in a survey running through ten chapters of Methods III. Surprisingly, he only briefly mentioned (ME 253, 348) what many today consider the core of a deontological view: Ross’s principle of non-maleficence, which makes it more seriously wrong to directly harm another than to fail to give her benefits, for example, more seriously wrong to kill her than to fail to save her life. The issues this principle raises barely figured in his discussion. One aspect of his claim that everyday versions of the principles are unclear concerned cases where they conflict. A common-sense deontology must weigh duties such as benevolence and promise-keeping against each other but has no precise rules for doing so. It therefore cannot always make decisive judgements about acts and is in that way partly indeterminate. The unclarity here is not in the individual principles, as Sidgwick’s statement of his first condition would suggest, but in the theory containing them as a whole. A second aspect did concern the individual principles. Here he argued that they often have vague boundaries, so it is unclear exactly which acts fall within their scope and which do not; there is ‘a sort of margin or dim borderland’ (ME 270) where we are unsure whether a given duty applies. Common sense says we should care more about those who stand in special relationships to us, but does someone’s belonging to your race give you stronger reason to promote his happiness? Do a childless person’s siblings have a special claim on his estate (ME 246–7)? Though it is sometimes clear that you ought to keep a promise, what if circumstances have changed radically and the person to whom you made it has died and cannot release you from it? Is the promise still binding (ME 306–7)? Some instances of this second unclarity involved versions of the first. Sidgwick often argued that a common-sense duty has conflicting internal elements. To assess the strength of a duty of gratitude, for example, should you look to the size of the benefit another gave you or to the degree of effort he expended (ME 261)? Distributive justice likewise has one side telling us to honour any expectations we have created and so to retain any social rules, for example about economic life, that have long been in place, and another proposing an ideal of justice and telling us to reform any practices that hinder its attainment (ME 273). Unless a deontological view can in each case choose one element in total preference to the other, its judgements about gratitude and justice must weigh competing claims and cannot always do so precisely. In elaborating these charges Sidgwick made many acute observations about common-sense morality and raised telling objections to particular deontological claims. But our question is whether his arguments successfully show the unacceptability of deontology as a whole, or a fundamental failing in it. Rashdall thought the arguments entirely successful: ‘The loose statements of Intuitionists as to the clearness, certainty, adequacy, and self-evidence of the ordinarily

152

british ethical theorists from sidgwick to ewing

received rules of conduct have never been subjected to so searching, so exhaustive, and so illuminating an examination. That task has been done once for all, and need not in detail be done over again’ (TGE I 83). F.H. Hayward agreed.1 But in fact Sidgwick’s critique had serious flaws, many of them due to his not having or clearly grasping Ross’s concept of a prima facie duty. We have already noted this in his failure to see that there is an ‘ought’ that is not ‘imperative’ but can be outweighed just as a ‘good’ can, but it was a recurrent feature of his discussion of deontology. He defined this as a view that assumes we can see certain acts as ‘right and reasonable in themselves, apart from their consequences’, or without considering consequences ‘except in so far as these are included in the common notion of the act’ (ME 200, 337). This last qualification was an astute response to the fact that, as he had noted earlier (ME 96–7), the same act can be described in different ways, some referring to consequences that are not mentioned in others; compare calling an act a ‘killing’ with calling it a ‘pulling of the trigger’.2 But his talk of seeing acts as ‘right and reasonable in themselves’ was ambiguous. If he meant seeing them as prima facie right or right other things equal, his definition of deontology was sound. If all acts of killing are prima facie wrong, then to know an act is prima facie wrong we need only know that it caused a death and need not know its longer-term effects. But if by ‘right’ he meant right all things considered, his description fit only absolute deontologies; a moderate deontology must always know a killing’s further effects, since they may contain enough good to outweigh its wrongness qua killing. Sidgwick did not distinguish these readings, which at the least made his characterization less than ideally clear. And several of his statements point to the all-things-considered one, which fits only absolute views. In one passage his examples of deontological principles were ‘that duty should be done “advienne que pourra”, that truth should be spoken without regard to consequences, that justice should be done “though the sky should fall”’ (ME 200; also EEM 25), all of which are absolute rather than other-things-equal. In another he said a deontological view makes ‘the practically ultimate end of moral actions their conformity to certain rules or dictates of Duty unconditionally prescribed’, where ‘unconditionally’ again means without considering all their consequences (ME 96, also 3). But our ultimate aim cannot be just to do acts that are right other things equal; it must be to do acts that are right all things considered. Here again he was assuming, as Rashdall and Moore at times also did, that the only alternative to consequentialism is an absolute deontology that forbids some acts whatever their consequences. That assumption is a mistake. 1 ‘Sidgwick’s discussion of this “common sense” doctrine is admitted by all critics to be extremely able, to be, in fact, the most irrefutable part of the book. The student . . . will thus come to see the weakness of popular intuitionism, and the necessity for a sounder ethical theory. . . . Most of the third book requires no commentary whatever; by common consent its chapters are so lucid, and the conclusions they embody are, for the most part, so indisputable, that even to point out their merits would be to gild refined gold’ (Ethical Philosophy of Sidgwick, pp. viii, xviii). 2 This has been called the ‘accordion effect’; see Feinberg, ‘Action and Responsibility’, p. 134.

consequentialism vs. deontology

153

There was a related error in his claim that in their initial formulations the deontological principles contradict each other. His official statement of his third condition said that any self-evident propositions must be mutually consistent. . . . any collision between two intuitions is proof that there is error in one or the other or in both. Still, we frequently find ethical writers treating this point very lightly. They appear to regard a conflict of ultimate rules as a difficulty that may be ignored or put aside for future solution . . . Whereas such a collision is absolute proof that at least one of the formulae needs qualification. (ME 341)

Here he took a conflict of duties to involve formal inconsistency and to demand the qualifications only such inconsistency requires. As Ross emphasized, however (FE 88–9), conflicting prima facie principles are not contradictory. That in his example you ought other things equal to keep your promise and ought other things equal to help the accident victims are perfectly consistent claims; each correctly describes one aspect of your moral situation. It might be problematic to say you ought all things considered to keep the promise and all things considered to help the victims, but the principles do not say that. In his more detailed discussions of deontological duties Sidgwick repeatedly said their ‘conflicts’ and ‘collisions’ require them to be restated with exception-clauses; as so stated, they are then neither intuitive nor widely accepted. But insofar as his demand for restatement rested on the claim that otherwise they contradict each other, it was in error. This mistake does not completely vitiate his negative argument. It just means he could not say conflicts of deontological duty involve both inconsistency and unclarity; at most they involve unclarity. But there is another point where he ignored the concept of prima facie duty. Anticipating the systematization argument of Book IV, he often argued that when common sense tries to remove its principles’ unclarities it looks to the consequences different acts will have, in particular for overall happiness. This suggests, he argued, that the principles are not really freestanding but are derived from the principle of utility, which is their ultimate ground. But his arguments for this claim often moved directly from the premise that common sense takes some account of consequences to the conclusion that it evaluates only by consequences, where the latter is not true of everyday thought. At one point he called the duty to compensate those we have harmed a ‘simple deduction’ from the utilitarian principle of benevolence, which we ‘approximately obey’ by reversing any reduction in happiness we have caused (ME 281). But if common sense really understood the duty to compensate that way, it would say that if you can either give a person you harmed a certain benefit or give an equal-sized benefit to someone else, it does not matter which you do. And common sense does not say that; it thinks the claim for compensation has priority, and does so even when you can give the stranger a somewhat larger benefit. If you can give him a vastly greater benefit, say by saving his life, you ought to, but that means only that its view is

154

british ethical theorists from sidgwick to ewing

moderate rather than absolute deontology; it is not consequentialism. He likewise argued that since common sense says it can be permissible to lie to a child or invalid for her own good, the only way to decide when lying is permissible is by ‘considerations of expediency’, or by weighing the total good and bad effects a lie will have (ME 316). Again, however, common sense does not think lying even to an invalid is permitted for just a small increase in overall good; it must be a large one. And Sidgwick’s line is especially implausible about the duty not to harm. In one passage he said that when we consider the ‘negative duty of abstaining from causing pain to others against their will’, which he was unsure whether to class under justice or benevolence, we find the only way to define its limits and determine when we may harm some for the sake of others is ‘the Utilitarian formula’ (ME 348). But common sense does not think it is permissible to kill one innocent person in order to save just a few more; in the well-known example it does not think you may push one person in front of a trolley to stop it from running over five. It thinks the duty not to kill has substantial weight against consequentialist considerations and is therefore not derived from them. It may say that if the only way to save a million people is to kill one, doing so is permitted and even required. But that is again moderate deontology, not consequentialism. That Sidgwick ignored the moderate view was Broad’s main objection to his critique: Sidgwick ascribed to common sense and then criticized an ‘extreme’ form of deontology, where there is a more restrained form that both fits everyday thought and escapes his objections (FT 217–23). If we set aside his charge that unqualified deontological principles contradict each other, Sidgwick’s remaining objection was that they are unclear, first because they cannot be weighed precisely against each other. How serious a defect is it if a deontological theory cannot always make determinate judgements? We need again to distinguish between metaphysical and epistemic indeterminacy. There is metaphysical indeterminacy if the moral truth itself is partly indeterminate, so in some cases it is true neither that one duty outweighs another, nor that the second outweighs the first, nor that their weights are exactly equal. There is merely epistemic indeterminacy if the truth is always determinate but we sometimes cannot know it. If a deontological view says there is metaphysical indeterminacy in the weights of some duties, it is hard to see how its own indeterminacy is any objection to it: it is simply reflecting the moral truth as it takes it to be. We might find the view’s incompleteness objectionable if we thought of morality as a device to help us make decisions we cannot otherwise make, but deontologists do not see morality that way; they see it as a matter of objective normative truth, as Sidgwick also did. And if the truth is not completely precise, a theory that describes it accurately cannot be so either. It may be argued on general metaphysical grounds that, whatever its subject matter, the truth is never even partly imprecise but is always fully determinate. But Sidgwick did not make this argument, and in any case it does not apply to a view that says there is only epistemic indeterminacy.

consequentialism vs. deontology

155

This was, recall, Ross’s view. He thought there always is a precise truth about the weight of competing duties, though it is often beyond our power to detect, and it is again hard to see what the objection to this view could be. There might be one if we knew in advance that we can apprehend all the moral truth there is, but Sidgwick did not make this claim, and it is surely not plausible. We do not know all the mathematical truths there are, and why think our access to the moral truth is any more complete? A deontological view like Ross’s is not metaphysically problematic, because it thinks the moral truth is fully determinate, and it rests the incompleteness of its own claims on the familiar fact of human cognitive limitations. What is remotely objectionable about that? It is actually puzzling that Sidgwick made his first charge of unclarity, since his own preferred theories have the very same feature. They say that what is right is what will result in the most pleasure possible, either for the agent or for all sentient beings, and that, as he often emphasized, is not something we can always know. His canonical method for applying the theories was ‘empirical hedonism’ (ME 123), which identifies the consequences different acts will have and then compares the amounts of pleasure and pain they contain. But even apart from the difficulty of knowing exactly what consequences an act will have, he did not think we can compare hedonic states precisely. When he considered two pleasures of his own of the same type, he said, ‘it is only to a very limited extent that I can obtain clear and definite results from such comparisons’ (ME 143), and the difficulties increased when he considered pleasures of different types or of different people. There had been proposals from Spencer and others to measure pleasure ‘scientifically’, using laws derived from psychology or evolutionary biology, but he thought they all fail (ME 176–95, 470–3; GSM 146–77) and hedonistic theories must use the empirical method with all its ‘perplexity and uncertainty’ (ME 460). But then they cannot always say either that one of two acts has better consequences than the other or that their consequences are exactly equally good. The puzzle is why he did not consider this any objection to these theories. He thought the difficulty of weighing deontological duties requires that they be reformulated to include exception-clauses, but the indeterminacy in hedonistic theories was just a practical inconvenience. Why this radical difference in response? He might have justified the difference if deontological theories involve metaphysical indeterminacy and hedonistic theories only an epistemic one, but he did not make this claim, and it does not apply to a deontology like Ross’s that takes the moral truth to be fully determinate. He could also have argued that the hedonistic indeterminacies concern empirical facts, while the deontological ones concern moral ones and would remain even if we had full empirical knowledge. But he again did not make this claim, and would need to give an argument why it matters. Why should indeterminacy resulting from one cognitive limitation be a serious objection while that resulting from another is not? Not only was his first charge of unclarity

156

british ethical theorists from sidgwick to ewing

unpersuasive in itself, it criticized deontological theories for a feature he did not find at all problematic in his own.3 Sidgwick’s second objection was that deontological principles are unclear individually, or have vague boundaries. When this is so we again cannot judge acts decisively, but the reason is now internal to one or more principles rather than due to any conflicts between them. Though Ross effectively answered the charge about conflicts with his concept of prima facie duty, he did much less to address this second one. On the contrary, he tended to state his prima facie duties in just the vague way Sidgwick found unsatisfactory, saying only that there is a duty to keep promises (but exactly when?), show gratitude (for exactly what?), or not harm others (which differs how from not benefiting them?). This second vagueness is more problematic. It is one thing for a moral theory to be indeterminate where it says we cannot have self-evident knowledge, as Ross thought we cannot about conflicts of duty. It is another for it to be indeterminate where it says we can have such knowledge. If we have a self-evident intuition that a property is right-making, should the intuition not should tell us exactly what the property is and when it is present? Sidgwick said that if deontological duties ‘can be referred to independent and self-evident principles, the limits of each must be implicitly given in the intuition that reveals the principle’ (ME 247). Where there is genuine intuition, in other words, there should at least after reflection be full determinacy, and with the proposed deontological intuitions there is not. Ross did address an objection of this type about promises. His response to a series of problem cases raised by W.A. Pickard-Cambridge4 was that the boundaries of a promise are set by conditions attached to it when it was made and accepted, if only tacitly, by both parties. Thus, if they both made it a condition of a promise that circumstances not change radically, and circumstances have changed, the duty to keep the promise is void (FE 94–8). But it is not clear that this clarification fully meets the objection. Some may agree that promises can have conditions but insist that they be explicitly rather than tacitly accepted. Others may allow tacit acceptance but disagree about what is required for it. Ross described a case where if A had stated a condition openly B would have accepted it, and said the condition therefore ‘implicitly determines the nature of the understanding’ between them (FE 94). But some may deny that this merely hypothetical fact suffices for B’s acceptance; he must do something to actually accept the condition. Yet others may deny that both parties must understand the promise in the same way. If A honestly believes B accepted a condition, then the condition is part of the promise even if A’s belief is false; Alan Donagan reports this as Whewell’s view.5 And of any fully determinate account of the duty to keep a promise we can ask Sidgwick’s questions: does a duty about promising 3 4 5

See also Phillips, Sidgwickian Ethics, pp. 101–3. Pickard-Cambridge, ‘Two Problems About Duty’. Donagan, ‘Sidgwick and Whewellian Intuitionism’, pp. 457–8.

consequentialism vs. deontology

157

specified in this way really seem self-evident, and is it really universally accepted? Similar issues arise about other deontological duties, such as the duty not to harm. As the voluminous discussion of trolley cases has shown, it is far from easy to say just what makes for the morally more objectionable case of harming as against the less objectionable one of not benefiting, and thus far from easy to say exactly when the duty applies. But should a self-evident intuition that harming is wrong not make it self-evident what harming is? There might not be a difficulty here if we could always judge particular acts decisively. Then a deontologist could say the situation is like that with a concept and its analysis. In judging particular acts we rely implicitly on certain criteria but cannot always state them explicitly; doing so is the task of ethical theory. But we cannot always judge particular acts. When we consider difficult cases about promising we are often uncertain whether a past pledge is still binding, and though we may be confident that it is permissible to throw the switch in the original trolley case, most of us are less certain about one involving a looping track and can disagree about it. Sidgwick’s second problem was not just about making implicit criteria explicit. A deontologist could try accepting this second indeterminacy as he does the first, though this would involve abandoning Sidgwick’s clarity condition. The later Ewing took this line, saying intuitions can be valuable even if they are somewhat confused, and citing Whitehead, Hegel, and Heidegger as philosophers who offer such intuitions (VR 46, 127). The resulting indeterminacy would then be metaphysical if there was no determinate truth about which precise properties are right-making and epistemic if there was one but we could not always know it. On either view there would be core cases where we are certain a right-making property is present, for example when a promise has explicitly been understood in the same way and all its conditions are satisfied, and peripheral cases where we are not. But both possibilities are problematic if the truth in question is supposed to be self-evident. Mathematical truths have precise contents and are known by us, when they are known, as precise. If moral truths have a similar status, should they not be knowable in the same way? Sidgwick’s second objection was therefore a serious one, but the question is whether it applies only to deontological theories. He thought his own consequentialist principles satisfy the clarity condition, but they need to be supplemented by claims about the good and there can be indeterminacies in these, either about the comparative weights of different goods or about the boundaries of a single one, for example, about exactly what counts as a virtue. These difficulties do not arise if, as Sidgwick thought, the only good is pleasure, but others remain. As he recognized (ME 415–16), a consequentialist principle can tell us to maximize either the total good in a population or the average good per member, and it is not obvious which it should do.6 May there then be core cases of consequentialism where the same act maximizes 6 Ross denied that it is self-evident that we should prefer promoting the greatest total good to promoting the greatest average (FE 69–71; also Broad, FT 249–50), and there are similar difficulties about the distribution of happiness, e.g. is it better if it is equal, and if so, how much weight does equality have?

158

british ethical theorists from sidgwick to ewing

both the total and the average and more contestable ones where the two diverge? There can be similar indeterminacies about individual pleasures. Sidgwick thought a pleasure that is twice as intense is always twice as good, but one can instead hold that the value of an increase in a pleasure’s intensity gets smaller the more intense the pleasure is; the choice between this view and Sidgwick’s can again generate indeterminacies. Sidgwick raised a serious challenge in his argument that if we can intuit a moral truth as self-evident we should be able to intuit it with a precise content. But it is unclear that the challenge applies only to deontological theories rather than more widely, to all claims to self-evident moral knowledge. If it applies more widely, it cannot support consequentialism over rival views.

7.2 Sidgwick: For Consequentialism The positive part of Sidgwick’s argument, given in Methods III 13, was that the principles defining consequentialism do satisfy his conditions. It too had flaws. The principles in question were his axioms of justice, prudence, and rational benevolence. The first says, ‘if a kind of conduct that is right (or wrong) for me is not right (or wrong) for some one else, it must be on the ground of some difference between the two cases, other than the fact that I and he are different persons’ (ME 379); it expresses the idea of universalizability or supervenience. The second initially says ‘one ought to aim at one’s own good’, but Sidgwick worried that if ‘good’ is analysed as ‘what one ought to aim at’ this claim is tautological. He therefore supplemented it with a demand for ‘impartial concern for all parts of our conscious life’, saying ‘the mere difference of priority and posteriority in time is not a reasonable ground for having more regard to the consciousness of one moment than to that of another’ (ME 381). He derived his third axiom, of benevolence, in stages. First he said ‘the good of any one individual is of no more importance, from the point of view (if I may say so) of the Universe, than the good of any other’. Adding the further claim that ‘as a rational being I am bound to aim at good generally . . . not merely at a particular part of it’ yielded the axiom in its full form: ‘each one is morally bound to regard the good of any other individual as much as his own, except in so far as he judges it to be less, when impartially viewed, or less certainly knowable or attainable by him’ (ME 382).7 Combining this axiom’s demand for impartiality across persons with the temporal impartiality required by the axiom of prudence yields a fully impartial consequentialist principle. It is unclear whether Sidgwick thought his first, or universalizability, axiom is substantive or follows analytically from the concept ‘right’. His calling it the axiom of ‘justice’ suggests the substantive reading; his saying we can arrive at it ‘by merely 7

A principle derived from two others cannot really be self-evident, but Sidgwick seems to have ignored this point.

consequentialism vs. deontology

159

reflecting on the general notion of rightness’ (ME 208, also 34; ME1 364) suggests the analytic one. But as he recognized (OHE 230, 232), the axiom is not distinctively consequentialist; it is accepted by most deontologists and cannot support consequentialism against them. Some commentators take the second axiom to make no positive claim but only the negative one that you should not care more about goods at one time in your life than at another, which is compatible with not caring about your good at all.8 But though Sidgwick did say the negative claim is ‘[a]ll that the principle affirms’ (ME 381), this is not on balance the most persuasive reading of his view. A merely negative axiom would not deserve the name ‘prudence’, which he elsewhere associated with positively seeking your good (ME 7, 25–6, 36, 327). The negative claim would also then not be just an ‘addition’ to the axiom needed to remove the appearance of tautology; it would be the axiom. And Sidgwick gave the axiom positive content when he restated it one chapter later as a ‘precept to seek . . . one’s good on the whole’ (ME 391, also 119–20) and when he considered a challenge to it in IV 2 (ME 418–19). Most importantly, a negative principle would not have the connection he thought his axiom has with egoism. In earlier editions of The Methods he said, ‘The axiom of prudence, as I have given it, is the self-evident principle on which, according to me, rational egoism is based’ (ME3 388; ME4 387), but that can only be so if the axiom tells us positively to pursue our good. In the final edition he said only that the axiom is ‘implied in Rational Egoism’ (ME 386), but a merely negative axiom is no more implied in egoism than it is in utilitarianism. And, crucially, only a positive axiom can generate the dualism of the practical reason. There can only be a ‘contradiction in our apparent intuitions of what is Reasonable in conduct’ (ME 508) if one apparently self-evident principle tells us to maximize just our own good and another tells us to maximize the good of all, and that is certainly how he described the dualism elsewhere (EEM 43). He stated his self-evident axioms in III 13, and there the one that can ground egoism is the axiom of prudence; moreover that axiom is precisely what he said must be harmonized with the axiom of benevolence if the dualism is to be avoided (ME 498). There has also been dispute about the axiom of benevolence. It cannot plausibly be read as just negative, but Shaver has argued that it describes only what is true ‘from the point of view of the Universe’, so an egoist can accept it as true yet not be committed to act as it commands so long as he refuses to adopt that point of view.9 This reading has the merit of making Sidgwick’s axiom more likely to satisfy his consensus condition, but again it does not fit crucial aspects of his text. It attaches great weight to a ‘point of view of the universe’ phrase he used only twice in the body of The Methods (ME 382, 420), introduced tentatively (‘if I may say so’), and elsewhere analysed reductively, so the good from the point of view of the universe 8 9

Phillips, Sidgwickian Ethics, p. 96; Shaver, Rational Egoism, pp. 74–7, and ‘Sidgwick’s Axioms’, p. 179. Shaver, ‘Sidgwick’s Axioms’, pp. 174–84.

160

british ethical theorists from sidgwick to ewing

is just what everyone ought to desire and pursue (EEM 27). The phrase occurs only in one of the premises from which he derived his axiom and neither in the axiom itself nor in his later restatements of it as, for example, an unrelativized ‘precept to seek . . . others’ good no less than one’s own’ (ME 391–2, also 496, 500, 507; EEM 25, 43, 44). And whereas Shaver says an egoist can accept the axiom of benevolence, Sidgwick explicitly denied that, saying a consistent egoist must make no claims about the universal good, since if he makes them he can be argued into impartialism (ME 420–1, 497–8). Shaver’s reading also makes the axiom massively fail the clarity condition; whereas Sidgwick criticized deontological principles for not always giving determinate guidance, a relativized axiom by itself gives no guidance. For that it must be supplemented by ‘You ought to adopt the point of view of the universe’, a claim Sidgwick, tellingly, never discussed. A relativized axiom also does not by itself yield consequentialism, as Sidgwick said it does (ME 387–8); that too requires the ‘ought’ claim he never discussed. And, again crucially, a stronger axiom is needed to generate the dualism. A principle merely describing what is true from an optional point of view does not and cannot contradict egoism; only an unrelativized principle telling each person to maximize the good of all does. For a multitude of reasons, he must have intended a principle of that kind. In the first edition of The Methods Sidgwick took a claim about your own good to be primary and derived the axiom of benevolence from it as a kind of corollary (ME1 359–60), but later editions treated the two axioms as coordinate and of equal selfevidence. More specifically, they took both to rest on a more abstract thesis ‘obtained by considering the similarity of the individuals that make up a . . . Mathematical or Quantitative Whole’ and saying it is irrational to care more about one part of a whole, say, the pleasure in one temporal part of your life, than about another (ME 380–3).10 This thesis cannot, however, give independent support to the axioms. Imagine someone who argues that the total pleasure in your life is only part of a whole comprising all your hedonic states, both pleasures and pains, so it is irrational to promote only your pleasures; you ought to promote your greatest total of both pleasures and pains. Sidgwick would object that a whole comprising both pleasures and pains is not a relevant one for moral purposes, but someone who rejects the demand for temporal impartiality can likewise deny that a whole comprising all the times in his life is relevant, and an egoist can do the same about a whole including all people. To apply the thesis about wholes we must first know what is a relevant ‘similarity’, which is precisely what the axioms address. Though he had rigorously tested deontological principles, he did not subject his own axioms to anything like the same scrutiny. On the contrary, his treatment of them was extremely lax.

10

The role of this thesis in Sidgwick’s arguments is emphasized in Irwin, Development of Ethics, vol. 3, pp. 496–505.

consequentialism vs. deontology

161

For one thing, he did not seriously consider possible counterexamples to them. The axiom of prudence requires full temporal impartiality, but Parfit has objected that it is not unreasonable to care less about pleasures and pains in your past than about ones in your present or future. If while suffering temporary amnesia you are told that you either had a long and very painful operation yesterday or will have a shorter, less painful one today, you will, perfectly reasonably, hope you had the longer operation yesterday.11 The same point was made by McTaggart. He asked us to consider two people, each with a perfect memory of his last ten years but no memory of earlier ones, and a perfect forecast of his next ten years but none beyond that. If the first person has had ten years of agony and will now have ten of pleasure while the second has had ten years of pleasure and will now have agony, will the first not feel much happier today (NE II 348–9; also Carritt, TM 26; Ewing, VR 281)? If so, the two do not value pleasure time-neutrally. This is a philosophically subtle objection, but there are more straightforward ones to the impartiality of the axiom of benevolence. Sidgwick acknowledged one when he said, after stating the axiom, ‘the duty of Benevolence as recognised by common sense seems to fall somewhat short of this’, because it does not require you to sacrifice your good for just slightly greater benefits to others; it is less demanding (ME 382, also 252–3, 499). His response was that if ‘a “plain man,” in a modern civilised society’, were ‘fairly brought to consider . . . whether it would be morally right for him to seek his own happiness on any occasion if it involved a certain sacrifice of the greater happiness of some other human being’, he ‘would answer unhesitatingly in the negative’. But he gave no evidence for this claim about the plain man—he had not done an opinion survey—and it is highly dubious; surely many would reject his axiom as too strict.12 Nor did he consider the even more telling objection from nonmaleficence: that the axiom is wrong to make it permissible to kill one innocent person if that will produce slightly greater benefits for others. Here the claim that a plain man would ‘unhesitatingly’ take the consequentialist line is not credible. In addition, he did very little to substantiate his claim that his axioms satisfy the four conditions. Applying the second condition, he said they ‘do present themselves as self-evident’ as rival principles do not (ME 383), but he gave no evidence for this claim, as, to be fair, it would be hard for him to do. He seems simply to have assumed they satisfy the clarity condition, not considering difficulties like the ones about the measurement of goods or total versus average good, and he did not discuss the question of their consistency. This may be because he thought his axioms of prudence and benevolence contradict each other and wanted to leave that topic until he addressed the dualism in his final chapter. But if he was saying his axioms satisfy all the conditions whereas deontological principles do not, should he not have 11 Parfit, Reasons and Persons, pp. 165–7; Parfit’s example was anticipated in Kraut, ‘Rationality of Prudence’, p. 354. 12 Parfit calls Sidgwick’s claim about the plain man ‘simply false’ (On What Matters, vol. 1, p. 453).

162

british ethical theorists from sidgwick to ewing

acknowledged this conflict earlier? And his argument that there is consensus on his axioms was disturbingly thin. His claim about the common-sense duty of benevolence was in part an attempt to show that his third axiom is widely accepted, but, as I have said, it is not persuasive. And his more explicit claims about consensus mentioned the opinions only of philosophers, and then only of two, Samuel Clarke and Kant, as if their assent could stand in for all moral theorists’ and then for all people’s generally (ME 384–6). How is an appeal to two people’s views a serious test of agreement?13 He did say he was considering the opinions of those moralists who have been ‘most in earnest’ in seeking ‘genuine intuitions’, but his initial description of dogmatic intuitionism said its proponents seek just such intuitions (ME 101), and it can hardly suffice for denying that Whewell was ‘in earnest’ that he did not share Sidgwick’s conclusions. Sidgwick did not, therefore, make a serious case that his axioms satisfy all his conditions. And what he did say reflects a fundamental unfairness in his argument, one that again turns on his lacking the concept of prima facie duty. Though he did not put it this way, his argument about conflicts of duty rested at bottom on the charge that deontologists equivocate between other-things-equal and all-things-considered claims. The unqualified principle ‘you ought to keep your promises’ makes acts falling under it right only other things equal, since some other principle can make them on balance wrong; it therefore does not yield decisive verdicts. Clear verdicts do follow from a principle that contains exception-clauses and so makes acts falling under it all things considered right, but that principle neither seems self-evident nor is accepted by all. But Sidgwick’s statements of his axioms equivocate just as much between otherthings-equal and all-things-considered claims. This is evident, first, in his axiom of prudence, which he needed to play two roles. On the one hand, it had to supply part of the framework of impartial consequentialism, its temporal neutrality, and for that purpose had to have an other-things-equal form, saying you ought to promote your own greatest good through time only when that will not prevent you from promoting more good for others. Then it could be extended to require temporal impartiality about others’ lives when it was combined with the axiom of benevolence. But it also had to be the basis for rational egoism or the self-evident element within it, and there it had to make an all-things-considered claim, saying that, no matter what else will result, you ought to promote your own good. That is certainly how he interpreted its central claim when he attributed it to Butler and Clarke (ME 119–20), restated it in an article of 1889 (EEM 43), and discussed it in connection with the dualism (ME 498).

13 The first edition of The Methods introduced its axioms through a discussion of Kant and Clarke (ME1 357–64), and though in later editions Sidgwick gave a ‘more direct’ statement of his views (ME ix), the material on them remained.

consequentialism vs. deontology

163

The axiom of benevolence did not play two roles, and therefore was needed only in its all-things-considered form. But this cannot be how Sidgwick was understanding it when he said it is the object of a consensus including Kant. For Kant clearly did not think you ought always to do what will result in the most good for all people. As a deontologist—in fact, the best-known one—he accepted the principle of benevolence (minus its claim about your own pleasure) only in an other-thingsequal version, so it can be and often is outweighed by other principles; thus he did not think benevolence permits you to lie or kill. Whereas Sidgwick’s defence of consequentialism required the axiom in a strong form, Kant accepted it only in a much weaker one.14 In fact Sidgwick’s axioms are open to the same general objection he made to deontological principles. Only when they are stated as other-things-equal is it at all plausible to say they appear self-evident and are generally accepted, though even then there can be objections to them like those mentioned above. But in that form they cannot ground either egoism or utilitarianism, as Sidgwick needed them to do. To do that they must make all-things-considered claims, but then they are less intuitive and will be widely rejected: many will deny that you should always do only what is best for you, and many will also deny that you should always do what has the best consequences for all. As he said of deontological principles, Sidgwick’s axioms may satisfy some conditions in one form and others in a different form, but in no form do they satisfy them all. It may be said in his defence that he did not need all-things-considered axioms. He could accept only other-things-equal ones and arrive at consequentialism by combining them with the negative claim that there are no other self-evident principles, which claim his critique of deontology supplies. By itself each axiom would then have a weaker form, but adding the critique would make it de facto stronger, or de facto all-things-considered.15 Like Shaver’s interpretation of the axiom of benevolence, this one gives the axioms a better chance of satisfying the consensus condition; it also fits Sidgwick’s later claim that he had only shown the consequentialist principle to be one moral axiom but not that it is sole or supreme (ME 421). But it has other and serious interpretive flaws. First, it makes the axioms fail Sidgwick’s clarity condition in the very way for which he criticized deontological principles, since as other-things-equal they do not by themselves yield decisive verdicts. It also means consequentialism does not have a fully self-evident basis: its ground now includes a negative claim that is not intuited but depends on a complex critical argument, carried on through eleven chapters of Methods III. Though Sidgwick claimed to be certain of consequentialism, that is not

14 This point about Kant is also made in Irwin, Development of Ethics, vol. 3, p. 518. If Kant lacked the concept of prima facie duty, his principle of benevolence would have included exception-clauses but would still have been weaker than a consequentialist one. 15 Skelton, ‘Sidgwick’s Philosophical Intuitions’, pp. 203–4; Shaver, ‘Sidgwick’s Axioms’, pp. 185–94, 200.

164

british ethical theorists from sidgwick to ewing

possible if one of its grounds is a discursive argument that as such cannot justify certainty. Nor on this reading can the dualism involve, as Sidgwick said it does, a contradiction between apparently self-evident judgements, since as other-thingsequal the axioms do not contradict each other. In fact, on this interpretation there cannot even be a dualism, central though that was to Sidgwick’s thought. That would require both a demonstration that there is no self-evident axiom other than the axiom of prudence—so it can be affirmed not only as other-things-equal but also as all-things-considered—and a demonstration that there is no self-evident axiom other than the axiom of benevolence—so it can be affirmed in the same way. Since there cannot be both those demonstrations—they contradict each other—there cannot be both an all-things-considered egoistic principle and an all-things-considered utilitarian one. And whereas this interpretation assumes that Sidgwick understood the concept of other-things-equal duty, we have seen that he did not, since he repeatedly failed to use it when it was relevant. How then could he have had it in mind when stating his axioms? I believe the only conclusion is that Sidgwick equivocated. Just as he tended to assume that any deontological principles must be absolute or all-thingsconsidered, so he often read his axioms as strong enough to ground fully determinate views such as egoism and utilitarianism. But he also wanted them to meet his four conditions, and when he thought of them that way he slid into imagining them as other-things-equal. He thought of them in one way when drawing conclusions from them and in another when saying they are intuitive, as he could not have done had he clearly grasped Ross’s concept. Sidgwick is often considered a paradigmatically fair-minded philosopher. McTaggart spoke of his ‘scrupulous fairness to opponents’ (‘EHS’ 412), Broad called The Methods a ‘conspicuously honest’ book (EHP 72), and even critics like Donagan and Bernard Williams have called him ‘a man of heroic disinterestedness’ and noted the ‘marked scrupulousness of his arguments’.16 But while this trait is often evident in his discussions of smaller topics, it can be missing in his treatment of larger ones. (A general truth about The Methods is that the more important a point is, the less time Sidgwick spent on it.) He took it to be a serious objection to pluralist deontology that it cannot always judge acts determinately because it cannot weigh duties precisely, but no objection to hedonistic theories that they cannot compare pleasures precisely. He tested deontological principles rigorously with his four conditions but barely applied those conditions to his own principles. And though one of his main objections to deontological principles turned on the difference between their other-thingsequal and all-things-considered forms, his defence of his axioms equivocated on the same point and involved the same ambiguity. In arguing against deontology and for consequentialism he applied, and not just once, a double standard.

Donagan, ‘Sidgwick and Whewellian Intuitionism’, p. 456; Williams, ‘Point of View of the Universe’, p. 154. 16

consequentialism vs. deontology

165

7.3 Prichard, Carritt, Ross: Defending Deontology Despite the flaws in Sidgwick’s case consequentialism may be true, and the best defence of it may be the simple assertion, as in Moore, that its principle is selfevident. This defence has to drop Sidgwick’s consensus condition, since not all people or even all philosophers accept consequentialism, and say its truth can be seen only by those with the right preparation, for example, the right background in moral theory. But similar claims can be made on the other side, by deontologists who say their principles are self-evident. They too cannot claim consensus, since some do think right acts maximize the good, but Prichard, Carritt, Ross, and to a lesser extent Broad, gave two related intuitive arguments for deontology. The first appeals to our intuitive judgements about particular cases. In an early paper Prichard said consequentialism makes the wrongness of a lie proportional to the badness of the error it will produce and a widow’s duty to give a mite vastly less than a millionaire’s to give a million, both of which he found counterintuitive. But consequentialism ‘most plainly’ violates our moral convictions about how to distribute goods between people. If a landslide has damaged a village, it says the one healthy person has an equal duty to find food and shelter for himself, for his parent, and for a stranger. But even if he has a right to consider himself as much as any other, Prichard said, no one thinks he has a duty to; even if he could help himself most, he would not be wrong if he helped others. And his duty to help his father is stronger than his duty to help the stranger. Consequentialism also says your duty to suppress bad thoughts or develop courage in yourself is no stronger than your duty to help others do the same, but in fact each person’s virtue is his own ‘special business’ (MW 2). He was here defending the common-sense view of benevolence against an impartial one, but he also made an objection about justice, saying consequentialism is indifferent between giving the whole of some total good to one person and dividing it equally among many. Carritt and Ross repeated this charge and added one about desert, saying consequentialism is indifferent between promoting the happiness of a virtuous person and that of a vicious one (TM 38; EPT 62–3, 68; RG 35; FE 70–2). But these objections do not apply to a consequentialism that recognizes a good of just distribution, either equal or proportioned to desert. Ross’s own theory included a good of that sort and used it to ground its duty of justice (RG 27). As Ross eventually recognized (FE 73–4), at least many intuitions about justice can be captured by a consequentialism that treats certain distributions as intrinsically good. Other objections concerned cases where consequentialism recommends acts that violate an intuitive deontological constraint, such as against breaking promises. Carritt wrote of ‘our inexpungable belief in the rightness of keeping some promises, paying some debts . . . even when more happiness might be produced by not doing these acts’ (TM 40). Ross gave an example where fulfilling a promise to A will produce 1000 units of good for him while another act will produce 1001 units of good for B, to whom you have made no promise; here he said you ought to keep the

166

british ethical theorists from sidgwick to ewing

promise (RG 34–5). Consequentialists sometimes try to justify the ban on promisebreaking by pointing to long-term effects on the institution of promising, but Carritt thought this assumes ‘a very remarkable stupidity in men’, as if they will be encouraged by promise-breakings that do promote overall good to perform ones that do not. (Though stupid, this happens.) He also said the consequentialist argument does not apply when the breach of promise will not become known, as in an unwitnessed promise to a dying creditor to pay his children (TM 39–40; EPT 64–5). Ross too cited deathbed promises (FE 104–5, 110) and argued that sometimes promise-breaking has good long-term effects; thus ‘the breach of Belgian neutrality by Germany in 1914 strengthened rather than weakened in all impartial observers the sense that treaties ought to be observed’, yet it still was wrong (FE 93). More generally, he thought consequentialists tend to exaggerate the bad effects of a single broken promise. A utilitarian ‘is apt, in order to bring the rigour of his view about promises up to that of the ordinary moral consciousness, to make more of the utilitarian reasons for keeping promises than the facts warrant’ (FE 113). Whereas Sidgwick had said common sense is unconsciously utilitarian, Ross was saying utilitarians are unconsciously deontological. Like Sidgwick, these deontologists said less about non-maleficence than we might expect. Ross illustrated this duty by saying we do not think it ‘justifiable to kill one person in order to keep another alive, or to steal from one in order to give alms to another’ (RG 22). This claim is unnecessarily weak, since he could equally well have said we do not think it right to kill one to keep several alive. In his later book he made the better claim that whereas utilitarians see no difference between inflicting pain and failing to give pleasure, ‘we should, in fact, . . . always judge that the infliction of pain on any person is justified only by the conferment not of an equal but of a substantially greater amount of pleasure on some one else’ (FE 75, also 106). Carritt likewise said that on a utilitarian view, ‘if we thought the torture of a few victims increased the sum of happiness we should think it right’ (TM 40), and together with Ewing the two developed what became a common objection to consequentialism: that it can favour deliberately punishing an innocent person.17 This possibility had not been discussed by earlier writers on punishment such as Kant, Bentham, and Bradley. It was first suggested by Ewing, who in a 1927 article asked, ‘If the justification of punishment lies solely in its future consequences . . . why is it wrong to punish an innocent man?’18 A year later Carritt said that if utilitarianism is true, we should ‘hang a man, guilty or not, if we think the example likely to prevent two or more murders’ (TM 108, also 109; EPT 65), and in 1929 the objection was repeated by Ewing (MP 54) and given an especially clear statement by Ross in a passage reprinted in The Right and the Good: ‘a government which found some offence against the law prevalent, and in its inability to find the offenders punished 17 18

The following account is drawn from Sverdlik, ‘Origins of “The Objection” ’. Ewing, ‘Punishment as a Moral Agency’, p. 295.

consequentialism vs. deontology

167

innocent people on the strength of manufactured evidence, would still be able to justify its action’.19 From then the objection became standard: repeated by Joseph (SPE 93), J.D. Mabbott, Rawls, H.J. McCloskey, and G.E.M. Anscombe.20 Our deontologists made the objection mostly against consequentialist accounts of punishment rather than against consequentialism more generally. In that larger context they relied more on examples involving such duties as to keep promises and pay debts. Promise-keeping was seen as an especially clear non-consequentialist duty (Sidgwick, ME 353; Ross, FE 113; Carritt, EPT 101), and in central cases it does have a fairly definite content. But it is also a comparatively weak duty and easily outweighed by consequentialist considerations in examples like Ross’s of the accident victims. The duty not to harm, and especially not to kill, is much weightier and has therefore played the leading role in present-day critiques of consequentialism. But Ross placed non-maleficence last on his list of prima facie duties, illustrated it initially with the weak example of killing one to save one, and never elaborated its content as he did the duty to keep promises. In his second book Carritt gave a list of prima facie duties that mostly followed Ross’s though with more detail about each, but he unaccountably omitted non-maleficence—it is not on the list (EPT 96–112). He included a duty of justice, but with only distributive and retributive sides, and subsumed the ‘disobligation to cause “wanton” pain’ under that of beneficence; he also mentioned a duty of non-interference with freedom but classed that too with beneficence. The idea that there is a duty not to harm that is stronger than any to give benefits was, surprisingly, not much stressed by our group. As we will see in the next chapter, that may in part be because this duty does not fit easily into their general understanding of deontology. Their second argument appealed to explanatory intuitions and said that even if consequentialism correctly identifies some acts as right, it gives the wrong explanation of why they are right. After his first paper Prichard argued almost exclusively in this vein, saying in ‘Mistake’ that the reason we think we ought to pay our debts or tell the truth is not that we will thereby originate something good (MW 10, also 29–30). Ross’s first objection to the utilitarian account of promise-keeping was that when we think we ought to do what we have promised, it is not from any thought about total consequences but simply because we made the promise (RG 17); he repeated this objection often (RG 24, 36–9; FE 65, 68–9). Such points were less common from Carritt, who thought the most reliable intuitions concern particular acts. But it fit Prichard’s and Ross’s view that the primary moral intuitions are that certain properties are right-making and led them to reject all theories that posit a

Ross, ‘Ethics of Punishment’, p. 205; also RG 56–7, FE 102. Mabbott, ‘Punishment’, p. 152; Rawls, ‘Two Concepts of Rules’, pp. 3, 4, 18 (with an attribution to Carritt’s EPT); McCloskey, ‘Examination of Restricted Utilitarianism’, pp. 468–9; Anscombe, ‘Modern Moral Philosophy’, p. 10. 19 20

168

british ethical theorists from sidgwick to ewing

single ground for all duties, not just consequentialism but also Kant’s deontology (MW 50, 62; RG 18–19; FE 189; KET, 28–9; also Broad, FT 215). These two arguments, about particular cases and about explanations, reinforce each other. If a consequentialist says his theory does capture intuitive judgements about particular acts of, say, promise-keeping once long-term effects are considered, the explanatory objection says some intuitions are not captured, about why the acts are right. Conversely, examining particular cases can confirm our explanatory intuitions. After discussing a deathbed promise Ross asked why he would consider such an artificial example, where many of the normal good effects of promise-keeping are absent. He said he did not need the example to convince himself or most thinking people that there is a prima facie duty to keep promises, since that duty ‘seems to me to stand out as a salient fact in the moral situation, even when the moral situation bristles with good utilitarian reasons for fulfilling promises, or for breaking them’. But faced with theorists who do not accept that duty, an example that isolates the issue by eliminating other considerations ‘affords the best hope of convincing them’ (FE 105). The main argument for deontology remained an appeal to our intuitions about right-making properties, but considering particular cases could for some be a useful supplement. A consequentialist may reply to these arguments by saying it is good from a consequentialist point of view if people have deontological intuitions, since they are more likely to keep promises when that will maximize the good if they think there is an underivative duty to keep promises than if they think doing so is justified only by its effects. But, read one way, this reply does not take the intuitions at face value, as ones that claim to be true; it only gives a reason why, even if they are false, it can be beneficial for people to have them. It discounts the intuitions from an allegedly superior point of view, and, since the question at issue is precisely whether that point of view is superior, it begs the question. Read another way, the reply questions the intuitions’ evidentiary force. Sidgwick’s second condition says we should check that what we think is an intuition is not a mental state with some other cause, and the claim now is that our explanatory intuitions originated in and are sustained by their beneficial effects. But this claim needs to be supplemented by an argument that this was the intuitions’ origin, and Sidgwick did make arguments of this sort when he said common sense is unconsciously utilitarian (ME 456). As we will see in Chapter 8, however, his arguments were often too cursory to be persuasive, and if the reply assumes that the intuitions must have an alternative cause because they cannot just be graspings of moral truth, it again begs the question. Ross summarized his objections to consequentialism by saying it ‘ignores, or at least does not do full justice to, the highly personal character of duty’ (RG 22). It thinks ‘the only morally significant relation in which my neighbours can stand to me is that of being possible beneficiaries by my action’, whereas they can also stand in that of ‘promisee to promiser, of creditor to debtor, of wife to husband, of child to parent, of friend to friend, of fellow countryman to fellow countryman, and the like,

consequentialism vs. deontology

169

where each of these relations is the foundation of a prima facie duty’ (RG 19; also ‘BO’ 125; Prichard, MW 13; Carritt, TM 41, 43). Others can relate to you in several ways that give you an additional duty to promote their good over and above that derived from general beneficence. The claim that duty is ‘personal’ can be used to make a different point, about the agent-relativity of deontological duties. Your primary duty concerning promises, it can be said, is not to minimize the total amount of promise-breaking in the world; it is not to break your promises, and you have only a weaker if indeed any duty to prevent promise-breaking by others. That is why, if by breaking one of your promises you could prevent three other people from breaking theirs, it would not be right to do so. The school’s deontologists occasionally recognized this point. Carritt thought he might have a stronger duty to remove his own lunch-papers from a mountain than to remove other people’s, condemned executing an innocent person even to prevent two murders, and said two people trying to thwart each other’s efforts may both be acting rightly, since each is trying to pay his own debts (TM 41, 108, 75; also Ross, RG 56–7; Broad, FT 211). But he did not emphasize the point and tended to assimilate it to the one about other people’s relationship to you. Ross made even less of it, and neither he nor Carritt appealed to agent-relativity when it would have been most natural to do so, in response to the later versions of ideal consequentialism proposed by Joseph and Ewing. These versions try to capture deontological intuitions by making act-types such as promise-keeping intrinsically good or, as Ewing thought more plausible, by making promise-breaking evil (‘SN’ 9–10; DG 187). (If promise-keeping is good, you have a duty to make and keep promises; if promise-breaking is evil, your only duty is, more reasonably, not to break promises when you have made them.) They yield the right conclusion in Ross’s example of producing either 1000 units of happiness by keeping a promise or 1001 by breaking it, since adding the evil of promise-breaking to the second act’s effects makes it on balance wrong. But they do not do so when breaking your promise will prevent three other people from breaking theirs; then they tell you to break the promise. I suspect that if Carritt and Ross were given this case they would agree that it is wrong to break the promise, but they never considered this type of example or wielded it against ideal consequentialism. They did not clearly grasp the kind of relativity it illustrates, and neither did Ewing. Addressing an objector who says he has a stronger duty to promote his mother’s good than a stranger’s, he replied ‘What I am choosing in cases like this is not just my mother’s good but my mother’s good as produced by myself. . . . It does not follow, because the good of A and that of B are equally worth pursuit per se, that therefore the whole—good of A plus the pursuit and attainment of A’s good by me—will have a value equal to the other whole—good of B plus the pursuit and attainment of B’s good by me’ (DG 191–2; also ST 108). But this reply, which assigns extra value to a mother’s being benefited by her son, does not cover the case where Ewing’s choice is between his promoting his mother’s good and allowing three other sons to promote their mothers’ good;

170

british ethical theorists from sidgwick to ewing

then his view tells him to forgo promoting his mother’s good. Yet neither this case nor the issue it raises seems to have occurred to him. Prominent though this kind of agent-relativity has been in recent discussions of deontological duty, it was not clearly understood by our school.21 Carritt and Ross had a different objection to the later ideal consequentialism. The strongest version of the objection simply denies that acts as such have intrinsic value; motives aside, a world in which promises are kept is no better than one in which they are broken (TM 73; FE 141). This argument reflects their view that an act’s goodness is entirely a function of its motive, but they also had a fall-back; if we do judge an act to be good apart from its motive, it is only because we first judge it to be right. Carritt said ‘the state we ought to bring about is not first judged to be best in any other sense than that it is the one we ought to bring about’. If we take money from a poor good man to pay his debt to a bad rich man, ‘the “goodness” then seems to consist not in his having the money but in the rightness of paying it to him’ (TM 72–3.) Ross likewise said: a morally good spectator will find satisfaction in seeing these things [acts of reparation or gratitude] take place; but in this case the satisfaction seems to me to depend on the previous thought that it was A’s duty to make such compensation to B. . . . What a morally good spectator thinks to be a worthy object of satisfaction is B’s getting the advantage or pleasure by A’s action, by A’s giving it to him; and that thought rests upon the prior thought that . . . A has a duty to give it to him. (FE 289, also 142–3)

Even if we do judge acts to be good apart from their motive, their goodness cannot explain why they are right, because the rightness explains the goodness. Shaver has challenged this objection, saying its claim about the order of explanation is hard to sustain.22 But Carritt and Ross could reply that if we think about what makes the world better as a question in itself, without an eye on implications for the right, we will focus just on states such as pleasure and knowledge rather than on acttypes such as promise-keeping; that is certainly what classical writers on the good did. And there is a further argument. To capture the agent-relativity of deontological duties, an ideal consequentialism of Ewing’s type must make agent-relative claims about value, so what is evil from your point of view is not just promise-breaking by anyone but, or at least especially, promise-breaking by you. This kind of claim is possible given a fitting-attitudes analysis of ‘good’, but Ewing’s view also needs timerelative values, since common sense does not think that if breaking one promise now will prevent you from breaking three promises next year it follows that you should break the promise now. But now the distinctive character of consequentialism is under threat. ‘Good’ differs from ‘ought’ mainly in that it is not governed by ‘“ought”

21 To my knowledge the first works to recognize this agent-relativity were Williams, ‘Critique of Utilitarianism’, pp. 89, 99 and Nozick, Anarchy, State and Utopia, pp. 28–30. 22 Shaver, ‘Birth of Deontology’, pp. 138–40.

consequentialism vs. deontology

171

implies “can”’, but if what is evil from your point of view now is your breaking a promise now, that is under your voluntary control, or is if it is true that you ought to keep the promise. And then the difference between the claim about value and the one about ‘ought’ vanishes. To fully capture deontological intuitions, an ideal consequentialism of the later type must make claims about value that are indistinguishable from ones about duty. If we ask whether Prichard’s, Carritt’s, and Ross’s deontological claims meet Sidgwick’s conditions for self-evidence, the answer must be no. Despite what Ross said, they cannot be held with the certainty that he, Sidgwick, and Prichard thought genuine intuition involves; that would require assigning zero probability to consequentialism, which is unreasonable. (Recall, though, Ewing’s view that intuition can involve just high-probability belief.) Nor do the claims meet Sidgwick’s consensus condition. They are not accepted by all philosophers, nor by all outside philosophy; like consequentialist principles, they are disputed. Nonetheless, the deontological claims are closer to common-sense morality than consequentialism is. They fit better with everyday verdicts on particular acts such as keeping a promise when breaking it will produce a little more happiness or not killing one person to save a few others; they also fit better with beliefs about why these acts are right. If consistency with common sense is desirable in a moral theory, as Ross for one thought, deontology has it to a higher degree. It does not follow that deontology is correct. It may be that the true moral principles are consequentialist, but it takes a special kind of reflection, which Sidgwick, Rashdall, and Moore carried through but Prichard, Carritt, and Ross did not, to see that. Still, its plurality of ultimate duties makes deontology of Ross’s type a better fit with everyday moral beliefs and at least on that basis the preferable view.

8 Act-Consequentialism, Pluralist Deontology As well as defending consequentialism or deontology, the school’s members developed specific versions of these theories. Even with its one principle consequentialism raises issues of interpretation; pluralist deontology raises even more.

8.1 Consequentialism: Act- and Indirect Sidgwick, Rashdall, and Moore had comparatively simple consequentialist theories. They were, first, maximizing theories, equating what is right with what will result in the most good possible; our school were not familiar with the satisficing view that right acts need only produce consequences that are ‘good enough’.1 The theories were also act-consequentialist, saying an objectively right act itself produces the most good possible, as against rule-consequentialist views that say an act with the best outcome can be wrong if the result of everyone’s doing acts of that kind would be bad, or if it is forbidden by the set of rules whose acceptance by all would be best.2 Sidgwick considered something like the first view when he discussed the question ‘what if everyone did that?’; he said it does not yield distinctive verdicts if an act can be described very specifically, for example as ‘lying when few other people are lying or will be caused to lie’ (ME 318–19, 486–7). Broad made a similar point (CE 56–9) and asked, more generally, why it is relevant, when evaluating an act, to ask what would be the case in a situation that is not actual (CE 43–5).3 He might similarly have asked, of the second theory, why it is relevant to consider the verdict of rules no one may accept. Like others’ in the school, his principle about consequences was act-focussed. In applying their principle some seem to have assumed the share-of-the total view Prichard used in arguing that your duty is only to set yourself to produce an effect. Addressing the worry that an act may have consequences far into the future, Moore

1

Slote, Common-Sense Morality, Chapter 3. For the first view see Harrod, ‘Utilitarianism Revised’; for the second, Brandt, ‘Toward a Credible Form’, and Hooker, Ideal Code, Real World. 3 Ewing discussed the Kantian question more sympathetically in ‘What Would Happen?’ 2

act-consequentialism, pluralist deontology

173

said its effects should be ‘neutralised by lapse of time’, since as we get further after a choice ‘the events of which either action would be part cause become increasingly dependent on those other circumstances, which are the same, whichever action we adopt’ (PE 153). This seems to assume that the more causes combine to produce an effect, the less any one’s responsibility for that effect. And Ewing explicitly said temporally remote consequences matter less because ‘an act can never be more than part-cause of these’, and ‘the further removed they are the greater will be the number of other causes which contribute to them, and so the smaller the share of our act in their production’ (MP 170; also ‘U’104; Broad FT 201–2). As I argued in Chapter 3, this is a mistake. If an act makes all the difference between an effect’s occurring and not, which it can do even if the effect is far in the future, it is responsible for all that effect, and is so even if later events also make all the difference and are likewise fully responsible. In practical deliberation we can usually ignore remote effects, because although they can be very good or very bad, we have no reason to believe they will more probably be the one than the other. But if we are thinking of objective rightness, as Moore was, an act with good consequences in the short term can turn out many years later to have been horribly wrong. Alongside an objective consequentialist principle, some discussed a subjective one, which is relativized to your beliefs or evidence and considers not an act’s actual effects but its probabilities of having different effects. A subjectively right act need not be most likely to be objectively right. If one act has a .51 probability of producing a slightly greater good and a .49 probability of causing disaster while another is certain to produce a slightly lesser good, the second act is less likely to be objectively right but is subjectively preferable. This is because it has the greater expected value, where that is calculated by multiplying the value in each of its possible outcomes by the outcome’s probability and adding the results. Russell’s otherwise fine discussion blurred this distinction. He first said the subjectively right act is ‘the one which will probably be most fortunate’, where ‘most fortunate’ was his term for objectively right. But he then said the subjectively right act ‘gives us the greatest expectation of good on the balance, or the least expectation of evil’ (‘EE’ 31), which suggests the expectedvalue view. Broad was sceptical of the latter view, asking whether mathematical expectation ‘is really a measure of anything in the world’ (CE 28), but it is hard to see how that is a relevant objection. Ewing, however, explicitly accepted it (‘U’ 102), thereby accepting the correct subjective principle. In hedonistic utilitarianism an act’s good consequences usually follow it in time, but non-hedonistic views allow goods simultaneous with an act, and the ideal consequentialists often said there can be intrinsic value in an act itself (Rashdall TGE I 96–7; Moore, EE 83, PE 24–5, 147, 177–9; Ewing MP 5, 170–1, 198, 200, 202, 214, 217–18, ‘SN’ 9–10, DG 187, ST 97–8, 139–40; also Broad, CE 18–19). Rashdall and Moore thought this value depends entirely on the act’s motive, and in particular

174

british ethical theorists from sidgwick to ewing

on whether it is virtuous; this was also Prichard’s and Ross’s view.4 But as Rashdall pointed out, considering the value of the motive you will act from now can never make a difference in first-person deliberation, in the sense of changing what you should conclude is right. Imagine that you can produce either what you believe is a greater good A or what you believe is a lesser good B. The most virtuous motive you can act from now is the desire for A, but the fact that makes it more virtuous—that you believe A is better—already makes choosing A subjectively right. Considering your present motive can give you an additional reason to do an act that is already subjectively right, or magnify the difference in value you believe will result if you choose it rather than some alternative. But it cannot change what you should think is the right act now (TGE II 42–3, E 72; also Ewing, DG 196–7). This point does not hold for third-person objective assessments. If your belief that A is the greater good is false, your acting from the desire for A will still be more virtuous, and if A is only slightly less good than B, that can make your choosing A all things considered better; something similar holds if you falsely believe an act is right (Carritt, EPT 67; Broad, CE 38). And the value of your own virtue at other times or that of other people can make a difference in first-person deliberation. If an act now will make you more virtuous in the future or encourage virtue in others, you should count that in the act’s favour and allow it sometimes to outweigh other effects; if it will corrupt someone’s character, you should count that against it. And Joseph’s and Ewing’s goods of promise-keeping and returning benefits, which inhere in acts but do not concern their motive, can always matter in deliberation. Still, Rashdall was right that the good of your own virtuous motivation now cannot. If there are goods simultaneous with an act, consequentialism cannot look only at what will follow the act. It must also consider the present, and in addition may need to consider the past. This will be so if there are goods that involve organic unities across time, like the good of desert. If we believe, with Moore and Ross, that it is good for virtuous people to enjoy pleasure and for the vicious to suffer, we cannot know how good it is to give someone pleasure unless we know how virtuous or vicious she has been. Consequentialism now requires knowledge of the past, and it can in principle require knowledge of all the past, so its most accurate statement says a right act is one that will result in the most value through all time. Moore said right acts ‘will produce the greatest possible amount of good in the Universe’ or best affect the Universe’s ‘value as an organic whole’ (PE 147–9), while Broad said that, as far as consequences go, the right act makes ‘the total state of the universe best’, where on

4

In his later E Moore denied that motives can be relevant to rightness (E 77–80, 97), but he there took an act’s rightness to depend only on consequences that follow it in time; its motive can affect ‘the goodness or badness of the whole state of affairs’ (E 78–9) but nothing deontic. Why he then took this narrower view of consequences is unclear; it may be partly that he wrote E for a broader readership and partly that he began it by discussing hedonistic utilitarianism, which does evaluate mostly by later consequences.

act-consequentialism, pluralist deontology

175

the widest view that involves ‘the intrinsic value of the whole universe, past, present, and future’ (CE 36). Despite defending act-consequentialism, Sidgwick and Moore did not think we should use it in everyday decision-making. For most choices we should set consequentialist reasoning aside and follow a set of easier-to-apply rules such as ‘you ought to keep promises’ and ‘you ought not to kill’. Though these rules are not strictly true, the result of following them will usually be better than if we tried to judge individual acts by their effects. They therefore defended versions of ‘indirect’ or ‘two-level’ consequentialism, which says that instead of applying a consequentialist principle directly to acts we should use it to identify those rules whose acceptance will have the best consequences and for the most part follow them.5 An act’s being in accordance with these rules need not make it right; if it has less than the best consequences it is objectively wrong. But following the rules is the best practical policy. Sidgwick and Moore defended this indirect view by emphasizing the difficulty of consequentialist act-assessment and the high probability that we will get it wrong (ME 460–1; PE 149–70); this is especially likely, Moore added, given our tendency to bias, or to exaggerate benefits to ourselves and minimize harms to others (PE 162). We can more reliably judge the general effects of act-types such as promise-keeping and killing, and acting on those judgements takes less time. Sidgwick also thought motives other than the consequentialist one are more directly pleasing: we will be happier if we have impulses to action other than conscious calculation (ME 344–5, 405–6, 433). And less demanding rules may be more fully obeyed. We may produce better consequences by mostly following a more lenient code than if we set ourselves a higher standard and completely fail to meet it (ME 434). Williams has objected that indirect consequentialism requires an unsustainable division within the self. We are both to believe the right act always has the best consequences and, for example, that we should keep promises as such. But the second belief will have good consequences only if it is not accompanied by a temptation to act on the first belief, which must therefore be absent from our minds. So the consequentialist principle in effect removes itself from the scene.6 This objection is partly met if the two beliefs are present at different times. When making concrete decisions we mostly follow the rules; we either do not think of the consequentialist principle or do not learn the empirical facts needed to apply it. We use the principle mainly when we are selecting the rules or reflecting on past acts. And sometimes the two beliefs can be present simultaneously. We can have a rule about keeping promises but believe that in a particular case the outcome will be better if we break a promise. If we have reason to distrust the second belief, because effects are hard to predict and we tend to be biased, we can suppress it and stick by the rule. The ‘two-level’ terminology comes from Hare, Moral Thinking. Williams, Morality, pp. 105–7; ‘Critique of Utilitarianism,’ pp. 131–5; Ethics and the Limits of Philosophy, pp. 106–9. 5 6

176

british ethical theorists from sidgwick to ewing

This is not unfamiliar; compare having doubts about a female job applicant but setting them aside because you fear they are influenced by sexism. Williams also worried about a social division of moral beliefs, especially given Sidgwick’s suggestion that ‘it may be right to teach openly to one set of persons what it would be wrong to teach to others’, so the consequentialist principle and the exceptions it sometimes allows to the rules are known only to a minority (ME 489). The ‘Government House utilitarianism’7 he envisaged here does seem objectionable; it involves a utilitarian elite intentionally inducing false moral beliefs in the majority and suppressing the true ones, so there is both inequality and manipulation. But no class division is needed if each person can keep the two sides of an indirect view in his mind. And Williams’s objection now appeals to intuitions that are nonconsequentialist or at least non-utilitarian, so his critique is no longer internal to a view like Sidgwick’s but external. Though it may lead us to reject utilitarianism, it does not show the view to fail by its own lights. Though indirect consequentialism can recommend following the existing common-sense rules, it need not do so. It can, first, recommend revising common sense. It can say some current rules are such that following them has the best consequences but others are not. It is then reformist about the rules. It can also say that, though you should often follow the rules, when you have strong evidence that a different act will have better consequences, you should do that act: the reasons to distrust direct consequentialist assessments can sometimes be outweighed. An indirect view is more radical the more it recommends changes to the rules or departures from them; it is more conservative the less it does so. Moore’s view was very conservative. He thought philosophy should not propose moral reforms because ‘it seems doubtful whether Ethics can establish the utility of any rules other than those generally practised’ (PE 161). He also thought we should never make exceptions to the rules: our limited knowledge and biases mean we will always more probably maximize the good if we follow the existing code (PE 162–4). He did think that when no rules apply we should do what we believe will produce the best outcome (PE 164–7), and it has been argued that he thought these cases are frequent, so his view was actually reformist.8 But this reading is hard to square with his extreme conservatism about the rules themselves. Sidgwick’s view was also conservative. He argued that, given how common-sense morality developed, its rules are unlikely to coincide fully with those utilitarianism would recommend (ME 463–7). But a would-be reformer faces people who are already accustomed to an existing code; though not ideal, its rules may be ‘the best that they can be got to obey’ (ME 469, also 480–5). A utilitarian should therefore recommend a form of society ‘varying but little from the actual, with its actually established code of moral rules’, and contemplate that code ‘with reverence and 7 8

Williams, Ethics and the Limits of Philosophy, p. 108. Regan, Bloomsbury’s Prophet, Chapter 8.

act-consequentialism, pluralist deontology

177

wonder as a marvellous product of nature’ (ME 474–6). And though direct consequentialist reasoning is needed when everyday principles are vague or conflict, the times when it is right to violate a clear rule ‘must certainly be rare’ (ME 488). Sidgwick was more sympathetic to exceptions for ‘a class of persons defined by exceptional qualities of intellect, temperament or character’, such as his utilitarian elite. But the larger community should almost always follow established norms (ME 489–90). Sidgwick’s and Moore’s conservatism contrasts with the radicalism of Bentham and Mill, the one primarily about the law, the other about liberty and the status of women.9 Broad applauded this side of Sidgwick: whereas earlier utilitarians had talked as if their theory ‘could and should produce a new ethical heaven and earth’, he ‘pours buckets of cold water on the reforming fires of such Utilitarians’ (FT 157). More liberal critics were less impressed. In a review of Sidgwick’s Elements of Politics David G. Ritchie wrote: He nowhere arrives at any conclusion which would differ very widely from that of the average man of the professional and commercial class of the present day. The method is Bentham’s; but there is none of Bentham’s strong critical antagonism to the institutions of his time . . . If this is Benthamism, it is Benthamism grown tame and sleek.10

And the conservatism often had a questionable basis. Why think the rules that do best by consequentialist standards coincide even roughly with those of common sense? Sidgwick’s and Moore’s arguments for the coincidence were often thin. Sidgwick tried to give a utilitarian justification of the common-sense view that you have a much stronger duty to promote the good of your family members than of strangers by saying our affections for particular people are a powerful source of happiness, a more impartial duty would not be fulfilled, and we can best help people we know best (ME 434–9). Moore likewise said each of us should aim mostly at ‘goods affecting himself and those in whom he has a strong personal interest’, because he is most likely to succeed in securing them (PE 166–7). These are relevant points, but on the other side, if you have a middle-class income your family are already reasonably well off, and if there is diminishing marginal utility of money, as Sidgwick believed (EP 9, 153, 173; PE 110), it is unlikely that giving them more will make them much better off. At the same time, there are many far worse-off people in other countries and even in your city to whose happiness a small financial contribution will make a large difference. Is it really so difficult to benefit these people? And may the size of the benefits to them not outweigh any disadvantages to you of a more 9 Sidgwick rejected Mill’s harm principle on the ground that any act can have effects on other people, and there is no utilitarian reason to discount harms that are ‘constructive or presumptive’ rather than direct (ME 477–8). He supported women’s education at Cambridge, but not full university membership for them. 10 Ritchie, Review of EP, p. 255. Hayward also criticized the conservatism of Sidgwick’s indirect view (Ethical Philosophy of Sidgwick, pp. 252–5).

178 british ethical theorists from sidgwick to ewing demanding rule? As Williams said of Sidgwick’s justifications of existing rules more generally, his argument here seems ‘slightly desultory’ and to make ‘pretty cavalier use of what are supposed to be evident matters of fact’.11 Sidgwick also tended to forget his view’s impartiality about times when advising against reforms to existing rules. Given a present generation’s habits, changing the code may indeed be difficult for them and may cause them some distress. But if more enlightened rules really would have better consequences, are the costs to people today of introducing them not outweighed by the greater benefits to the large number of future generations? It is plausible that in practical contexts act-consequentialism should take the indirect form Sidgwick and Moore recommended, but a serious selection of rules by consequentialist standards may produce ones considerably more radical than any these two envisaged.

8.2 Pluralist Deontology: Consequentialist Overlaps Though originated by Prichard and shared by Carritt and Broad, the school’s deontology was most fully developed by Ross. It contains a plurality of underivative prima facie duties, with none absolutely prior to any other. Ross listed seven such duties (RG 21–2). The first was a duty to keep promises, which he thought implies a duty to tell the truth, since any time you speak you implicitly promise not to lie. He called this the duty of fidelity. Then came a duty to compensate those you have wronged; like that of promise-keeping, this duty of reparation rests on previous acts of yours. The duty of gratitude rests on previous acts of another person and is in a way the reverse of reparation: the latter requires you to benefit someone you have harmed, the former to benefit someone who has benefited you. His next three duties involved promoting the good. The duty of justice requires you to bring about distributions of pleasure that are in accordance with people’s merits; Ross thought these distributions are intrinsically good. Beneficence requires you to promote the virtue, knowledge, and pleasure of other people, which he again thought are good, and self-improvement requires you to promote your own virtue and knowledge. He did not include a duty to promote your own pleasure. Though he went on to argue that there must be such a duty (RG 24–6), he never found it intuitive and eventually abandoned it (FE 272–4). On his final view your only duty concerning pleasure is to promote it in other people. The last duty on his list was non-maleficence, a duty not to injure others that is separate from the duty of beneficence and stronger, or ‘prima facie more binding’.

11 Williams, ‘Point of View of the Universe’, pp. 163–4, citing Sidgwick’s defence of the double standard in sexual morality (ME 451–2). Donagan likewise found Sidgwick’s utilitarian justification of a commonsense claim about promise-keeping ‘perfunctory in the extreme (‘Sidgwick and Whewellian Intuitionism’, p. 458).

act-consequentialism, pluralist deontology

179

Though Ross contrasted his deontology with consequentialism, it shares several features with that view. Some have found this problematic, saying it prevented him from developing a genuine alternative to consequentialism. How far is this true? The duties of justice, beneficence, and self-improvement are instances of a more general duty to promote the good that is a consequentialist element in Ross’s theory. One objection is that admitting this element at all was an error, since the consequentialist duty threatens to swamp all the others or make them mere exceptions to it.12 But there is no basis for this charge. The duty to keep promises does not swamp all others, and the duty to promote the good is no different in status; it too is just one prima facie duty among others. It has been said in Ross’s defence that he understood the duty to promote the good differently than consequentialists. They see an act as right because it produces something separate from itself, the best possible outcome, whereas he, perhaps again following Prichard (MW 140–2), said, ‘That which is right is right not because it is an act, one thing, which will produce another thing, an increase of the general welfare, but because it is itself the producing of an increase in the general welfare’, or itself has the property of promoting the good (RG 47; also KET 21–2). This, it has been argued, assimilates his duty to deontological ones and prevents it from swamping them.13 But Ross was here finding significance in the fact, which Sidgwick rightly saw is trivial, that an act can be described in different ways, some mentioning consequences that others do not. Surely nothing substantive follows from that.14 A different objection says Ross was wrong to state his duty of beneficence using a Moorean concept of intrinsic goodness rather one of what is ‘good for’ people or constitutes their ‘welfare’.15 As I argued in Chapter 1, our school denied that there is a ‘good for’ concept with independent normative significance, and were right to do so; any good worth promoting is ‘good’ period. A more serious charge concerns Ross’s commitment to maximizing, and especially his claim that ‘if we are ever under no special obligation such as that of fidelity to a promisee or of gratitude to a benefactor, we ought to do what will produce the most good’ (RG 39; also FE 252). This implies that if by sacrificing your life you could save two strangers’ lives, you are morally required to do so, and in general must do anything if it will produce even slightly greater benefits for others. This is not the common-sense view. Everyday morality permits you to make the large sacrifice but does not require it, considering it supererogatory or beyond duty. This is because it allows you to care somewhat more about your own good, or prefer somewhat lesser benefits for yourself to greater ones for others; in doing so it grants what have been called ‘agent-favouring’ moral

13 Wiggins, ‘The Right and the Good’, pp. 266–8. Dancy, ‘Wiggins and Ross’, pp. 284–5. A contemporary of Ross’s said his restatement of the consequentialist duty involved a ‘logical artifice’ rather than any point of substance (Metz, ‘Recent Trends’, p. 304; also Shaver, ‘Birth of Deontology’, pp. 131–2). 15 Darwall, ‘Under Moore’s Spell’, pp. 289–90. 12 14

180

british ethical theorists from sidgwick to ewing

permissions.16 Ross’s endorsement of maximizing was therefore a departure from the common sense he said was his touchstone and means his theory avoids only one of the two main objections to consequentialism.17 One says consequentialism permits too much, because it makes acts such as killing right whenever they produce the most good. The other says it demands too much, because it makes acts that are merely supererogatory obligatory. The deontological duties in Ross’s avoid the first objection, but his claim that when no such duties apply you must maximize opens it to the second. Nor was he alone in this. Prichard’s writings do not mention supererogation, while Carritt found it a ‘puzzle’. Considering the case, which is not actually the central one, where you sacrifice your own greater pleasure for the somewhat lesser pleasure of other people, he asked how you can deserve praise for doing what you have no obligation to do and concluded that you cannot (EPT 115–16); Ewing too called sacrifice in this case wrong (‘U’ 107–8). At work here was an assumption shared by many present-day philosophers: that the basic normative factors are all ones that count positively in favour of an act, such as a prima facie duty or reason to do it. Some have tried to reconcile this assumption with the permissions that ground supererogation by saying there are both reasons to promote your own good and reasons to promote the good impartially, the two cannot always be compared determinately, and when they cannot, you may do either what is best for you or what is best for others, even if the benefit to them will be somewhat greater.18 But this approach only generates real rather than apparent moral permissions if there is metaphysical rather than just epistemic indeterminacy about the reasons’ weights, and it is hard to believe the existence of supererogation turns on that abstruse metaethical issue. Surely when we think you may care somewhat more about yourself we are not assuming any view about the determinacy of moral truth; we just think the demand always to maximize is excessive. A better way to avoid the second objection is to stop assuming the ultimate factors are all positive and recognize, alongside underivative prima facie duties, some underivative prima facie permissions, including a prima facie permission to pursue your own good. This permission has to be weighed against the prima facie duty to promote the good impartially, and when the benefit you can give others is much greater, it will be outweighed and you will be required to produce the greater good. But when the benefit to them is only somewhat greater, the permission will outweigh the duty and you will be allowed to prefer the lesser good, though also allowed to produce the greater one. The latter choice may be more admirable, but, given the permission, it is not your duty.19

16 17 18 19

Scheffler, Rejection of Consequentialism, Chapter 3; Slote, Common-Sense Morality, Chapter 1. Kagan, Limits of Morality. Scheffler, Rejection of Consequentialism, pp. 56–67; Parfit, On What Matters, vol. 1, pp. 137–41. Hurka and Shubert, ‘Permissions to Do Less Than the Best’.

act-consequentialism, pluralist deontology

181

The concept of a prima facie permission may not be familiar, but it can be understood by analogy with that of a prima facie duty. You have a prima facie duty to do acts of type F if an act’s being F tends to make it your duty all things considered, and you have a prima facie permission to do acts of type F if its being F tends to make an act all things considered permitted. Just as the prima facie duty to do F is stronger than the prima facie duty to do G in a situation if, taking the two together, what you ought to do is F, so the prima facie permission to do F is stronger than a prima facie duty not to do F if, combining the two, you are permitted to do F. Thus your permission to pursue your own good is stronger than your duty to promote the good impartially whenever you are permitted pursue your lesser good. Adding prima facie permissions would help with another problematic feature of Ross’s theory. His final view was that there is no duty to pursue your own pleasure, but as Ewing pointed out, this implies that if you can give either yourself a vast amount of pleasure or another person a little, say, yourself 1000 units or her one unit, you must give her the one. That act is supported by the prima facie duty to promote others’ pleasure while the alternative is supported by nothing (‘SN’ 20).20 A prima facie permission to pursue your own pleasure avoids this implication without introducing any duty to pursue your pleasure: you can be permitted to give yourself the 1000 but not required to. And a related move is possible. If we admit a duty to promote pleasure impartially, as Ross did not, we can add a second prima facie permission, this time not to pursue your pleasure. This will imply that in some cases of the kind discussed by Carritt and Ewing you may forgo a greater pleasure for yourself in order to give a somewhat lesser one to another, or have an ‘agentsacrificing’ permission to do so; it will also imply that you have no duty to seek just modest increases in your pleasure. But if this second permission can be outweighed, as is plausible, you are not permitted to prefer another’s one unit to your 1000; that would be excessive. And you do have a duty to seek large increases in your pleasure. The resulting view is more moderate than Ross’s, agreeing that in some cases you do not have a duty to pursue your own pleasure but holding that in others, where the amount of your good at stake is greater, you do. Though adding underivative permissions is consistent with Prichard’s, Carritt’s, and Ross’s general framework, it is not something they did; like most present-day theorists, they took the ultimate normative factors all to be positive. But if a theory is persuasively to avoid what many see as the excessive demands of consequentialism, it needs such permissions. There is another point where our deontologists were close to consequentialism. In an early paper Prichard said, ‘unless the effect of an action were in some way good, there would be no obligation to produce it’, adding that any moral principle must mention ‘(a) a good thing which the action will produce, (b) a definite relation

20

Also Stocker, ‘Agent and Other’, p. 208.

182

british ethical theorists from sidgwick to ewing

in which the agent stands either to another or to himself ’ (MW 2, 4). In ‘Mistake’ he weakened this claim, saying only that ‘it may be urged’ that right acts must produce something good (MW 10), and he seems later to have abandoned it, since he then held that only virtue is good but we ought to promote others’ pleasure. Claims like Prichard’s were also made by Carritt and Ross. In The Theory of Morals Carritt said, ‘every right act must afford some satisfaction to somebody’ (TM 41, also 42, 54, 68, 85–6, 139); since he equated a person’s ‘good’ with her satisfaction, this meant every right act must be good for someone, and it led him to doubt whether a promise to a dying person that will not satisfy anyone else, such as a promise to bury him in a particular spot, is binding (TM 41–2). A year before The Right and the Good Ross said our duties ‘are all instances of the production of something that is considered good by the agent, and that is at the same time not his own pleasure’ (‘NM’ 267–8), and he repeated this claim in the book: when we think we ought to keep a promise, we think doing so will produce something good for the promisee, and likewise for other duties (RG 162–3). The claim is absent from Carritt’s and his later writings, so like Prichard they may have abandoned it. But all three held at some time that a necessary condition for an act’s being right is that it produce some good. Combining this claim with Ross’s charge that consequentialism’s main failing is to ignore the many relations you can stand in to other people yields a distinctive general account of deontology. It recognizes an initial duty to promote the good, but says your relations to particular people can make their good more important from your point of view than other people’s, or channel your duty of good-promotion more toward them than toward others; that is why some acts that maximize the good are wrong. As Ross put it, whereas consequentialism says we should ‘produce the maximum good, irrespective of the question of who is to possess it’, on his view ‘it makes a great difference who is to possess the good’, so your duty is stronger if you have made someone a promise or been benefited by him (‘BO’ 125). The resulting deontology has important overlaps with consequentialism. Though it does not hold that right acts always maximize the good, it shares the weaker claim that they all promote the good to some degree. It also joins consequentialism in making the good in an important sense prior to the right. Though we can know what is good without knowing what is right, we cannot know what is right without first knowing what is good, since to be right it must promote some good. This account of deontology fits some of Ross’s prima facie duties, especially reparation and gratitude, but it does not fit others. It does not fit the duty to keep promises if a promise to produce a result with neutral value, such as burial in a particular place, is to some degree binding. Nor, more importantly, does it fit nonmaleficence. If it is wrong to push one stranger in front of a trolley in order to save the lives of five others, what makes it wrong is not some special relation between the one and you; he is no less a stranger than the five. Nor can we say that pushing him will create the relation, because that is supposed to exist before the act and make it

act-consequentialism, pluralist deontology

183

wrong. Though non-maleficence is a key non-consequentialist duty, it is owed to all people equally; its non-consequentialism therefore cannot rest on some special relation. Nor need acts of non-maleficence promote any good. (They avoid causing an evil, but that is different.) That is because, unlike Ross’s other prima facie duties, non-maleficence is a negative duty, telling us not to act in a certain way rather than to do something positive; it forbids rather than requires. So on that more general ground too it does not fit his account. It is striking that in both places where Ross said right acts must promote some good he illustrated his claim by listing the six other prima facie duties but did not mention non-maleficence, perhaps recognizing that it does not fit. I said earlier that Prichard, Carritt, and Ross emphasized non-maleficence less than we might expect, and one reason may be that it did not fit a general account of deontology they at least initially held; alternatively, their downplaying nonmaleficence may have made it easier for them to accept that account. Either way, a duty that is central to present-day deontologies fits poorly in a general framework they proposed, and does so because of an overlap between it and consequentialism.

8.3 Pluralist Deontology: Conflicts of Duty A key feature of Ross’s prima facie duties is that when they are outweighed they can leave what I earlier called traces. One trace is a duty to compensate the person to whom the duty was owed; another is the appropriateness of feeling ‘compunction’ for infringing the duty (RG 28; also Ewing, DG 194–5). You should not feel guilt, because you did not do anything all things considered wrong. But you should not feel just regret that you were in a situation where you could not do everything you ought. You should feel badly about your choice, though as failing to fulfil an other-things-equal rather than an all-things-considered duty. Though Ross allowed conflicts between prima facie duties he did not consider the possibility of your having a duty proper to do a given act and a duty proper not to do it, so no matter what you do you do something wrong. Williams has argued that such ‘tragic moral conflicts’ occur and are typified by Agamemnon’s dilemma in the play of Aeschylus, where Agamemnon ought all things considered to sacrifice his daughter and ought all things considered not to.21 Ross may have thought the combination of a duty proper to do A and a duty proper to do not-A is contradictory, but it is not. The contradictory of the duty to do A is the claim that there is no duty to do A, and that does not follow just from the duty to do not-A. The combination of the two duties proper would violate the ‘ “ought” implies “can” ’ principle if it implied a duty to do (A and not-A), but the ‘agglomeration principle’ that underwrites that inference can be rejected. Then each of the two duties is consistent with ‘ “ought”

21

Williams, ‘Ethical Consistency’; also Nussbaum, Fragility of Goodness, Chapters 2–3.

184

british ethical theorists from sidgwick to ewing

implies “can” ’, and though their conjunction is not, it can be denied. While Ross did not contemplate tragic conflicts, there is no reason in principle why his theory could not include them. Williams also argued, more strongly, that such conflicts must be included if we are to make sense of our feelings of regret after a choice. Writing specifically of Ross, he said a prima facie duty ‘must surely be a claim for consideration as the only thing that matters, a duty, and if a course of action has failed to make good this claim in a situation of conflict, how can it maintain in that situation some residual influence on my moral thought?’22 But this was just to assert that an outweighed prima facie duty cannot leave traces, and why believe that? An outweighed physical force has effects, and Williams’s own view of the issue faces a dilemma. It can say tragic conflicts are ubiquitous, but then it implies that in Ross’s example of the accident victims you should feel not just compunction about breaking your promise but full-blown guilt at having done something all things considered wrong; that is not plausible. Alternatively, it can say tragic conflicts are rare, found only in unusual situations like Agamemnon’s; but then you need have no bad feeling about your choice in Ross’s example. Ross’s intermediate view has the merit of calling for some negative feeling in many conflict cases but not for all-out guilt. If duties leave traces, there is a difference between outweighing cases and ones where a duty that might otherwise apply does not because it contains an exception. Ross recognized one case of the second kind when he said a promise with tacit conditions imposes no duty if the conditions are not satisfied (FE 94–9); in his discussion of punishment he also said that someone who violates another’s life or liberty loses his right to have his own life or liberty respected, so the state does no prima facie wrong when it punishes him (RG 60–2). The same is true in cases of selfdefence. The duty of non-maleficence is not engaged if you kill an attacker because that is the only way to stop him from killing you; that is why you may kill ten or a hundred attackers if that is necessary to save your life and why, if you do, you need feel no compunction nor pay any compensation afterward. Ross did not discuss selfdefence, and in general said less about exception-cases than about ones where a prima facie duty is outweighed. But the former are just as important. He seems also to have assumed that all outweighed duties leave traces, and that can be questioned. Imagine that in the case where the only way to save five people from being run over by a trolley is to push another in front of it you decide not to. I do not think either compunction about not saving the five or any compensation is called for; you just did the right thing. We could say the duty of beneficence has an exception-clause for cases where fulfilling it will violate a stronger duty, but that was apparently not Ross’s view; he seems to have held that even when beneficence

22

Williams, ‘Ethical Consistency’, p. 176.

act-consequentialism, pluralist deontology

185

yields to non-maleficence it is still present. Given that view, only some duties seem to leave traces, and a theory like his must specify which they are. There are other points where his view may have been too simple. His duty of beneficence had a simple consequentialist form, so what conflicts with deontological duties and sometimes outweighs them is just an act’s total good effects; thus it is when the total good achieved by killing an innocent person is large enough that doing so is permitted. But some deontologists have proposed more complex views. One counts the total good an act will cause but puts a constraint on what can contribute to it: to count toward the relevant total an effect must be above a threshold of goodness that is higher the more serious the duty that will be infringed. This can imply that there is no number of mild headaches preventing which could justify killing an innocent person, though preventing some number of more serious injuries could.23 A more radical view does not consider any total but, rejecting aggregation across persons, decides conflict cases by pairwise comparisons between the goods of different individuals.24 Carritt suggested something like the first view when he said, ‘we feel a stronger obligation to benefit or improve one person greatly than to distribute any number of minute satisfactions among different people’, so ‘we should be more bound to cure one man of blindness than to save any number the bother of spectacles’ (EPT 98).25 But Ross never considered this idea and seems to have allowed headaches in principle to justify killing. If so, this may be another point where his theory was too close to consequentialism. He seems also to have assumed that the weighing of prima facie duties is always simply additive. For any act there are some prima facie duties tending to make it right and some tending to make it wrong; each duty has a weight that is independent of any other duties; and to determine the act’s overall deontic status we add the rightmaking weights together, do the same for the wrong-making ones, and see which total is greater. But this ignores the possibility of interaction effects among duties, where, for example, the right-making tendency of two prima facie duties together is greater than the sum of those they would have on their own. Ross should have recognized this possibility, since it applies in a deontic context the principle of organic unities, which he accepted and used to state his views about desert (RG 69–73). And there are cases where it may be deontically relevant. In the most compelling counterexamples to consequentialism, of punishing an innocent person or throwing him in front of a trolley, two deontological distinctions cut in the same direction. The intuitively wrong act both directly causes an evil rather than allowing it and intends the evil rather than merely foreseeing it. But in some cases where only one of these distinctions applies, for example when you turn the trolley away from five and toward one, it has less or no weight. Could the prima facie wrongness of an 24 Brennan, ‘Thresholds for Rights’. Scanlon, What We Owe, pp. 229–41. Carritt cited Ross’s FE 69–72 as agreeing with him, but there Ross defended just the average rather than the total good in a population as what we should maximize. 23 25

186

british ethical theorists from sidgwick to ewing

act that both directly causes and intends an evil be greater than the sum of the wrongnesses it would have if it did only one?26 Whether there are such interaction effects must be decided case by case, but if there are, they can again be described in two ways. We can say the doing/allowing and intending/foreseeing distinctions have constant weights but an additional factor of doing-and-intending is introduced when the two go together, or that the weight of one or both is increased. Either way, accepting the effects again makes deontology more complex than Ross, despite his non-consequentialism, imagined it needs to be.

8.4 Pluralist Deontology: Elaborations Ross listed just a few prima facie duties and stated them mostly in the vague way Sidgwick found unsatisfactory. His theory can be elaborated in three ways: by specifying some duties more precisely, by giving further explanations for some, and by adding extra prima facie duties to capture elements of common-sense morality he missed. The duty Ross described most fully was promise-keeping, saying its boundaries are set by conditions attached to a promise when it was made and accepted, even if tacitly, by both parties. This is not the only possible specification. Earlier I mentioned Whewell’s view that the boundaries of a promise depend only on the promissor’s beliefs; it yields a different verdict if he believes falsely that the promisee accepted a condition. (This view cannot change what is subjectively right for the promissor to do, but can change what is objectively right.) Ross would need to defend his specification against this and other alternatives. Another issue, raised by Sidgwick (ME 443), concerns promises extracted by force or fraud. Ross said a promise based on fraud is not binding, because the promisee broke his implicit promise to tell the truth and so destroyed the basis on which the promise was made (FE 97). He could likewise have said a promisee who uses or threatens force to extract a promise forfeits any right he might acquire or, alternatively, that a verbal promise is binding only when it is made freely. Ross took the strength of the duty to keep a promise to depend on two factors. One is ‘the value of the promised service in the eyes of the promisee’; the other concerns how and when the promise was made, so one ‘made casually in a moment of halfattention is less binding than one made explicitly and repeatedly’ and a recent promise more binding than one made long ago (FE 100–1). We may question whether the value of what is promised really is relevant. If you have a stronger duty to keep a promise to give a larger benefit, is this not because doing so will have better consequences, so the added strength comes from beneficence rather than promise-keeping? Ross said that when you promise a benefit you do not have two The possibility of deontic interaction effects is defended in Kagan, ‘Additive Fallacy’; Kamm, Morality/Morality, vol. 2, pp. 51–7. 26

act-consequentialism, pluralist deontology

187

duties but ‘a single responsibility’, to give the promisee what you promised, and its strength depends in part on what that is. But if beneficence and fidelity are distinct duties, why can they not both be present here? We may also question the factor of recency. Can repeatedly failing to keep a promise make the duty it imposes weaker? Ross thought a promisee will more confidently expect a promise to be fulfilled if it is more recent, but that again seems to bear on the effects of keeping the promise rather than on the duty to do so as such. This leaves the manner in which the promise was made, which is relevant. There are conventional devices, such as swearing on a Bible, that make a promise more serious and strengthen the resulting duty. We tend to use these devices more when our promises concern greater goods, so these promises are in two ways more urgent to keep. Ross grounded the duty of reparation on ‘a previous wrongful act’ (RG 21), but he cannot have meant ‘wrong all things considered’ since he thought you can owe compensation when a prima facie duty is outweighed. That aside, reparation seems not to be a fully independent duty, since it is engaged only when some other duty has been infringed and is arguably just a trace of that duty.27 Nor, as we saw above, is it engaged by all other duties. Only promise-keeping and non-maleficence and not, say, beneficence, seem to trigger a duty to compensate, as Ross may have recognized when he said reparation is owed for ‘injuries done by oneself to another’ (RG 26). This duty raises many issues of detail. Does the compensation you owe depend on the amount of harm you actually caused or on the possibly smaller amount you could reasonably have foreseen would result? If your act contributed to harm only given a later wrongful choice by someone else, does that reduce what you owe? The law of torts addresses issues like these; Ross made only the general claim that you owe some compensation for harm. Similar issues arise about gratitude. Ross would have denied that you owe a benefactor a feeling of gratitude; your duty is only to give him some benefit. But what determines the size of this benefit? Sidgwick mentioned two factors, the size of the benefit the other gave you and the amount of sacrifice it cost him (ME 261). But you may also owe more gratitude if your benefactor acted from a virtuous motive rather than just from hope of a reward, or when his act was supererogatory rather than a duty. A complete account must identify all the factors relevant to gratitude and how they weigh against each other. Also needing specification is the duty of non-maleficence. Ross and the other deontologists seem to have thought the difference between harming and failing to benefit is straightforward and rests on the distinction between actively causing and merely allowing a bad effect. They were unaware of cases where actively causing harm does not seem morally wrong, such as diverting a trolley away from five people 27

Should there not then also be a prima facie duty to feel compunction after infringing a prima facie duty? Ross would have denied that there can be a duty to have a mental state.

188

british ethical theorists from sidgwick to ewing

toward one or giving a terminally ill patient a pain-killing drug that as a side effect will hasten his death. Nor did they consider cases where merely allowing harm seems wrong, such as intentionally letting someone die so you can use his organs for transplants. These cases are better handled by the doctrine of double effect, which says it is worse to intend a bad consequence than merely to foresee that it will result. This doctrine contradicts Ross’s view that rightness never depends on mental states, but we saw earlier that his arguments for that view do not touch a thesis that forbids rather than requires certain intentions. And our school seem simply not to have heard of double effect. Sidgwick explicitly collapsed the distinction between intending and foreseeing (ME 202),28 and neither he nor later members mentioned it when discussing the duty not to harm. This is presumably in part because double effect is a Catholic doctrine and their backgrounds were Protestant. But a full specification of non-maleficence must say exactly what distinguishes the acts it forbids from those it finds less objectionable, and in doing so must consider more than just the one distinction between doing and allowing. Ross grounded the duty not to lie in ‘the implicit undertaking not to tell lies which seems to be implied in the act of entering into conversation’ (RG 21; also FE 112–30). This nicely explains why lying is other things equal more seriously wrong than deception that does not involve lying, such as intentionally leaving misleading evidence for someone (think of Iago and Othello) or making true but incomplete statements in the hope of inducing him to draw a false conclusion: to the prima facie wrong in deception it adds the further wrong of promise-breaking. But Ross’s view can be refined. He seems to have taken the promise implied in conversation to be a general one made to all listeners, but you lie to a particular person and do so through the speech-act of assertion. When you assert a proposition you intend to get your audience to accept it on the basis of your asserting it, and do so by tacitly assuring them that you believe the proposition and do so on the basis of evidence. A lie violates that assurance, but the assurance was given only to the person you were addressing. If you assert an untruth to A with the aim of getting B, who is eavesdropping, to accept it, you have deceived B but not lied to him; you lied only to A.29 Ross could also have given some of his duties a further explanation. One way is by adopting a general account of deontology that also overlaps partly with consequentialism but does so in a different way. A variant of the neo-Thomist deontology defended by John Finnis,30 it starts by identifying some states of affairs such as pleasure and knowledge as intrinsically good and their contraries as evil. It then says that for each pair of a good and evil there are two prima facie duties. One is a positive

28

Anscombe rebuked him for this (‘Modern Moral Philosophy’, p. 11). This account is taken from Fried, Right and Wrong, Chapter 3. 30 Finnis, Natural Law and Natural Rights, Chapter 5; Fundamentals of Ethics, Chapter 5. Finnis derives the second principle from a claim about the incommensurability of fundamental goods; the variant I propose makes no such claim and treats the two principles as underivative. 29

act-consequentialism, pluralist deontology

189

duty to promote the good and prevent the evil; the other, which is separate and stronger, is a negative duty not to destroy the good or directly cause the evil. This second duty can be explicated using the doing/allowing, double effect, or some other distinction, but whichever the account uses, it accommodates the duty of nonmaleficence as Ross’s does not. Throwing a stranger in front of a trolley is wrong not because he stands in some special relation to you but because it violates the stronger duty not to cause the evil of death rather than the weaker duty to prevent it. And the stronger duty is explicitly negative, unlike the positive ones that were all some of Ross’s statements allowed. This account also unifies some duties more clearly than Ross did. He initially listed beneficence and non-maleficence as separate duties, with no closer ties to each other than, say, beneficence and promise-keeping. But both concern the causing of good or evil, as the neo-Thomist view recognizes by seeing them as mandating different responses to the same good or bad effects.31 It also accommodates some duties Ross did not. He said little about deception that does not involve lying and may have thought it falls under non-maleficence: if knowledge is good, to deceive someone is to harm him by putting him in the bad state of false belief (RG 55). But the neo-Thomist view yields this result explicitly, and also yields some others. If just distributions are good, it says there is both a duty to promote them and a stronger duty not to create unjust distributions. This means that even if creating a small injustice now will prevent greater injustices in the future, doing so can be wrong. Ross did not say this. His duty of justice was only a positive duty to promote just distributions, and his duty of non-maleficence forbade only harm to individual persons; it follows that any act producing a net increase in justice was by his lights right. By extending its negative duty to include all direct causing of evil, the neo-Thomist view can affirm a deontological duty of justice, which many will find intuitive, and it can affirm similar duties about the self. Ross’s only self-regarding duty was the positive one of self-improvement, but some may think it is more seriously wrong to cause false beliefs in yourself, for example by deliberate self-deception, than to fail to acquire true ones, and more seriously wrong intentionally to corrupt your character than to fail to improve it. The neo-Thomist approach can explain these views. More generally, by condemning all direct causation of evil it can affirm more deontological duties than Ross did and give them a unifying rationale. Ross could also have further explained the duty to keep promises, in a way suggested by Hart and using the concept of a natural right.32 Hart argued that before you make any promises you have the right to make certain choices, for example about who you will have lunch with tomorrow, where this right has three components. The first is a moral permission to choose with whom you will have lunch; you have no duties in this area. Next is a duty other people have not to interfere with your choice,

31

Ross did briefly connect them on this basis at RG 26.

32

Hart, ‘Are There Any Natural Rights?’

190

british ethical theorists from sidgwick to ewing

either by forcing you to have lunch with them or by preventing you from lunching with someone else. Last is a further permission you have to resist people who try to interfere with your choice, or to enforce their duty to respect it. Most importantly, along with this right comes a moral power to give it up or transfer it to another person, and that is what you do when you make a promise. If you promise to have lunch with me, it is now I who am permitted to choose that you lunch with me, you who ought not to interfere with my choice by refusing to lunch, and I who may enforce your duty; the moral factors that used to be yours are now mine and vice versa. On this account the duty to keep a promise is not completely underivative, as Ross thought, but follows from the exercise of a moral power to transfer a right, or bundle of permissions and duties, from yourself to someone else. This account has several merits. It explains why the duty to keep a promise is enforceable as, say, a duty of beneficence normally is not: it results from the transfer of a right, and rights are in general enforceable. It also explains why a promisee can release you from your duty: you transferred a two-sided permission, either to lunch together or not, which he can exercise in either way. And it explains why a promise to do something seriously immoral, such as committing murder, imposes no prima facie duty whatever. This is hard to explain on Ross’s view. ‘I promise to murder A’ looks like a standard promise, and though the duty it generates may be outweighed by the duty not to murder, it seems it should still exist. But if promising involves transferring a right, you cannot transfer what you do not have, and you have no right to murder. Hart’s account uses the concept of a right, which Ross set aside as a confused way of talking about duties (RG 53–4). But while its claims concern permissions and duties, they concern a distinctive set of them, including in particular a power to transfer them to another. Its claims can, moreover, be underivatively true. If we ask why you have the permission to choose with whom you will lunch or have the power to transfer that permission, there may be no answer: you just do. The account could therefore be added to Ross’s theory without changing the latter’s basic character. It would be just another place where a duty Ross thought completely underivative can be grounded in another, more truly fundamental, claim. His theory can also be supplemented with further prima facie elements. Two are the prima facie permissions to promote your own good and not to promote it discussed above, but others are prima facie duties. Some of these involve the concept of rights just used to explain the duty to keep promises. Ross’s list contains no duty concerned with freedom of choice, or making it wrong to interfere with another’s freedom. He could therefore have no objection to paternalistic acts, where you force someone to do something now because it will benefit him in the future, for example by making him happier or more virtuous. These acts are approved by his duty of beneficence, since they promote another’s good, and are not opposed by any other duty. Yet many find something at least prima facie wrong in paternalistic interference.

act-consequentialism, pluralist deontology

191

One way to capture this wrong is to add free or autonomous choice to the list of intrinsic goods and use the neo-Thomist view to generate two duties: a weaker one to promote free choice and a stronger one not to destroy or reduce it. This has initially attractive implications: we ought to increase others’ range of options, for example by giving them education and resources, but have a stronger duty not to remove options from them, by force or threats. But it does not capture a key point discussed above: that removing another’s option is not even prima facie wrong if he transferred his right to choose it to you, by making a promise. The better way of condemning interferences with freedom is to posit a general right to freedom of the kind Hart proposed, with the various permissions and duties that involves. Its elements will all be prima facie, so there is a prima facie duty not to interfere with another’s choice that must be weighed against other duties. If a paternalistic act will give her a small benefit, non-interference outweighs beneficence and the act is on balance wrong. But if the benefit is large, as when restricting her freedom will save her life, paternalism can be right. A right to freedom also explains why an act that would otherwise be wrong can be permitted if a person consents to it. If she has the right to choose whether or not she is punched in the arm (assuming someone willing to punch her), then if she asks you to punch her and you do, you do not interfere with her freedom but help her to exercise it. Deception can likewise be permitted if it is consented to, as in a poker game or in bargaining. (Everyone knows ‘This is my final offer’ need not be true.) A permission due to consent is again just prima facie, so if punching someone who asks to be punched will injure her seriously, it is forbidden. A right to freedom also connects with issues about self-defence. If all rights are enforceable, and to a degree proportioned to their importance, then anyone with the right to choose to continue living may, if this is the only way to save his life, kill an attacker who is trying to kill him. Here the self-defence exception to the duty of nonmaleficence follows from a general feature of rights, their enforceability, as does the similar exception for punishment. This explanation needs to show why the various elements of a right go together, and in particular why the initial permission and duty must be joined by a further permission to enforce. But it can be argued that the general idea of a right to freedom requires that permission.33 A final implication concerns duties about property. Ross seems to have included these under non-maleficence, since one act he took that duty to forbid is theft (RG 22). But this treatment is inadequate to cases where wrongful use of another’s property does not harm him in the ordinary sense, as when you occupy his house without his consent when he is away but put everything back as it was before you leave. Nor does it capture the fact that ownership includes a power to transfer your property to others, for example by selling it. The details of a right to property are

33

I gave one argument for this claim in ‘Rights and Capital Punishment’.

192

british ethical theorists from sidgwick to ewing

controversial, but if there are underivative duties concerning property, they fit better in a rights framework than in Ross’s simpler one of individual prima facie duties. Duties concerning rights are often called ones of justice, as is that of nonmaleficence. Ross’s duty of justice concerned distributions that proportion happiness to merit, and I have suggested supplementing his positive duty to promote these distributions with a stronger duty not to cause unjust ones. A further addition may be needed. Ross did not discuss discrimination by race, gender, or religion, and would presumably have said that giving one racial group more happiness than another is unjust because it violates proportionality. But he would not have found this distribution more unjust than one that is disproportionate to the same degree but was arrived at randomly, with no differential treatment of races. If we think the first distribution is worse, we should make race-based distribution an additional evil, which there is an additional duty to avoid, and likewise for other categories. Intentionally disadvantaging a race, gender, or religious group will then be more unjust than doing the same to a random collection. A further addition concerns the idea that we have stronger duties to people who are close to us, such as our spouses, children, and friends. Applied to beneficence this yields Broad’s self-referential altruism, which says you should do more to promote a person’s happiness the closer she is to you (CE 279–80). But the same idea can be extended to deontological duties. However objectionable it is to lie or break a promise to a stranger, it seems worse to do so to a spouse or friend. Her relation to you makes those acts more seriously wrong, so she can ask not only ‘How could you do that?’ but ‘How could you do that to me?’ Ross accepted self-referential duties, since among the relations he said consequentialism ignores are those ‘of wife to husband, of child to parent, of friend to friend, of fellow countryman to fellow countryman’ (RG 19). But he did not make them a separate item on his list because he thought they derive from other, more basic duties; thus your special duty to your parents and friends ‘in the main depends’ on that of gratitude (FE 76). As Ewing argued, however, gratitude cannot explain all duties of this type, for example the special duty of parents to their young children (E 78–9; ST 138–9). It also does not explain our stronger deontological duties to intimates or have the right form to explain self-referential duties. A duty of gratitude is usually circumscribed: someone has given you a specific benefit, you owe him something specific in return, and once you have given him that your duty is fully discharged. But self-referential duties are open-ended, calling for a continuing pattern of greater concern for another’s good and greater respect for her rights. Later Ross suggested that parents have a special duty concerning their children because of ‘a promise, or an implicit promise’ they made, presumably to care for them (FE 273). But to whom was this promise made, and what is the evidence that it was implied? Carritt rested parents’ duties, differently, ‘on the belief that life is a doubtful benefit’, saying ‘we have put our children in an awkward situation which we ought to help them deal with’; he rejected the idea that children owe their parents for

act-consequentialism, pluralist deontology

193

‘the gratuity of birth’ as excessively ‘optimistic’ (EPT 105–6). But if a child’s life is on balance good, as we can hope it often is, this justification does not apply. Ewing, finally, grounded parental duties in the related idea that one ‘has a special responsibility to see that what he has voluntarily brought into existence fares satisfactorily’, which also applies to institutions and policies one has created (ST 139). But this idea does not fit our duties to siblings or co-nationals and in any case seems too weak to capture the very strong duties parents have to their children. A more promising approach abandons these attempts to derive self-referential duties from something else and supplements Ross’s list with the claim that all your duties are stronger when the person they concern is closer to you; though not an additional prima facie duty, this is an across-the-board strengthener of duties and, as such, underivative.34 The closeness it refers to can be understood in different ways, but one rests it on a shared history that involved causal interaction between you and another person and benefited either the two of you or outsiders (think of parents raising a child). Then your duty concerning another is stronger when you and she have interacted more and when your doing so has done more good. Members of your family score highly on both these dimensions, since you have lived with and presumably shared significant benefits with them. Friends usually score lower, because your interactions with them have been less intimate, and distant strangers do not score at all. This gives you greatly strengthened duties to your family, weaker ones to friends, and only generic prima facie duties to strangers.35 These are some possible additions to Ross’s prima facie duties, and there may be others. Though a pluralist about the right, he shared some of Sidgwick’s preference for a smaller number of ultimate factors, as reflected in his deriving the duty not to lie from promise-keeping and self-referential altruism from gratitude. But it is consistent with his general approach to supplement his list if there are moral concerns his duties do not capture, as I have argued. His pluralistic deontology has the right form to capture our everyday moral thought, but needs more than just his seven duties to capture all its content.

34 It may be argued that the duty of gratitude is not strengthened by closeness; if anything, you owe less gratitude for a benefit from an intimate. But given self-referential altruism, an intimate has some duty to give benefits of that kind, and duty is arguably a factor that reduces the gratitude you owe. 35 I elaborate this account of self-referential duties in ‘Justification of National Partiality’, pp. 148–55.

9 Non-Moral Goods Our school also discussed questions about the good, or about which states of affairs have intrinsic value. These questions are vital for consequentialism, whose duty to maximize the good has no content until they are settled, but they also matter for a deontology like Ross’s, which likewise contains a duty to promote the good. Its connections to the right aside, the theory of the good matters in itself. Even when states like pain, knowledge, and distribution by desert come about independently of anyone’s will, so no one acted rightly or wrongly in producing them, we can ask whether they are good or evil and, if so, how good or evil they are. Whereas the school were evenly split between consequentialists and deontologists, only Sidgwick was a hedonist about the good while the others defended pluralist theories valuing perfectionist states such as knowledge and virtue. Pleasure was, however, a generally accepted good and is our first topic.

9.1 What Is Pleasure? On the simplest, or ‘internalist’ view, pleasures are mental states with an introspectible quality of pleasantness that cannot be defined or described but must be experienced directly. Moore took this view when he said pleasure is ‘a certain definite feeling’ and ‘absolutely indefinable, some one thing that is the same in all the various degrees and in all the various kinds of it that there may be’ (PE 12–13, also 78, 80). So did Rashdall and McTaggart (TGE II 11; SHC 108–9; NE II 168, 480) and, arguably, Prichard, Carritt, and Ross. Broad gave it an especially sophisticated statement; he said ‘there is a quality, which we cannot define but are perfectly well acquainted with, which may be called “Hedonic Tone”’, and which has ‘the two determinate forms of Pleasantness and Unpleasantness’. It always attaches to other mental states and can belong ‘both to Feelings and to those Cognitions which are also Emotions or Conations’, so a pleasure is ‘any kind of mental event which has the pleasant form of hedonic tone’. A sensation, say of taste, that has pleasant tone is a pleasure, as is a thought, for example of an imagined triumph. It follows that ‘There is not a special kind of mental events, called “pleasures and pains”; and to think that there is is as if one should solemnly divide human beings into men, women, and blondes’ (FT 228–30).1 1 For later statements of the hedonic-tone view see Duncker, ‘On Pleasure, Emotion, and Striving’; Smuts, ‘Feels Good Theory’.

non-moral goods

195

A ‘heterogeneity’ objection to internalism says pleasures are too varied for there to be any one quality they all share.2 But several members said pleasantness cannot be experienced on its own. It is felt only in conjunction with other mental qualities, and its instances therefore differ introspectibly because of what accompanies them (Rashdall, TGE I 26, II 8, 11; McTaggart, SHC 108–9; Ross, RG 145). Broad’s view in particular underwrites this claim, since if pleasantness is a ‘tone’ that attaches to other mental states it cannot be found apart from such states. And he thought this tone itself is affected by a state’s non-hedonic qualities. Headaches and toothaches are both pains, but the one has a specific introspectible quality of ‘headachiness’ and the other of ‘toothachiness’; as Ross put it, ‘Pleasant feeling is coloured through and through by what we are taking pleasure in’. An analogy is with the loudness of sounds. Loudness cannot be experienced on its own, apart from a certain pitch and timbre, and loudness in a trumpet therefore differs qualitatively from loudness in a bass drum. But loudness is still a quality of sounds and one we can rank them in, so we can say the trumpet is louder than the drum.3 Broad recognized both sensory pleasures, which however complex introspectibly do not have an intentional object, and intentional pleasures that something is the case, which do. In the former, hedonic tone attaches to a sensation like the taste of chocolate; in the latter to a thought such as one of a triumph. But this account may not quite fit intentional pleasures. A thought of a triumph with positive hedonic tone is a pleasant thinking of a triumph, but is that quite the same as a pleasure in a triumph? If not, intentional pleasures are sui generis hedonic states. Sidgwick’s view of pleasure was surprisingly unstable. He officially rejected internalism, saying he could find no ‘common quality’ in the great variety of his pleasures (ME 127). He is sometimes taken to have held the externalist view that pleasures are those feelings we want to have given just their qualities as feelings, but in fact he rejected this view, which he attributed to Spencer and Bain. He argued that the degree to which we want a pleasure is not always proportioned to its intensity, citing the sensation of being tickled, which we want very much to avoid though it is not very painful (ME 125–7). For a simpler counterexample, imagine that blue is someone’s favourite colour, so he wants to have sensations of blue. On the externalist view these sensations are pleasures, and they are so even if he has no feelings whatever about them, his mind containing just an awareness of blue and a desire that it continue. This does not seem right; though a sensation of blue may cause pleasure in someone who likes blue, it is itself just a sensation of blue. Sidgwick sometimes identified pleasures as ‘desirable feelings’, where ‘desirable’ means what we ought to desire or what is good (ME 129, 397–8, 402, 404). Bradley and Green took this to be his view and objected that it reduces the hedonist thesis

2 3

Griffin, Well-Being, p. 8; Parfit, Reasons and Persons, p. 493. Kagan, ‘Limits of Well-Being’, p. 172; Kagan attributes the analogy with loudness to Leonard Katz.

196

british ethical theorists from sidgwick to ewing

that pleasure is the good to the tautology ‘good feeling is the good’.4 Sidgwick replied that even as so understood the thesis is not empty, since it implies that the only good things are feelings (ME 129, GSM 129); though that is true, the thesis does not tell us which the good feelings are. Sidgwick also complained that his critics misstated his view. In his most careful statements he said pleasure is ‘a feeling which, when experienced by intelligent beings, is at least implicitly apprehended as desirable or—in cases of comparison—preferable’ (ME 127, also 129, 131), or is the subject of a ‘judgment of value’ based on its qualities as feeling. And the thesis that a feeling someone apprehends as good is good is not a tautology (GSM 130). But this definition too is problematic. It seems to imply that beings without the concept of the desirable, such as most animals, cannot feel pleasure. It still faces the objection about blue: if someone thinks sensations of blue are desirable, that makes them pleasures even if he has no feeling about them. And, contrary to the spirit of Sidgwick’s realism, it makes some judgements of value, given one further assumption, self-validating. Imagine that you judge a certain feeling to be desirable or good. On his definition that makes the feeling a pleasure; and given the hedonist thesis that pleasure is good, it makes the feeling in fact desirable. But surely he did not believe thinking something good can make it so. Moreover, despite his definition he often made internalist-sounding claims. He several times said we should prefer pleasures in proportion to their intensity or pleasantness (ME 121, 127, 399). This is a substantive claim given internalism, but on his official view a pleasure’s intensity just is the degree to which we think we should desire it, which makes his claim tautologous. He also said a feeling’s degree of pleasantness is ‘directly cognisable by the individual who feels it at the time of feeling it’ or is ‘definitely given in experience’ (ME 128, 129, also 398); that too makes sense if pleasantness is an introspectible quality, but not if it depends on an evaluative judgement. And he thought every pleasure has a precise degree of intensity (ME 123–4); on his official view that would require us always to make precise quantitative judgements about pleasures, which we do not.5 Despite his formally giving an externalist account of pleasure, internalism often fits his claims about it best.6 However he understood pleasure, Sidgwick thought its value depends on just two factors, its duration and its intensity. By a pleasure’s intensity he did not mean its intensity as a sensation, since, as he nicely put it, ‘a pleasant feeling may be strong and absorbing, and yet not so pleasant as another that is more subtle and delicate’ (ME 94). And he thought intensity has to be ‘balanced against’ duration or made ‘commensurable’ with it (ME 124, 124n), that is, weighed against it. But the two factors are

Bradley, ‘Mr Sidgwick’s Hedonism’, pp. 79–80, 83, 93; Green, Prolegomena, sec. 366. He also said all pleasures have ‘certain cerebral nerve-processes as their causes’ (ME 178–9), which again makes more sense if pleasures share an introspectible quality than if they involve value-judgements that can be made about different feelings by different people. 6 See further Crisp, ‘Pleasure and Hedonism’, pp. 32–7. 4 5

non-moral goods

197

not added together, since then a long-lasting sensation with no hedonic intensity, or at what he called the ‘hedonistic zero’ (ME 124), would have positive value, which it does not. Instead they are multiplied, which eliminates the need for weighing: doubling a pleasure’s intensity while its duration stays constant doubles its value, as does doubling its duration when its intensity is constant. Because he valued just these two factors, Sidgwick rejected Mill’s view that hedonism can recognize ‘higher pleasures’, such as the pleasure of reading poetry as against an equally intense one of playing pushpin. This view introduces ‘a nonhedonistic ground of preference’ and is ‘a perplexing mixture of Intuitionism and Hedonism’; in a genuinely hedonistic theory ‘all qualitative comparison of pleasures must really resolve itself into quantitative’ (ME 94–5, also 121). Later members shared this view, though usually accepting the non-hedontic values Sidgwick rejected. Rashdall said ‘the admission of qualitative differences in pleasure abandons the hedonistic point of view’ (TGE I 27, also II 31–2), and Moore, Carritt, Ross, and Ewing agreed (PE 77–81; TM 21, 43, EPT 63–4; RG 145; E 42–3). Broad was a little more sympathetic. He thought there might be different determinate forms of pleasantness, as there are different shades of a colour, and though nothing is relevant to the value of an experience except its hedonic quality, ‘two experiences which had exactly the same degree of pleasantness . . . might differ in value because they had this pleasantness in different determinate forms’. It was ‘just conceivable’ that Mill meant this, but Mill ‘was so confused that he probably did not himself know precisely what he meant’ (FT 232).7 Sidgwick thought the goodness of pleasure and evil of pain are symmetrical, so a pleasure of intensity n is exactly as good as a pain of intensity n is bad: his ideal was ‘the greatest possible surplus of pleasure over pain, the pain being balanced against an equal amount of pleasure, so that the two contrasted amounts annihilate each other for purposes of ethical calculation’ (ME 413). This claim again seems to assume an internalist view of pleasure. On his official view the symmetry claim says a pleasure that is good to degree n is exactly as good as a pain that is evil to degree n is evil, which tells us nothing about which those pleasures and pains are. To have content his claim must understand intensity in some other, probably internalist, way.8 Most utilitarians value pleasure and pain symmetrically, but some in our school took a different view. Moore thought pain is ‘a great evil’ but said ‘pleasure, however intense, does not, by itself, appear to be a great good, even if it has some slight intrinsic value’; pain is therefore ‘a far worse evil than pleasure is a good’ (PE 212, also 91, 94–6, 213–14). He took this asymmetry to be one thing a mistaken desire for ‘system’ or ‘symmetry’ can lead one to ignore (PE 222), and in an earlier paper

7 Bradley and Green likewise thought Mill’s admission of higher pleasures abandons hedonism (Ethical Studies, pp. 116–20; Prolegomena, secs. 162–8, 238). 8 It may be objected that the intensities of pleasure and pain cannot be compared independently of comparing their values. I try to answer this objection in ‘Asymmetries’, pp. 201–3.

198

british ethical theorists from sidgwick to ewing

suggested a ground for it. Rejecting McTaggart’s early view that ‘a greater quantity of pleasure [is] always also proportionally greater in value’, he suggested that the value of a fixed increase in a pleasure’s intensity gets smaller the more intense the pleasure is (‘MME’ 358). If this value diminishes toward zero, there is an upper limit on the goodness of any pleasure; if the disvalue of increases in pain does not diminish, that makes pain a greater evil. Broad was likewise ‘more inclined to think that pain is intrinsically evil than that pleasure is intrinsically good’, while for Carritt ‘Pain seems more obviously bad than pleasure is good’ (FT 134, EPT 94; also Ross, FE 275). If this view is taken into the theory of the right, it makes the duty to relieve pain stronger than the duty to promote pleasure, and inflicting pain worse than preventing pleasure.9 The good of feeling is sometimes called ‘happiness’ rather than ‘pleasure’, but for Sidgwick the two terms were essentially equivalent. While noting that ‘happiness’ is sometimes used for a ‘calmer and more indefinite’ feeling (ME 92), he took it to connote a whole composed of pleasures, so ‘by “greatest possible Happiness” we understand the greatest attainable surplus of pleasure over pain’ (ME 120, also 41n, 92, 92n2, 413). This was a common view, but Rashdall drew a distinction. He thought happiness involves ‘satisfaction with one’s existence as a whole—with the past and the future as well as with the immediate present’, so it is ‘possible to get an enormous amount of pleasure into one’s life . . . and yet to be on the whole unhappy through the presence of desires which are unsatisfied, dissatisfaction with the past, anxiety as to the future, unfulfilled aspirations, baffled hopes and the like’ (TGE II 57). He here defended the ‘life-satisfaction’ view of happiness, on which it is a good feeling with a particular object, your life as a whole,10 and he thought happiness as so understood is ‘far more valuable’ than other pleasures (TGE II 59). We may question this last claim. Why, duration aside, should a good feeling of intensity n about your life as a whole be better, just as a feeling, than one of the same intensity got from eating chocolate? Intentional pleasures are not in general better than sensory ones, and it is hard to see why this one should have special value. Still, Rashdall usefully saw happiness as, though a pleasant feeling, one with a distinctive object.

9.2 Hedonism: For and Against Not all in the school thought pleasure is good. In his article on McTaggart, Moore said ‘it does seem very probable that pleasure has no value at all’ (‘MME’ 358), though in Principia he said only that it has little value; Joseph too denied that a world containing only pleasure would be good (SPE 89–91). But most saw pleasure as at

9 A recent defence of pleasure/pain asymmetry is in Mayerfeld, Suffering and Moral Responsibility, Chapter 6. 10 See Nozick, Examined Life, pp. 110–14.

non-moral goods

199

least one intrinsic good; the question then was whether it is, as hedonism holds, the only good. Sidgwick argued that any view that supplements pleasure with other goods must weigh their values against each other and cannot do so systematically (ME 406). But this argument can be met in the same way as his critique of pluralistic theories of the right. A pluralist about the good can say the truth about value is only partly determinate, or, as Ross did, that the truth is fully determinate but we cannot always know it. Either way there is no special problem in his pluralism. Early in The Methods Sidgwick said nothing can be good ‘out of relation to human existence, or at least to some consciousness or feeling’ (113). Moore famously denied this when he said a beautiful world with no minds in it would be good (PE 83–5; also EE 71–3, 90–1, 192), but later he allowed that beauty on its own may have no value (PE 202, 203) and in Ethics he accepted that view, saying all good things involve some states of consciousness (E 107). McTaggart, Ross, Joseph, Carritt, and Ewing too denied that beauty by itself is good (‘IV’ 433–7; NE 399; RG 70, 106, 130, 141; SPE 84–6; EPT 86; ‘BVG’ 233); the early Moore aside, there was general agreement that the intrinsic goods all involve mental states.11 Sidgwick’s argument that no mental states other than pleasure are good began with a critique of the value of virtue that we will consider in Chapter 10. He then considered the possibility that the values of some mental states depend on ‘objective relations of the conscious being, not strictly included in his present consciousness’, such as ‘Cognition of Truth, Contemplation of Beauty, Volition to realise Freedom or Virtue’ (ME 399). Aesthetic contemplation and virtue involved conformity to an ideal (ME 400), which is not really an object outside the mind, so his more central cases were freedom from domination and especially knowledge, or ‘the mental state of apprehending truth’. He recognized that many prefer knowledge to false belief even when the latter is more pleasant; what they value is not a purely internal state such as pleasure but a belief that stands in the relation to non-mental reality of representing it accurately. He replied that, though intelligible, this view ‘ought not to commend itself to the sober judgment of reflective persons’, because ‘it seems clear after reflection that these objective relations of the conscious subject, when distinguished from the consciousness accompanying and resulting from them, are not ultimately and intrinsically desirable’. If we consider knowledge that is not accompanied by pleasure and does not produce pleasure, we do not find it good; likewise for freedom and aesthetic contemplation. But if all these states are valueless apart from pleasure, pleasure must be the sole good (ME 400–1).

11 Jones likewise denied that beauty on its own is good (‘Mr. Moore on Hedonism’, pp. 435–7). Welchman reports that Moore’s claim about beauty was the most discussed part of PE in the decade after its publication and was almost universally rejected (‘G.E. Moore and the Revolution’, pp. 318–19); McTaggart likewise said Moore’s claim ‘has very few supporters’ (‘EHS’ 407).

200

british ethical theorists from sidgwick to ewing

As Moore pointed out, this argument ignores the possibility of organic unities: ‘from the fact that no value resides in one part of a whole, considered by itself, we cannot infer that all the value belonging to the whole does reside in the other part, considered by itself ’ (PE 92). Even if knowledge by itself has no value, once there is pleasure—and it must be pleasure in the knowledge rather than an unrelated pleasure of, say, eating chocolate—the fact that its object is knowledge can give it more value than an equally intense pleasure in something else. Bradley had made the same objection in 1877,12 and the organic-unities view it proposes gives an attractive interpretation of Mill’s doctrine of higher pleasures. On this interpretation a reading of poetry unaccompanied by positive feeling is not a ‘pleasure’ and so has no value. Once there is pleasure, however, a pleasure in poetry is better than an equally intense pleasure in pushpin, because a quality with no value by itself can increase the value of a state that involves pleasure. Given these antecedents, it is surprising that Sidgwick never addressed the organic-unities objection to his argument, but it was repeated after Moore by Rashdall and Broad (TGE I 75–6; FT 235–7). Sidgwick did not claim a consensus on hedonism, recognizing that many accept ideal goods such as knowledge. But he tried to explain away their dissent, on several grounds: the term ‘pleasure’ is mostly used for ‘coarser’ feelings, to get pleasure you must often aim at other goals, and hedonism is often wrongly associated with an egoistic view of duty (ME 402–6). But his principal claim was that the ideal goods not only produce pleasure but ‘seem to obtain the commendation of Common Sense, roughly speaking, in proportion to the degree of this productiveness’ (ME 401). Just as common sense is unconsciously consequentialist, so it is unconsciously hedonistic. Moore replied by emphasizing explanatory intuitions. Sidgwick was accusing common sense of confusing means and ends, of thinking a pleasure is better in itself because it will produce more pleasure later. But Moore rejected the charge: common sense distinguishes immediate pleasantness and conduciveness to pleasure, believes that sometimes a less immediately pleasant feeling is better, and does not do so because of the feeling’s consequences. It would not think it a sufficient justification for ‘refined pleasures’ now that they would lead to a heaven in which the greatest possible pleasure is obtained ‘by a perpetual indulgence in bestiality’ (PE 94–5). There were also positive critiques of hedonism. Most radically, Bradley and Green challenged the coherence of the hedonist ideal. Pleasures, they argued, cannot be added, largely because they are fleeting and cannot provide ‘an abiding satisfaction of an abiding self ’; the idea of a greatest sum of pleasures is therefore self-contradictory.13 Sidgwick, Rashdall, and McTaggart had little difficulty refuting this particular 12 ‘It is a logical error to argue, Because A is not desirable without B, therefore B by itself is desirable; or again, Because A + B and B (by itself) are both desirable, therefore nothing but B is desirable’ (‘Mr Sidgwick’s Hedonism’, p. 91; also 92–3, Appearance and Reality, p. 358). Bradley’s criticism was repeated by Hayward, again before Moore (Ethical Philosophy of Sidgwick, pp. 189, 200, 225, 231). 13 Green, Prolegomena, sec. 234, also secs. 176, 219–39, 358–61, 369, 371; Bradley, Ethical Studies, pp. 95–9, 103–4, 111, 125.

non-moral goods

201

argument, noting that, no less than pleasures, virtuous actions and aesthetic experiences succeed each other in time (ME 132–4, 407n; EEM 96–8, 247–9; GSM 107–15, 119–23; TGE II 2–22; SHC 109–18). But there is a more interesting objection Bradley and Green could have been making. It starts from the fact McTaggart and Parfit noted: that we care much less if at all about pleasures and pains in our past. It then says a true good must appear good or important from all temporal perspectives, not just ones that precede it in time but also ones that follow it; it must in that sense provide ‘abiding satisfaction’. This claim tells against the value of pleasure if past instances of other goods are not discounted, as they appear not to be. A past virtuous act or achievement can still be a source of pride and a past failing prompt shame; both remain important after they are gone.14 Carritt made this contrast when he said that, whereas the desire for pleasure is never satisfied because past pleasures are nothing to us, ‘an honest act, a scientific discovery, the creation of a beautiful thing . . . are in a sense joys forever’ (TM 26). Only a joy forever, the objection says, can be a true good. Someone who values pleasure can try to meet this objection in several ways. He can say that, though in one sense we care less about our past pleasures, we still believe them to have been good; however much our emotions are affected by our temporal location, our judgements of value are not. He can also say we do not discount other people’s past pleasures,15 and can give our discounting an evolutionary explanation. The biological function of pleasure and pain is to attract us to activities that promote our survival and repel us from ones that threaten it. Since this function concerns the present and future, it would be hindered by strong feelings about our past; hence natural selection has not given us such feelings. But the appeal to evolution here is double-edged. When we try to judge the value of pleasure we usually do so by imagining it from the inside, as if we were experiencing it now. This raises the possibility that the resulting assessment reflects in part a biologically-based urge rather than a pure judgement of value. Could the biological salience of present pleasure be leading us to overvalue pleasure? Others rejected hedonism more directly. Broad thought it can be refuted by a ‘single convincing contrary instance’ and cited malice, or pleasure in another’s undeserved misfortunes. ‘Is it not perfectly plain that this is an intrinsically bad state of mind, not merely in spite of, but because of, its pleasantness?’ (FT 234; also Rashdall, TGE I 98–9; Ewing MP 69). This affirmed the goodness of virtue and evil of vice, and Ross did the same when he asked us to imagine two worlds containing equal

14

Imagine that in a variant of Parfit’s example you are told that you are either a scientist who made a major discovery last year or a scientist who will make a minor discovery next year. Surely you will hope you are the scientist who made the major discovery last year. Or as the workout T-shirt puts it, ‘Pain is temporary, pride is forever’. 15 Parfit, Reasons and Persons, pp. 181–4.

202

british ethical theorists from sidgwick to ewing

amounts of pleasure but in one of which people are virtuous while in the other they are vicious. Is the first world not better (RG 134–5)? He gave a similar argument for the value of knowledge, describing two worlds with equal amounts of pleasure and virtue but in one of which people have more understanding of nature and its laws; is this world, again, not better (RG 139)? Moore countered hedonism with the positive claim that the two greatest goods are aesthetic appreciation and personal affection (PE 188). Like knowledge, these goods’ best forms involve what Sidgwick called an ‘objective relation’ to something nonmental, either a real beautiful object or a real person. And a further objection to hedonism is that it can find high value in a life completely without such relations and enjoying only pleasurable illusions. This is the point of Nozick’s example of an ‘experience machine’ that, by electrically stimulating your brain, can give you the experience and therefore the pleasure of anything you like: it can be just as if you were climbing Mt Everest or making love to Brad Pitt, though in fact you are not. Nozick thought life on this machine is not ideal because it lacks ‘contact with reality’,16 and Moore made essentially the same point without the science fiction apparatus: We can imagine the case of a single person, enjoying throughout eternity the contemplation of scenery as beautiful, and intercourse with persons as admirable, as can be imagined, while yet the whole of the objects of his cognition are absolutely unreal. I think we should definitely pronounce the existence of such a universe, consisted solely of such a person, to be greatly inferior in value to one in which the objects, in the existence of which he believes, did really exist just as he believes them to do. (PE 197)

The school’s later members often took hedonism about the good to have been decisively refuted. Ewing said he simply assumed that ‘the doctrine that pleasure is the only good, is false’ (MP 5), while Ross said ‘After all the able refutations of hedonism that have been published in recent years, it seems to me unnecessary to tread once more on this rather hackneyed ground’ (RG 99). And for Broad hedonism was ‘far too simple’, since ‘the intrinsic goodness or badness of any state of affairs will depend on many different factors’ (FT 239). These anti-hedonist claims do not meet Sidgwick’s conditions for self-evidence, most obviously, there is no consensus on them. But, like the deontologists’ intuitive claims against consequentialism, they are closer to common sense than hedonism is, as Sidgwick implicitly recognized when he tried to explain them away. And Rashdall, Moore, and Ross did not just reject hedonism; they elaborated alternative theories valuing additional goods, including aesthetic appreciation and knowledge.

16

Nozick, Anarchy, State, and Utopia, pp. 42–5.

non-moral goods

203

9.3 Aesthetic Appreciation Rashdall often cited the value of aesthetic appreciation in his critique of hedonism (TGE 74–5, 177–83, 191; II 32) but did not explain how it relates to his principal goods of virtue, knowledge, and pleasure. Did he think, like Ross (RG 141; FE 270), that it combines some of these? Ewing too valued aesthetic experience (MP 127n, 160; DG 154, 164; E 42–3, 72), but the members who wrote most about it were Moore and Carritt. Moore’s discussion was hampered by his definition of beauty as ‘that of which the admiring contemplation is good in itself ’ (PE 201). This definition turns the seemingly substantive thesis that it is good to admire beauty into the near-tautology that it is good to admire what it is good to admire, and the difficulty was increased by his denial that there is any ‘criterion of beauty’ (PE 202), or one property shared by all beautiful things. Taken strictly, this reduces our judgements of beauty to a set of unconnected assertions that particular objects are good to admire. He would have done better to treat beauty either as a distinct non-natural property or as a natural one. What Moore valued in this area was only the contemplation of beauty and not, as the Romantics had emphasized, its creation.17 This aspect of his thought fits Keynes’s claim that in the ‘religion’ he and his circle learned from Moore ‘Nothing mattered except states of mind’ and that these ‘were not associated with action or achievement or consequence. They consisted in timeless, passionate states of contemplation and communion.’18 Ross did recognize the value of artistic creation and gave a tentative account of its basis (FE 270–1), but for Moore the only aesthetic good was appreciation. His more specific claims about this good were a mixture of the naive and the insightful. Naively, he thought admiring a painting of a natural scene is less good than admiring the scene itself, as if there is no point in viewing a Monet of water lilies if you can see the flowers themselves (PE 195). He did not consider any possible cognitive benefits from reading literature, such as insights into ethics or psychology; his only good relating to art concerned beauty. But he gave a wonderful analysis of the difference between the classical and romantic styles in art, saying the former aims to maximize the value of a work as a whole, or residing in the relations between its parts, whereas the latter sacrifices this value as a whole to increase the value in some part. He thought either style can produce a result that is equally good on the whole, but said the ‘distinctively aesthetic temperament’ prefers the classical style (PE 215–6).19

This point is noted in Baldwin, ‘Editor’s Introduction’, PE: Revised Edition, p. xxiv. Keynes, ‘My Early Beliefs’, p. 53. This last view is a little hard to reconcile with Moore’s fondness for the novels of Walter Scott (‘A’ 26). Carritt questioned the importance of the classical/romantic distinction in IA, Chapter 14. 17 18 19

204 british ethical theorists from sidgwick to ewing His account of aesthetic appreciation involved two organic unities. The first concerned the mental state of appreciation, apart from whether its object really exists. He took this to be a compound of two elements, a mental representation of beautiful qualities and an appropriate emotion toward them, each having little or no value on its own but together having considerable value (PE 190–1). This was a mistake. Your admiration for something beautiful is not two states but one: a feeling directed at a particular representation that is its object and made the specific feeling it is by that object; you cannot admire without admiring some thing. Moore would have benefited here from Brentano’s views about intentionality. The second concerned the real existence of the beauty. He thought that if you admire a beautiful object you believe exists, its actually existing makes the situation significantly better, and by more than any value the object has in itself (PE 192–9). He described this value holistically, saying the truth of your belief makes for a whole combining your mental state and its external object that has value as a whole additional to any values in its parts. But we could equally well say your admiration has more value when its object is real. (It is less plausible to say the object has more value when it is admired.) Either way, this was a paradigm organic-unities claim, and Moore elaborated it further. The value of aesthetic appreciation as such is present, and is all that is present, when you admire a beautiful object you merely imagine and know does not exist. If you believe the object exists, that makes the situation better if your belief is true and worse if your belief is false, but there are two ways your belief can be false. You can admire qualities you believe are beautiful but are in fact ugly, or you can admire qualities that are beautiful but believe falsely that they exist in the world. The first he called an ‘error of taste’ and the second an ‘error of judgement’, and he seems to have thought the first error, where you admire something ugly, is worse (PE 193). But both cases of false belief make for less value than if you admire beautiful qualities you know you are just imagining, which in turn is less good than admiring beautiful qualities you correctly believe are real. Moore made similar claims about virtuous feelings, saying compassion for what you know is just imaginary suffering, such as that of King Lear, is less good than an otherwise similar compassion for suffering you correctly believe is real (PE 219). Here too he held that a false belief in the reality of the object makes for less value, but we may wonder whether that is always true. Taking pleasure in what you falsely believe is real pain, such as that of the victim in what is only feigned torture, seems no less evil if the pain is real. That detail aside, Moore’s treatment of issues about emotion and belief is an exemplary piece of philosophical analysis. Carritt included the appreciation of beauty on his list of intrinsic goods (EPT 86–7) and discussed it at length in his books on aesthetics. He rejected views that value aesthetic appreciation as a means either to pleasure or to moral improvement, saying that given the first, we should prefer ‘a certain quantity of tobacco or conjuring’ to Shakespeare, and given the second, should prefer sermons (TB 61–2, 119). He also denied that recognizing beauty gives us knowledge of things outside us

non-moral goods

205

(TB 105–6), saying the philosophical claims in works like Paradise Lost are irrelevant to their aesthetic value: ‘the moment that we ask if any single statement there propounded is logically true we have ceased . . . to treat them as works of art’ (TB 211; also IA 31–6). His positive account was an ‘expressive’ one, drawn from Croce but without the Idealist metaphysics. It says a beautiful natural object or work of art—and he thought an adequate aesthetics must accommodate both—is so in virtue of expressing an emotion or aspect of our will. More specifically, a beautiful object arouses in us feelings we were only dimly aware of and can now contemplate and know more fully; in a sunset, for example, we see ‘all the passions of life’ (TB 109–11, 138–9, 181–3, 279; also IA 36–40, 56–67). It therefore gives us insight into ‘our own inner nature and processes’ (TB 111, also 135) and, by giving our feelings ‘form and pressure’ makes them ‘determinate’ as they previously were not (TB 279, 182, also 284). Since only the appreciation of beauty has these effects, it is an irreplaceable good. Carritt did not explain further why aesthetic appreciation as so understood has value. If it gives us knowledge of our feelings, why is that knowledge so good? If it makes the feelings more determinate, why is that worth seeking? But, unlike Moore, he did try to explain what beauty is, and he also explained why some aesthetic experiences are more valuable than others. They are the ones with more beautiful objects, and here three features are relevant. An object is more beautiful if it expresses a more complex or comprehensive state of mind; thus Measure for Measure is more beautiful than any shorter selection from it (TB 215). An object can also express an emotion more adequately or precisely; thus a poet’s revisions to his first draft may result in a better poem, because his new words better achieves his expressive goal (TB 215–17). Finally, there is a feature of ‘depth’, found when an object unites in an expressive whole elements that on their own seem recalcitrant to expression, or ugly; his examples included the beggars of Rembrandt and dwarfs of Velasquez (ST 217–18, 256–8). But he did not share Moore’s concern for the real existence of contemplated beauty. Since what is good in aesthetic experience is just the knowledge and improvement of our inner states, it is ‘without existential judgment’ (TB 288) and is as good when its object is imaginary as when it is real. The materials of Carritt’s account differed from those normally used in our school, since they came from an Idealist source. But he alone tried to go beyond the claim that appreciating beauty is good by saying more precisely what beauty is.

9.4 Knowledge and Achievement The philosophical tradition has often valued knowledge highly. Plato and Aristotle thought theoretical contemplation is the greatest good, as did Aquinas and Hegel; knowledge has often vied with virtue for the title of the highest good.

206

british ethical theorists from sidgwick to ewing

In Principia Moore discounted the value of knowledge, saying it derives mostly from the extra value in contemplating really existing beauty. Knowledge ‘is an absolutely essential constituent in the highest goods, and contributes immensely to their value’, but has ‘little or no value in itself ’ (PE 199). This view makes knowledge mainly contributively good, and then only when it is of a beautiful object; knowing scientific laws or historical facts has little worth. Both earlier and later Moore valued knowledge more, citing it alongside aesthetic appreciation as a key non-hedonic good (EE 188–9, 192; E 22, 102), and that was certainly others’ view. Prichard in ‘Mistake’ thought knowledge is intrinsically good (MW 10), as did Laird, Carritt, Broad, and Ewing (SMT 49; EPT, 87–8; CE 262–3, 275; MP 160; DG 154, 164, 209; E 42–3, 47–8). Its value was especially emphasized by Rashdall and Ross, who thought it one of the two main non-moral goods, alongside pleasure, and higher or better than pleasure. They did not think it infinitely better than pleasure; Rashdall said it is not always wrong to give up study for the sake of enjoyable conversation. But they both thought knowledge is generally preferable to pleasure and thus the greatest non-moral good (TGE I 47–9; RG 151–2). Like pleasures, instances of knowledge can be more or less good. Most obviously, knowing the number of grains of sand on a beach or of storeys in a building has little or no value (TGE II 65; RG 139; also Carritt, EPT 87), while understanding the origins of the universe has more. Ross discussed this issue at length. His account was shaped by his and Prichard’s view that knowledge is a completely different state from true belief, involving both certainty and the right to be certain. But he thought a true belief too can be good and that its degree of goodness depends initially on two factors: its degree of groundedness, or the degree to which it is justified by your evidence, and the extent to which your degree of belief or credence in it corresponds to its degree of groundedness (RG 147). Knowledge has both these factors to the highest degree, since it is completely grounded and complete certainty corresponds to that. But a true belief can have them to differing degrees. Ross’s talk of degree of groundedness seems to assume that you have an all-out belief in a proposition rather than just assigning it a probability. If you did only the latter, it is hard to see why having a credence of 0.9 in a proposition your evidence gives a probability of 0.9 would be better than having a credence of 0.3 in a proposition your evidence gives a probability of 0.3. But his second factor of correspondence assumes that all-out beliefs can be accompanied by degrees of belief less than 1.0.20 There is presumably a lower limit on these; you cannot all-out believe a proposition you give a credence of 0.3. But above that limit you can have more or less confidence in a belief, and Ross wanted your degree of confidence to match your degree of justification, so you are more confident in a belief your evidence makes 0.9 probable than in one it makes 0.8 probable. He considered the view that it is worse to

20

For a more recent defence of this view see Foley, Working Without a Net.

non-moral goods

207

hold a belief with more conviction than it deserves than to hold it with less, for example to have a credence of 0.95 in a proposition your evidence makes 0.9 probable than to have a credence of 0.85 in it. While granting that the first error is more common and more dangerous, he denied that it is worse in itself: the two mismatches are equally bad (RG 147). Though that is a reasonable view, so, I think, is the one that makes overshooting worse. Ross assumed there is value only in true beliefs, and we can question that. Imagine that your evidence makes a proposition 0.99 probable and you believe it with credence 0.99, but through bad luck it turns out to be false. Your belief lacks the connection to reality that is required for knowledge, but you have exercised your cognitive faculty well and even as well as you can. Should there not be some value in this, and thus some value in merely justified belief? That aside, Ross was surely right that the value of a belief or instance of knowledge depends in part on its relation to evidence. Its value also depends in part on its object, or on properties of the proposition you believe, which should not concern just storeys in a building. Carritt said what has value is not knowledge as such but ‘ “intelligent activity” or “understanding”, which includes “the apprehension of logical implications or causal connexions” ’ (EPT 87–8; also Ewing, ‘BVG’ 229). Ross developed a similar view more fully. With an acknowledgement to Bradley21 he wrote: knowledge of general principles is intellectually more valuable than knowledge of isolated matters, and . . . the more general the principle—the more facts it is capable of explaining—the better the knowledge. Our ideal in the pursuit of knowledge is system, and system involves the tracing of consequents to their ultimate grounds. Our aim is to know not only the ‘that’ but the ‘why’ also, when the ‘that’ has a ‘why’. (RG 147–8)

The generality criterion he proposed contrasts with the view sometimes taken by Plato and Aristotle that the value of an item of knowledge depends on the value of its object, so the best knowledge is of the best things in the universe. That view makes understanding the causes of pleasure or virtue intellectually better than understanding the causes of pain or vice, which is implausible; more specifically, it introduces considerations relevant to an account of virtue into one of a cognitive and therefore non-moral good. Grounding the value of knowledge in its explanatory power, as Ross did, makes that value just cognitive. In his review of Ross’s book, Price questioned the generality criterion, asking whether ‘a physicist’s well-grounded opinions concerning the increase of Entropy are more valuable than a historian’s well-grounded opinions concerning the Principate of Augustus’.22 But we need to separate two senses of generality that Ross conflated. He first said a truth is more general the more facts it is ‘capable of

21

Bradley, Principles of Logic, vol. 2, pp. 685–8.

22

Price, Critical Notice of RG, p. 24.

208

british ethical theorists from sidgwick to ewing

explaining’, whether we use it to explain them or not. He then said our ideal is ‘tracing consequents to their ultimate grounds’, which involves actual explanation. The truths the historian knows may be less general in the first sense, because they apply less widely than the physicist’s and so cannot explain as much. But if he knows many facts about Augustus’s reign and has connected them to each other, so he understands why events at that time unfolded as they did, his knowledge may be more general in the second sense if the physicist has not used the law of entropy to explain much else. These two senses of ‘general’ mean there are really two goodmaking properties of an object of knowledge: its intrinsic explanatory power, regardless of how much you have used it to explain, and how much it has actually added to your understanding. The physicist’s knowledge has more of the first property; the historian’s may have more of the second.23 Rashdall and Ewing used a generality criterion to argue that the best knowledge is of philosophy and especially of metaphysics, because they concern ‘the Universe as a whole’ (TGE II 414, also 453–4; VR 36). Philosophers have often philosophized that philosophy is the best activity, but there are several grounds for doubt. Though metaphysical truths apply widely, they are not very explanatory; certainly general truths about being do little to explain why the world contains the particular objects it does. It is also unclear how far metaphysical beliefs pass the tests of being justified and true, given the many competing metaphysical theories and the difficulty of deciding between them. Though philosophical speculation may have virtues, it is questionable how far explanatory knowledge is among them. There may be other relevant features of an object of knowledge besides its generality. If your lack of knowledge on Nozick’s experience machine is disturbing, it is not so much because you do not know general truths; you may retain your beliefs about scientific laws. It is rather that you are mistaken about where you are in the world: you think you are climbing Mt Everest, if that is the content of your fantasy, but in fact you are sitting motionless in the machine. The subject of your delusion is not general in either of Ross’s senses; it is a very particular fact, yet the delusion diminishes the value of your condition significantly. This suggests that, generality aside, it is especially important to have true rather than false beliefs about how you relate to your immediate environment, because this concerns how you relate. And an extension of this view says it is important to have any true beliefs about yourself, including about your internal psychology. I myself question this extension. Selfknowledge is often valuable instrumentally, as a means to achievement or other goods, but if some self-deception would promote such goods I think its disvalue is fairly easily outweighed. But the more restricted claim that it is important to have true rather than false beliefs about your place in the world does point to a factor bearing on the value of knowledge other than generality.

23

See my discussion of ‘extent’ and ‘hierarchical dominance’ in Perfectionism, Chapter 9.

non-moral goods

209

There is a related issue here. Your condition on the experience machine is disturbing not only because you lack some valuable knowledge, so your condition is less good than it could be, it is also that you are positively deluded. This raises the question whether, just as pleasure has the corresponding evil of pain, knowledge has a corresponding evil of false or unjustified belief. Some members occasionally discussed this question. Ross said ‘badly grounded opinion’ is bad (RG 102) but said nothing about false belief. McTaggart listed ‘error’ as an evil (NE II 412), which makes sense if it involves not just an absence of connection to reality but a positive mismatch, while Carritt had a more nuanced view. He thought ‘some ignorance seems bad and some false opinion’, with false opinion ‘generally worse than ignorance’. But where knowledge has no value, for example, of trivial facts, ignorance or false opinion is also not bad (EPT 88–9). And he made the interesting asymmetry claim that, whereas pain is more evil than pleasure is good, ‘knowledge [is] more plausibly good than either ignorance or error is evil’ (EPT 94). This seems right for scientific and other impersonal knowledge. That Aristotle had false beliefs about the laws of motion does not seem bad, and certainly not worse than having no opinion about them, like most of his contemporaries; if his beliefs about the laws were justified given his evidence, they seem positively good. But false beliefs about where you are in the world, as on the experience machine, do seem bad, and much worse than knowledge about so specific a topic is good. On some subjects knowledge seems more good than error is evil, whereas on others there is an opposite asymmetry. At one point Ross considered the value not of knowledge but of the search for knowledge, which many have also thought a good. He said this involves the desire for knowledge, which is a form of virtue and ‘may be much more fully present in many an unsuccessful search for difficult knowledge than in the successful attainment of knowledge that is easy to get’ (RG 152). The love of knowledge is indeed virtuous, but it is not found only in the search for knowledge. It is also present when you take pleasure in knowledge you already have, and it need be no more present in successful scientific research that yields a major discovery than in research that fails. If successful research is more valuable, as it seems to be, that must be for some other reason. We come now to a gap in the school’s thinking about the good, one they shared with many earlier philosophers. They did not see an additional intrinsic good in achievement, or in having a goal for how the world should be and then realizing it. This good is the converse of knowledge. In knowledge the world is a certain way and you make your mind match the world, by forming a true belief about it. In achievement you start with a goal in your mind and make the world match your goal. Just as for knowledge a belief cannot be true by luck but must be justified by your evidence, so an achievement should not be lucky but should have been predictable in advance given how you pursued it. Each of the two therefore involves a non-lucky match with reality, and each intuitively has value. While part of the external contact that is missing on the experience machine is knowledge, another is achievement, since while

210

british ethical theorists from sidgwick to ewing

plugged in you do not make any real changes in the world. You may think you are placing yourself atop Everest or making a movie star love you, but you in fact do none of those things; your intentions to produce those effects all fail. Finally, the factors that make for more valuable knowledge are paralleled by those for achievement. Just as knowing a truth that concerns more of the world is more valuable, so is achieving a goal that extends further. This is partly why the achievement of a politician who introduces laws affecting millions of people for years to come is greater than that of someone who ties a shoelace. And just as knowing a truth that explains many others is better, so is achieving a goal that requires you to achieve many others as means to it and is in that way difficult; that is another reason why the politician’s achievement is better. Both explanatory knowledge and difficult achievement involve a hierarchy of other mental states subordinate to one that explains or generates them, with the more such states the better.24 Achievement figures prominently in our everyday thinking about the good. We think part of a good life is doing things: achieving goals and making a difference in the world. Something like it was a central value for Marx and Nietzsche, but our school engaged little with them, and achievement was not part of the tradition they inherited from Plato and Aristotle. Aristotle thought there are goods of theoretical and practical reason, knowledge of the first and moral virtue of the second. Rashdall, Ross, and Ewing similarly affirmed goods of feeling, cognition, and conation, with virtue the good of the latter; since they recognized just one good per faculty, that again left no room for achievement. Achievement occurs on its own when you realize a goal that, though difficult, has no value in itself, so there is no virtue in pursuing it. Such goals are found preeminently in games There is no intrinsic value in a ball’s going into a hole in the ground or in your standing atop a mountain, but if you achieve those goals in the challenging way golf and mountain climbing require your doing so has considerable worth.25 The same holds in business. The world is no better if people drink Coke rather than Pepsi or use Apple rather than Microsoft products, yet someone who founds a business and makes it successful is rightly admired. The admiration is not for the end product of his activity but for the process of achieving it, given its demands on skill. And the same value can be present when you achieve an intrinsically good goal: then you combine the good of virtuously pursuing a valuable end with the good of succeeding at something difficult. That explains the extra value in a successful search for knowledge: though sharing the good of loving knowledge with unsuccessful research, it adds the separate good of achievement. Rashdall came close to recognizing this good when he said, ‘ “Sport” has been well defined as the overcoming of difficulties simply for the sake of overcoming them’ (TGE II 105; also ICE 154). But he valued sports only instrumentally, as promoting 24 25

Hurka, Perfectionism, Chapter 9. On games see Suits, The Grasshopper; and my ‘Games and the Good’.

non-moral goods

211

intellectual and moral qualities though with appropriate scepticism about the latter. (‘A football player who excels in “combination” is quite as likely as other men to play for his own hand in real life’; this against ‘the enormous exaggeration of the moral value of athletics which is popular at the present moment’ [TGE I 191–2n].) Laird too said that sometimes a series of acts with an end is ‘justified principally by the means. In a contest of skill, for example, the end is victory, but victory in itself is barren. It is the quality of the contest that is excellent’ (SMT 47). But he too left achievement off his official list of goods (SMT 49); it should be a central non-moral good, alongside and paralleling knowledge, but was mostly ignored by our school.

9.5 Aggregating Goods Having identified its intrinsic goods, a moral theory must combine their values into a measure, however rough, of the aggregate value in a single life and then in the world as a whole.26 Our school also had views on this topic. Because a life is a whole, it can involve organic unities. Moore, surprisingly, did not consider any of these—Keynes said he was oblivious not only to the life of action but also to ‘the pattern of life as a whole27—but others did. Broad and Ewing discussed the view that a life that begins badly and improves is better than one that, though containing the same total of other goods, is better at the start and deteriorates. Broad considered this view in the context of hedonism and wondered whether our tendency to accept it may just reflect our recognition that the first life will contain additional pleasures of anticipation and the second additional pains; he was therefore hesitant to accept it (FT 225–6). Ewing discussed the different case of a life that becomes more virtuous as against one that deteriorates morally. Here the first life may involve the additional good of struggle against vice and the second an evil of laxness about virtue, but Ewing did not rest the difference in their values entirely on that fact but saw it as in part a case of organic unity (VR 218–19). And improvement in time can in principle be valued whatever one’s view of the goods being improved. A different view values variety of goods in a life. It says a life with significant amounts of pleasure, knowledge, and achievement is other things equal better than one in which a single good predominates. A similar view can be applied to an individual good, so it is better to have knowledge of physics, Roman history, and your friends’ personalities than to concentrate your understanding in one area, or to have achievements in a great many fields than in just one. This view can say that in addition to the value in its components a life can have value as a whole, given its 26 I here assume that a theory will aggregate first across the times in a life and only then across persons; the alternative is to aggregate first across the states of all persons at each time and then across times. If a theory’s principles are both summative, the order in which it applies them makes no difference; if either or both are non-summative, it can change the result. But I will assume that a theory emphasizes the unity of individual lives by aggregating first within them. 27 Keynes, ‘My Early Beliefs’, p. 58.

212

british ethical theorists from sidgwick to ewing

degree of variety, or that the comparative value of a given increase in a good depends on the comparative amount of it you already have, being greater when that is smaller. Either way, it finds some more value in a life that is well rounded than in one that is specialized. Rashdall suggested this view when he said that in an ideal life different goods are ‘harmonized and combined’ or stand in their ‘true proportion’, so ‘No amount of one kind of good can compensate for the absence or deficiency of another’ (TGE I 221, II 40n, 39; also E 70–1). But he interpreted it in a restrictive way. He did not think it sufficient for the good of ‘proportion’ that your day contain seven hours devoted to virtue, six to knowledge, and four to pleasure; your individual activities must combine these goods, so you get pleasure from virtuous activities and have enjoyments that promote knowledge (TGE II 40). Ewing, who also valued variety, did not impose this condition. He thought ‘the more different kinds of good there are the better the world’, but the goods ‘need not all be realized simultaneously’ (VR 216, 221, also 219–20). He seems also to have valued variety not only in a single life (‘BVG’ 229) but also in a population, as we can value the varied achievements of a civilization like ancient Greece even if the lives in it were specialized; in an early work he denied that ‘the intrinsic value of a community as a whole is necessarily equal to the sum of the values of the individuals who compose it’ (MP 180; also Broad, EMP II 667–8). That a well-rounded life is other things equal good does not mean you should always on balance seek it, since pursuing too many goods can result in your not achieving much of any. Rashdall recognized this, saying the attempt at ‘an equal, all-round development of one’s whole nature’ can involve ‘a very apotheosis of mediocrity, ineffectiveness, dilettantism’ (TGE II 63–4). But if variety has some value, the ideal life for everyone will involve some more balance among goods than without that value, and for people with special talents a well-rounded life may be a realizable goal.28 Other organic-unities views value a pattern of distribution across lives, such as equality or proportionment to desert; we will discuss them in Chapter 11. Such views aside, there is the question of how the value in a life is a function of the values at particular times in it, and the value in a population of lives a function of the values in each one. Sidgwick noticed the difference between two answers to the second question when he said maximizing the total happiness of human beings differs from maximizing the average happiness; in particular, it favours population increases that lower the average so long as they increase the sum (ME 415–16; also EEM 7). He thought utilitarianism is committed to the total view, but Broad called acts that increase the sum of happiness while lowering the average ‘quite certainly wrong,’ and Ross agreed

28

See Hurka, Perfectionism, Chapter 7.

non-moral goods

213

(FT 250; FE 69–71). But total and average views on this topic are both possible, as are ones that compromise between the two.29 One objection to the total view concerns what Parfit calls ‘the repugnant conclusion’. Imagine an ideal world in which billions of people lead spectacularly good lives. If the total view is correct, there is another world that would be better even though the lives in it are barely worth living, with just a tiny surplus of pleasure over pain and no other goods. If there are enough additional people in the second world, its sum of goods will be larger.30 McTaggart discussed a single-life version of this objection, comparing a brilliant life that lasts a million years with a merely ‘oyster-like’ one that contains just minimal pleasures and no virtue or love but, if it lasts long enough, has a greater sum of goods. He even called the view that the second life is better a ‘conclusion’ that will be ‘repugnant to certain moralists’. But he thought we should accept the conclusion, and tried to explain away our resistance to it as reflecting our inability to imagine great lengths of time and our tendency to discount goods in the far future (NE II 452–3). As Dennis McKerlie argued, this was a surprising line for McTaggart to take.31 He thought love is the greatest good, and in The Nature of Existence explained how by extending a view Moore had proposed against him. He argued that, whereas the value of a fixed increase in love is always the same, so the potential value of love is unlimited, the values of increases in other goods diminish toward zero, so there is an upper limit on their values. Love whose value is above the limits for pleasure and knowledge is therefore better than any instance of those goods could be, making love in that way the highest good (NE II 436–9); the same would also follow if the value of increases in love approached a limit, but one higher than for other goods. The puzzle is why McTaggart did not apply this view to the repugnant conclusion and say the love in the brilliant life is better than any amount of pleasure in the oyster-like one, making the former on balance preferable. One possibility is that he thought only the value of increases in the intensity of a pleasure diminishes; increases in its duration have constant value. But then the puzzle is why he did not consider the view that the marginal value of duration diminishes and at least argue against it. Broad proposed the latter view in his commentary on McTaggart, saying the value of a state ‘increases with each increase in its duration, yet it increases less for each successive increment of duration and approaches an upper limit as the duration increases indefinitely’ (EMP 684). But, strangely, almost immediately afterward he conceded, at least for lives with a constant level of quality, a premise that yields the repugnant conclusion, apparently ignoring the claim he had just made about duration (EMP 685–6). And he confined his case against that conclusion to lives whose level of quality is not constant but goes up and down. 29 30 31

I proposed one such view in ‘Value and Population Size’. Parfit, Reasons and Persons, pp. 387–90. McKerlie, ‘McTaggart on Love’, pp. 77–85.

214

british ethical theorists from sidgwick to ewing

When there is that variation, he argued, a life has a ‘temporal pattern’ that may be ‘highly good-making’. But the pattern needs room to display itself, and has more of that in a life whose maximum quality is very high than in one whose quality is never more than barely positive. The brilliant life has ‘an elaborate temporal pattern’ that is or may be ‘like the performance of an opera. The oyster’s life can have no such pattern. It is like a single note played with hardly any variation on a single very simple instrument such as a tin-whistle’ (EMP 686–8). This argument seems to appeal to two organic unities, one of variety at a time (the many instruments in an opera vs. the tin-whistle) and one of variety through time (the changes in a melody vs. the single note), with the second perhaps more prominent. But we can question the importance of both. Though the oyster-like life cannot vary as much in absolute terms as the brilliant life, it can do so proportionally; even if its average pleasure is barely positive, it can have moments that are ten times as pleasant (though still barely positive) and others that are one-tenth so. Why should the absolute degree of variation matter more than its degree considered proportionally? We can also ask whether up-and-down variation is really a good. Is a life that alternates between great virtue and modest vice in any way better than one that is always moderately virtuous? The former may be more interesting to observe, but the claim that it is better is dubious. And both of Broad’s goods seem very much ones whose value diminishes through time. Having five goods rather than one at a particular time may be significantly good, but having the same five for a million years is not much better; likewise for repeating the same up-and-down pattern for a million years. If those claims are true, and the value of pleasure does not diminish through time, the organic goods will not outweigh it; but if the value of pleasure does diminish, we do not need specifically organic goods to avoid the repugnant conclusion. The simplest way to do that is to have the value of longerlasting lower goods approach but never exceed a limit that higher goods can exceed. McTaggart did not consider this possibility and Broad, though at one point affirming it, did not apply it to the repugnant conclusion. A final topic is the good in animal lives and how that relates to human goods. Sidgwick followed Bentham and Mill in thinking pleasure is good and pain evil in all sentient beings, so we have the same duties concerning non-human animals as we do concerning ourselves (ME 406n, 414). But he thought no one would find perfectionist value in the lives of animals (ME 114–15), and that was also Rashdall’s and Ross’s view. Rashdall agreed that animal feelings matter in themselves. He rejected the view of ‘the school of Green’ that we ought not to harm animals ‘not in the interest of animal well-being, but in that of our own humanity’, saying ‘If the suffering of animals is no evil, it cannot be inhumane in me to cause it’ (TGE I 214, also I 98–9, II 49, 441–2; E 76). But he thought the well-being of animals is ‘of very small value in comparison with that which we set upon human well-being’, because the former contains only pleasure while the latter adds the greater goods of knowledge and virtue; he therefore rejected Henry S. Salt’s proposal that there is value in animal

non-moral goods

215

‘self-realization’ (TGE I 215, also 239).32 Ross too thought animals’ pain is bad in the same way as humans’, so we have a duty not to cause it (RG 49, 137–8), but ‘human life is in itself something more valuable than animal life’, because it contains higher goods (RG 149). This again allowed him to give duties to animals less weight than in a purely hedonistic theory like Sidgwick’s; his perfectionist goods supply reasons to benefit humans that we cannot have about animals. There is another way our duties to animals may be less. In addressing the theological problem of evil Ewing considered the suffering of animals and said the principle that the suffering of one cannot be justified as a means to another’s good may not apply in their case (VR 235). Ross made a similar suggestion when he said that although we have duties to animals, animals do not have rights, because rights are things one must be able to claim (RG 50). The general view these remarks point to is that our duties concerning animals are limited in kind. In one version of the view, they include only the consequentialist duty to maximize the surplus of good over evil and no deontological duties that constrain that; then any treatment of an animal that does more to benefit humans than it does to harm the animal can be right. This view is in fact hard to motivate in Ross’s deontology. If beneficence and non-maleficence both concern good and evil effects, and the pain of animals is no less evil than that of humans, it is hard to see why non-maleficence should apply less to them than to humans, or why deliberately making an animal suffer should be less wrong than doing the same to a human. Nonetheless, just as animal lives may contain only some of the goods found in human lives, so our duties to animals may include only some that govern our treatment of each other.

32

Salt, ‘Rights of Animals’, p. 209.

10 Moral Goods Alongside non-moral goods such as pleasure and knowledge, most in the school valued moral virtue. This was Rashdall’s chief intrinsic good, the later Prichard’s only good, and Ross’s highest one; it was also implied in Broad’s ‘single convincing contrary instance’ to hedonism. Closely related to it are the goods of personal love, which McTaggart and Moore emphasized, and moral desert, which several members affirmed. These three goods are the subjects of this chapter.

10.1 Virtue: For and Against As part of his defence of hedonism Sidgwick argued against the view that virtue is intrinsically good, first by arguing that it cannot be the only such good. He took virtue to be a disposition to act rightly, or in conformity with common-sense rules. But his examination of those rules had shown their ultimate basis to be promotion of the good. The claim that virtue is the only intrinsic good therefore involves ‘a logical circle’: virtue consists in promoting the good, which in turn consists just in promoting the good (ME 392; also EEM 93, 187, 256–7, 263; GSM 72–7).1 His argument here can be generalized. It need not assume that virtue involves causally promoting the good; any relation to the good will do. And it need not assume that all forms of virtue relate to the good. So long as some do, and the good they relate to is not just other forms of virtue, there must be some values other than virtue. In early editions of The Methods Sidgwick proceeded, surprisingly, as if showing that virtue cannot be the only good suffices to show that it is not a good at all. In an 1885 article Rashdall pointed out the non sequitur in this by defending a view on which virtue is just one good among others. Sidgwick had refereed an earlier version of this article for Mind and said that, since he was preparing a third edition, the author should wait and address his criticisms to that. He then pre-emptively answered one of Rashdall’s objections in that edition (ME3 401–2; ME 404) but did not address Rashdall’s main charge until the fifth edition, and then did not really understand his view.

1

Hume had made a similar argument; see Treatise of Human Nature, p. 478.

moral goods

217

Rashdall first questioned whether the argument that virtue cannot be the only good is decisive. One can in principle characterize virtue by relation to states such as the pleasure of others but deny that they are good. He thought this was the Stoic view and also Green’s: certain states are ‘preferred’, and therefore the targets of virtue, but only virtue itself is good. But he found this view ‘in a high degree paradoxical’, asking how states can be the objects of virtue unless they themselves are good. If they must be good, Sidgwick’s conclusion follows and virtue cannot be the only good (‘PS’ 207). Rashdall then gave an account of virtue on which it involves a non-causal relation to other goods. There are goods prior to virtue, such as pleasure and knowledge, but whenever something is intrinsically good, certain attitudes to it, such as desiring and pursuing it for itself, are also intrinsically good and constitute virtues. Virtue involves pursuing a good ‘of the whole’, or of all persons, and ‘the pursuit of this good of the whole’ is ‘a good to the individual’, or makes his life better (‘PS’ 219). As he later put it, the goodness of virtue rests on the ‘idea of the intrinsic worth of promoting what has worth’: if certain things are good, virtuously seeking them is another intrinsic good (TGE I 59, also 63, 70–3, 76, 125–9, 136–7, II 41–2). Sidgwick had argued that if virtue is a disposition of character, it is not plausibly ultimately good: as a tendency to certain acts or feelings, it ‘appears to me clearly not valuable in itself but for the acts and feelings in which it takes effect, or for the ulterior consequences of these’ (ME 393). Rashdall replied that ‘it is the settled bent of the will towards that which is truly or essentially good, and not a mere capacity or potentiality of pleasure-production such as might be supposed to reside in a bottle of old port, which constitutes the “goodness” or “virtue” which is regarded as a “good” or “end-in-itself ” by the school criticised by Prof. Sidgwick’ (‘PS’ 224; also TGE I 65). His point was partly that his account values not mere dispositions but occurrent states of virtuous desiring and willing,2 but also, and more importantly, that the relation it uses to characterize virtue is not Sidgwick’s causal relation but the intentional one of being psychically oriented toward good by wanting or willing it for itself. Virtue involves not causing good but having the right attitude to it, and whatever may be true of causing a good, the intentional relation of willing it is a possible and plausible intrinsic good. Stated more fully, Rashdall’s account rests on four recursive principles. One says that if something is intrinsically good, the intentional attitude of loving it for itself, that is, desiring, pursuing, or taking pleasure in it for itself, is also intrinsically good and a form of virtue. If another person’s pleasure is good, wanting it as an end in itself is also good and virtuous; more specifically, it involves the virtue of benevolence. If knowledge is good, wanting knowledge as an end is virtuous, and since the principle’s application recurs, being pleased by another’s desire for knowledge is virtuous too. A second principle says that if something is evil, loving it for itself is intrinsically evil

2

This point is more to the fore in the TGE version of the passage than in the earlier ‘PS’.

218

british ethical theorists from sidgwick to ewing

and therefore vicious. Desiring another’s pain as an end is evil, as is being pleased by his pain; both involve the vice of malice. It is likewise evil to love your malice, for example, to be pleased by your pleasure in others’ pain. Two further principles say that hating something evil, as in desiring or pursuing its absence and being pained by its presence, are good, and that hating something good is evil. The former makes compassion for another’s pain a virtue while the latter makes enviously wanting to destroy another’s pleasure a vice. Underlying all four principles is the idea that attitudes whose orientation matches the value of their object, either positive to positive or negative to negative, are virtuous and good, while ones whose orientation opposes it are vicious.3 Sidgwick finally addressed the possibility that virtue is one good among others in his fifth edition. He now included virtue among the goods he said common sense values in proportion as they promote pleasure (ME 401–2)4 and added a special argument against its being good in itself. Though it is possible for the same thing to be instrumentally and intrinsically good, he wrote, ‘it seems difficult to conceive any kind of activity or process as both means and end, from precisely the same point of view and in respect of precisely the same quality’ (ME 396). But this was still to characterize virtue by a causal relation to the good, or as instrumentally good, and that was not Rashdall’s view nor one a valuer of virtue must take. If virtue involves an intentional relation to the good, the property that makes it good as an end is different from the one that makes it good as a means. Sidgwick then said, ‘so far as we judge virtuous activity to be a part of Ultimate Good, it is, I conceive, because the consciousness attending it is judged to be in itself desirable for the virtuous agent’ (ME 397), or pleasant. (Recall his definition of pleasure as consciousness that is judged desirable.) That again was not Rashdall’s view, which values virtue apart from any accompanying pleasure. He also asked ‘whether Virtuous life would remain on the whole good for the virtuous agent, if we suppose it combined with extreme pain’, and said it would not. But a view that values virtue need not make it infinitely greater than other values; it can allow that it is sometimes outweighed by evils such as pain, so the life Sidgwick described is on balance bad. His discussion of the view that virtue is one good among others never met it on its own terrain. Though Rashdall was the first to state the recursive account, it was shared by others. Brentano arrived at it independently,5 and it also figured, though without the word ‘virtue’, in Moore’s theory of the good. Like Sidgwick, Moore defined virtue instrumentally, as a ‘habitual disposition to perform certain actions, which generally produce the best possible results’, and denied that as such it is intrinsically good (PE 172–4).

3

I elaborate this account in Virtue, Vice, and Value. He had actually included it in this discussion in ME1 but dropped it in ME2 when he introduced the ‘logical circle’ argument. 5 Brentano, Origin of Our Knowledge, pp. 22–3. 4

moral goods

219

But he accepted the four principles, saying the intrinsic goods include ‘the appreciation of what has great intrinsic value’ and the ‘hatred of what is evil’, while among the evils are the ‘enjoyment or admiring contemplation of things which are themselves . . . evil’, as in cruelty, and ‘hatred of what is good’ (PE 203–4, 217, 208–9, 211). If your motive in acting is love of a good consequence you may produce or hatred of a bad one, your motive is intrinsically good, so the exercise of a virtuous disposition to have such motives is good (PE 177). McTaggart too equated virtue with volitions ‘directed towards what is good, or at least, to what appears to [one] to be good’ (NE II 421, also 422–3, 431–2, 466–8; Russell, ‘EE’ 17–18).6 The four principles relate virtue to other values, but there is another form of virtue that relates to right action: conscientiousness, or wanting to do acts because they are right. Having in Principia analysed ‘right’ in terms of ‘good’, Moore there denied that conscientiousness is a distinct good; it contains a ‘more or less vague cognition’ of the goods or evils that will result from an act and derives whatever value it has from that fact (PE 218–19). But the Principia analysis is implausible, and an additional virtue is especially needed if some right acts do not have the best consequences. It was therefore affirmed by Ross. He recognized the non-moral goods of pleasure and knowledge and took virtue to have three forms: conscientiousness, or action springing from the desire to do one’s duty, action from the desire to bring into being something good, and action from the desire to produce some pleasure, or prevent some pain, for another being (RG 160). The second and third forms echo Rashdall’s principles, but the first is an addition, both when a right act maximizes the good and, more clearly, when it does not. Then a desire to do what is right conflicts with a desire for the greatest good and should ideally be stronger. Prichard, Carritt, and Broad also thought some attitudes are intrinsically good and others evil, and though they did not discuss them systematically, their examples— conscientiousness, sympathy, and benevolence on the one side, and malice on the other—fit the view that an attitude’s value depends on the moral quality of its object.7 Ewing too tied virtue to loving good and hating evil (MP 34–5, 107, 111, 155; DG 120; E 48, 148; ST 98, 125; VR 216–18, 220) but thought the point so familiar it did not need elaboration. Rashdall’s principles make virtue a higher-level intrinsic value, involving attitudes to lower-level values. Ross’s view makes it, more generally, a higher-level moral property, involving attitudes to items with other, previously given moral properties. It does not quite follow that virtue cannot be the only intrinsic good. As McTaggart noted, in a world where nothing else is good a person can believe falsely that, say,

6 Hayward also affirmed the recursive idea in his perfectionist critique of Sidgwick’s hedonism (Ethical Philosophy of Sidgwick, pp. 227–8, 230). 7 Like Moore, Carritt characterized virtue instrumentally (TM 137; EPT 85–6) but thought motives like charity and sympathy are intrinsically good.

220

british ethical theorists from sidgwick to ewing

pleasure is good; if so, his virtuously desiring pleasure can be good even though nothing else is (NE II 421). False beliefs aside, however, virtue can be good only if some other things are good or right; it is in that way a derivative moral property. And its being so allows a distinctive argument for other intrinsic values: if the desire for x is virtuous and good, then x must be good as well. Rashdall and Ross used this argument to confirm that pleasure is good and pain evil: if kindness is good and cruelty evil, their objects must likewise be too (TGE I 136–7, 214, II 41–2; RG 135; also KET 11–12). And at one point Moore used it to argue that beauty is good (EE 90–1). The account is nicely illustrated by Rashdall’s discussion of humility (TGE I 204–7). This virtue is sometimes equated with underestimating your character or achievements,8 but he recognized that a theory that values knowledge cannot find intrinsic value in a false belief. The virtue must instead involve your attitudes, and here what it opposes is ‘any habitual dwelling with satisfaction’ on your merits, where such satisfaction, we may add, is consistent with self-underestimation. If the thirdbest physicist in the world mistakenly believes he is tenth-best but nonetheless constantly boasts about his physics achievements and pats himself on the back for them, he is not humble. As for the positive good in humility, that, Rashdall said, rests on two other attitudes. A truly good person is too concentrated on becoming better, and too aware of the flaws in himself that hinder that, to take much pleasure in his current condition. And he cares too much that others become better to be pleased by the gap between himself and them. If he does underestimate his merits, that is because his attention is focussed, virtuously, on more important things. Rashdall sometimes argued for the goodness of virtue on a very abstract ground. He said a hedonistic theory like Sidgwick’s is irrational because it assigns ‘a different end to the individual and to the race’. It tells each person to subordinate his own good or pleasure to duty, or to promoting the good of all, but equates their good with the sum of their pleasures, ignoring the very duty it says they should give priority; it therefore fails Kant’s test of being ‘ “capable of serving for law universal” ’. This defect can be avoided, he argued, only if each person’s virtuous pursuit of the good of all is a good in his life, so in promoting the universal end he also promotes his own (‘PS’ 215–17; TGE I 54–7; E 63–4). This is the argument Sidgwick addressed pre-emptively, and he did so effectively. Since an individual is distinct from the larger whole, he said, there is no reason why his good may not differ from that of the whole and his duty be to prefer the latter (ME 404); right action need not make your life best.9 In later writings Rashdall allowed that this position is not contradictory but added a further argument: that we cannot be motivated to do our duty unless we believe it will promote our good (TGE I 57–9; E 64–5). But this was a strange argument for him to make, since it seems to assume a 8 9

Driver, ‘Virtues of Ignorance’. In giving this reply Sidgwick set aside his later concerns about the dualism (ME 404n).

moral goods

221

version of the psychological egoism he had argued incisively against in his first chapter (TGE I 23, 28–9, 38–43). If we can be motivated by ‘disinterested desires’, as he there held, why not by a desire to do what is right?10 The more persuasive, though as always not compelling, arguments for the value of virtue are intuitive, and Rashdall made them too, appealing to ‘one’s own moral consciousness and that of others’ (TGE I 70). Moore, characteristically, presented the recursive principles as simply self-evident, whereas Ross gave his example of the two worlds with equal amounts of pleasure, in one of which people are virtuous and in the other of which they are vicious. To the objection that the virtue in the first world will lead its inhabitants to act in ways that increase each other’s pleasure, he replied that we can imagine this increase countered by less favourable natural conditions, such as more disease. Even so, he claimed, the first world is better (RG 134). There were also simpler intuitive claims, such as Broad’s that malicious pleasure is evil (FT 234), and the idea that virtue is intrinsically preferable to vice is attractive on several levels. The principles defining virtue and vice, for example that it is good to love what is good, are intuitively appealing in the abstract, as are their implications and the way the principles explain them. Certainly the view that virtue is good in itself has been embraced by many philosophers from ancient Greece to the present.

10.2 Forms of Virtue An important aspect of the school’s view appears in Ross’s distinction between his second and third forms of virtue, the desire to produce something good and the desire to produce pleasure for another person. Since another’s pleasure is good, the third form may seem to be included in the second, but Ross distinguished them as follows. Though you can desire another’s pleasure as something good (the second form), you can also desire it without thoughts of goodness but just as a pleasure (the third). Then you desire something for a property that makes it good, its being pleasant, but do not think of the property as good-making; even so, your desire is virtuous (RG 161–2). Given this distinction, it is surprising that Ross did not say it is also virtuous to desire virtue, knowledge, and distribution in accordance with merit without thinking of them as good. He may have thought that, just as virtue and knowledge call for a specific emotion of admiration that requires a belief that they are good, so they call for a special form of desire. But even if they do, they can surely also be objects of ordinary desire, and he denied that distribution by merit calls for admiration (FE 323). He would have been more consistent if he had held that whenever something is good, it is virtuous to love it either because it is good or without any thoughts about goodness. He could have taken a similar view about the right: if promise-keeping is

10

For a similar point see Skelton, ‘Ideal Utilitarianism’, p. 56.

222

british ethical theorists from sidgwick to ewing

right, it is virtuous to want to keep a promise both because it is right and from a simple aversion to promise-breaking; Broad thought both these motives are good (CE 138). In affirming his third form of virtue Ross rejected Kant’s view that the only good motive is the motive of duty. He thought the motive of duty is the intrinsically best motive (RG 164; FE 116–17, 206, 303–5; KET 3) but others such as love for a particular person are almost as good; in one discussion he imagined the motive of duty having ten units of value and love eight (RG 170–1). Complete virtue requires both kinds of motive, so someone who cares only about duty is not fully virtuous. This anti-Kantian view was common in the school. Sidgwick said common sense recognizes other virtuous motives than duty (ME 204–5, 223, 225–6, 239, 345, 395), while Rashdall rejected the ‘revolting and inhuman Stoicism’ of Kant’s ideal and commended acts ‘in proportion as they are inspired by a desire of objects which Reason pronounces intrinsically good, although the man may not pursue them consciously because Reason pronounces those objects to be good’ (TGE I 119–20, 123, also 121–2, 124–9, II 73–4; ICE 117; and McTaggart, SDR 158–9, NE II 466). Moore said the New Testament values motives such as pity that Kant termed ‘mere “natural inclinations” ’, and thought the motive of duty ‘certainly has not more value’ than them (PE 178–80; also Russell, ‘EE’ 34–5). Prichard, Carritt, and Ewing likewise said the best acts are motivated both by duty and by some non-moral motive such as sympathy (MW 3, 11, 15–16, 59–61, 159–60, 216; TM 47–8; EPT 85–6; ‘PKE’ 50–2; E 53–4, 147–8; ST 139–40; VR 182–3). At one point Moore went further, saying the motive of duty, and even the desire for a good because it is good, is the least good of good motives and ‘only to a slight degree a good thing in itself ’ (EE 154); others thought acting from duty is sometimes less good than acting from a non-moral motive such as love. Ewing said this about a father looking after his children (E 53; ST 125), and Broad held, more generally, that some acts are better when done from conscientiousness and others from personal affection (CE 79, also 40, 73–4). Some who took this view thought conscience should always be present in a background role. Rashdall thought it best to act from some motive other than duty, so ‘an ideal love of mankind would supersede all sense of duty as such’ (TGE I 128). But we less-than-ideal agents should have the sense of duty ‘in the . . . “fringe” of consciousness’, where it can act as ‘a consenting party to all our actions’ (TGE I 121, also 126–7) by monitoring our other motives and approving some while inhibiting others. We need not think consciously of duty if non-moral motives will lead us to act rightly, but it should be standing by, ready to block any impulse that would lead us astray.11 Ewing had a similar view (‘PKE’ 50–2; E 53–4, 147–8), but he and Rashdall justified this ‘filtering’ role for conscience mainly instrumentally, saying non-moral motives often overvalue some goods at the expense A similar view is proposed from a Kantian direction in Herman, ‘On the Value of Acting’; and Baron, ‘The Alleged Moral Repugnance’. 11

moral goods

223

of others and so need correction. If this is the only justification, however, it does not tell against acts done just from non-moral motives when those do weigh all relevant considerations proportionally. Carritt and Ewing thought the value of acting from duty is greatest when there are contrary impulses to overcome (EPT 85; ‘PKE’ 54–6), but Broad said we would hardly admire someone who had continually to struggle against temptations to murder and rape. He suggested the advantage of successful struggle is only epistemic: only then can we be certain a person’s commitment to duty is strong enough to defeat competing desires (FT 199–200). This was also Ross’s view. He thought a person can have a ‘surplus’ of motivation to do the right thing, as when his sense of duty is strong enough to lead him to act rightly but he also has a supporting non-moral motive (RG 172). Here what matters is just how strong the motive of duty is in itself, and it can be equally strong with or without support. The overall value of a person’s motives is just the sum of the values they have individually, and though the strength of any one can be hard to assess when it does not conflict with others, there is a truth about what it is. Another issue concerns the comparative importance of occurrent virtuous desires or feelings and lasting traits of character. Aristotle gave priority to the latter, saying that to be virtuously done an act must issue from a ‘firm and unchangeable character’.12 Sidgwick too made virtue primarily a matter of permanent dispositions (ME 222), as did Rashdall when he valued the ‘settled bent of the will’ or acts ‘in which character or disposition takes effect’ (‘PS’ 224; TGE I 63). But Moore emphasized the value of particular desires for a good outcome or admirings of a good object (PE 177, 203–4); he thought dispositions as such have no value, and Ewing did too (ST 180; VR 218). Ross alternated. At one point he explained ‘morally good’ as ‘good either by being a certain sort of character or by being related in one of certain definite ways to a certain sort of character’ (RG 155), but elsewhere he said virtuous action is action from certain occurrent desires, with no reference to dispositions (RG 156–7, 160, also 134–5; FE 291–2). The second view fits better with his general account. If virtue involves responding appropriately to items with other moral properties, it should be found mainly in individual occurrent responses.13 There may be some additional intrinsic value in virtuous dispositions, and Ross thought there is. He said that alongside the value in ‘actual felt desires and emotions’ there is some value in ‘relatively permanent modifications of character even when these are not being exercised’. But our understanding of a virtuous character is derivative, following from a prior understanding of the individual virtuous attitudes it is a disposition to have (FE 291–3). Different members emphasized different forms of virtue. Rashdall, Prichard, and Ross spoke most of desiring or actively willing the good, Moore most of admiringly 12

Aristotle, Nicomachean Ethics, 1105a32–3. I argue more fully for the primacy of occurrent virtuous attitudes in ‘Virtuous Acts, Virtuous Dispositions’. 13

224

british ethical theorists from sidgwick to ewing

contemplating it. But most recognized all three forms of virtue and vice, and therefore faced a special issue about malicious pleasure, or pleasure in another’s pain. The claim that such pleasure is evil has a weaker and a stronger form. The weaker says malicious pleasure is evil insofar as it is malicious but good insofar as it is pleasure; though it may be and usually is evil on balance, it is good in one respect. The stronger form says malicious pleasure is entirely evil; any good-making tendency its being a pleasure might have is cancelled by its being a pleasure in pain. Since they did not address this issue directly, the school’s statements are often ambiguous between these two forms. Ross has been read as making the stronger claim when he said, ‘a state of pleasure has the property not necessarily of being good, but of being something that is good if the state has no other characteristic that prevents it from being good’ (RG 138).14 But ‘good’ here could mean just ‘good on balance’, and several features of his discussion point to the weaker view. He compared the relation between a state’s goodness on balance and its good-making properties to that between a duty proper and the prima facie duties it rests on (RG 137–8); but if an outweighed duty remains present so should an outweighed property of goodness. The weaker claim also better fits his view that intrinsic value depends only on intrinsic properties, since a natural extension of that view says a property’s good-making tendency is independent of any other properties it is co-instantiated with. And at one point he positively assumed the weaker view. Arguing for the infinite superiority of virtue over pleasure, he cited the pleasure of cruelty and said, ‘If the goodness of pleasure were commensurable with the goodness or badness of moral disposition, it would be possible that such a pleasure if sufficiently intense should be good on the whole’ (RG 151, also 153–4; FE 274); this requires malicious pleasure to still be good as pleasure.15 Some may find this weaker view less intuitive and say there is nothing at all good about malicious pleasure. But the stronger view is hard to generalize coherently while the weaker one better fits a recursive structure where we first identify base-level values such as pleasure and pain and only then consider the values of attitudes to them.16 Like other values, virtue and vice admit of degrees, and a full account must explain how. Here some began with an emendation. The claims that it is good to love the good and evil to hate it suggest that the intermediate state of being indifferent to a good or evil has intermediate or neutral value. But indifference to another’s pain involves callousness, which is a vice, and indifference to goods is sloth. Rashdall, Ross, Broad, and Ewing therefore held that indifference to goods or evils is evil (TGE I 73; FE 302, 326; EMP 704; E 148), and though they did not say this, it follows that very mild loves of good and hatreds of evil are also evil. In Foundations Ross rejected his earlier view that an action’s moral value depends just on the motives it is done 14 15 16

Stratton-Lake, ‘Pleasure and Reflection’, p. 126. Ewing too endorsed the weaker view in MP 69. I elaborate these last points in Virtue, Vice, and Value, pp. 144–52.

moral goods

225

from and said it can also be affected by motives the action lacks; if some that should be present are not, the action can be bad (FE 306–8). Ewing used a claim about indifference to argue for the necessity of the motive of duty. If this motive is not present in at least a background role, you show an indifference to right and wrong that deprives your act of value (‘PKE’ 50–1); this is a better argument for the necessity of conscience than the instrumental one about overvaluing. More generally, the degree of value of a virtuous or vicious attitude depends on two factors: its intensity and the degree of value of its object. Other things equal, a more intense love of a good or hatred of an evil—for example, more intense compassion for another’s pain—is better. Also other things equal, love of a greater good or hatred of a greater evil is better. And the interaction of these factors is plausibly governed by an ideal of proportionality, so a fully virtuous person loves goods and hates evils in proportion to their degrees of good or evil, feeling, for example, twice as much compassion for a pain that is twice as great. Rashdall affirmed this ideal when he said our love of others should include ‘a desire of various goods for them in proportion to their relative value’ (TGE I 128, also 125–7, 273–4, II 82; Moore, EE 148–51), as did Broad when he said it would be unfitting to desire the pleasures of the table as intensely as the beatific vision (FT 245–6; also CE 173, 204–5, 293; Ewing VR 211–12). A parallel view says you should care about deontic factors in proportion to their right-making tendency, so if you have a stronger duty to help accident victims than to keep a promise you should want more to do the former. These claims are in fact needed if there is to be an appropriate overlap between the theories of virtue and of the right. It should be that an ideally virtuous person with all relevant information will always do what is objectively right, and that follows if he always desires greater goods more than lesser ones and to fulfil stronger duties more than to fulfil weaker ones. Ross noted this overlap (FE 308–9), and it is what led Rashdall to say that considering your own virtue now cannot affect what you think is now right. There was some dissent from the proportionality ideal. Ewing said it was not unfitting for him to care more about his mother’s happiness than about a stranger’s (DG 159; ST 108), but that view can be accommodated by the agent-relative claim that from his point of view his mother’s happiness is a greater good. Ross objected that we do not think worse of a soldier whose duty it is to rest before battle if his desire for the pleasure of resting is stronger than his desire to fulfil that duty. It is a key feature of this example that the soldier’s disproportionate desire will not lead him to act wrongly; if wanting pleasure prevents you from doing your duty or from producing a greater good, Ross thought your motives are bad (RG 167). But it is hard to see how this fact about a motive’s effects bears on its intrinsic value, and the simpler way to make virtue and rightness overlap is to apply the proportionality view across the board. Then the soldier’s desires are less than ideally virtuous—which does not mean they are vicious—but in the circumstances excusably so.

226

british ethical theorists from sidgwick to ewing

A final question is how the value of virtue compares to that of other goods. Ross took the extreme view that virtue is infinitely or incomparably better than pleasure or knowledge. In The Right and the Good he said virtue starts at a point higher on the scale of value than pleasure ever reaches (RG 150); in Foundations he said it has a different kind of value and is on a different scale (FE 274–5). But his argument for this conclusion from the badness on balance of the pleasure of cruelty, which we discussed above, is fallacious. Imagine that a pleasure in another’s pain is on balance evil. Increasing its intensity will increase its goodness as pleasure, but if its evil as malicious also increases at the same or a greater rate, it will remain evil on balance. So we need not say vice is infinitely more evil to always condemn malicious pleasures. In Foundations Ross worried that if virtue and pleasure are on the same scale, some large amount of pleasure will outweigh some perhaps small amount of virtue, which he thought it cannot do. But others defended more moderate views on which virtue can be outweighed. Rashdall associated the view that virtue is infinitely better with Cardinal Newman, who said it would be less evil for all humankind to die ‘in extremest agony’ than that ‘one soul . . . should commit one venial sin’.17 Rashdall presented several cases against this ‘tremendous judgement’, beginning with ones involving other people’s virtue. If virtue is promoted by curates, knowledge by education, and relief of pain by hospitals, must we give all our charitable contributions to an Additional Curates Society? If the only way to save several people from torture is to bribe a Chinese mandarin and thereby further corrupt his character, is it wrong to do so (TGE II 43–4)? Carritt gave a similar example to this last, involving bribing a committed Nazi (EPT 83–4), while other examples of Rashdall’s concerned one’s own future virtue. He said it can sometimes be right to adopt a profession that in the long run will worsen your character (TGE II 47). If you become a nurse or a surgeon you may do much to relieve others’ pain, but the constant exposure to pain may harden your character and make you less warmly sympathetic; that does not make this career choice wrong. Broad and Ewing gave similar examples involving a daughter’s developing moral defects from tending her peevish invalid mother or a politician’s being corrupted by the power he needs to accomplish his worthy goals (CE 273–4; E 29). Ewing thought it ‘priggish, and indeed selfish in a bad sense’ to decide these cases primarily by their effects on your own character (E 29). Though Rashdall denied that virtue is infinitely better than pleasure or knowledge, he nonetheless thought it the highest good in the following sense: the value of a virtuous or vicious attitude to an object is always greater than the value of that object. Though virtue must aim at some good, ‘the will which wills that good is a greater

17

Newman, Certain Difficulties, vol. 1, p. 240.

moral goods

227

good than the good which it wills’, or, given an act that causes pain, ‘the character or disposition which this act shows is worse than the pain which it causes’(TGE II 106, I 94, also I 72, 76, 101, 136–7, 174–5, 190, II 42; E 64, 72; ICE 115). But Moore at least sometimes took the contrary view, and gave a compelling argument for it. If we consider the combination of a real evil and virtuous hatred of it, for example one person’s pain and another’s compassion for it, there ‘seems no reason to think that . . . the total state of things is ever positively good on the whole’. On the contrary, ‘the amount of this real evil is always sufficient to reduce the total sum of value to a negative quantity’. The combination of pain and compassion for it is evil, and that means the compassion must be less good than the pain is evil. Moore therefore rejected those solutions to the problem of evil that say evils such as pain exist because they are necessary conditions for certain virtues: ‘It is not a positive good that suffering should exist, in order that we may compassionate it; or wickedness, that we may hate it’ (PE 220). The virtues cannot justify the evils because they have less value than their objects. Moore did not apply this view across the board, because he thought the appreciation of beauty is a greater good than the beauty itself. But the view has many attractive implications. If a torturer is taking malicious pleasure in his victim’s pain and you can eliminate either his pleasure or the victim’s pain, you surely ought to eliminate the pain. If a teacher teaches his students from a virtuous desire for their knowledge and they acquire knowledge, you should be more pleased by their knowledge than by his virtue—it is the point of the exercise—and the same goes for him. If he cared more about his virtue, he would be not virtuous but priggish or morally self-indulgent.18 That follows if his virtue is less good, since then his priggishness involves a disproportionate preference for a lesser over a greater value. Finally, if the value of an attitude is always less than the value of its object, the results of repeated applications of the recursive principles, for example to a love of a love of a love of . . . a good, have progressively less value, perhaps diminishing toward zero.19 Ross and to a lesser extent Rashdall thought virtue is the greatest good, but Moore suggested a better view on which virtue is in a certain sense a lesser good. It is not that every instance of virtue has less value than any instance of another good or evil; your compassionate efforts to relieve another’s great pain may be much better than a minor pleasure for you. But the idea that a virtuous or vicious attitude has less value than its object fits virtue’s status as a derivative moral property, involving attitudes to items with other, previously given properties. Your virtue is an important element in the value of your life, but in any particular situation you face it is not the most important thing and should not be your primary concern.

18 19

Williams, ‘Utilitarianism and Self-Indulgence’. For a fuller presentation of these arguments, see my Virtue, Vice, and Value, Chapter 5.

228

british ethical theorists from sidgwick to ewing

10.3 Personal Love Love, or ‘personal affection’ (PE 203), was one of Moore’s two chief goods and McTaggart’s highest good. But it was also discussed by others. Sidgwick took love to have two main elements: ‘a desire to do good to the object beloved’, and, even more centrally, ‘a pleasurable emotion, which seems to depend upon a certain sense of union with another’ and includes ‘a desire of the society of the beloved’ (ME 244–5). This account fits the common view that the object of a desire or emotion is always a state of affairs, in this case that the person you love be happy or that she spend time with you; to love her is to want certain states because they involve her. Sidgwick thought the two elements can conflict, as when your insistent desire for her company interferes with her pursuit of her good. But an unselfish love balances them properly. I have mentioned Ross’s view that love combines the goods of virtuous concern for another, knowledge of her character, and pleasure arising from that concern and knowledge (RG 141); strangely, he did not mention the pleasure of her company. Given his view that virtue is the greatest good, he presumably thought the virtue love involves is the main source of its value, and that in effect was Moore’s view. He introduced the good of love immediately after discussing aesthetic appreciation and said it involves the same elements of ‘appropriate emotion, cognition of truly beautiful qualities, and true belief ’, but with the difference that now ‘the object itself is not merely beautiful, . . . but is itself, in part at least, of great intrinsic value’ (PE 203). Love was his main example of the goodness of loving the good, and its more specific object is some of the beloved’s mental qualities, especially as given bodily expression ‘by looks, by words, or by actions’. Moore thought loving her mental qualities alone, if that were possible, would have just limited value, perhaps less than appreciating her physical beauty; the mental qualities’ main role is to combine with their outward expression to form a whole that is the object of truly valuable love (PE 203–4). What are these mental qualities? The only good Moore had discussed to that point was aesthetic appreciation, and he therefore said they ‘consist very largely in an emotional contemplation of beautiful objects; and hence the appreciation of them will consist essentially in the contemplation of such contemplation’, so what you love is primarily your beloved’s love of beauty. The very best attitude to her appreciates her appreciation of other people; but what she appreciates in them is again their appreciation of beauty, so a reference to beauty remains fundamental (PE 204). We can see the appeal of this view to the Bloomsbury group, as if the supreme expression of love is admiring another’s taste in paintings and her appreciation of others’ taste. But it leads to a highly truncated picture of love, which at its best surely involves a positive attitude to many other aspects of her good such as her happiness, intellectual development, and virtue.

moral goods

229

There was another place where Moore’s picture was truncated. As in his account of aesthetic appreciation, he equated love primarily with the admiring contemplation of another’s good qualities, or with one of Keynes’s ‘timeless, passionate states of contemplation’, and spoke little of desiring or actively pursuing her good. He nowhere suggested that you might want to do things with your beloved, such as have and raise children or decorate a house. This is partly because he did not recognize achievement as a good and could not see joint achievement as a desirable aim, but also because he emphasized contemplative over active forms of love. (In an 1899 Apostles paper he said, ‘Love of others is a thing of very great value, and it may be very strong although it does not lead to any actions. There is, indeed, no reason why it should’.20) His account of love therefore lacked any element of the erotic. He seems to have held that part of loving another is admiring her physical beauty, but even in romantic love you only gaze at her beauty, as it were from across the room. You are not aroused by it or try in any way to engage with or possess it. This was reflected in his condemnation of ‘lasciviousness’, by which he meant any sexual desire. He thought such desire involves ‘cognitions of organic sensations and perceptions of states of the body, of which the enjoyment is certainly an evil in itself ’, because it includes ‘an admiring contemplation of what is ugly’. And its worst forms involve enjoyment of the same state of mind in another, which is a love of something evil (PE 209–10). But given his reductive analysis of the aesthetic concepts, to say it is evil to enjoy certain bodily sensations because they are ugly is to say it is evil to enjoy them because it is evil to enjoy them, which does nothing to explain why enjoying them is wrong. The young Moore had been extremely puritanical about sex, reading papers to the Apostles condemning masturbation and sex for any purpose other than procreation,21 and those attitudes are reflected both in Principia’s claims about erotic desire and in its more general picture of love as essentially contemplative rather than active. Moore also found no intrinsic significance in the partiality that is characteristic of love. You are to admire qualities of your beloved with a property of goodness that is agent-neutral, which means that if someone else appears with the same good qualities, you have the same reason to love her, and if she has better qualities, you have more reason to love her and should switch your affection to her. Moore could say you know your current love better and therefore can appreciate her qualities more fully. But that is only an instrumental reason to continue your love and does not capture the view many have that attachment to a particular person is essential to love, so you should not ‘trade up’ the minute someone with better qualities appears.22

20

Quoted in Levy, Moore, p. 203. Levy, Moore, pp. 139–46; Regan, Bloomsbury’s Prophet, p. 39. These views seem to have persisted; after spending time with him in the 1940s, Blanshard noted ‘a strong strain of puritanism in his attitudes to morals’ (‘Autobiography’, p. 86). 22 Vlastos, ‘The Individual as Object’; Nozick, Examined Life, Chapter 8. 21

230

british ethical theorists from sidgwick to ewing

This view can be captured with agent-relative claims about value. We can say a shared history with another person gives her happiness, knowledge, and achievement more value from your point of view, so it is proportional and therefore virtuous to care about them more; to abandon her for a slightly more attractive stranger would be positively disproportionate. But Moore’s account of ‘good’ as unanalysable does not allow agent-relativity, and his picture of love is therefore wholly impersonal, with no intrinsic value for particular attachments. It was influential within the Bloomsbury group, who discussed intricate questions about the value of love;23 it also lies behind the emphasis the Schlegel sisters in E.M. Forster’s Howards End place on ‘personal relations’.24 But it is a picture from which the ‘personal’ is largely absent. This is not true of McTaggart’s picture of love (SHC 212–13, 278–80; NE II 147–61). Unlike Sidgwick and Moore, he did not think the object of love is a state such as a person’s being happy; it is the person herself, and involves a sense of union with her. (In unrequited love the union is felt only on one side, but it is still felt.) Love is often accompanied by a desire for her happiness and admiration for her qualities, but it need not be; thus it is possible to love someone while desiring her ill-being, and the view that love should always involve moral approbation is ‘utterly mistaken’. You can love someone you know is wicked, and, if you do, your emotion is no less truly love than if she were virtuous (NE II 150), nor is it any less good. Whereas admiration should be proportioned to the value of its object and can be criticized if it is not, ‘if the love does arise, it justifies itself ’ and is beyond criticism. Your love for a particular person may be caused by some of her qualities, but, contra Moore, these qualities need not be good or important ones. ‘To love one person above all the world for all one’s life because her eyes are beautiful when she is young, is to be determined to a very great thing by a very small cause’ (NE II 153; also SHC 279). Though you love her ‘because of ’ a property, you do not you love her ‘in respect of ’ it; you just love her. And your love’s value is independent of its initial cause. For McTaggart these points had an important implication. If, having started to love someone because you believed she had certain qualities, you learn that she no longer has or never had them, you have no reason to abandon your love. Though admiration is no longer justified if someone loses the properties you admired her for, ‘love, if it were strong enough, could have resisted, and ought to have resisted’ (NE II 154; also SHC 279–80); it should, in other words, be forever. His account of love was therefore at the opposite extreme to Moore’s. Whereas Moore’s was entirely impersonal, equating love with an appreciation of qualities anyone could share and implying that it would be right to abandon a beloved for someone a little bit better, McTaggart’s was entirely personal, saying love is for an individual apart from any qualities. He did not think you cannot love more than one person; in the timeless ‘absolute reality’ his metaphysics posited, each self will love 23 24

Keynes, ‘My Early Beliefs’, pp. 53–8. See Sidorsky, ‘Uses of the Philosophy of G.E. Moore’.

moral goods

231

every other self it directly perceives, of which there may be many (NE II 155). But he did think that, once love has begun, it should be unaffected by changes in the one loved. Surely the sensible view lies between these extremes. Love is not just for general qualities but involves attachment to an individual and a degree of loyalty to her. But it is also not independent of her qualities. If someone who was public-spirited and generous becomes callous and greedy or who was lively becomes morose, that can be a reason to change or end a relationship. We accept this with non-romantic friendships, which we move in and out of as our interests and enthusiasms change, and it applies to romantic love too. That should not be affected by every change in your beloved; you should often find new qualities endearing because they are hers. But it need not survive every change, nor must your specific tie to her be, as McTaggart thought, independent of all her qualities; it can be based on some uniquely individuating ones that arise from your shared history with her. Once you and she raised those children and decorated that house, no one else can be the very person who did those things with you. Her having participated in those aspects of your past is a property no one else can now have, and to love her for that property is to love only her.25 Broad made this point about history when, in a critique of McTaggart’s view, he said a mother can love her children ‘in respect of the relational quality of being her children’, and that sexual love ‘at later stages, is determined, in the main, not by the present qualities of the beloved, but by the traces of innumerable actions and experiences in common’ (EMP II 123, also 117–18). But he mostly objected that McTaggart’s view was too high-minded. In saying this he felt like ‘one of the Protestant underworld brawling in a high church during a celebration of the mass’, but ‘love is, in some respects, so sublime, and, in others, so ridiculous . . . that it is not easy to keep a just mean between cheap cynicism and muddled mysticism’. McTaggart, he thought, approached dangerously near the latter extreme (EMP 129). Against McTaggart’s view that love involves just a sense of union, Broad said it is compound of many elements, ‘cogitative, conative, and emotional’ (120), where these are different in different kinds of love; thus, sexual love differs from a mother’s love for her child. Even the same love is, and should be, different at different times. In romantic love the sexual desire that was central in early phases becomes less important through time. In a lovely passage he wrote: If we may compare prolonged and successful sexual love for a person to the course of a river from its source to the sea, it begins as a violent torrent in a narrow bed full of rocks and shallows; in its middle reaches it receives many tributaries; and in its later stages it becomes a calm wide deep stream.

25

For a fuller discussion see my Best Things in Life, Chapter 7.

232 british ethical theorists from sidgwick to ewing But he could not end on this positive note, adding, ‘Too often, of course, there is no such happy ending, and the stream peters out into the shallows of mere habitual toleration or the swamps of mutual irritation and frustration’ (EMP 122). He also rejected McTaggart’s view that love is beyond criticism. He noted that we talk of ‘blind’ or ‘doting’ affection and call certain men ‘uxorious’; these are not compliments, while falling in love with someone you know has a bad character is ‘definitely unfitting’ (EMP 124–5). Nor must love, once begun, never cease. This is true to a degree of maternal love, whose basis is primarily historical. But if a romantic partner’s character changes radically soon after your relationship with her begins, you are not to blame for ending it. Broad thought you would be to blame if the relationship was long-standing, and we may find that extreme: important though history is, it can be outweighed by a sufficient loss of other qualities. Nonetheless, Broad rightly saw that love is not directed just at an individual as an individual but also concerns general qualities and can change when they do. That love and friendship are intrinsic constituents of a good life is a claim many will endorse, and the school’s writings contain the elements of an attractive account of them. From Sidgwick and Broad comes the idea that love involves a set of desires and emotions about a person, such as a desire for her happiness and pleasure in her company. Broad added that what ties you specially to her and makes you want her happiness more than a stranger’s is her having participated with you in a shared history. Your love is not based just on qualities others could share, as Moore thought, but also on historical ones that, once she has them, no one else can have; this gives a less grandiose explanation than McTaggart’s of why you will not trade up for someone better. Ross saw that loving relationships are not a distinct good but, at their best, instantiate other intrinsic goods to a high degree: pleasure in a loved one’s company, understanding of her character, virtuous desires for her, and, though he did not mention this, joint achievements with her. The most important of these may indeed be virtue: it is toward those you love that you most have the altruistic concern you should have for everyone. And, though none in the school said this, your partiality to those you love can itself be a form of virtue. If there are agent-relative values, their pleasure and achievement can be greater goods from your point of view, so your wanting them more than a stranger’s is positively proportionate and good. Then the partiality that is essential to love is also an element in its value.

10.4 Moral Desert The good of desert involves getting what, or being treated as, you deserve. You can deserve different things on different grounds: income for economic contribution or effort, criminal punishment for breaking the law, or happiness for being virtuous. We will discuss economic and criminal desert in Chapter 11; here our focus is moral desert, or desert based on your moral character. It is the central form of desert and, like love, involves the good of virtue.

moral goods

233

What you deserve morally is happiness or suffering, and you deserve it for your virtue or vice. The good here is the combination of the two in one life and involves an organic unity, as in Moore’s account of ‘vindictive punishment’. If a vicious person suffers pain, his vice is evil and so is his pain, but the whole composed of the two is good as a whole, and more good than the pain is evil, so his pain makes the situation all things considered better (PE 214–15, also 221). Moore’s statement was holistic, so the bad person’s pain remains evil as pain, and we could instead say its value is changed from evil to good. (It would not be plausible to say that when vice is accompanied by pain the vice ceases to be evil.) But here the holistic view is preferable. What should our attitude be to a vicious person’s pain? On the variability view it should be just pleasure: his pain is purely good, and the appropriate attitude to it is therefore purely positive. But that seems wrong. Our attitude to deserved pain should be sombre or subdued, and the holistic view explains why. Though we should be pleased by a vicious person’s pain as deserved, and more pleased by it than we are pained, we should still be to some degree pained by it as pain. Moore’s case involves desert of suffering, but you can also deserve happiness for being virtuous; then your virtue and happiness are both good, but there is a further holistic good in their being combined in one life. There are also desert-evils. The combination of virtue and suffering is evil as a combination, so there are two reasons why a good person’s pain is evil, as is the combination of vice and happiness. Ross understood desert in this way and defended its value by describing two worlds, each containing the same totals of virtue and vice and pleasure and pain, but where in the first the virtuous are happy and the vicious miserable and in the second the reverse is true. Is the first world, he asked, not better (RG 138, also 72)? And though he rejected many of Moore’s applications of the doctrine of organic unities, he accepted this one, saying ‘The surplus value of the first whole arises not from the value of its elements but from the co-presence of goodness and happiness in one single person, and of badness and unhappiness in another’ (RG 72). Carritt and Broad too valued moral desert (TM 109–10, 139; FT 77, 133–4, 204–5). Though no members noted this, the principles generating these desert-values parallel those for virtue. In the latter case positive attitudes to positive values and negative attitudes to negative ones are good and their opposites evil. Here the combination of two positive or two negative values, pleasure with virtue or pain with vice, is good, while combining a negative with a positive or a positive with a negative value is evil; again matching is good and its opposite evil. This supplies a further reason to favour the holistic formulation of desert. If a vicious person’s pleasure is still good as pleasure, we can explain the badness of his enjoying it by saying an evil is being joined with a good, as we could not if his pleasure ceased to be good by being undeserved.26 26 Lemos, Intrinsic Value, pp. 43–4; Slote, ‘Virtue in Self-Interest’, pp. 273–4. It may be replied that the badness of undeserved pleasure can be explained by the person’s getting what is ‘good for’ him even if it is not simply good. But our school denied that there is a relevant ‘good for’ concept.

234

british ethical theorists from sidgwick to ewing

The desert-values also parallel virtue in being governed by proportionality, so the best division of a fixed amount of happiness among people is proportioned to their degrees of virtue. Deriving ultimately from Aristotle,27 the association of desert with proportionality was recognized by Sidgwick and Rashdall (ME 279; TGE I 256) and affirmed by Ross, who called this good ‘the proportionment of happiness to virtue’ (RG 27, also 21, 26, 58–9, 153; FE 286, 288, 319, 323). And just as the value of a virtuous attitude is less than that of its object, so the value of getting what you deserve should be less than the value of what you deserve it for. Moore suggested this view when he said that if a vicious person suffers deserved pain, the value on the whole of the situation is negative (PE 215; also Ewing, VR 213); it is not better for there to be vice and punishment for it than for there to be no vice. It does not quite follow that the goodness of his getting what he deserves is less than the evil of his vice; the former could be somewhat greater, with the evil of his pain as pain tipping the balance. But it is again plausible that the value of a response to a value is less than the value of what it responds to, and Ross gave a more complex argument for this conclusion. He asked us to imagine someone whose life contains just the amount of pleasure appropriate to her degree of virtue, but some of whose pleasures are morally bad, for example, are malicious. He said we would think her life was better if she was without these pleasures and the vice they involve, even though there would then be less proportion in her life because she would have more virtue and less pleasure. But then the virtue must be better than the desert (RG 153–4). We may wonder why Ross valued only the proportioning to virtue of pleasure and not of other goods such as knowledge. Desert involves a kind of fittingness, where one good calls for or is complemented by another, and whereas pleasure seems a fitting reward for virtue, knowledge somehow intuitively does not. Knowledge may be deserved for a specific form of virtue: if two scientists seek a discovery, one from love of knowledge and the other only from a selfish desire for fame, it may be better if the first makes the discovery. But this is a circumscribed form of desert of limited importance, and there is likewise no fit in the ideas that people deserve pleasure on the basis of their knowledge or virtue in the future on the basis of their virtue in the past. Among the intrinsic goods the one that calls for a complement is virtue, and the one complement it calls for is happiness. Not all in the school valued desert. Sidgwick did not, while McTaggart left it off his list of goods (NE II 412) and rejected a Moorean account of retributive punishment as ‘patently false’ (SDR 163; also SHC 131, 140–1). Rashdall likewise dismissed the ‘old-world cry of blood for blood’ (TGE I 284), denied that there will be punishment for its own sake in the afterlife, and argued that the justification of criminal punishment must rest on its consequences.28 We could in principle reject 27

Aristotle, Nicomachean Ethics, 1131a30–b16. In a passage that cited Moore’s organic unities view, he did say that even when punishment fails to reform the offender, ‘Wickedness humbled and subdued, though it be only by external force, is a healthier 28

moral goods

235

retributive desert but hold that the positive good of happiness should be distributed in accordance with virtue (TGE I 256), but he denied this too, saying it is ‘childishly unreasonable in so far as it mean[s] that every individual should be assigned sugarplums in proportion to his moral or other “merit” ’ (TGE I 263). If a good person will use some resource as a means to higher goods, he has a greater claim on it than someone who will use it only for sensual enjoyment. But in the distribution of lower pleasures there is no reason to favour the more virtuous; you should not offer a morally better guest champagne and a less good one small beer (TGE I 257–8, 252). Ewing had a more complex view. He thought that if a virtuous or innocent person is punished and suffers pain, that is a significant injustice and a significant evil. But if a vicious person suffers or a virtuous one is happy, that is just a minor good, if it is good at all (MP xiv, 109, 110–11; ST 136–7). There is an asymmetry in value whereby being treated as you deserve is at best slightly good, but being treated as you do not deserve is seriously evil. The recent literature has distinguished two different forms of desert, often called ‘non-comparative’ and ‘comparative’.29 Non-comparative desert concerns the value of an individual’s having what he on his own deserves, regardless of what other people have. For any degree of his virtue there is an amount of happiness it makes ideally deserved; his enjoying that amount is non-comparatively best, while his having either more or less is less good and can in extreme cases be evil. This form of desert may be governed by an ideal of proportionality, so its further specification makes the best division of happiness among a group always proportioned to their degrees of virtue. But its subject is the value of combinations of hedonic and virtuestates within single lives. Comparative desert, by contrast, values a pattern of distribution across lives in which happiness is proportioned to virtue, and values it as a pattern, regardless of how happy any individual on his own is. Its concern is just that those who are more virtuous are proportionally happier, for example, that those who are twice as virtuous are twice as happy. The two forms of desert yield similar verdicts in cases where the total happiness to be distributed among a group is constant; thus both imply that the best distribution of the total is proportional. But their verdicts can differ when the total varies. Because it values only a pattern, comparative desert is indifferent between equally proportional distributions at different levels. Thus, if A is twice as virtuous as B, it is indifferent between their getting, respectively, 20 and 10 units of happiness, 200 and 100, and 2000 and 1000. But non-comparative desert may think one of these distributions is best, because it gives each what he ideally deserves, while the other

moral condition than wickedness successful and triumphant. This is the extremest point to which we can go with the advocates of the vindictive theory’ (TGE I 294). His meaning here is unclear, but it may be just that successful action from a vicious desire is more evil than the desire on its own. Feinberg, ‘Noncomparative Justice’; Hurka, ‘Desert: Individualistic and Holistic’; Kagan, Geometry of Desert. 29

236

british ethical theorists from sidgwick to ewing

two are less good. Or imagine that A and B both ideally deserve 200 units, but whereas A has 200 B has only 100. Both forms agree that it would improve the situation to move B to 200 units, but what if we cannot do that and can only move A down to 100? Non-comparative desert condemns this move because it involves only a loss of non-comparative value, but comparative desert approves it, as replacing a disproportionate distribution with a proportionate one. Our writers did not draw this distinction. The desert Moore analysed was only non-comparative, as for the most part was the one Rashdall rejected. But Ross’s ‘proportionment of happiness to virtue’ can be read either way, and without noticing it he alternated between the two. When he spoke of ‘the co-presence of goodness and happiness in one single person’ (RG 72), he was describing non-comparative desert; likewise when he said, ‘pleasure when deserved is good and pain when undeserved bad’ (RG 136). But his describing the good of justice as involving a ‘distribution between other people in proportion to merit’ (RG 26, also 21, 59; FE 288) suggests the comparative view, as he did most explicitly when he called proportional distribution a ‘situational good’, where these are not ‘activities or enjoyments resident in individuals, but would involve relations between individuals’ (FE 286). Ewing’s remarks too sometimes fit one view and sometimes the other (MP 43, 180n; E 172). Had he and Ross had been shown the distinction between non-comparative and comparative desert, they might have accepted both, but they never clearly separated them. Ross had another relevant view. He thought desert-judgements, of whichever form, concern a person’s life as a whole, so what has value ‘is a condition of things in which the total pleasure enjoyed by each person in his life as a whole is proportional to his virtue similarly taken as a whole’ (RG 58). An alternative view says his happiness in each particular phase of his life, such as his teens or his forties, should match his virtue just in that phase. (The idea of fitting happiness to virtue in each second is too extreme.) Then it will be better if someone who is virtuous in the first half of his life and vicious in the second is first happy and then unhappy than if the unhappiness comes first. This was not Ross’s view. He considered only the total virtue and total happiness in a person’s life, and was followed in this by Ewing (MP 37–8; E 172; ST 136–7). One effect was to make the prima facie duty to promote distributions in accordance with merit harder to apply in practice. However difficult it is to assess someone’s degree of virtue at a particular time, it is more so to include her past and future virtue and do now what will make the total happiness in her life match that. This is one reason Ross thought claims about desert are not relevant to the justification of criminal punishment, which he gave a different basis. We will discuss punishment in the next chapter; here the point is that a whole-life view of desert like Ross’s makes the value of desert more important for assessments of the values of states of affairs than for determining what it is right to do now.

11 Self-Benefit, Distribution, Punishment This chapter will discuss three further normative topics: the duty to promote your own good, the distribution of goods, and the justification of punishment. The last is the applied topic the school discussed most.

11.1 Promoting Your Good The duty to promote your own good can involve different goods and can be discussed in two contexts: as following from a more general duty to promote the good of all and as the sole duty in an egoistic theory. Let us begin with the first. Many in the school derived a duty to promote your pleasure from a duty that is strictly impartial; Sidgwick, for one, did. He recognized that the common-sense view on this topic is not impartial. It allows both some agent-favouring, where you prefer your own lesser to another’s greater pleasure, and some agent-sacrifice, where you do the opposite (ME 253–4, 348–9, 382; 431–2). He said common sense will on reflection accept an impartial principle, but that claim is dubious. Rashdall approved some agent-sacrifice, saying there is ‘beauty and propriety’ in the act of a mother who willingly gives up more for her child than he will gain by her sacrifice, but he thought that outside special relationships like that of mother and child preferring less good to more is wrong (TGE II 76). It is unclear, however, how in an impartial theory like his these relationships can affect what is right. Moore’s occasional suggestions that pleasure has no value imply that there is no duty to promote your pleasure, and he at best gave this duty little weight. Prichard and Ross too denied a duty to promote your pleasure, though they affirmed one to promote others’ pleasure. Kant had said we have no duty to pursue our own pleasure because we all already desire it,1 but Ross found this argument unpersuasive: that a mother has a strong desire to care for her children does not mean she has no duty to do so (FE 278). He and Prichard seem just to have found the duty unintuitive. Their view not only permits agent-sacrifice of pleasure but positively requires it, and

1

Kant, Doctrine of Virtue, p. 46.

238

british ethical theorists from sidgwick to ewing

requires it even when the pleasure you can give yourself is vastly greater than what you can give another; I said in Chapter 8 that this is highly implausible. Carritt, Broad, and Ewing argued against Prichard and Ross that there is a duty to promote your pleasure (EPT 113–16; CE 274–5; ‘SN’ 3, 20; DG 133, 150–1, 160–3; ST 112–14). They often derived it from an impartial duty, but Broad also considered the possibility that your duty to pursue your pleasure is stronger than your duty to pursue other people’s; he thought this view may be closer to common sense than either pure egoism or pure impartialism (FT 243–4). It permits agent-favouring but again positively requires it: if your duty to promote your pleasure is three times stronger than your duty to promote another’s, you not only may but must prefer one unit for yourself to two units for her. None of these proposals capture the intuitive view that when you are permitted to depart from the impartial pursuit of pleasure, by favouring either yourself or another, you are also permitted to abide by it and produce the greatest pleasure. That requires supplementing a prima facie duty to promote the pleasure of all with, what none in the school considered, prima facie permissions to promote and not to promote your own. As well as denying a duty to pursue your pleasure, Ross said there is no virtue in desiring your pleasure. Of his three forms of virtue the only one concerned with pleasure was the desire to produce it for others, and the only self-regarding virtue the desire for your knowledge or virtue (RG 134, 160); he also explicitly denied that wanting your pleasure is virtuous (RG 168; FE 282, 288, 302, 322). This denial fits both common sense—we do not call someone morally good for being devoted to his own enjoyment—and his later view that from your point of view your pleasure is not good. But it was in tension with his official view in The Right and the Good that your pleasure is good: if virtue involves desiring what is good, how can desiring your pleasure not be virtuous? That there is no virtue in desiring your pleasure was also at times suggested by Ewing (E 149; ST 113) and was again in tension with his belief that your pleasure is good. Carritt was more consistent, both accepting a duty to seek your own pleasure and saying ‘if pleasure is good, the pleased contemplation of my own pleasure would be good; and I think we do find some goodness in the capacity, even among the discontents of age, to prize rather than regret the memories of a happy youth’ (EPT 121–2). Whether desiring your pleasure is virtuous becomes more complicated if there are agent-favouring and -sacrificing permissions. A view that grants these should say that if you prefer your lesser to another’s somewhat greater pleasure, then despite acting permissibly you are not as virtuously motivated as if you preferred the other’s greater good; only then would your desires be fully proportionate. This claim does not threaten the overlap between virtue and right action, since if you were most virtuously motivated you would also act permissibly. The harder case is where you prefer another’s lesser to your greater pleasure, since there we do not want to say your motivation is less virtuous; though not morally better than if you weighed the two pleasures equally, it seems as good. We could maintain the overlap by saying your

self-benefit, distribution, punishment

239

pleasure is good when you desire it but not when you do not, and as good as your desire for it is intense; then your desires are ideally proportionate and good in either case. This move also yields a version of Ross’s view that it is not virtuous to desire your pleasure, since you are not more virtuous when you do than when you do not. As well as deriving the duty to promote your pleasure from an impartial principle, Sidgwick affirmed it on its own, as the foundation of egoistic hedonism; here it helped generate his dualism of the practical reason. In his version the dualism involves just an egoistic principle and a single contrary one, his axiom of benevolence, but conflict can also arise between egoism and the many principles in a pluralistic deontology. And though Sidgwick thought his conflicting principles yield strictly contradictory verdicts, that is not so. As we saw in Chapter 8, ‘you ought to do A’ and ‘you ought to do not-A’ are not inconsistent; what contradicts the first is ‘it is not the case that you ought to do A’, which does not follow without further premises from ‘you ought to do not-A’. Instead, the dualism generates a great many tragic conflicts in Williams’s sense, where no matter what you do you violate an all-things-considered principle and do something all things considered wrong. This outcome is still problematic, however, since whenever the principles conflict they give no determinate guidance. And though Sidgwick seems not to have cared about with this, the principles are contradictory if each is read as giving the one explanation why acts are right. In Book IV of The Methods Sidgwick considered an ad hominem argument intended to dissolve the dualism by converting an egoist to utilitarianism. If the egoist says, implicitly or explicitly, that his happiness is good ‘not only for him but from the point of view of the Universe’, we can point out that his happiness cannot be a more important part of universal good than the equal happiness of anyone else and so persuade him that he should seek the happiness of all (ME 420–1). This was also Moore’s argument in Principia that egoism is self-contradictory. Though the egoist says he should promote only his own good, ‘his good’ can only mean what is simply good and a state of him, and if a state of him is simply good, others have just as much reason as he does to promote it. He likewise has just as much reason as they to promote their good, so the only possible goal of right action is the greatest good of all (PE 97–102; also Rashdall, E 63n2). This anti-egoist conclusion follows from two premises Moore accepted. One is that the right must be identified in terms of the good, either because ‘right’ means ‘will result in the most good’ or because the two are self-evidently connected. The other is that goodness is unanalysable and as such cannot be had ‘for one person’ but not ‘for another’: something either is good or is not. The conclusion can therefore be resisted in two ways. An egoist can say only that each person ought to promote his own pleasure, but make no claim about the good; his egoism is then purely deontic. Or he can use a relativized goodness-concept, saying each person’s pleasure is the only thing good ‘from his point of view’ and therefore the only thing he should pursue. Sidgwick did not consider the first response but thought the egoist can successfully

240

british ethical theorists from sidgwick to ewing

make the second. If he says only that his happiness is ‘good for him’ or is his ‘rational ultimate end’ but not that it is good universally, he cannot be argued out of his view (ME 420–1, 497–8). And the relativized concept is available if ‘good’ is analysable as what one ought to desire, as Sidgwick himself thought, since what each person ought to desire can be different. A position that is impossible given Moore’s understanding of ‘good’ is coherent and defensible given Sidgwick’s. In conceding that his ad hominem argument fails, Sidgwick held that egoism is logically consistent. But he also held, beyond that, that its principle is self-evident or at least apparently so. On what basis did he do so? In Book III of The Methods he affirmed an axiom of prudence, but as we saw, his statement of it equivocated between other-things-equal and all-things-considered forms, and his discussion of it concentrated on the negative claim that you should not prefer a good at one time to an equal good at another, which is not unique to egoism. So there he gave no clear defence of egoism. In an 1889 article he conceded that earlier editions of his book had not shown ‘the irrationality of the sacrifice of selfinterest to duty’ and said the ‘missing argument’ for the ‘rationality of Egoism’ should be supplied. He continued: The assumption is simply that the distinction between any one individual and any other is real and fundamental, and that consequently ‘I’ am concerned with the quality of my existence as an individual in a sense, fundamentally important, in which I am not concerned with the quality of the existence of other individuals. If this be admitted, the proposition that this distinction is to be taken as fundamental in determining the unique end of rational action for an individual cannot be disproved; and to me this proposition seems self-evident, although it prima facie contradicts the equally self-evident proposition that my own good is no more to be regarded than the good of another. (EEM 44)

He then added a shortened version of this passage to the last chapter of his book (ME 498). Some who read the axiom of prudence as making only a negative claim think this ‘distinction passage’ is the one place where Sidgwick positively affirmed egoism,2 but if the dualism involves self-evident principles from III 13 it is better to see it as giving additional support to an axiom stated there. Either way, however, the passage is problematic. Its article version included a qualification omitted in the book. Sidgwick granted that a critic can accept the metaphysical distinction between persons yet insist that ‘the preference of Virtue or general happiness to private happiness is a dictate of reason’, thereby refusing the egoistic principle. He found this refusal ‘impossible to myself ’ and ‘paradoxical’, but conceded that he could not see how to prove egoism to someone who did not share his intuitions (EEM 44–5); this was in effect to concede that it fails the consensus condition. In addition, as Phillips has argued, the passage 2

Shaver, Rational Egoism, pp. 83–4; Phillips, Sidgwickian Ethics, pp. 114–15.

self-benefit, distribution, punishment

241

does not uniquely support egoism. Its claim that the distinction between persons ‘is to be taken as fundamental’ is also satisfied by an intermediate view that says you should promote the happiness of everyone but count your own happiness, say, three times as much as anyone else’s.3 (This is the view Broad thought closer to common sense.) Most importantly, it is hard to find a persuasive argument in the passage. Its initial claim that there is a distinction between persons is descriptive and therefore cannot by itself support a normative conclusion; it is also a claim no one would deny. It needs to be supplemented by a normative premise, and Phillips has suggested ‘If the distinction between persons is real, then “I” ought to be concerned with the quality of my own existence more than with anyone else’s’.4 But the point about an intermediate view aside, this conditional premise seems very close to egoism; certainly it is hard to see anyone accepting it who does not already accept egoism. If so, the ‘missing argument’ Sidgwick took himself to be supplying begs the question.5 He could have just asserted that the egoistic principle is self-evident, as he did the two more abstract claims from which he derived his axiom of benevolence. The 1889 passage can be read that way, since what it says is self-evident is ‘that this distinction is to be taken as fundamental’. And he attributed all-things-considered egoism to Butler and Clarke on the basis of simple assertions of theirs (ME 119–20), and made a similar assertion of his own in his article (EEM 43; also ME xx). But a bare assertion of egoism is much harder to accept than the parallel one about benevolence. Is it selfevidently wrong to sacrifice a small amount of your own happiness to promote the much greater happiness of others? That claim is dubious. Prichard and Ross would certainly have rejected it, as would Rashdall and Moore; in Ethics Moore called egoism self-evidently false (E 98). Even Broad, Sidgwick’s greatest admirer, said egoism seems ‘flagrantly contrary to common-sense morality’ and ‘plainly false’ (FT 244–5); he could not find ‘the least trace of self-evidence’ in egoistic hedonism (CE 272), while to McTaggart the claim that common sense accepts egoism was ‘wrong’ (‘EHS’ 413–14). Sidgwick wondered whether our finding both egoism and impartial consequentialism self-evident may justify us in believing in the existence of God, as necessary to harmonize the two; he wisely concluded that it does not (ME 506–9). But a contrary hypothesis suggests itself: that only if we already believe in God will we find egoism credible, because only then can we be confident that obeying it will not lead to immorality.6 Though it was vital for Sidgwick’s dualism that there be an otherwise self-evident principle telling us to promote only our own good, he did little to support that highly contestable claim. For Sidgwick the duty to promote your good was just the duty to promote your pleasure, but others recognized further intrinsic goods and duties to promote them. Ross, while denying that there is a duty to promote your own pleasure, affirmed 3 5 6

4 Phillips, Sidgwickian Ethics, pp. 128–9. Phillips, Sidgwickian Ethics, p. 127. A fuller critique of the argument in the distinction passage is in Shaver, Rational Egoism, pp. 84–98. See also Shaver, Rational Egoism, p. 155.

242

british ethical theorists from sidgwick to ewing

duties to promote your knowledge and virtue, and on their basis said, ‘The doctrine that morality is entirely social, that all duty consists in promoting the good of others, seems to me a profound mistake’ (RG 153; also Prichard, MW 2, 216–17; Broad, CE 272). Since we do not all desire our knowledge or virtue, a duty to pursue them is just as categorical as any other-regarding one and has the same title to be called ‘moral’. For Rashdall and Moore the duty to promote perfectionist goods was impartial. Moore would therefore say that if you can either preserve a valuable friendship of your own or a slightly more valuable friendship of two strangers, you should do the latter; here again his account of personal relationships was not ‘personal’. But Ross had a similar view. He thought we have no more duty to promote our own knowledge or even virtue than anyone else’s: ‘if we feel a special responsibility to improve our own character rather than that of others, it is not because a special principle is involved, but because we are aware that the one is more under our control than the other’ (RG 26; also 168; FE 283, 302–3). Carritt agreed, saying that caring more about your own conscientious actions ‘cannot arise from love of what is good, but rather from a form of pride or rivalry, the desire to possess oneself what is esteemed by others’ (EPT 124). But in both early and late writings Prichard said your duty to promote your own virtue is stronger even if it is not more easily fulfilled; for each his own character is his ‘special business or duty’ (MW 2, 216–17, but contrast 13). Sidgwick went further, saying in his account of perfectionism, ‘no one has ever directed an individual to promote the virtue of others except in so far as this promotion is compatible with, or rather involved in, the complete realisation of Virtue in himself ’ (ME 11). These views can be grounded in the agent-relative claim that from each person’s point of view his own virtue is a greater or the only such good, but Ross would have rejected that idea. He thought that whereas it is right to feel satisfaction in the pleasure of others, knowledge and virtue call for admiration, where that involves the thought that its object has an unanalysable property of goodness that, as unanalysable, must be agent-neutral (FE 278–9). Given the kind of value knowledge and virtue have, he held, the duty to promote them must be impartial. Just as there is a question whether desiring your own pleasure is virtuous, so we can ask whether caring about your virtue is virtuous. Wanting to become virtuous when you are not is certainly virtuous, as is feeling shame about a vicious impulse or feeling. Here Prichard’s view is especially plausible: surely you ought to feel worse about a malicious desire of your own than about a similar desire of a stranger. But what about being pleased by your virtue? Is that virtuous? It is not virtuous if it is disproportionate, as when you are more pleased by your virtuous pursuit of a student’s knowledge than by the student’s knowledge. But even proportional love of your virtue does not seem something it is a failing to lack. A teacher who is so focussed on his students’ knowledge that he does not notice or take pleasure in his virtuous desire for it is not lacking in virtue, nor is a saint who is unaware of his saintliness. We can accommodate these claims by saying that from each person’s

self-benefit, distribution, punishment

243

point of view his virtue is good when he loves it but not otherwise; then his coming to love it, though not worsening his attitudes, does not make them better. Again, however, this claim requires the goodness of virtue to be analysable, which Ross would resist. Any view on which it matters that you should promote others’ good assumes that people’s goods can conflict. In his discussion of the dualism Sidgwick considered whether, God aside, promoting others’ pleasure always most promotes your own and concluded, sensibly, that it does not (ME 498–503). Green agreed and took this to show that pleasure is not ultimately good. ‘The distinction of good for self and good for others’, he wrote: has never entered into that idea of a true good on which moral judgments are founded. The idea of the good implies interest in an object which is common to all men in the proper sense,—in the sense, namely, that there can be no competition for its attainment between man and man; and the only interest that satisfies this condition is the interest, under some form or other, in the perfecting of man or the realisation of the powers of the human soul.7

Though Sidgwick found this claim striking, he did not think it can be sustained. That people’s goods do not conflict might be true if the only good were virtue, but Green’s ideal included the development of our faculties for art and science, and in ‘the material conditions of human existence . . . as they are now’ there can be competition for the means to that development. People need money to buy and leisure to read books, admission to theatres and art galleries, and more (GSM 69–72; also Rashdall, TGE II 99–100; Carritt, TM 62; Broad, FT 39–44).8 So for these perfectionist goods there can be conflict. And though Sidgwick did not say this, there can also be conflicts with others’ virtue. Just as a career relieving others’ pain can harden and thereby worsen your character, so working as a moral guide, helping others identify and correct their vices, can make you arrogant or complacent about your own virtue. Given where most people find their pleasure, it is plausible that, as Rashdall held, perfectionist goods such as knowledge and aesthetic appreciation conflict less often than hedonic ones, so it is less often the case that the means to them cannot be shared (TGE II 99). And the point is strengthened if virtue is good, since then in desiring another’s pleasure or knowledge you realize a good in yourself.9 But this fact does not completely eliminate interpersonal conflict: if a virtuous attitude has less value than its object, then in preferring another’s greater to your lesser pleasure you may lose more in pleasure than you gain in virtue. Moreover, the ultimate explanatory

7

Green, Prolegomena, secs. 232, 281; also 244, 245, 286. Note how when Sidgwick wanted to reduce perfectionism to intuitionism, he emphasized the importance of virtue in its account of the good, but when questioning Green’s claim about noncompetition he emphasized non-moral goods. 9 Hayward too thought perfectionist goods conflict less than hedonic ones, but thought Sidgwick’s objection to Green on this point was the most serious objection to the latter’s theory (Ethical Philosophy of Sigdwick, pp. 214–15). 8

244

british ethical theorists from sidgwick to ewing

principle tells you to pursue the good of all, and must do so if other-regarding desires are to be virtuous. Nonetheless, a theory that recognizes perfectionist goods may face fewer conflicts between people’s goods than if it valued only pleasure.

11.2 Distribution: Intrinsic and Instrumental Goods The simplest principle about promoting the good tells us to maximize the total good in a population, but some views care how the good is distributed and can prefer an outcome with less good if it is better distributed. They are often said to concern distributive justice and can take two forms. They can say a certain pattern of distribution is an intrinsic good alongside, say, pleasure, or without making any claim about the good just say we should distribute in that pattern. On the second approach departures from the pattern that occur naturally are not unjust; only ones that result from human choice are. One such view is Ross’s valuing of desert. If his claim that happiness should be proportioned to virtue was only non-comparative, it did not directly concern distribution but had distributive implications. But when he called desert a ‘situational’ good (FE 286), he valued a pattern of distribution as such. A different view of this type values equality in distribution and equates that with justice; it was suggested by some in our school. If utilitarianism accepts Bentham’s formula of ‘everybody to count for one, and nobody for more than one’, it is egalitarian to the extent of weighing the equal goods of all people equally. But Sidgwick thought it should go a little further. When the imprecision of hedonic measurement means we cannot say that either of two outcomes is better, it ‘becomes practically important to ask whether any mode of distributing a given quantum of happiness is better’, and here Bentham’s formula favours ‘pure equality’, so then—but only then—we should prefer a more equal outcome (ME 416–17, also 447). It is unclear, however, why utilitarianism needs to be supplemented in these cases. Why not say that if we cannot distinguish two totals, we act subjectively rightly whichever we choose? It is also puzzling why, if equality is important enough to break ties, it cannot sometimes outweigh the total good and make an outcome with less pleasure preferable. Ross thought it can, though he valued equality only between people of equal merit (FE 71–2). In one passage Rashdall also gave equality more weight, saying that if sacrificing a small amount of total good will secure greater equality in its distribution we should do so (TGE I 265, but contrast 270). But at the same time he denied that ‘so abstract a thing’ as equality in distribution can be intrinsically good; all true goods are states of individuals (TGE I 266, also 268, 281, 301). He justified his egalitarianism differently, by saying that each person’s good includes virtue, justice is a virtue, and the value of this virtue limits how far we may promote other goods at the expense of equality (TGE I 267). But it is hard to see how this argument fits with his general account of virtue. If virtue involves loving what is good, and equality is not good, how can

self-benefit, distribution, punishment

245

desiring equality be virtuous? A better view says the value of a fixed increase in a person’s pleasure or other good gets less the more of it she has; this ‘priority’ view favours equal distributions without valuing anything so ‘abstract’ as a pattern.10 Moore and McTaggart both briefly mentioned this view (‘MME’ 358; NE II 437–8) but neither explored its distributive implications. More central for Rashdall was a reading of Bentham’s formula on which everyone’s good is of equal value with ‘the like good’ of everyone else (TGE I 222, also 224, 226–7, 233), but people’s differing abilities can make their goods unlike. Just as humans’ capacity for perfectionist goods makes it more important to promote our good than that of other species, so those few who are capable of ‘the highest intellectual cultivation’ have a stronger claim on resources: there are ‘superior rights of the superior kind of Well-being, and therefore of the superior man who is capable of enjoying it’ (TGE I 234, 239, 280). On this basis Rashdall defended the existence of ‘more or less hereditary classes enjoying a certain superiority of wealth, culture, and consequently of opportunity’ (TGE I 232, also 234; Sidgwick, EP 153–4) and said, notoriously, that ‘sooner or later, the lower Well-being—it may be ultimately the very existence—of countless Chinamen or negroes must be sacrificed that a higher life may be possible for a much smaller number of white men’ (TGE I 238–9). He seems not to have based this last claim on a belief in the innate inferiority of non-white races, since he hoped the development of European culture will eventually enable a higher life for blacks (TGE I 241); he also hoped progress will reduce the differences between social classes in Britain. But despite occasional remarks favouring equality he more often held that sometimes ‘the higher Well-being of the few’ must take precedence over ‘the lower Well-being of the many’ (TGE I 236). He presented this view as a simple corollary of the consequentialist concern to maximize the good, and to some extent it is. If some people’s talents mean they can achieve more good given the same resources, it will most promote the good if they have more. This implication may be limited if there is diminishing marginal utility of resources, which then contribute less to a person’s good the more of them she has, and Rashdall did worry that too much wealth can undermine your character (TGE I 259). But he did not consider diminishing marginal utility more widely, and more often emphasized how unequal capacities can justify unequal distributions. We may wonder, however, whether that is all that motivated his view. Imagine that the ‘higher life’ has a hundred units of value and we can either secure this life for a thousand superior people or make a one-unit improvement in the lives of many more lesser people. If there are enough lesser people total consequentialism favours the second choice, but Rashdall seems not to have considered numbers and to have preferred the higher good regardless. He may have been considering benefits for future generations, and did say that slavery in ancient Athens may have been justified

10

Parfit, Equality or Priority?

246 british ethical theorists from sidgwick to ewing by the city’s contributions to later culture (TGE 241). But he may also have been tacitly guided by a different view some associate with perfectionist goods: that it is important that certain forms of knowledge or artistic expression exist somewhere in society or the world, even if they are not widely shared and do not benefit other people.11 At one point Sidgwick argued against a principle giving significant weight to equality of happiness by saying it would require us ‘to give less to cheerful, contented, self-sacrificing people than to those who are naturally moody and exigent, as the former can be made happy with less’; he found this ‘too paradoxical to recommend itself to Common Sense’ (ME 284n2). The latter comment suggests that, like recent views that value equality ‘of resources’ rather than ‘of welfare’,12 the common sense of his day sought equality only in instrumental goods such as money and not in happiness. But he did not notice that this view will object just as much to utilitarianism, which likewise recommends giving less to those who can be happy with less, and did not consider it further. Rashdall did consider it, imagining a critic who rejects favouring the more talented by saying, ‘Let the conditions be equally distributed; for the rest, the individual must take care of himself ’. He replied that this proposal violates the ‘principle of equal consideration’ he derived from Bentham, since our goal should be ‘not equality of conditions, but equal Well-being, or . . . so much equality as is consistent with there being as large as possible an amount of good to distribute’ (TGE I 228–9). This was in effect to say, like some present-day critics of resource-equality,13 that our ultimate principles must concern the distribution of what is intrinsically rather than just derivatively good. Equality of resources is not the only distributive view focussed on instrumental goods. Another is the Lockean or libertarian view that people have natural rights to acquire property, by mixing their labour with it, and then to transfer it to others, so a distribution is just if it was arrived at by legitimate exercises of these rights. Its clearest exemplar is the entitlement theory in Nozick’s Anarchy, State, and Utopia. Sidgwick discussed this view in connection with the broader libertarian theory that justice involves respecting rights, and that the one natural right is the right to freedom. He objected that, far from supporting rights to property, this theory positively excludes them. While you may have a right not to be interfered with while you are using an object, the right to prevent others from using it at other times, which a full property right in it involves, would unjustly restrict their freedom and so is ruled out (ME 276–7). Though perceptive, this objection is not, I think, decisive. There are two consistent libertarian theories, one giving you only the right to noninterference when you are using an object and the other allowing you full ownership in it. In each case some exercises of your rights limit the freedom of others—in the 11 12 13

Nagel, ‘Fragmentation of Value’, p. 132. Dworkin, ‘What is Equality? Part 1’ and ‘What is Equality? Part 2’. Sen, ‘Equality of What?’, pp. 215, 217; Nussbaum, ‘Aristotelian Social Democracy’, pp. 209–10.

self-benefit, distribution, punishment

247

first, their freedom to use an object you are now using—but do not do so improperly given how their rights are limited by yours. What Sidgwick’s objection shows is that neither view, and especially not the second, follows directly from the broader libertarian theory. If that theory grants full property rights, it is because of a stipulation with no deeper libertarian rationale. He also questioned the right of initial acquisition, or what determines its limits. If merely hunting on land does not give you the right to exclude others from pasturing sheep on it, as most Lockeans hold, why should a shepherd have the right to exclude those who want to till the land or someone who uses only the surface be allowed to exclude would-be miners (ME 277; also EP 68)? More generally, since ‘a man does not create matter by his labour, but only modifies it’, how can his labour give him ownership of the matter (EP 66, also 140, 156)? He concluded that what justifies property-rights is the consequentialist consideration that only with the prospect of ownership will people have sufficient incentive to labour productively (EP 45, 64, 68–9). Rashdall rejected libertarianism on the general ground that all duties derive from that of promoting the good (‘PTP’ 47), but he too noted that labour requires material and asked, ‘May a man mix ever so little labour with ever so much material’, for example by merely putting a fence around a thousand acres of land, ‘and yet secure a right over that material for himself and his heirs for all time’ (‘PTP’ 44–5)? This is like Nozick’s question whether pouring a can of tomato juice into the sea is a way of acquiring the sea or of losing your tomato juice.14 Sidgwick and Rashdall also noted Locke’s condition that for legitimate initial acquisition there must be ‘enough and as good left for others’, and questioned how it can still allow ownership of land once all the land in a territory is occupied (ME 277; EP 46, 68; ‘PTP’ 44). Nozick argued that the rights of those who can no longer acquire unowned land are not violated if the sequence of past acquisitions has left them better off,15 but Sidgwick said this is not always so and, in any case, reading the condition this way introduces a utilitarian element into what was supposed to be a theory centred on freedom (ME 278). A better reply says that though libertarian theories have often included the Lockean condition, they need not do so. It does not in general follow from the fact that you have a right that you are able to exercise it. The right to life is the right to choose to continue living, but if you have a terminal illness you cannot do that. A libertarian theory can likewise say that, though everyone has the right to acquire unowned land, past a certain point in history no one can exercise that right. This less compromising theory may be harder to accept intuitively, but it is internally consistent and does not forbid land-ownership today. A different view focussed on instrumental goods concerns desert. It says you deserve income on the basis of either or both of your economic contribution and economic effort, where your contribution is the good your labour does for others and

14

Nozick, Anarchy, State, and Utopia, pp. 174–5.

15

Nozick, Anarchy, State, and Utopia, p. 177.

248

british ethical theorists from sidgwick to ewing

your effort includes everything costly about it for you, including its duration, how taxing it is, how much training it requires, and so on. This economic desert differs importantly from moral desert. The latter says you deserve the intrinsic good of happiness for the intrinsic good of virtue; here you deserve the instrumental good of income for work that is either instrumentally good by benefiting others or instrumentally evil by costing you. Economic desert therefore does not care how your work was motivated. If two people make beneficial inventions, one out of altruism and the other from a selfish desire for riches, the second deserves no less income than the first. Nor does it matter how happy the deserved income makes you; if it leaves you unsatisfied, you have no claim for more. Just as in moral desert one intrinsic good is the fitting reward for another, so here one instrumental value calls for another. As Sidgwick recognized (ME 279–80), a desert view can, like libertarianism, justify free-market distribution. Contribution, it can argue, is rewarded on the demand side of the labour market, where employers pay more for work they think will benefit them more. And on the supply side workers require higher pay for work that is harder for them. A desert view can also justify some initial acquisition, since by labouring on an object you improve it and so deserve something for that contribution. But it also motivates Locke’s ‘enough and as good’ condition. If there is unlimited land for others to use, the value to them of any land you have improved depends entirely on your improvement. If that is not so, some of the land’s value is independent of your labour, and you have no right to that part of its value. Carritt included a principle about economic desert in a pluralistic theory that also values equality and total happiness (EPT 169), but Sidgwick and Rashdall criticized economic desert. Their arguments often anticipated Rawls’s influential ones in A Theory of Justice and shared those arguments’ flaws. Both objected, like Rawls, that how much you contribute to others depend on factors outside your control, such as your innate abilities, which, they said, are not a legitimate ground of desert (ME 283–4; TGE I 250–2).16 But this argument blurs an important distinction. The desert idea is not that you deserve income for having abilities. You deserve income for benefiting others, and though your abilities may be a necessary condition for doing that, they are not sufficient. You have to choose to exercise your abilities, and, given the abilities, your choosing makes all the difference between your contributing and not. But if your choosing makes all the difference to your contributing, why can it not make you deserve for all you contribute? The objection about abilities is, at the least, too quick. Sidgwick thought a view that rests desert on voluntary effort fails if determinism is true (ME 284), but in Chapter 6 we saw reasons to doubt this. He then made a further point that Rashdall also made and again anticipates Rawls. Both said that while your effort depends in part on how you choose, it also depends on factors outside your

16

Rawls, Theory of Justice, pp. 103–4.

self-benefit, distribution, punishment

249

choice, such as an innate character that is favourable to effort. And just as Rawls said there is no way to discount for these factors, so the idea of rewarding effort is ‘impracticable’,17 Sidgwick said ‘it does not seem possible to separate in practice that part of a man’s achievement which is due strictly to his free choice from that part which is due to the original gift of nature and to favouring circumstances’ (ME 285), while Rashdall said we cannot ‘discriminate between the portion of the work produced which is due to superior goodwill, to industry, perseverance, integrity, and that which is due to superior capacity’ (TGE I 260, also 255, 262). Rawls thought his point about impracticability justifies abandoning effort-based desert. Sidgwick likewise said we must ‘leave to providence’ the rewarding of effort and content ourselves with trying to reward contribution (ME 285), while Rashdall said we can only reward ‘the actual quantity of work done’, regardless of its cause, where by ‘work’ he meant ‘contribution to social good’ (TGE I 260). But all these conclusions are unwarranted. As Sidgwick recognized, that we cannot in practice determine how much a person deserves for her effort does not mean there is not a truth of the matter, or an income that is objectively right for her to receive. Even subjectively, how can our inability to separate the roles of two factors in her effort, one supposedly able to ground desert and the other not, justify us in acting as if only the second is present, which is in effect what Sidgwick, Rashdall, and Rawls did? And that you have a character favourable to effort is again not sufficient for effort; you have to choose to use that character. If your doing so makes all the difference between your exerting effort and not, why can you not deserve for all your effort? This is especially true if by effort we mean everything costly or disadvantageous about work. If one person has enough innate strength of will that he can do two hours of painful work while another can do only one, why can the first not deserve more for that extra painful hour? After concluding that only contribution can in practice be rewarded, Rashdall rejected economic desert on the ground that ‘the ideal of just reward’ is not satisfied when incomes depends on factors outside people’s control (TGE I 260). But Sidgwick raised further difficulties about how contribution is measured, in particular whether it is accurately measured by the free market. He noted that the market rewards work that gives people what they believe will benefit them, but their belief may be false (ME 287). They can be wrong about what their good consists in, especially if it includes perfectionist elements, but he was thinking mostly of mistakes about the means to their good, such as what will actually give them pleasure. Either way, the market can reward what is not a real contribution. 17 Rawls, Theory of Justice, p. 312. The bulk of Rawls’s longest discussion of economic desert (pp. 310–15) is directed at the view that income should be proportional to moral virtue, which no one has held. Rashdall too sometimes conflated economic and moral desert, in particular sliding between talk of moral merit and of what looks like effort (TGE I 243); he also used ‘work’ sometimes for contribution and sometimes for effort (TGE I 260; TGE I 247). Sidgwick more clearly separated economic from moral desert and the two forms of the former from each other.

250

british ethical theorists from sidgwick to ewing

But a desert view should not say the market perfectly rewards contribution; it should say only that it roughly or largely rewards contribution, so distribution by desert is what Nozick called a ‘major patterned strand’ in a market economy.18 It can also recommend interventions to minimize consumers’ false beliefs, for example by forbidding misleading advertising; this will make the market more often reward real rather than apparent contribution. Sidgwick and Rashdall also noted that how much the market rewards a type of work is affected by numbers. You earn more for providing a benefit that more people happen to want or when the number of others who can supply it is smaller; how, they asked, can these accidents affect your desert (ME 288; TGE I 250–1)? They cannot affect your moral desert, but that is not the issue here. And they can affect your contribution if that is understood in a marginalist way, so we compare the outcome when you provide a benefit with what would have been the case had you not; this may involve another, less talented person’s doing so. On this view, the exercise of a rare talent does benefit others more, as does doing what more of them want; it makes more of a difference to their good. And a desert view can again recommend limiting the effect of numbers, for example by forbidding monopolies, as Sidgwick urged (ME 288; EP 555–6). Marginalist ideas can also be used to assess an individual’s contribution to a co-operative project, an issue Sidgwick and Rashdall both raised (ME 289; ‘PTP’ 46–7). Her contribution is the difference between what the group produced with her and what it would have produced without her, and is reflected in how much they will pay to have her rather than someone else participate. Sidgwick’s most powerful objection denied that ‘the social worth of a man’s service is necessarily increased by the fact that his service is rendered to those who can pay lavishly’, where in the free market ‘his reward is certainly likely to be greater from this cause’ (ME 288). The market pays more to those who contribute more, as desert requires. But they can then use their greater wealth to outbid in the consumer market those who have less, even though they will not be benefited more. This is especially problematic if there is diminishing marginal utility of money. Then the market rewards you more for providing luxuries to the rich, which benefit them only a little, than for supplying necessities to the poor, which benefit them a lot. This is the opposite of what desert approves. This objection highlights an internal tension in an ideal of desert by contribution: to reward contributions proportionally at one time is to give some people market power that will undermine proportional distribution at later times; it follows that the ideal cannot be perfectly realized through time. More specifically, to prevent radical departures from desert in the future there must be some departures in the present, in the form of redistribution from those who deserve more to those who do not. The

18

Nozick, Anarchy, State, and Utopia, p. 158.

self-benefit, distribution, punishment

251

result may be something like Carritt’s mix of a desert principle with one about equality, though now justified just in desert terms. Having rejected Lockean and desert views, Sidgwick and Rashdall both thought the correct approach to economic questions is consequentialist, so, their minor interest in equality aside, the best distribution of income or wealth is whichever results in the most total good for all; Rashdall called this consequentialist view ‘the general tendency of modern political philosophy’ (‘PTP’ 57) and Sidgwick made similar claims (EP 35–6, 45–6). Since Sidgwick identified the good with pleasure, his specific economic proposals depended on some competing empirical claims about pleasure. On one side was the diminishing marginal utility of money for happiness (EP 9, 153, 173; PE 110), which gave his view an egalitarian thrust and supported some state action to protect people from ‘extreme indigence’ (EP 157–9). On the other side were the need for market incentives to encourage thrift and ‘the most energetic and enlightened application of labour’ (EP 68, also 64, 69, 97, 101), as well as a duty to fulfil the expectations created by existing economic institutions. He seems to have found the latter claims more pressing, so his final view approved a broadly laissezfaire economy like that of Victorian Britain (EP 139, 151–2); thus he opposed legislation limiting the length of the working day as interfering improperly with the freedom of employers and employees to decide what arrangement of work will benefit them most (EP 154–5). Rashdall thought those with more capacity for the higher goods should have more resources when they will use them for that purpose. He also emphasized the need for freedom in the use of one’s resources if they are to contribute to higher goods; this was a further justification for some inequality (TGE I 233, 275–7; ‘PTP’ 59–63). But he recognized that considerations about freedom justify the ownership of personal property much more than they do private capital or an unlimited right of bequest; those aspects of the free market he rejected (TGE I 276–7; ‘PTP’ 62–3). He also said ‘the present distribution of good things is excessively and arbitrarily unequal’ and affirmed a ‘duty of trying to reduce the present enormous differences between the highest and the lowest standard’ to eliminate the ‘outrageous inequalities . . . produced . . . by excessive wealth and excessive poverty’ (TGE I 225, 272; also ‘PTP’ 64). These are stronger egalitarian claims than any of Sidgwick’s, but Rashdall’s final view, which he too did not specify precisely, probably allowed more inequality than most philosophers today would. Economic distribution was not discussed much by our school. McTaggart, Moore, Prichard, Ross, Broad, and Ewing barely mentioned the topic, and even Carritt gave it just a few pages in a book half about political philosophy (EPT 166–71). Sidgwick and Rashdall wrote more, and anticipated many influential objections to libertarian and desert views while defending a simpler consequentialist approach; they did not, however, attend as much to egalitarian views as the recent literature has.

252

british ethical theorists from sidgwick to ewing

11.3 Criminal Punishment The school wrote more on the topic of criminal punishment. After substantial discussions by Sidgwick it was the subject of dedicated articles or chapters by Rashdall, McTaggart, Carritt, and Ross and of Ewing’s The Morality of Punishment. Sidgwick’s was the standard utilitarian view that punishment is justified because it can prevent crime by deterring the criminal and others, incapacitating him by imprisonment, and perhaps reforming his character, though he thought reform is more likely with juveniles than with adults (ME 71–2, 290–3, 446–7; EP 107–8, 117–20). The correct severity of punishment is whichever will have the best overall effects, counting the criminal’s pain as a negative, though often more deterrence is achieved by increasing the probability of punishment than by increasing its severity (EP 118; also Ewing, MP 57–8). Sidgwick rejected retributivism, saying he had ‘an instinctive and strong moral aversion to it’, though he acknowledged that, while ‘gradually passing away from the moral consciousness of educated persons’, it ‘is still perhaps the more ordinary view’ (ME 281). He also argued, characteristically, that retributive views cannot specify deserved punishments precisely (ME 291–3; also Rashdall, TGE I 289; Ewing, MP 36–40), but he did not consider the objections that utilitarianism can approve punishing the innocent or excessive punishment, such as execution for petty theft if that does enough to prevent theft. Much of Rashdall’s discussion in The Theory of Good and Evil reprinted an 1891 article19 and continued his polemic against retributivism, especially as defended by Bradley.20 Just as it is false that moral vice demands suffering in the afterlife, he said, so the legal punishment of criminals is not an end in itself (TGE I 284–91, 300–12). Rashdall echoed Sidgwick on the imprecision of the retributive calculus but added a further objection: if it turns out that harsh imprisonment only encourages repeat crimes whereas gentler treatment leads to reform and more deterrence, will the retributivist still demand harsh punishment (TGE I 286; also Ewing MP 18–19)? His positive justification again cited deterrence and reform but with more emphasis, given his valuing of virtue, on the latter than in Sidgwick. (He dismissed the worry about punishing an innocent by saying ‘as a general rule, no public purpose is served by hanging the wrong man’ (TGE I 290).) A prisoner who is reformed by a prison chaplain or by reading a book from the prison library is not reformed by his punishment, but punishment can also contribute to reform directly, since ‘if the lower self is kept down by the terror of punishment, higher motives are able to assert themselves’; thus punishment can be ‘at least the condition of moral improvement’ (TGE I 292–3). The threat of prison can also strengthen moral motivation in those who are not criminals: ‘there are few of us, perhaps, whose conduct would not fall still further behind our own ideal than it actually does, if our better selves were not sometimes reinforced by fear of punishment’ (TGE I 293). And that an act is singled 19

Rashdall, ‘Theory of Punishment’.

20

Bradley, Ethical Studies, pp. 26–33.

self-benefit, distribution, punishment

253

out for punishment reinforces the general belief that it is wrong. ‘Popular ideas as to the moral gravity of many offences depend largely upon the punishment which is awarded to them by the criminal courts’; this illustrates ‘the enormous importance of the criminal law in promoting the moral education of the public mind’ (TGE I 296). Here Rashdall made an important point. Though likewise a perfectionist, Green had argued that the state should not try actively to promote the good of its citizens. Their good consists in ‘character’, and ‘no one can convey a good character to another. Every one must make his character for himself.’ The state’s business is therefore ‘not indeed to promote moral goodness, for that, from the very nature of moral goodness, it cannot do’, but is ‘confined to the removal of obstacles’.21 Green here anticipated an influential recent argument that the state must be neutral about the good and not try to promote some activities over others on the ground that they are intrinsically better. To be truly good, this argument says, an activity must be one you ‘endorse’ or choose because it is good. Since the state cannot make you believe anything is good, it cannot successfully promote what truly has worth.22 As stated, this argument is incomplete.23 It shows at best that state efforts to promote the good will not succeed and therefore lack a positive rationale; it does not show they are wrong. For that it needs a further assumption: that the threat of punishment will introduce a less valuable motive that suppresses the more valuable endorsing one and makes your actions less good. Rashdall challenged this assumption. By attaching a punishment to murder the state gives us a self-interested reason to avoid murder, but that hardly suppresses our moral one. On the contrary, most of us have as our main motive for not murdering the belief that murder is wrong, and the existence of the punishment may even strengthen that belief. But if punishment has this effect with not murdering, there is no reason in principle why it may not do something similar for perfectionist goods (TGE I 297–300). Rashdall therefore thought that ‘with the necessary moralization of a community, the sphere of criminal law ought gradually to extend’ (TGE I 297) to cover more activities. We may not share all his views about what is worth promoting; thus we may not think drunkenness is intrinsically evil. We may also be less optimistic about the effectiveness of legally prohibiting bad activities. But his general claim that the criminal law can strengthen rather than undermine moral motivation surely stands. Starting from premises like Rashdall’s, McTaggart gave less weight to moral reform. His chapter on punishment in Studies in Hegelian Cosmology24 addressed Hegel’s theory, which as he read it takes the object of punishment to be repentance by the criminal because of, rather than just alongside, the pain of his punishment.

Green, Prolegomena, sec. 332; ‘Liberal Legislation’, p. 374; Lectures on Political Obligation, sec. 332. Kymlicka, Contemporary Political Philosophy, pp. 203–4; Dworkin, Sovereign Virtue, pp. 217–18, 269. 23 It also has other flaws, for example, not saying anything against laws that forbid intrinsically evil activities rather than requiring intrinsically good ones; see further my ‘Indirect Perfectionism’. 24 The chapter was originally published in 1896 as ‘Hegel’s Theory of Punishment’. 21 22

254

british ethical theorists from sidgwick to ewing

McTaggart thought repentance is the ‘highest end’ of punishment (SHC 150) and in principle achievable if the criminal sees the authority that punishes him as correctly expressing the moral law (SHC 137–44). But he said this cannot be the main justification of punishment today, which must instead be the prevention of crime. This is because ‘the interests of the innocent are to be preferred to those of the guilty—for there are more of them’; in addition, the deterrent effect of punishment is more likely than its purifying one, because the conditions favouring repentance do not for the most part obtain (SHC 145). Some criminals, for example those guilty of treason, think their acts are morally right and will not change their minds when punished; the larger number may know they are doing wrong but do not care and cannot be made to care by punishment (SHC 145–7). Though Hegel’s theory may justify punishment in an ideal state, it does not fit actual social conditions but reflects ‘his persistent attempt to identify the kingdom of Prussia with the kingdom of Heaven’ (SHC 150). Whereas Sidgwick, Rashdall, and McTaggart rejected retributivism, Moore embraced it in his organic-unities claim that the combination of pain and a bad state of mind in the same person is good as a whole (PE 214–16, 221). It is not clear from his brief remarks whether he meant to affirm only moral or also criminal desert, where the latter, like economic desert, differs importantly from the former. What makes you deserve criminal punishment is not only or all vice; you are not punished for vicious feelings or desires that do not issue in action. Its basis is instead a wrong act prohibited by a publicly promulgated law, and in the best version of retributivism what you deserve is not some amount of pain but an objectively specified punishment, such as so many years in jail. If you find imprisonment especially unpleasant, that does not warrant a reduction in your sentence; if you do not mind it, that does not call for extra years. Again there is a kind of match between a more objectively described desert-base and a more objective response to it. Even if Moore did endorse criminal desert, he thought crime-prevention matters too (PE 164). Carritt likewise had a pluralistic theory, with a ‘retributive element’ central but weight also given to deterrence and reform (TM 108–13; EPT 70–6). In his first book he wrote as if we have a duty to inflict deserved punishment, as something ‘due’ the criminal (TM 113); in his second he affirmed only a right to punish, which we should not exercise when doing so will neither reform the criminal nor deter others (EPT 73). But both books rebutted the charge that retributivism rests on a desire for vengeance, saying that, unlike revenge, retribution is only for acts deemed wrong and should be imposed by an impartial tribunal rather than personally by the victim (TM 111; EPT 71; also Ewing, MP 32–4).25 The most impressive contributions were those of Ewing and Ross. Their theories too were pluralistic, but with more developed accounts of the relations between their elements.

25

See further Nozick, Philosophical Investigations, pp. 366–70.

self-benefit, distribution, punishment

255

A cornerstone of Ewing’s theory was the asymmetry between desert goods and evils mentioned in Chapter 10. He thought injustice in punishment ‘a very much greater intrinsic evil than justice is a good, especially if the injustice consists in punishing somebody for an offence of which he is not guilty or in excessive severity’ (MP 109, also 110–11); later he questioned whether punishing the guilty is good at all (ST 136; MP xiv). He concluded that retributive considerations play ‘only a small part’ in the positive justification of punishment, which turns more on its role in defending society against crime (MP 20, 46–7, also 18–21, 153; E 167–8, 174). But punishment must be only of the guilty, and here a retributive element enters: the significant evil of suffering pain when you do not deserve to makes it seriously wrong to punish the innocent (MP 19–21, 44–5, 60, 90–1, 109) and also wrong to punish too severely, which is a greater intrinsic evil than punishing too leniently (MP 109). So the retributive desert that cannot make a punishment right can make it wrong. If the evil of unjust punishment is just one value among others, it can be outweighed by sufficient utilitarian gains. But this only means Ewing’s theory was moderately rather than absolutely anti-utilitarian; it still gives more protection to the innocent than one that counts only effects on happiness. His theory involved a division of labour. Utilitarian considerations explain why we ought to punish, while retributive ones determine who we ought and especially ought not to punish. It therefore anticipated theories developed in the 1950s by Rawls and Hart and featuring a similar division. What Hart called the ‘general justifying aim’ of punishment is utilitarian and looks to the future benefits of punishment, but its ‘principle of distribution’ is retributive and forbids punishing the innocent. The dispute between utilitarian and retributive theories is therefore misconceived, because each addresses a different question.26 Rawls grounded this division in the indirect-consequentialist argument that an institution that punishes the innocent will have overall bad effects. Hart expressed doubts about this argument, as did Ewing in 1970 (MP xii–xiv). But whereas Hart grounded his view in a pluralist deontology with a principle forbidding punishing the innocent as a side-constraint on one about promoting the good, Ewing’s justification was different. It was not that questions about aim and distribution are fundamentally distinct; it was that a consideration about desert that could in principle answer both is weak in the form that might positively justify punishment but strong in the form that restricts punishment to the guilty. Ewing’s positive justification cited the many ways punishment can prevent crime, but emphasized what recent writers have called its ‘expressive function’,27 where by punishing an act the state publicly declares that the act is bad and communicates society’s disapproval of it, so punishment is a ‘kind of language’ (MP x–xi, 61, 83, 90–1, 93–5, 105). This expression, which arguably distinguishes punishment from a 26 27

Rawls, ‘Two Concepts of Rules’; Hart, ‘Prolegomenon’. Feinberg, ‘Expressive Function of Punishment’; Hampton, ‘Moral Education Theory’.

256

british ethical theorists from sidgwick to ewing

mere penalty, has three justifications. It can encourage moral reform if an offender is led by it to see that his act was wrong (MP 80–93), though Ewing shared McTaggart’s view that this effect is infrequent (MP 85, 87, 92). The expression is also good in itself. Applying the idea that virtue is a higher-level value, he said ‘a right mental attitude to evil’ is intrinsically good, and its expression ‘shares in the intrinsic value, especially if is viewed as a solemn moral judgment on behalf of the community’ (MP 107). This value, however, is comparatively small (MP 109). The most important justification of expression, as in Rashdall, is that it can educate the rest of society (MP 93–107). It is not that without a law forbidding, say, stealing, people will not think stealing is wrong. But they put wrong acts into two classes. Some they know are wrong but will do if they are sufficiently tempted; afterward they will not feel very sorry and may even view their act as excusable. Others they see as very seriously wrong and not on any account to be done. What public legal condemnation can do is move an act from the first class to the second, so its wrongness is more fully appreciated. This will reduce the act’s frequency, since people will now not even consider it, and can also reduce the frequency of other acts. If stealing were just in the first class, acts that come next to it in badness, such as certain ‘sharp practices’, might seem less bad in comparison, and doing them might seem a resisting rather than a yielding to temptation, since you do not do something you are tempted to, namely steal. If stealing is entirely ruled out, however, the sharp practices become the worst of your temptations and are more easily resisted (MP 95–8). Ewing thought the expressive function of punishment gives it retributive features, so it is directed at the past—condemning a particular past act—and permitted only of the guilty (MP 100, 103–4). The last feature is supported by the first two justifications. You cannot reform someone who has not done wrong, nor is there intrinsic value in condemning what is not wrong. But it does not follow from the third justification. It is in principle possible to educate the public by punishing a person for an act you only pretend he did, by fabricating evidence against him, or to deter by punishing him. Ewing here needed his simpler claim that punishing the innocent is seriously intrinsically evil. But its emphasis on expression is his theory’s second notable feature, alongside its asymmetry-based combining of utilitarian and retributive elements. Ross’s theory likewise combined these elements, though in a different way. He first presented it in ‘The Ethics of Punishment’ of 1929, which he then reprinted, with some changes, as an Appendix to the second chapter of The Right and the Good. He too denied that retributivism positively justifies punishment. In his article he doubted whether vice punished is in itself better than vice unpunished and thought ‘intuitionists’ should reject Moore’s account of punishment as wrongly deriving the duty to punish from claims about the good (‘EP’ 208, 206). But in his book, which affirmed a good of desert, he emphasized a different point. What is good is the proportionment of the total pleasure in a person’s life as a whole to his total virtue. But punishing particular criminal acts does not reliably promote this good, since

self-benefit, distribution, punishment

257

those who are punished may already be enjoying less happiness than they deserve given all their behaviour till now. The criminal law also addresses only the small fraction of wrong acts it expressly forbids and does not try to reward good ones, so it cannot produce everything desert demands. And since the state lacks the information it would need to assess people’s total virtue, its aiming at true justice is ‘impracticable’ (RG 58–9). This argument mistakenly assumes that the desert relevant to criminal punishment must be moral desert and does not consider the possibility, defended later by J.D. Mabbott,28 that criminal desert mandates objectively described punishments for individual prohibited acts. What is more important is Ross’s positive justification of punishment. It involved claims about rights and separated two questions: what justifies the state in making a law that attaches a punishment to a crime, and what justifies it in actually inflicting that punishment when a crime has been committed (RG 61)? Ross held that someone who violates the life, liberty, or property of another loses his own rights to those things, so the state violates no prima facie duty if in punishing him it takes them away. But someone who has respected others’ rights retains his rights, so there is a strong prima facie objection to punishing him even if that will do social good (RG 60–1). These claims give the theory a retributive element, forbidding punishment of the innocent, but they do not make punishment positively deserved. They mean the state is permitted to threaten punishment, since what it threatens to do will not be wrong. Its positive reasons to make the threat are consequentialist, deriving from its essential duty to protect the rights of its citizens; it should pass those laws and create those legal institutions that will best prevent rights-violations (RG 60–3). Even here, however, there is a retributive concern, since the threatened punishment must not be disproportionate to the offence it is trying to prevent, for example by being too severe (RG 62–3). If consequentialist considerations mean the state should also not threaten more punishment than is needed to prevent crime, its right to punish is limited in the same way as on most views the right of self-defence is. Just as a defender may not use disproportionate force or more force than is necessary, so the state may not threaten excessive punishment or more punishment than will do good. It may not, for example, threaten capital punishment for crimes other than, say, homicide, or threaten capital punishment if life imprisonment will deter equally well. When he came to the question of why the state may inflict punishment, Ross could again have said just that the criminal has lost his right not to be punished and punishing him will do social good. But he did not say that, because he apparently took the right to threaten punishment to be primary. He insisted that punishment not be retroactive; instead of coming ‘like a bolt from the blue’ it must be ‘preceded

28

Mabbott, ‘Punishment’.

258

british ethical theorists from sidgwick to ewing

by the making of a law in which a penalty is affixed to a crime’ (RG 61). (In his article he noted that this condition does not follow from utilitarianism (‘EP’ 205), but it is also missing from many retributive views.) The state’s right to inflict punishment is therefore not underivative but follows from a prior right to threaten punishment, presumably because it is permissible to do what you have permissibly expressed an intention to do. The positive reasons to exercise the right to punish then include consequentialist ones, such as that only if the penalty a law prescribes is actually imposed can the law successfully deter. But there is also a deontological consideration. When the state enacts a law it in effect promises victims and the community that it will punish anyone who breaks the law, both to give them satisfaction and to protect them from further offences. This promise helps turn the right to inflict punishment into a duty (RG 63–4). Ross’s theory combines retributive and utilitarian elements in another distinctive way. It holds that punishment must be only of those guilty of violating rights, must not be excessive, and must be only for violations of previously promulgated laws; the last follows because the primary justification is of the right to threaten punishment. But punishment is also justified only when it has good consequences, primarily by preventing more rights-violations than any alternative; when that is not so, punishing is wrong. This is an attractive combination of claims, but arrived at using concepts of rights and forfeiture unlike those our school usually employed. They were, however, underivative moral claims rather than derived from ones on a supposedly more fundamental topic and so fit the school’s general approach.29

29 I defended a theory of punishment very like Ross’s in ‘Rights and Capital Punishment’, though I was unaware then of echoing him. Another account that gives priority to the right to threaten punishment is in Quinn, ‘The Right to Threaten’.

12 Historians of Ethics As well as developing their own views, our school wrote about the history of ethics. Sidgwick contributed an entry on this subject to the 1878 edition of The Encyclopedia Britannica that was later published as Outlines of the History of Ethics;1 Broad’s Five Types had chapters on Spinoza, Butler, Hume, and Kant; Prichard wrote on Plato, Aristotle, and Kant; and Ross discussed Aristotle’s ethics in Aristotle and late in life published Kant’s Ethical Theory.2 Most members also sprinkled their writings with comments on earlier moralists; these were often critical but reflected the school’s distinctive views.

12.1 Ancient Ethics Most in the school knew ancient ethics very well. Their early schooling had centred on Greek and Latin and, McTaggart and Broad aside, they continued to study classics, including philosophy, as undergraduates; Ross in particular was a leading classicist. But to many of them ancient ethics had serious flaws. Sidgwick thought a genuinely self-evident principle will help us identify particular right acts. He therefore warned against ‘sham-axioms’ that ‘appear certain and selfevident because they are substantially tautological’, and thought ancient ethics in particular often centred on such claims (ME 374–6). Plato identified virtue with knowledge of the good and a harmony within the soul that accords with this knowledge but said nothing about what the good in question is, while Aristotle’s claim ‘that the Good in conduct is to be found somewhere between different kinds of Bad . . . only indicates the whereabouts of virtue: it does not give us a method for finding it’ (ME 376; also OHE 63–4). The Stoic system was ‘a complicated enchainment of circular reasonings’, saying that the ultimate end of action is ‘Life according to Nature’, which for us is a life directed by reason, where reason directs us ‘To live according to Nature’. The Stoics sometimes called certain ends ‘in a manner preferable,

Earlier, however, he had said, ‘I hate the history of philosophy even more than any other history; it is so hard to know what any particular man thought, and so worthless when you do know it’ (M 140). 2 Though his Kant book appeared when Ross was aged 77 it reads as if based on Oxford lectures, and Blanshard reports attending lectures by Ross on Kant’s ethics in the 1920s (‘Autobiography’, p. 59). The book’s content, then, may date from much earlier. 1

260

british ethical theorists from sidgwick to ewing

though not strictly good’, but if they were asked how we identify these ends or weigh them against each other, they could only say we should choose what is reasonable or act in accordance with nature (ME 377–8). This last was a characteristic objection of Sidgwick’s and again not decisive: why can it not be underivatively true that pleasure and knowledge are preferable, as it was for him that pleasure is good? But his main charge was that the Greek formulae do not give concrete guidance as an adequate moral theory should. He did not object to the ancients’ general methodology. Socrates and Aristotle (though less so Plato) took their premises from the common-sense morality of their day and tried to systematize it (OHE 29, 55–8; GSM 88); this was his own procedure and one he learned from them (ME xix–xx, 375n). His complaint was that they did not carry it through by arriving at substantive rather than tautological principles: ‘there is probably no treatise so masterly’ as Aristotle’s Ethics ‘that yet leaves on the reader’s mind so strong an impression of dispersive and incomplete work’ (OHE 70). He also thought Aristotle merely reported common-sense views rather than trying to improve them. His own indirect utilitarianism tended to conservatism but made some reforming claims, for example about punishment (ME 281) and sexual morality (ME 329–31, 357–9), and his general view was that a philosopher should ‘tell men what they ought to think, rather than what they do think’ (ME 373). But Aristotle did not do this: his ‘unrivalled intellect was continually misemployed in throwing into scientific form the unsifted products of common sense’ (EEM 125; also OHE 64; GSM 91, 96),3 as was predictable given the ‘sham’ nature of his axioms.4 Sidgwick also thought the ancients failed to address what to him was the central issue in ethics: the potential conflict between doing what is best for you and doing what is best for everyone, or between doing what is best for you and doing what is all things considered right. This is because they never clearly distinguished the concepts ‘own good’ and ‘right’, as an adequate ethics must. Formally their theories were egoistic; they assumed that each person’s ultimate end is his own good, and the question was only how to achieve it (ME 91–2; also EEM 92; OHE 20–1, 24–5, 26–8, 197–8). They did not understand this good hedonistically, so it is misleading to translate their ‘eudaimonia’ as ‘happiness’ (ME 92n2; OHE 48n1). Instead they took it to involve, wholly or largely, the exercise of moral virtue, and they therefore held that an act can promote your good even if it causes you pain, for example if you make the ‘heroic exchange of a life full of happiness for a painful death at the call of duty’. But Sidgwick found this last claim implausible, and attributed it: partly to a confusion of thought between what it is reasonable to an individual to desire, when he considers his own existence alone, and what he must recognise as reasonably to be desired 3 Ross too thought Aristotle’s description of the particular virtues ‘presents a lively and often amusing account of the qualities admired or disliked by cultivated Greeks of Aristotle’s time’ (A 202). 4 The connection between Sidgwick’s two charges—of conservatism and tautology—is noted in Irwin, ‘Eminent Victorians’, p. 289.

historians of ethics

261

when he takes the point of view of a larger whole: partly, again, to a faith . . . that there cannot be really and ultimately any conflict between the two kinds of reasonableness. (ME 404–5)

That conflict was of course a central theme of his thought. The ancients did sometimes argue that right acts always promote your good, but Sidgwick found their arguments fallacious. Plato’s Gorgias contained ‘more or less dextrous sophistry, playing on a confusion of thought latent in the common notion of good’ (ME 405n1), while The Republic argued that it benefits you to have a harmonious soul governed by reason but assumed without argument that this harmony necessarily involves the conventionally moral treatment of others. Why, Sidgwick asked, can an egoistic hedonist not unify his soul around the pursuit of his pleasure, so all his more specific desires harmonize with that (ME 171–2)? Sidgwick’s version of this objection assumed a hedonistic account of the good, but that can be dropped. If there are perfectionist goods other than virtue, such as the knowledge Plato and Aristotle valued, why can rightly choosing death not cost you more in future knowledge than it gains you in present virtue? The same holds for virtue. Someone who dies in battle can sacrifice not only a life full of happiness but also one containing years of virtuous action. How then can his act improve his life? This is an especially pertinent question given Aristotle’s claim that eudaimonia requires ‘a complete life’,5 and it was pressed by later members. Ewing granted that sacrificing your life when that is your duty is ‘very virtuous and therefore a very great good’, but: can we possibly say that it is such a great good as to outweigh all the goods that the person who sacrificed his life would, if he had continued to live, have attained and enjoyed? Five minutes or an hour’s virtuous action in which he laid down his life could not outweigh the good of years of virtue which he might still have had if he had not made the sacrifice in question. (E 28; also Moore, E 99)

It is hard not to conclude that in saying the sacrifice would outweigh any loss the ancients were conflating ‘own good’ and ‘right’ in just the way Sidgwick alleged. Sidgwick charged them with an omission, of not addressing a key issue. A stronger criticism is that their egoism was misguided, giving moral duties the wrong explanation. This became Prichard’s main criticism. In ‘Mistake’ he commented on ‘the extreme sense of dissatisfaction produced by a close reading of Aristotle’s Ethics’ (MW 17). J.O. Urmson called this a ‘surprisingly imperceptive’ remark about ‘a work I regard as being still the greatest in its field’ (MW xiii), but he needed to read further! For in the next paragraph Prichard denied that he was criticizing Aristotle. We are dissatisfied with the Ethics, he said, because we want it to answer the ‘why be moral?’ question, and it does not. And his main point was that this question should not be asked; in this article Aristotle was on the right side.

5

Aristotle, Nicomachean Ethics, 1098a18–19.

262 british ethical theorists from sidgwick to ewing This was not, however, Prichard’s later view. In ‘Mistake’ he said Plato did try to justify morality in self-interested terms (MW 8), and in ‘Duty and Interest’ he elaborated this claim. The Republic, he said, contained two different arguments. The first shared with the Sophists the assumption that to be ‘just’ or ‘right’ an act must be advantageous to the agent, or promote his happiness, and then claimed that the acts we think are right, such as keeping promises, really are right, because the state of the soul they issue from does promote our happiness (MW 22–30). The second argument assumed that we do not just think but know that acts like promisekeeping are right. It then argued that these acts promote our happiness, but did so because Plato assumed that unless we believe that we will not do the acts, since our one ultimate desire is for our happiness (MW 30–3). But both arguments justified acts like promise-keeping on the ground that they promote our good (also MW 172–81),6 and Prichard now attributed a similar argument, though comparatively ‘disguised and weak’, to Aristotle (MW 22). Later, in ‘The Meaning of ’agathon in the Ethics of Aristotle’, he argued that by ‘good’ (agathon) Aristotle almost always meant ‘conducive to our happiness’ or ‘conducive to our pleasure’ (MW 102–13), so he too grounded morality in the agent’s happiness. Prichard thought it false that right acts always promote one’s good, and he later echoed Sidgwick’s objection to the Republic’s argument about harmony in the soul: the just actions, the profitableness of which [Plato’s] speakers are considering, are certain acts of a man to others, and unless he holds that the activity within the soul, which he maintains that just action consists in, shows itself outwardly in these acts, his argument is broken-backed. (MW 178)7

Plato did hold this, but Prichard found it implausible: what gives us happiness depends on what we want, and not everyone wants above all to act rightly (MW 180). But his more central objection was, characteristically, explanatory. Plato’s first argument implies that what makes an act right is its being advantageous to the agent, and ‘the fatal objection to maintaining this is simply that no one actually thinks it’ (MW 29). We think keeping a promise or looking after our parents is right because it keeps a promise or looks after our parents, not because of something about our interest. So even if it acting rightly did always maximize our happiness, Plato’s arguments would not prove, as they were intended to, ‘that the moral convictions of our ordinary life are true’ (MW 30). Prichard’s hedonistic interpretation of Aristotle was effectively criticized by Austin8 and is, frankly, an embarrassment; his similar interpretation of Plato is likewise dubious. But within the school not only Sidgwick but also Rashdall, Ross, and Ewing

For a similar objection to the arguments early in Plato’s Republic see Brown, ‘Glaucon’s Challenge’. Sidgwick and Prichard therefore anticipated the objection in Sachs’s famous, ‘Fallacy in Plato’s Republic’. 8 Austin, ‘Agathon and Eudaimonia’. 6 7

historians of ethics

263

recognized that the standard view of the good in ancient ethics was perfectionist, with virtue playing a central role (TGE I 216; A 190; FE 4; E 28; ST 105). And that Prichard’s interpretations were false does not mean we can just set his explanatory objection aside. We have to ask whether it still has force when separated from his claims about pleasure, and the same holds for a related objection by Carritt. Whereas Prichard said Aristotle (almost) always meant by ‘good’ ‘conducive to our pleasure’, Carritt thought Plato and Aristotle alternated between two meanings without noticing that they are different. In some passages they used ‘good’ to mean ‘to the advantage of ’, which he took to mean ‘conducive to the satisfaction of ’. But in others they used it to express something like Moore’s concept of intrinsic goodness, which is unrelated to satisfaction. Their discussions of the good therefore equivocated between two concepts they mistakenly treated as identical (‘AG’ 52–6). If they were not hedonists, Plato and Aristotle cannot have used ‘good’ in quite Carritt’s first sense. But many present-day commentators think Aristotle was ultimately concerned with the ‘good for’ a human, in an evaluative sense connected to a concept of welfare our school did not use. Yet Carritt seems right that some of Aristotle’s claims—that the prime mover is the best thing in the universe, that the theoretical intellect is the best part of the soul, that the good of a city is ‘greater and more complete’ than that of an individual—require a more Moorean concept.9 How did Aristotle relate these two different kinds of claim, and how in particular did he move from premises using the Moorean concept to conclusions about the ‘good for’ a human? Carritt put his finger on a serious difficulty for Aristotle’s theory. Prichard often said Plato and Aristotle resolved the moral ‘ought’ into the nonmoral one, which for him was just hypothetical. If they were not hedonists they cannot have recommended just or virtuous acts as means to the agent’s pleasure, but they often made psychological claims, such as that everyone above all wants his own good, or more specifically in Aristotle, his eudaimonia. Could they then have been recommending acts as means to some non-hedonic end and thus still recommending them hypothetically? They would have been if their substantive arguments were empirical, saying virtuous acts cause a state of eudaimonia external to them. But Aristotle’s arguments look normative: eudaimonia does not consist in pleasure because that is not distinctive of humans, but does involve virtuous activity because that constitutes ‘living well’. He can therefore be read as recommending virtuous action categorically and so as escaping Prichard’s objection. But neither he nor Plato distinguished clearly between telling us how to achieve an end we already desire and telling us what end to pursue whether we desire it or not. Insofar as they suggested the first argument, this version of the objection has force. A different version of the objection says that, even if Plato’s and Aristotle’s claims were categorical, they gave particular duties a mistakenly egoistic explanation. 9

Aristotle, Nicomachean Ethics, 1094b6–10. Carritt’s view that Aristotle used both a relativized and a non-relativized goodness-concept is endorsed in Irwin, ‘Mistakes About Good’, pp. 116–23.

264

british ethical theorists from sidgwick to ewing

Prichard made this objection about promise-keeping and looking after your parents but, as I said in Chapter 6, it applies especially to beneficence. Plato’s and Aristotle’s views imply that the ultimate explanation why you ought to benefit others is that this will make your life better, whereas the right explanation is that it will make their lives better. Ross made this point when he said, ‘It is for their sake that we feel bound to [benefit others], not for our own’ (FE 276; also ‘IME’ 97–8; RG 16), as did Rashdall and Ewing (TGE I 40, 42; E 29). And the objection was common in their period, made even by philosophers outside our school. G.C. Field, for example, wrote: according to Aristotle’s own statement, the final end of each man is some condition of himself. . . . he really wants to do good to others because in doing this he exercises his own nature . . . So however much for the good of others the actual results of his action might be, he does them primarily because he is thinking of and aiming at a condition of himself. . . . Aristotle’s idea of the final end or good [is] in this very real sense a selfish one.10

A formally egoistic theory need not approve action from egoistic motives. It can say that to exercise virtue in the way eudaimonia requires you must pursue others’ good for its own sake rather than as a means to your benefit.11 But many in the school thought Aristotle did not say this, and gave an unattractively self-centred picture of the virtuous agent’s motives. One target of their critique was his account of the proud or great-souled person (megalopsychos). This person has every virtue and is pleased by this fact, but seems especially pleased that he is more virtuous than other people. He likes to give benefits but not to receive them, because ‘the one is the mark of a superior, the other of an inferior’. He is disdainful and will do only great and notable deeds, finding ordinary favours beneath his dignity. Rashdall noted ‘Aristotle’s revolting picture of the highsouled man’ (TGE I 205, also II 157), while Ross said the description of the megalopsychos ‘betrays somewhat nakedly the self-absorption which is the bad side of Aristotle’s ethics’ (A 208). Equally problematic was the self-lover (philautos) of Ethics IX.8. He will share his wealth with a friend, since while the friend gains financially ‘he himself achieves nobility’ and thereby ‘assign[s] the greater good to himself ’. He will likewise stand back and let his friend do a virtuous deed, but his reason is that ‘it may be nobler to become the cause of his friend’s acting than to act himself ’, so he again ‘assign[s] to himself the greater share in what is noble’.12 Carritt rightly condemned ‘the egoistic self-righteousness of Aristotle’s philautos’ (‘AG’ 69).13 Also relevant was Aristotle’s claim that a virtuous person chooses his acts ‘for their own sakes’ or for the sake of the ‘noble’ (kalon). Choosing an act for itself is better than choosing it as a means to your pleasure and is appropriate for acts that fulfil a Field, Moral Theory, pp. 110–11; see also Stocks, ‘Golden Mean’, p. 177. See e.g. Annas, Morality of Happiness, pp. 118, 127–8, 224. 12 Aristotle, Nicomachean Ethics, 1169a28–36. 13 Taylor likewise notes the ‘self-referentiality’ of Aristotle’s megalopsychos, philautos, and other supposedly virtuous agents in Aristotle: Nicomachean Ethics Books II–IV, pp. 88–92. 10 11

historians of ethics

265

deontological duty such as to keep promises. But often a virtuous person chooses an act as a means, though to a good state of someone else. She gives her child a pain reliever, for example, as a means to the child’s being free from pain; her primary concern is that end-state, and she values her act mainly because it will produce that. Aristotle never recognized this type of motive. He said a courageous person’s goal when fighting in war is not victory or the safety of his community but ‘conformity to the corresponding state of character’,14 or the exercise of his own courageous disposition. He also said a virtuous person is pleased and pained by the right things, but gave as his only example of virtuous pleasure pleasure in one’s own virtuous activity;15 he never said a virtuous person is pleased by things that happen independently of his agency, as when another’s pain is relieved by someone else or goes away by itself. On the contrary, he said that just as poets are especially fond of their own verses, so a virtuous person is pleased by what he produced because he produced it: ‘to the benefactor that is noble which depends on his action, so that he delights in the object of his action’.16 Instead of caring first about another’s being free of pain and only second about his act as producing it, Aristotle’s virtuous agent cares first about his own virtuous act and only derivatively about its result. Sidgwick criticized this side of Aristotle’s view. He noted that Aristotle’s list of virtues did not include benevolence; the closest virtue to it, liberality, did not distinguish between acts benefiting others and ‘graceful profusion in self-regarding expenditure’ (OHE 122, also 66; GSM 96–7). He took Aristotle’s concept of the noble to have aesthetic connotations, as the Greek kalon often does, and said his virtuous agent ‘makes a deliberate choice of virtuous acts for the sake of their intrinsic moral beauty, and not for any end external to the act’ (OHE 59), so: [t]he limits of Aristotle’s Liberality are not determined by any consideration of its effect on the welfare of its recipients, but by an intuitive sense of the noble and graceful quality of expenditure that is free without being too lavish; and his Courageous warrior is not commended as devoting himself to his country, but as attaining for himself, even amid pains and death, the peculiar kalon of a courageous act. (EEM 90; also GSM 90–4)

This particular criticism was too quick. Because ‘noble’ is a supervenient property, an act cannot be noble unless it has other properties that make it so, and in choosing an act because it is noble a virtuous agent chooses it for those properties. But Aristotle never said clearly what they are; it apparently was not important to him to explain how a virtuous person decides how to act. And one influential interpretation only heightens the self-absorption in his view. On this interpretation an act is made noble by its motive. You have an initial motive to do an act, say, a desire to relieve another’s pain; you see that an act done from that motive will be noble; and you then have an

14 15 16

Aristotle, Nicomachean Ethics, 1115b20–1. Aristotle, Nicomachean Ethics, 1099a10–20, 1104b4–8, 1110b12–13, 1175b24–1176a3. Aristotle, Nicomachean Ethics, 1167b34–1168a18.

266

british ethical theorists from sidgwick to ewing

additional motive to do it, namely to act nobly, which then becomes your primary motive.17 So again your main concern is your act’s relation to your virtue rather than how it will benefit the other. Bernard Williams called an agent ‘morally selfindulgent’ if ‘what the agent cares about is not so much other people, as himself caring about other people’, or if he ‘focuses disproportionately upon the expression of his own disposition’.18 To Sidgwick Aristotle’s virtuous person was self-indulgent in precisely this sense. Ross was a pre-eminent Aristotelian scholar and translator of the Nicomachean Ethics. What did he think of Aristotle’s ethics? Like Sidgwick, Ross thought Aristotle took his start from the everyday ‘moral consciousness’, and was right to do so (FE 1–2; A 189, 219). Ross also cited Aristotle when denying that we can know a particular act is right; ‘the decision’, he quoted the Ethics as saying, ‘rests with perception’ (RG 42; KET 35).19 Some anti-theoretical interpreters have argued that for Aristotle moral thinking never involves the application of rules, just the immediate perception of what is right in a particular situation.20 Ross was not himself an anti-theorist, nor is it clear that Aristotle was. His explanation why ethics cannot be completely precise was that it deals with ‘things which are only for the most part true’,21 but he said the same of generalizations about the physical world, which likewise hold only ‘for the most part’. This leaves room for moral generalizations that hold ‘for the most part’, and that is one way of describing principles of prima facie duty. In Aristotle Ross did not interpret him this way. He imagined the Aristotelian agent applying rules, but ones that directly yield final verdicts on acts as prima facie rules do not (A 195, 215, 221). But in that book he seems not yet to have had the prima facie concept. Ross thought the formal structure of Aristotle’s view was ‘teleological’: ‘morality for him consists in doing certain actions not because we see them to be right in themselves but because we see them to be such as will bring us nearer to the “good for man” ’. Though Ross recognized that sometimes an act’s contribution to the Aristotelian good is ‘immanent’ rather than causal, ‘in the sense that it forms an element in the ideal life’, he thought this teleological structure a mistake, both because it is teleological and because it is egoistic (A 187–8, 188n, 230; KET 12). But later he argued that Aristotle did not actually follow it in his more detailed views. When Aristotle said a courageous person faces danger ‘for the sake of the noble’, he could mean either ‘because the action, the facing of danger, is itself noble’, or ‘for the sake of the noble object to be attained’. Though his teleology called for the second meaning, other remarks of his show that he intended the first:

17 18 19 20 21

Korsgaard, ‘From Duty and for the Sake of the Noble’. Williams, ‘Utilitarianism and Self-Indulgence’, pp. 45, 47. Aristotle, Nicomachean Ethics, 1109b23; also 1126b4. See, e.g. McDowell, ‘Virtue and Reason’; Wiggins, ‘Deliberation and Practical Reason’. Aristotle, Nicomachean Ethics, 1094b.

historians of ethics

267

in the actual treatment of the virtues Aristotle somewhat forgets his formal view; he nowhere attempts to deduce the necessity of any single virtue from the supreme end to be attained. He treats the agent as being moved to action by the contemplation of the ‘fineness’ of the good act itself, and thus becomes in his detailed treatment an intuitionist (A 204–5; also FE 5).

This reading of Aristotle was close to Sidgwick’s, and Ross elsewhere compared the Greeks’ use of kalon for acts to his own concept of fittingness to a situation (FE 53). But even if Ross’s later reading was correct, Aristotle’s virtuous person lacked the motivation mentioned above, of choosing an act as a means to another’s good, and Ross may have had that in mind when he criticized this person’s ‘self-absorption’. Our school were more critical of ancient views of the right than recent commentators have been, finding their egoism not harmless or merely ‘formal’ but fundamentally misguided. But most were more sympathetic to ancient accounts of the good. Sidgwick aside, the school were all largely perfectionists about the good and applauded the similarly perfectionist views of Plato, Aristotle, and the Stoics. Their valuing of virtue echoed the ancients’, as did the value many found in knowledge. It is hard to say how far their views derived from ancient models and how far they came to them independently. In Principia, for example, Moore quoted Plato’s Philebus at length and repeated one of its arguments against hedonism (PE 87–90).22 Was he influenced by Plato here or just citing someone with a similar view? Ross’s list of goods—pleasure, knowledge, virtue, desert—was close to Aristotle’s and may have been derived from it, though Ross never said so. Whatever the causal relation, many members had views about value close to those of the Greek philosophers; though they thought those philosophers wrong about the right, they often shared their views about the good.

12.2 The British Moralists Many in the school were interested in earlier British moral philosophers, especially of the seventeenth and eighteenth centuries. They responded similarly to their metaethical views, but differed about their normative ethics. All rejected the psychological hedonism of Hobbes and Locke and applauded what they took to be its refutation first by Butler and then by Hutcheson, Hume, and others. Sidgwick said he learned from Butler that we do not desire only our own pleasure (ME xix), and repeated Butler’s arguments against that view in Methods I 4. Moore too rejected psychological hedonism (EE 54–9, 66; PE 68–72), while Rashdall made it the subject of the first full chapter of his book; he also addressed nonhedonistic versions of egoism, such as the claim that all desire is for one’s ‘personal good’ or ‘self-realization’ (TGE I 38–43, II 61). Broad said psychological egoism:

An 1897 edition of this dialogue thanked ‘Mr. G.E. Moore’ for ‘notes and criticisms’ (The Philebus of Plato, ed. Bury, pp. vii, 215–17); Moore therefore had been studying it closely. 22

268

british ethical theorists from sidgwick to ewing

has naturally been popular both with vicious persons who wanted a philosophical excuse for their own selfishness and with decent people who felt slightly ashamed of their own virtues and wished to be taken for men of the world. One of Butler’s great merits is to have pointed out clearly and conclusively the ambiguities of language which make it plausible. As a psychological theory it was killed by Butler; but it still flourishes, I believe, among bookmakers and smart young business men whose claim to know the world is based on an intimate acquaintance with the shadier side of it. In Butler’s day the theory moved in higher social and intellectual circles, and it had to be treated more seriously than any philosopher would trouble to treat it now. The change is very largely the result of Butler’s work; he killed the theory so thoroughly that he sometimes seems to the modern reader to be flogging dead horses. Still, all good fallacies go to America when they die, and rise again as the latest discoveries of the local professors. So it will always be useful to have Butler’s refutation at hand. (FT 54–5; also Carritt, EPT 45; Ross, FE 293–4; Ewing, E 25–7)

Butler said psychological hedonism gets the order of explanation backward. We do not first desire pleasure and then desire ends such as honour and power because we think they will give us pleasure. We first have particular desires for particular external objects, and are pleased when we attain those objects because pleasure follows on all fulfilment of desire. Butler made the strong claim that we never desire pleasure as such but enjoy it only as a side effect of fulfilling some other desire. Sidgwick, Rashdall, Ross, and Broad countered that we do sometimes desire pleasure; even if a starving person wants just to eat, a gourmet aims at enjoyment (ME 44–5, 136; TGE I 15–16, 18–19, 23–4, 36; FE 293–7; FT 66–70). Rashdall granted the hedonist that when fulfilling a desire gives us pleasure, that strengthens the desire, whereas when it does not, the desire weakens (TGE I 33). And Ross thought many desires are, if not for pleasure as such, then for particular pleasures such as those of a specific dish (FE 295–6). These qualifications aside, all thought Butler had made a major contribution. If we can have ‘disinterested’ desires, we can benevolently desire others’ good and want to do what is right because it is right. To Broad, Butler had, in contrast to Kant’s ‘moral fanaticism . . . the solid common-sense and the sweet reasonableness of an English bishop of the eighteenth century’ (FT 53).23 In metaethics the earlier moralists divided into two groups. Cudworth, More, Cumberland, and Clarke and, later, Butler, Price, Reid, and Stewart, had nonnaturalist views close to our school’s: moral judgements can be objectively true, are not derivable from non-moral judgements, and are known by the same faculty of reason that knows mathematical truths. By contrast, Shaftesbury and then Hutcheson, Hume, and Smith asked empirical questions about the origin of our moral judgements, which they found in a ‘moral sense’ and ultimately in feelings such as

23 Both Broad and Carritt cited Butler’s refutation of Hobbes’s hedonistic account of pity as, respectively, ‘a model of philosophical reasoning,’ and as ‘lucid and successful a philosophical reputation [sic]’ as he could recall (FT 64; Fifty Years a Don, p. 90).

historians of ethics

269

sympathy; this often led to the view that moral judgements report feelings. Our school applauded the first group but were critical of the second. When Sidgwick abandoned Mill’s naturalism and became a utilitarian ‘on an Intuitional basis’, he adopted the view of ‘the earlier English Intuitionists, More and Clarke’ (ME xix). Rashdall said Cudworth stated as ‘emphatically’ as anyone before Sidgwick that ‘good’ is indefinable (TGE I 136n), while Prichard quoted at length from Cudworth and Clarke when disputing Green’s view that we can have obligations only in a politically ordered society. Those two had denied that the will either of God or of a sovereign can create moral obligation, which instead is ‘natural’; to Prichard they ‘performed an outstanding service’ and have ‘been far too much neglected’ (MW 245). He could also have quoted Price’s anticipation of his thesis in ‘Mistake’: To ask, why are we obliged to practise virtue, to abstain from what is wicked, or perform what is just, is the very same to ask, why we are obliged to do what we are obliged to do? It is not possible to wonder at those, who have so unaccountably embarrassed themselves, on a subject that one would think was attended with no difficulty.24

Price may have been the most sophisticated of these moralists and was much admired. Sidgwick said that, whereas Clarke made misleading analogies between moral and mathematical or physical truths, Price saw that ‘right’ and ‘wrong’ are ‘“single ideas” incapable of definition or analysis’ (OHE 224). Rashdall called his Review of the Principal Questions and Difficulties in Morals ‘the best work on Ethics published till quite recent times’ (TGE I 81n), Carritt listed it as one of the four books in ethics that had influenced him most,25 and Broad said that until The Right and the Good, ‘there existed, so far as I know, no statement and defence of what may be called the “rationalistic” type of ethical theory comparable in merit to Price’s’ (CE 188). Shaftesbury, by contrast, rejected non-naturalism in favour of a sentimentalist or subjectivist metaethic and thereby founded the moral-sense school. Sidgwick thought the tendency his ideas gave a period of English ethical thought ‘to take a psychological rather than an ethical turn’ was unfortunate. It made the fundamental questions ‘what is right?’ and ‘why?’ drop into the background, ‘not without manifest danger to morality’ (OHE 223). He and later members had three main objections to the moral-sense theorists. The first was that by grounding morality in feelings that can vary from person to person they destroyed its universality. Hutcheson had wondered why the moral sense might not differ in different people but did not see the ‘peril’ in this, or the way it can make different acts right for different people (ME 104; also Rashdall, TGE I 145; Carritt EPT 38–9, 41–2). This objection is telling against an equation of ‘right’ with ‘causes a certain feeling in me’ but not against one that equates ‘right’ with ‘causes a certain feeling in most people’; Prichard thought the latter was Hutcheson’s view and 24

Price, ‘Review of the Principal Questions’, sec. 714.

25

Carritt, Fifty Years a Don, pp. 89–90.

270

british ethical theorists from sidgwick to ewing

Broad thought it was Hume’s (MW 116, 169; FT 85, 260). But of this view we can still ask whether different acts would be right if most people had different feelings, and also make the other objections to subjective naturalism surveyed in Chapter 4. The second objection was that moral-sense views cannot account for what Butler called the ‘authority’ of moral judgements, or their superiority over other motives to action. The feeling these judgements are said to rest on can conflict with another feeling, say a desire for pleasure. If it does, why should we act on the former rather than the latter? Sidgwick said ‘the mere existence of a moral sentiment is in itself no reason for obeying it’ (OHE 223), while Rashdall asked, ‘If moral approbation is a mere feeling, how can it claim any superiority over other feelings?’ (TGE I 143, also II 195; Carritt, EPT 185; Ewing DG 57–8). To say conscience has authority is really to say we ought to act on it when it conflicts with other motives. But what is the status of this ‘ought’? If it merely reports another feeling, the question why we should act on it rather than on others arises again. The third objection said moral-sense views cannot capture what is distinctive in moral judgements. Hume said we approve a virtuous character because we feel sympathy with the pleasures it causes others, but a convenient house can also cause pleasure, as can non-moral traits of persons such as intelligence and wit. Of inanimate objects like a house Hume said it is just a fact that some emotions take only rational beings as their objects,26 but Sidgwick found that remark ‘unsatisfactory’ (OHE 212). And Hume positively assimilated non-moral traits and virtues, including items of both types on a list of ‘useful and agreeable qualities’. Sidgwick thought this ‘obliterate[s]’ the common-sense distinction ‘between virtues and other excellences of behaviour’ (ME 220; also OHE 210, 212; Carritt, EPT 40; Ewing, DG 165), a distinction he thought we draw on the basis of voluntariness. But he had to grant that virtue is not always in our power (ME 220), and, given that, it is unclear how his own instrumental account of virtue avoids his objection. A better view is Rashdall’s and Ross’s that virtue is intrinsically good, as against the merely instrumental goodness of houses, wit, and intelligence, and differs from other intrinsic goods by being the higher-level good of responding fittingly to other goods and evils. In Chapter 4 I said objections like the first two are not decisive against a noncognitivism that takes moral judgements to express attitudes that are universal and categorical, and that view can be combined with the higher-level account of virtue. But the moral-sense theorists never clearly stated a non-cognitivist view, and the school’s objections to them had force. In normative ethics Sidgwick applauded those whose views were proto-utilitarian, though not without reservations. He thought the earliest of them, Cumberland and Shaftesbury, wanted so much to show that benevolence toward others promotes your own happiness that, like the ancients, they did not really distinguish ‘own good’ and

26

Hume, Enquiry Concerning Morals, p. 213n; also Treatise of Human Nature, p. 617.

historians of ethics

271

‘right’ (OHE 174–5, 186–9). Butler was the first to make this distinction clearly, by naming reasonable self-love and conscience, or universal and egoistic reason, as separate regulative principles (EEM 92; ME 405; OHE 195–8), and his doing so was ‘an important step’, highlighting the ‘fundamental difference between the ethical thought of modern England and that of the old Greco-Roman world’ (OHE 197). But there were competing interpretations of Butler. In one passage Butler said that if conscience and self-love were to conflict, as he thought they cannot, conscience would have to give way, since we cannot justify any action ‘till we are convinced that it will be for our happiness’.27 Prichard took this to show that Butler’s view was really egoistic and involved an argument like the second in Plato’s Republic: we know what is right but can be motivated to do it only if we believe it is for our interest (MW 30, 34–5, 125). Broad, by contrast, emphasized Butler’s claims about the supreme authority of conscience and took the above passage not to assert his own view but to make just ‘a hypothetical concession to an imaginary opponent’ (FT 80; also CE 269). Sidgwick’s reading was between these two. He thought Butler recognized the two principles but with a ‘guarded optimism’ assumed they never conflict (OHE 197). Butler therefore did not really face the issue their duality creates, and neither, Sidgwick thought, did later writers such as Bentham and Mill. Sidgwick himself was the first to do so, and his praise of Butler was ultimately for pointing the way to his own philosophy. Sidgwick also thought the early utilitarians did not address the conflict between their theory and common-sense deontology. Cumberland formulated his consequentialist principle ‘not to supersede but to support’ common sense, and his moral conservatism was shared by Shaftesbury, Hutcheson, and even Hume; only later, in Paley and Bentham, did the reforming side of utilitarianism come to the fore (ME 86, 86n2, 423; OHE 174, 200–1). But such conservatism seems inevitable if one is trying to explain why we make the moral judgements we do rather than to determine which ones are true. On the other side, Sidgwick thought Clarke, though commonly considered a deontologist, recognized only universal benevolence as, alongside universalizability, constituting our social duty (ME 86; OHE 180–1), while Butler’s earlier writings too saw no conflict between conscience and benevolence. Only in his later ‘Dissertation on Virtue’ did Butler deny that benevolence is the whole of virtue and affirm independent duties forbidding ‘falsehood, unprovoked violence, injustice’ (ME 86n2; OHE 199–200).28 Ross applauded these ‘weighty words’ of Butler’s, saying they ‘answer better to what we really think’ than does consequentialism (FE 77–9); Cudworth, Price, Reid, and Stewart too were deontologists. Sidgwick found all these writers philosophically unsatisfying. Cudworth contented himself with ‘simply reaffirming’ the absoluteness of common-sense principles rather than giving them a ‘systematic exposition’ (ME 103–4n; OHE 171); Butler did not try

27

Butler, Fifteen Sermons, Sermon XI, sec. 423.

28

Butler, ‘Dissertation’, para. 434.

272

british ethical theorists from sidgwick to ewing

to reduce his dictates of conscience ‘to self-evident intuitions’ (OHE 199); and Price, Reid, and Stewart never penetrated to ‘profounder principles’ (ME 104; OHE 226, 230–1). These objections reflected Sidgwick’s general hostility to pluralist deontology and seem to insist that it be grounded in a non-deontological principle. By contrast, Ross endorsed Butler’s pluralism and also applauded Clarke and Price for equating rightness with ‘fitness . . . to a certain situation’ (FE 52, 54n), while Broad said Price saw that what is self-evident is not that an act is right or wrong all things considered but that it ‘tends to be right in respect of being an act of truth-telling and to be wrong in respect of being one of betrayal. . . . These points were made clearly enough by Price, but have since been made much more clearly by Ross’ (CE 202). To Prichard, Ross, and Broad the deontologists Sidgwick dismissed had essentially the correct normative view. Combined with their non-naturalism, it made them come closest of all previous moral philosophers to the truth.

12.3 Kant’s Ethics Our school’s attitude to Kant was mixed. He was often seen as the founder of philosophical Idealism and was therefore reviled by some Cambridge realists. Moore, for example, said a false equation of ‘to be true’ with ‘to be thought in a certain way’ was essential to Kant’s ‘Copernican revolution’ and ‘renders worthless the whole mass of modern literature, to which that revolution has given rise’ (PE 133; also Broad, FT 10).29 All the school rejected Kant’s substantive views about the good, the right, and what explains what is right. But many admired him as perhaps the greatest defender of the autonomy and rationality of morality. Prichard wrote: [Kant’s] moral philosophy is of course open to many obvious criticisms. Nevertheless he always strikes me as having, far more than any other philosopher, the root of the matter in him. . . . He will have nothing to do either with the idea that the rightness of action depends on its being for our own good, or with the idea that we think of it as so depending, or with the idea that desire for our own good is our only motive. And it is, I think, for this reason that in spite of his obvious mistakes he retains so close a hold on his readers. (MW 48–9, also 62–3, 126; also Rashdall, TGE I 108; Ross, RG 157; KET 12; Broad, FT 123)

Their commentaries centred on the Groundwork of the Metaphysics of Morals, which Kant opened by arguing that the only thing good without qualification is a good will, one that wills duty as duty. All our school accepted goods other than good motivation, such as pleasure and knowledge, and good motives other than duty, such as compassion. Some attributed Kant’s denigration of these motives to his partial acceptance of psychological hedonism, his belief that all motives other than that of duty aim at our pleasure (Rashdall, TGE I 122–3, 130–1; Carritt, TM 85; Ross, FE 29

See also Prichard’s vigorous realist critique of Kant’s epistemology in Kant’s Theory of Knowledge.

historians of ethics

273

227–8). (If he was woken from his dogmatic slumbers in epistemology by Hume, he could have used Butler’s alarm clock in moral psychology.) But all found his view of the good too narrow. Ross took him to have two arguments for his view: any things other than a good will that might be thought good, such as understanding and courage, can have bad effects if they are not combined with a good will, and they can make for a bad whole if combined with something other than a good will. He found these arguments ‘curiously indirect’: the natural way to decide whether something is good without qualification is to contemplate it itself, not what happens when it combines with other things (KET 10). He also questioned the arguments. The first is obviously fallacious, since it does not follow from the fact that something can be instrumentally bad that it is not good in itself (KET 10–11). And, as Moore also noted, even a good will can have bad effects if combined with false beliefs about your duty (PE 180). As for the argument about wholes, Ross said it ignores the principle of organic unities. It does not follow from the fact that the combination of moral vice and happiness is bad that the happiness is not good (KET 11; also Broad, FT 117); in fact, one explanation why the combination is bad is that in it a bad person enjoys something good. Many read Kant’s view as especially strict. They thought he found worth only in motives involving explicit thoughts about duty, and Ross and Broad thought he held that an act done from duty and some other motive such as compassion has less value than if it was done from duty alone; Ross thought the resulting value was the average of the values of the two motives, which, if compassion has no value, is half the value of duty alone (RG 170–2; FE 305, 325; KET 17–18; FT 121). But Kant could have held, less strictly, that an act done from a non-moral motive has worth so long as it is regulated by duty, so you would not do it if you had reason to believe it is wrong. This is the view Rashdall proposed, and Ewing thought it may have been Kant’s (‘PKE’ 50–2).30 Kant could also have held that the presence of an additional motive, so long as it is not evil, does not affect the value of your action; what matters is only whether your motive of duty was strong enough to make you act rightly without supporting motives. Ewing again thought this may have been Kant’s view (‘PKE’ 53). But even if weakened in these ways, Kant’s view had two features our school rejected. Whereas he thought an act done only from explicit concern for duty has all the value it can have, they thought one done from duty and, say, compassion has more; for Ross its value was the sum of the values of its two motives, with the second’s only a little less than the first’s (RG 171–2; FE 305). And whereas Kant thought an act not regulated by duty has no value, our school thought choosing what is best or right for the properties that make it so is good even if in different circumstances the same motive could make you act wrongly. What matters is your attitudes’ fit with your actual situation, not how they would fit a different situation.

30

Baron, ‘The Alleged Moral Repugnance’; Herman, ‘On the Value of Acting’.

274

british ethical theorists from sidgwick to ewing

Some took Kant to derive his claims about the right directly from his claim about the good, so an act is right just in case it is done from the motive of duty. Prichard and the Ross of The Right and the Good thought this derivation confuses ‘right’ and ‘morally good’ and generates the difficulty discussed in Chapter 2: the motive of duty involves the thought that an act is right apart from any motive, so either that thought is true and contradicts Kant’s view or acting rightly requires a false belief (MW 11, 154–9, 219; RG 5–6, 156). In Foundations Ross revised his interpretation and said Kant grasped the distinction between ‘right’ and ‘good’ because he distinguished between an act that merely conforms with duty and one done from duty (FE 139). All rejected Kant’s absolute deontology, expressed most famously in his denial that it can ever be right to lie, even to a would-be murderer about the whereabouts of his intended victim. The consequentialists of course did so, with Rashdall accusing Kant of confusing a correct sense of ‘categorical imperative’ that allows no exceptions for inclination and an incorrect one that allows no exceptions for circumstances (TGE I 116). But the deontologists did so too. Prichard and Ross said Kant ‘went too far’ or ‘overshot the mark’ (MW 63–4; FE 313) in not recognizing that even deontological duties can conflict, and that when they do any duty can be outweighed. Even if Kant did not think right acts must be done from the motive of duty, he thought an ultimate moral principle cannot have any substantive content, such as a reference to promise-keeping or maleficence. That would make the principle ‘material’ and therefore ‘heteronomous’, and support only hypothetical, not categorical, imperatives. Hence his attempt to derive the content of the moral law from just its form as categorical and universal. Ross found Kant’s worry here groundless. A duty to keep our promises whatever we desire is perfectly categorical; its being material does not at all make it hypothetical, and there is no good objection to substantive ultimate principles (KET 44). Many in the school objected to Kant’s talk of a ‘moral law’. Moore said this language suggests an analogy between moral principles and laws of nature that blurs the distinction between ‘ought’ and ‘is’ (PE 126–7), while Prichard, Carritt, and Ross thought Kant equivocated between these two senses of ‘law’ when he argued that the action of a free will must be governed by the moral law (MW 71–2; TM 128; KET 71). A different objection said his language suggests a false analogy between moral principles and civil laws, which are commanded by an external authority and do not by themselves generate moral obligations (Moore, PE 127–8; Prichard, MW 52–3, 116, 127, 169; Ross, KET 25; Ewing, ST 97, VR 186). Kant did not think the moral law is commanded externally, saying we impose it on ourselves. But Broad could find no ‘clear meaning’ in the idea of a self-imposed principle (FT 133), while Prichard said any suggestion that we ‘make’ moral principles as a legislator makes laws is a ‘fundamental mistake’: we ‘find’ moral principles (MW 57–8). Ross too said the moral law ‘is a thing which we do not make but find’; its rule may be self-imposed in the sense that we choose to act on it, but in knowing it we know an independent objective fact (KET 26, also 59–60).

historians of ethics

275

Many criticized Kant’s attempts to derive substantive moral conclusions from his three formulations of ‘the’ categorical imperative, which he took to express the form of the moral law. Their objections were often ones now seen as standard, and they seem also to have been common then. Thus Carritt said the failure of Kant’s deductions ‘has been so often exposed that the task of insulting over it is ungracious, but it must be faced’ (TM 79). The first or ‘universal law’ formulation says an act is wrong if we could not will the maxim or intention with which it is done to be a universal law, or to be acted on by everyone. Kant illustrated it by deriving from it two ‘perfect’ duties and two ‘imperfect’ ones, where the latter allow some choice in how they are fulfilled. With the perfect duties the universalization of the forbidden maxim cannot be thought, because it is self-contradictory; with the imperfect ones it can be thought but cannot be willed. His first perfect duty was not to commit suicide in order to escape pain. As Prichard and Carritt noted, he did not ask what would happen if everyone committed suicide (MW 61; TM 80–1);31 he considered one person and said it would be inconsistent for her to destroy her life by means of a faculty of self-love whose purpose is to preserve life. But this argument requires the dubious assumptions that self-love has a natural purpose and that it is, more specifically, to preserve life whatever its condition. Carritt and Ross asked why nature’s purpose might not be ‘to preserve the happy and successful, and let the others go hang themselves’ (TM 81) or to have us ‘improve our condition when we can’ but end our lives when we cannot (KET 46; also Ewing, ‘PKE’ 45–6). Kant’s second perfect duty was not to make lying promises, ones you do not intend to fulfil, and his argument was that if everyone did this no one would believe promises and promises could not be made. Many again were unpersuaded. Ross said that even if promises were known always to be broken they could still be useful, as telling us what a promissor will not do (KET 30). Insincere promising could also be universal if people never learned that promises had been broken, and even if promising would die out that does not, Broad said, involve any strict inconsistency, since it turns on contingent facts about how people would react (FT 130; also Rashdall, TGE I 114; Ewing ‘PKE’ 44–5). Prichard argued, differently, that though lying promises could not be successful, in the sense of deceiving anyone, they could still be made, by which he seems to have meant that they would still create obligations: a lying promise still binds (MW 60). It may be replied that a promise is binding only if it is accepted by a promisee, but then Kant’s argument rests on a normative premise about when promises oblige that itself needs justification. And whether or not we call it ‘promising’, it is surely possible to bind yourself unilaterally, for example by making a vow, and to do so even if no one thinks you will follow through. 31

Rashdall did think Kant asked this (TGE I 114–15), but as Ewing pointed out (‘PKE’ 45), this is a misinterpretation.

276

british ethical theorists from sidgwick to ewing

Like many since, our critics rejected Kant’s view that his test yields exceptionless deontological duties: it does so only given a specific act-description when other equally good descriptions are available. Sidgwick was the first to make this point (ME 318–19, 486–7), but Ross developed it more fully and illustrated it with the case of lying to the would-be murderer. (Kant applied his test only to lying promises, but if he thought there is a perfect duty not to lie more generally, he must have thought the test validates it.) Kant’s prohibition of this act requires evaluating it under the description ‘a lying’, but it can also be described more specifically, as ‘a lying to a murderous person’, or more abstractly, as ‘a statement’, and under neither of these descriptions is its universalization problematic. Kant’s argument, Ross concluded, ‘pitches, arbitrarily, on the middle one’ of these descriptions when there is no justification for doing so. More generally, since his test yields different results when combined with different actdescriptions, it is not a useable criterion of rightness (KET 32–3; also Joseph, SPE 99; Ewing, ‘PKE’ 49–50). This objection forgets that a Kantian maxim is a mental state of intending that for any person in particular circumstances has a specific content. If you lie to a would-be murderer you do not intend just to make a statement; your maxim is not that abstract. But if you lie to him only because he is a murderer and would not do so otherwise, your maxim is also not just to lie. It is arguably to lie to a would-be murderer, and that maxim can be universalized. These objections say Kant’s test does not yield the conclusions he wanted, but others say it yields unwanted ones. Rashdall said that if making a lying promise is wrong because everyone’s doing so would bring promising to an end, then philanthropy, peacemaking, and celibacy are wrong because universalizing them likewise removes the conditions that make them possible (TGE I 114–15). Carritt and Ross repeated these examples (TM 79–80; KET 32), and there are simpler ones. Because it is impossible for everyone to act on maxims like ‘lend money but never borrow it’ or ‘sell lettuce but never buy it’, Kant’s first test seems to make such policies seriously wrong. Kant’s derivations of his imperfect duties used the weaker test that forbids an act if the universalization of its maxim cannot be willed. Thus you have a duty to develop your talents because they are means to all possible ends, and as a rational being you will the means to your ends. And you have a duty to promote others’ happiness because, though a situation in which no one does so is conceivable, you cannot will it since you will sometimes need and want others’ aid. Ross noted that, like the argument about suicide, the one about developing your talents does not universalize but considers just a single person. We can wonder, however, whether it yields all the results we want: do our talents for pure mathematics or aesthetic appreciation really serve further ends? Even if it does, Ross thought it gives the duty the wrong explanation. Whereas scientific knowledge and the appreciation of beauty are good in themselves, Kant’s derivation values them only as means to other ends and so yields just a hypothetical not a categorical imperative to pursue them (KET 46–7).

historians of ethics

277

Several members raised an objection to the derivation of the duty of beneficence that has since become standard. Whereas Kant thought you cannot will universal non-aid, Sidgwick said, ‘We can certainly conceive a man in whom the spirit of independence and the distaste for incurring obligations would be so strong that he would choose to endure any privations rather than receive aid from others’ (ME 389); Carritt and Broad mentioned a strong person or one committed to an egoistic principle (TM 82; FT 130–1). And Kant’s argument may again assume his faulty moral psychology. Given his belief that all desires not connected to duty aim at your happiness, he may have thought you cannot avoid wanting the outcome in which your happiness is best secured. Without that assumption a person who cares more about self-reliance becomes possible. Many also had explanatory objections to the universal-law test. Prichard and Ross thought Kant erred, like the consequentialists, in proposing a single ground for all duties (MW 62; RG 18–19; FE 189; KET 28–9), and Prichard had a more specific objection. ‘No one’, he wrote, ‘could suppose that the reason why an act ought to be done consists in the fact that every one could do it’. (MW 59). Thomas E. Hill, Jr has echoed this objection, saying ‘What is wrong with slavery, for example, is not adequately explained by saying that it is impossible for everyone to act [on] the maxim of a would-be slave-owner’.32 The charge is that the universalization test diverts our attention from what really matters in many wrong acts: their effect on their particular victim. As Broad said, if a friend asks if it would be right to murder a wealthy uncle who has remembered him in his will: I shall not consider it appropriate to point to him that, if everyone murdered his wealthy uncle from whom he had expectations, a deplorable insecurity would prevail among a deserving class of men and an excuse would be provided for them to leave their money to missionary societies. I should rather insist on the loss of well-being to Uncle Joseph himself. (CE 51)

The member most sympathetic to the universal-law tests was Ewing. He thought the first or contradiction test fails but a version of the second is more promising. Any maxim, he said, includes a purpose in light of which you act, and willing its universalization can be self-defeating in the sense that it undermines that purpose. Thus a liar aims to deceive others but could not achieve that goal if everyone lied, whereas a philanthropist aims to reduce poverty and that end would not be frustrated—it would be fulfilled—if so many people acted philanthropically that poverty vanished; contra critics like Rashdall, these two cases are different (‘PKE’ 44–5).33 Even Ewing, however, denied that Kant’s test provides a ‘strict proof ’ of duties; the argument about lying, for example, could not move someone, strange as he would be, who thought there is a categorical duty to lie (‘PKE’ 48–9). Nor does it

Hill, Jr, ‘Kantian Normative Ethics’, p. 488. He here anticipated the interpretations in Nell (O’Neill), Acting on Principle, and Korsgaard, ‘Kant’s Formula’. 32 33

278

british ethical theorists from sidgwick to ewing

give the only explanation of duties. Another reason to help others in distress ‘is simply that it produces good or at least lessens their pain (an evil)’ (‘PKE’ 47), and there is the practical difficulty of deciding which act-descriptions to use when applying it (‘PKE’ 49–50). While thinking better of the universal-law test than others, Ewing still did not find it the complete ground of duty. Kant’s second formulation of the categorical imperative says we must never treat another person only as a means but always also as an end. Ross thought this formula has ‘great homiletic value’, because the notion of the value of human personality makes ‘a much warmer appeal to the human heart’ than that of law or ‘the bare logical notion of universalizability’ (KET 53; also Ewing, E 62). It also avoids one of the explanatory objections to the first formulation, since it does make what is wrong with lying or killing the way it treats its victim. But our school still found it problematic. Many objected to the idea that persons are ends, saying that by ‘end’ we commonly mean ‘something to be realised’ (Sidgwick, ME 390) or a desired effect that leads to action (Prichard, MW 56; Ross, KET 51; Broad, CE 275–6). Kant took persons to be ends in a different sense, one in which they are ‘self-subsistent’ or ‘independently existing’ and therefore not to be acted against by being treated just as means. But they rejected this idea, with Ross saying ‘the notion of self-subsistent ends is nothing but an embarrassment to Kant’. This conclusion was too hasty. Even if ends must be states of affairs, you can have a stronger aversion to destroying existing ones than a desire to create new ones, as neo-Thomist deontology recommends. And would there be a decisive objection to Kant if he had said only that persons have intrinsic value and not used the word ‘end’? Even if only states of affairs can be good, as Ross for example held, that must be argued for rather than assumed. The more serious objections said the second formulation is too vague to yield determinate results. Rashdall and Ewing thought that to treat a person as an end is to promote his good, an idea that has no content until it is supplemented by claims about what the good is (TGE 131–3; E 62–3). Prichard took the ban on treating another only as a means to say he ‘has certain claims on us, i.e. rights against us, which we ought to respect’, so ‘we ought to perform our obligations’ to him; it therefore presupposes substantive duties and cannot ground them (MW 56–7). Broad and Ross too thought the point of Kant’s formula was that others have claims on us that it is wrong to ignore (FT 132; CE 276–7; KET 49). A more moderate version of this objection says only that the idea of treating another as an end admits of different interpretations with different implications. If you kill one innocent person in order to save five, a consequentialist can say you treat the one as an end so long as you count his good equally with that of the five; though outweighed, it is still valued in itself. A deontologist will deny that this suffices for treating him as an end, which requires not sacrificing him. But this dispute cannot be resolved by reflecting on what ‘treating as an end’ really means; it is a matter of substantive moral principle.

historians of ethics

279

Again there were objections to Kant’s derivations from this formula. If a person’s rational nature consists in her ability to govern her actions by moral principles, it would seem to follow that one duty is to promote her virtue or moral perfection. Kant denied any such duty, saying only your own perfection is in your power.34 But as Sidgwick and others argued, though you cannot cause another now to will morally, you can promote or cultivate her moral willing in the future no less than you can your own (ME 240; Rashdall, TGE I 112–13; Ross, RG 26, KET 56). Kant’s derivation of a duty to promote others’ happiness was also questioned. He equated happiness with the satisfaction of impulses from the non-rational side of a person’s nature, and Sidgwick and Carritt asked how respecting her rationality can require any concern for that (ME 390; TM 83). Ross seemed to grant that valuing her as an end does require not interfering with her (KET 51–2), but this too can be questioned. If she is acting morally, respect for her rationality may require leaving her alone. But what if she is acting on inclination against morality, or acting on inclination in a situation where morality does not speak, for example when she is choosing her breakfast? What does the value of her rational nature say against interfering with her then? The school did not much discuss Kant’s third formulation, which combines ideas from the first two, believing it inherits the flaws of the first two. And Ross and Ewing thought the first or universal-law formula was the main one Kant used for identifying right acts, with the second and third playing a more rhetorical role (KET 62, 64; ‘PKE’ 44). Kant was no metaethical naturalist, affirming as emphatically as any in our school the autonomy of ethics and the impossibility of deriving an ‘ought’ from an ‘is’. But like many naturalists he was leery of positing underivative substantive moral truths, at least about duty. Hence his attempt to derive the content of duty from the bare idea of moral principles as categorical and universal, or from the bare idea of a rational agent as able to govern her actions by such principles. Our critics thought his particular derivations fail: they do not yield the results he wanted, yield others that are unwanted, and give their results the wrong explanation. But Kant was hardly alone in pursuing this type of project. Many philosophers from Plato and Aristotle to the present have wanted to avoid positing underivative duties and have therefore tried to ground claims about how we ought to act in something that is not a claim about how we ought to act but is somehow more fundamental. Our school, by contrast, thought ethics does and must affirm underivative truths about both value and duty and that there is nothing problematic in its doing so. Their philosophical heroes were therefore moralists like Butler and Price who shared this view, but they themselves affirmed it especially clearly and emphatically. That is the core of their distinctiveness in the history of ethics: to hold that the most fundamental moral truths are true just in themselves and, ignoring more grandiose projects of justification, to explore their structure and implications.

34

Kant, Doctrine of Virtue, pp. 44–5.

Bibliography Adams, R.M. Finite and Infinite Goods: A Framework for Ethics (New York: Oxford University Press, 1999). Anderson, Elizabeth. Value in Ethics and Economics (Cambridge, MA: Harvard University Press, 1993). Annas, Julia. The Morality of Happiness (New York: Oxford University Press, 1993). Anonymous. Review of G.E. Moore, Principia Ethica, Athenaeum 3794 (Dec. 26, 1903): 847–8. Anscombe, G.E.M. ‘Modern Moral Philosophy’, Philosophy 33 (1958): 1–19. Aristotle. Nicomachean Ethics, trans. W.D. Ross and J.O. Urmson. In The Complete Works of Aristotle, vol. 2, ed. J. Barnes, pp. 1729–1867 (Princeton, NJ: Princeton University Press, 1984). Armstrong, D.M. Universals and Scientific Realism, 2 vols. (Cambridge: Cambridge University Press, 1978). Audi, Robert. The Good in the Right: A Theory of Intuition and Intrinsic Value (Princeton, NJ: Princeton University Press, 2004). Austin, J.L. ‘Agathon and Eudaimonia in the Ethics of Aristotle’. In Aristotle: A Collection of Critical Essays, ed. J.M.E. Moravcsik, pp. 261–96 (London: Macmillan, 1968). Austin, J.L. ‘A Plea for Excuses’, Proceedings of the Aristotelian Society 59 (1956–7): 1–30. Ayer, A.J. Language, Truth and Logic (London: Victor Gollancz, 1936). Ayer, A.J. Part of My Life (London: Collins, 1977). Baldwin, Thomas. ‘Editor’s Introduction’. In G.E. Moore, Principia Ethica: Revised Edition, ed. T. Baldwin, pp. ix–xxxvii (Cambridge: Cambridge University Press, 1993). Baldwin, Thomas. G.E. Moore (London: Routledge, 1990). Baron, Marcia. ‘The Alleged Moral Repugnance of Acting from Duty’, Journal of Philosophy 81 (1984): 197–220. Berlin, Isaiah. ‘Austin and the Early Beginnings of Oxford Philosophy’. In Essays on J.L. Austin, eds I. Berlin et al., pp. 1–16 (Oxford: Clarendon Press, 1973). Blackburn, Simon. ‘Moral Realism’. In Morality and Moral Reasoning, ed. J. Casey, pp. 111–29 (London: Methuen, 1971). Blackburn, Simon. Spreading the Word (Oxford: Oxford University Press, 1984). Blanshard, Brand. ‘Autobiography’. In The Philosophy of Brand Blanshard, ed. P.A. Schilpp, pp. 1–185 (La Salle, IL: Open Court, 1980). Blanshard, Brand. Reason and Goodness (London: George Allen & Unwin, 1961). Bosanquet, Bernard. Critical Notice of G.E. Moore, Principia Ethica, Mind 13 (1904): 254–61. Bouwsma, O.K. Wittgenstein: Conversations, 1949–1951, eds J.L. Craft and R.E. Hustuist (Indianapolis: Hackett, 1986). Boyd, Richard. ‘How to Be a Moral Realist’. In Essays in Moral Realism, ed. G. Sayre-McCord, pp. 181–228 (Ithaca, NY: Cornell University Press, 1988). Bradley, F.H. Appearance and Reality: A Metaphysical Essay, 2nd ed. (Oxford: Clarendon Press, 1897) [1st ed. 1893]. Bradley, F.H. Collected Essays, 2 vols. (Oxford: Clarendon Press, 1935).

282

bibliography

Bradley, F.H. Ethical Studies, 2nd ed. (Oxford: Clarendon Press, 1927) [1st ed. 1876]. Bradley, F.H. ‘Mr. Sidgwick on Ethical Studies, Mind o.s. 2 (1877): 122–5, reprinted in Sidgwick, Essays on Ethics and Method, 190–2. Bradley, F.H. ‘Mr. Sidgwick’s Hedonism’. In Bradley, Collected Essays, vol. 1, pp. 71–128; originally published separately (London: Henry S. King, 1877). Bradley, F.H. The Principles of Logic, 2nd ed. (London: Oxford University Press, 1922) [1st ed. 1883]. Braithwaite, R.B. ‘George Edward Moore’, Proceedings of the British Academy 47 (1961): 293–310. Brandt, Richard B. A Theory of the Good and the Right (Oxford: Clarendon Press, 1979). Brandt, Richard B. ‘Toward a Credible Form of Utilitarianism’. In Morality and the Language of Conduct, eds H.-N. Castaneda and G. Nakhnikian, pp. 107–43 (Detroit: Wayne State University Press, 1963). Brennan, Samantha. ‘Thresholds for Rights’, Southern Journal of Philosophy 13 (1995): 143–68. Brentano, Franz. The Origin of Our Knowledge of Right and Wrong, trans. R.M. Chisholm and E. Schneewind (London: Routledge & Kegan Paul, 1969). Brink, David O. ‘Common Sense and First Principles in Sidgwick’s Methods’, Social Philosophy and Policy 11/1 (Winter 1994): 179–201. Brink, David O. Moral Realism and the Foundations of Ethics (Cambridge: Cambridge University Press, 1989). Broad, C.D. ‘Analysis of Some Ethical Concepts’, Journal of Philosophical Studies (later Philosophy) 3 (1928): 285–99, reprinted in Broad’s Critical Essays, 63–81. Broad, C.D. ‘Are There Synthetic A Priori Truths?’, Proceedings of the Aristotelian Society, Supplementary Volume 15 (1936): 102–17. Broad, C.D. ‘Autobiography’. In The Philosophy of C.D. Broad, ed. P.A. Schilpp, pp. 1–68 (New York: Tudor Publishing, 1959). Broad, C.D. Broad’s Critical Essays in Moral Philosophy, ed. D. Cheney (London: George Allen & Unwin, 1971). Broad, C.D. ‘Certain Features in Moore’s Ethical Doctrines’. In The Philosophy of G.E. Moore, ed. P.A. Schilpp, pp. 41–68 (Chicago: Northwestern University Press, 1942). Broad, C.D. ‘Conscience and Conscientious Action’, Philosophy 15 (1940): 115–30, reprinted in Broad’s Critical Essays, 136–55. Broad, C.D. Critical Notice of H.A. Prichard, Moral Obligation, Mind 59 (1950): 555–66. Broad, C.D. Critical Notice of Julian S. Huxley, Evolutionary Ethics, Mind 53 (1944): 344–67, reprinted in Broad’s Critical Essays, 156–87. Broad, C.D. Critical Notice of Sir David Ross, Foundations of Ethics, Mind 49 (1940): 228–39. Broad, C.D. ‘The Doctrine of Consequences in Ethics’, International Journal of Ethics 24 (1914): 293–320, reprinted in Broad’s Critical Essays, 17–42. Broad, C.D. ‘Emotion and Sentiment’, Journal of Aesthetics and Art Criticism 13 (1954): 203–14, reprinted in Broad’s Critical Essays, 283–301. Broad, C.D. Ethics and the History of Philosophy: Selected Essays (London: Routledge & Kegan Paul, 1952). Broad, C.D. Examination of McTaggart’s Philosophy, 2 vols. (Cambridge: Cambridge University Press, 1933 and 1938). Broad, C.D. Five Types of Ethical Theory (London: Routledge & Kegan Paul, 1930).

bibliography

283

Broad, C.D. ‘Henry Sidgwick’. In Ethics and the History of Philosophy, 49–69 (London: Routledge & Kegan Paul, 1952). Broad, C.D. ‘Is “Goodness” a Name of a Simple Non-Natural Quality?’, Proceedings of the Aristotelian Society 34 (1933–4): 249–68, reprinted in Broad’s Critical Essays, 106–23. Broad, C.D. ‘John McTaggart Ellis McTaggart’. In Ethics and the History of Philosophy, 70–93 (London: Routledge & Kegan Paul, 1952). Broad, C.D. ‘Obligations, Ultimate and Derived’. In Festskrift tillagnad Karl Olivecrona (Stockholm: Norstedt & Soners Forlag, 1964), reprinted in Broad’s Critical Essays, 351–68. Broad, C.D. ‘On the Function of False Hypotheses in Ethics’, International Journal of Ethics 26 (1916): 377–97, reprinted in Broad’s Critical Essays, 43–62. Broad, C.D. ‘Ought We to Fight for Our Country in the Next War?’, Hibbert Journal 34 (1936): 357–67, reprinted in Broad’s Critical Essays, 124–35. Broad, C.D. ‘A Reply to My Critics’. In The Philosophy of C.D. Broad, ed. P.A. Schilpp, pp. 711–830 (New York: Tudor Publishing, 1959), partially reprinted in Broad’s Critical Essays, 302–23. Broad, C.D. ‘Self and Others’, The Herbert Spencer Lecture (1953), Oxford University, reprinted in Broad’s Critical Essays, 262–82. Broad, C.D. ‘Some of the Main Problems of Ethics’, Philosophy 21 (1946): 99–117, reprinted in Broad’s Critical Essays, 223–46. Broad, C.D. ‘Some Reflections on Moral-Sense Theories in Ethics’, Proceedings of the Aristotelian Society 45 (1944–5): 131–66, reprinted in Broad’s Critical Essays, 188–222. Broad, C.D. ‘Symposium on the Relations Between Science and Ethics’, Proceedings of the Aristotelian Society 42 (1941–2): 100A–100H. Broome, John. ‘Normative Requirements’, Ratio 12 (1999): 398–419. Broome, John. ‘Reasons’. In Reason and Value, eds R.J. Wallace, M. Smith, S. Scheffler, and P. Pettit, pp. 28–55 (Oxford: Oxford University Press, 2004). Brown, Lesley. ‘Glaucon’s Challenge, Rational Egoism and Ordinary Morality’. In Pursuing the Good: Ethics and Metaphysics in Plato’s Republic, eds D. Cairns, F-G. Hermann, and T. Penner, pp. 42–60 (Edinburgh: Edinburgh University Press, 2007). Bury, R.G., ed. The Philebus of Plato (Cambridge: Cambridge University Press, 1897). Butchvarov, Panayot. Skepticism in Ethics (Indianapolis: Indiana University Press, 1989). Butler, Joseph. ‘Dissertation of the Nature of Virtue’. In British Moralists, vol. 1., ed. D.D. Raphael, 378–86 (Oxford: Clarendon Press, 1969). Butler, Joseph. Fifteen Sermons. In British Moralists 1650–1800, vol. 1, ed. D.D. Raphael, 325–77 (Oxford: Clarendon Press, 1969). Carlson, Erik. ‘Organic Unities, Non-Trade-Off, and the Additivity of Intrinsic Value’, Journal of Ethics 5 (2009): 335–60. Carritt, E.F. ‘An Ambiguity of the Word “Good” ’, Proceedings of the British Academy 23 (1937): 51–80. Carritt, E.F. Ethical and Political Thinking (Oxford: Clarendon Press, 1947). Carritt, E.F. ‘Ethics in Philosophical Education’, Journal of Philosophy 25 (1925): 573–7. Carritt, E.F. Fifty Years a Don, typescript in University College Library, Oxford, 1960. Carritt, E.F. An Introduction to Aesthetics (London: Hutchinson’s University Library, 1949). Carritt, E.F. ‘Moral Positivism and Moral Aestheticism’, Philosophy 13 (1938): 131–47. Carritt, E.F. ‘Professor H.A. Prichard: Personal Recollections’, Mind 57 (1948): 146–8.

284

bibliography

Carritt, E.F. ‘The Sublime’, Mind 19 (1910): 356–72. Carritt, E.F. The Theory of Beauty (London: Methuen, 1914). Carritt, E.F. The Theory of Morals (London: Oxford University Press, 1928). Carritt, E.F. What is Beauty? A First Introduction to the Subject and to Modern Theories (Oxford: Clarendon Press, 1932). Chisholm, R.M. Brentano and Intrinsic Value (Cambridge: Cambridge University Press, 1986). Clark, G.N. ‘Sir David Ross’, Proceedings of the British Academy 57 (1971): 525–43. Cox, H.H. ‘Warnock on Moore’, Mind 79 (1970): 265–9. Crisp, Roger. ‘Pleasure and Hedonism in Sidgwick’. In Underivative Duty, ed. T. Hurka, 26–44 (Oxford: Clarendon Press, 2011). Crisp, Roger. Reasons and the Good (Oxford: Clarendon Press, 2006). Crisp, Roger. ‘Sidgwick and the Boundaries of Intuitionism’. In Ethical Intuitionism, ed. P. Stratton-Lake, 56–75 (Oxford: Clarendon Press, 2002). Cuneo, Terence. ‘Saying What We Mean: An Argument against Expressivism’. In Oxford Studies in Metaethics 1, ed. R. Shafer-Landau, pp. 73–98 (Oxford: Oxford University Press, 2006). Dancy, Jonathan. Ethics Without Principles (Oxford: Clarendon Press, 2004). Dancy, Jonathan. ‘In Defense of Thick Concepts’. In Midwest Studies in Philosophy XX: Moral Concepts, eds P.A French, T.E. Uehling, Jr, and H.K. Wettstein, pp. 263–79 (Notre Dame: University of Notre Dame Press, 1996). Dancy, Jonathan. Moral Reasons (Oxford: Blackwell, 1993). Dancy, Jonathan. ‘Should We Pass the Buck?’. In The Good, the True and the Beautiful, ed. A. O’Hear (Cambridge: Cambridge University Press, 2000), 159–73. Dancy, Jonathan. ‘Wiggins and Ross’, Utilitas 10 (1998): 281–5. Darwall, Stephen. ‘Moore, Normativity, and Intrinsic Value’, Ethics 113 (2003): 468–89. Darwall, Stephen. ‘Sidgwick, Concern, and the Good’, Utilitas 12 (2000): 291–306. Darwall, Stephen. ‘Under Moore’s Spell’, Utilitas 10 (1998): 286–91. Darwall, Stephen, Gibbard, Allan, and Railton, Peter. ‘Towards Fin de Siecle Ethics: Some Trends’, Philosophical Review 101 (1992): 115–89. Donagan, Alan. ‘Sidgwick and Whewellian Intuitionism: Some Enigmas’, Canadian Journal of Philosophy 7 (1977): 447–65. Dreier, James. ‘The Structure of Normative Theories’, The Monist 76 (1993): 22–40. Driver, Julia. ‘The Virtues of Ignorance’, Journal of Philosophy 86 (1989): 373–84. Duncker, Karl. ‘On Pleasure, Emotion, and Striving’, Philosophy and Phenomenological Research 1 (1941): 391–430. Dworkin, Ronald. ‘Objectivity and Truth: You’d Better Believe It’, Philosophy and Public Affairs 25 (1996): 87–139. Dworkin, Ronald. Sovereign Virtue: The Theory and Practice of Equality (Cambridge, MA: Harvard University Press, 2000). Dworkin, Ronald. ‘What is Equality? Part 1: Equality of Welfare’ and ‘What is Equality? Part 2: Equality of Resources’, Philosophy and Public Affairs 10 (1981): 185–246, 283–345. Edmonds, David and Edinow, John. Wittgenstein’s Poker: The Story of a Ten-Minute Argument Between Two Great Philosophers (London: Faber, 2001). Elga, Adam. ‘Reflection and Disagreement’, Nous 41 (2007): 478–502. Elstein, Daniel Y. and Hurka, Thomas. ‘From Thick to Thin: Two Moral Reduction Plans’, Canadian Journal of Philosophy 39 (2009): 515–36.

bibliography

285

Enoch, David. Taking Morality Seriously: A Defense of Robust Realism (Oxford: Oxford University Press, 2011). Ewing, A.C. ‘Blanshard’s View of Good’. In The Philosophy of Brand Blanshard, ed. P.A. Schilpp, pp. 222–36 (La Salle, IL: Open Court, 1980). Ewing, A.C. The Definition of Good (London: Routledge & Kegan Paul, 1947). Ewing, A.C. Ethics (London: English Universities Press, 1953). Ewing, A.C. The Fundamental Questions of Philosophy (London: Routledge & Kegan Paul, 1951). Ewing, A.C. ‘G.E. Moore’, Mind 70 (1962): 251. Ewing, A.C. Idealism: A Critical Survey (London: Methuen, 1934). Ewing, A.C. Kant’s Treatment of Causality (London: Kegan Paul, Trench, Trubner, 1924). Ewing, A.C. ‘The Linguistic Theory of A Priori Propositions’, Proceedings of the Aristotelian Society 40 (1939–40): 207–44. Ewing, A.C. ‘Meaninglessness’, Mind 46 (1937): 347–64, reprinted in Non-Linguistic Philosophy, 15–33. Ewing, A.C. The Morality of Punishment, with some Suggestions for a General Theory of Ethics (Montclair, NJ: Patterson Smith, 1970; originally published London: Kegan Paul, Trench, Trubner, 1929). Ewing, A.C. Non-Linguistic Philosophy (London: George Allen & Unwin, 1968). Ewing, A.C. ‘The Paradoxes of Kant’s Ethics’, Philosophy 13 (1938): 40–56. Ewing, A.C. ‘Punishment as a Moral Agency: An Attempt to Reconcile the Retributive and the Utilitarian View’, Mind 36 (1927): 292–305. Ewing, A.C. ‘Reason and Intuition’, Proceedings of the British Academy 27 (1941): 67–107, reprinted in Non-Linguistic Philosophy, 34–66. Ewing, A.C. ‘Recent Developments in British Ethical Thought’. In British Philosophy in the Mid-Century: A Cambridge Symposium, ed. C.A. Mace, pp. 65–95 (London: George Allen & Unwin, 1957). Ewing, A.C. Second Thoughts in Moral Philosophy (London: Routledge & Kegan Paul, 1959). Ewing, A.C. A Short Commentary on Kant’s Critique of Pure Reason (London: Methuen, 1938). Ewing, A.C. ‘A Suggested Non-Naturalistic Analysis of Good’, Mind 48 (1939): 1–22. Ewing, A.C. ‘Symposium on the Relations Between Science and Ethics’, Proceedings of the Aristotelian Society 42 (1941–2): 68–86. Ewing, A.C. ‘Utilitarianism’, Ethics 58 (1948): 100–11. Ewing, A.C. Value and Reality: The Philosophical Case for Theism (London: George Allen & Unwin, 1973). Ewing, A.C. ‘What Would Happen If Everybody Acted Like Me?’, Philosophy 28 (1953): 16–29. Feinberg, Joel. ‘Action and Responsibility’. In Doing and Deserving, pp. 119–51 (Princeton, NJ: Princeton University Press, 1970). Feinberg, Joel. ‘The Expressive Function of Punishment’. In Doing and Deserving, pp. 95–118 (Princeton, NJ: Princeton University Press, 1970). Feinberg, Joel. ‘Noncomparative Justice’, Philosophical Review 73 (1974): 297–338. Feldman, Fred. Doing the Best We Can (Dordrecht: Reidel, 1986). Field, G.C. Moral Theory: An Introduction to Ethics (London: Methuen, 1921). Finnis, John. Fundamentals of Ethics (Oxford: Clarendon Press, 1983).

286

bibliography

Finnis, John. Natural Law and Natural Rights (Oxford: Clarendon Press, 1980). Foley, Richard. Working Without a Net: A Study of Egocentric Epistemology (New York: Oxford University Press, 1993). Foot, Philippa. ‘Utilitarianism and the Virtues’, Mind 94 (1985): 196–209. Frankena, William K. Ethics, 2nd ed. (Englewood Cliffs, NJ: Prentice-Hall, 1973). Frankena, William K. ‘Ewing’s Case Against Naturalistic Theories of Value’, Philosophical Review 57 (1948): 481–92. Frankena, William K. ‘Obligation and Value in the Ethics of G.E. Moore’. In The Philosophy of G.E. Moore, ed. P.A. Schilpp, pp. 91–110 (Chicago: Northwestern University Press, 1942). Fried, Charles. Right and Wrong (Cambridge, MA: Harvard University Press, 1978). Gampel, Eric. ‘A Defense of the Autonomy of Ethics: Why Value is Not Like Water’, Canadian Journal of Philosophy 26 (1996): 191–209. Geach, P.T. ‘Assertion’, Philosophical Review 74 (1965): 449–65. Geach, P.T. ‘Good and Evil’, Analysis 17 (1956): 33–42. Geach, P.T. Truth, Love and Immortality: An Introduction to McTaggart’s Philosophy (Berkeley and Los Angeles: University of California Press, 1979). Gert, Bernard. Morality: Its Nature and Justification (New York: Oxford University Press, 1998). Gert, Joshua. Brute Rationality: Normativity and Human Action (Cambridge: Cambridge University Press, 2004). Green, Thomas Hill. Lectures on the Principles of Political Obligation. In Lectures on the Principles of Political Obligation and Other Writings, eds P. Harris and J. Morrow, pp. 13–193 (Cambridge: Cambridge University Press, 1986). Green, Thomas Hill. ‘Liberal Legislation and Freedom of Contract’. In Lectures on the Principles of Political Obligation and Other Writings, eds P. Harris and J. Morrow, pp. 194–212 (Cambridge: Cambridge University Press, 1986). Green, Thomas Hill. Prolegomena to Ethics, ed. A.C. Bradley (Oxford: Clarendon Press, 1883). Greenspan, Patricia S. ‘Conditional Oughts and Hypothetical Imperatives’, Journal of Philosophy 72 (1975): 259–76. Grice, G.R. ‘Alfred Cyril Ewing’, Proceedings of the British Academy 59 (1973): 499–513. Griffin, James. Well-Being: Its Meaning, Measurement, and Moral Importance (Oxford: Clarendon Press, 1986). Hampton, Jean. ‘The Moral Education Theory of Punishment’, Philosophy and Public Affairs 13 (1984): 208–38. Hare, R.M. The Language of Morals (Oxford: Clarendon Press, 1952). Hare, R.M. Moral Thinking: Its Levels, Method and Point (Oxford: Clarendon Press, 1981). Hare, R.M. ‘Wanting: Some Pitfalls’. In Agent, Action, and Reason, eds R. Binkley, R. Bronaugh, and A. Marras, pp. 81–127 (Toronto: University of Toronto Press, 1971). Harrod, R.F. The Life of John Maynard Keynes (London: Macmillan, 1951). Harrod, R.F. ‘Utilitarianism Revised’, Mind 45 (1936): 137–56. Hart, H.L.A. ‘Are There Any Natural Rights?’, Philosophical Review 64 (1955): 175–91. Hart, H.L.A. ‘Legal and Moral Obligation’. In Essays in Moral Philosophy, ed. A.I. Melden (Seattle: University of Washington Press, 1958), 82–107. Hart, H.L.A. ‘Prolegomenon to the Principles of Punishment,’ Proceedings of the Aristotelian Society 60 (1959–60): 1–26. Hayward, F.H. The Ethical Philosophy of Sidgwick (London: Swan Sonnenschein, 1901).

bibliography

287

Herman, Barbara. ‘On the Value of Acting from the Motive of Duty’, Philosophical Review 90 (1981): 359–82. Hill, Jr, Thomas E. ‘The Hypothetical Imperative’, Philosophical Review 82 (1973): 429–50. Hill, Jr, Thomas E. ‘Kantian Normative Ethics’. In The Oxford Handbook of Ethical Theory, ed. D. Copp, pp. 480–514 (New York: Oxford University Press, 2006). Hobart, R.E. ‘Free Will as Involving Determination and Inconceivable Without It’, Mind 43 (1934): 1–27. Hooker, Brad. Ideal Code, Real World (Oxford: Clarendon Press, 2000). Horgan, Terence and Timmons, Mark. ‘New Wave Moral Realism Meets Moral Twin Earth’, Journal of Philosophical Research 16 (1991): 447–65. Hudson, W.D. Modern Moral Philosophy (London: Macmillan, 1970). Hume, David. An Enquiry Concerning the Principles of Morals. In Enquiries Concerning the Human Understanding and Concerning the Principles of Morals, 2nd ed., ed. L.A. SelbyBigge (Oxford: Clarendon Press, 1902). Hume, David. A Treatise of Human Nature, ed. L.A. Selby-Bigge (Oxford: Clarendon Press, 1888). Hurka, Thomas. ‘Asymmetries in Value’, Nous 44 (2010): 199–223. Hurka, Thomas. The Best Things in Life: A Guide to What Really Matters (New York: Oxford University Press, 2011). Hurka, Thomas. ‘Desert: Individualistic and Holistic’. In Desert and Justice, ed. S. Olsaretti, 45–68 (Oxford: Clarendon Press, 2003). Hurka, Thomas. ‘Games and the Good’, Proceedings of the Aristotelian Society, Supplementary Volume 80 (2006): 217–35. Hurka, Thomas. ‘Indirect Perfectionism: Kymlicka on Liberal Neutrality’, Journal of Political Philosophy 3 (1995): 36–57. Hurka, Thomas. ‘The Justification of National Partiality’. In The Morality of Nationalism, eds R. McKim and J. McMahan, pp. 139–57 (New York: Oxford University Press, 1997). Hurka, Thomas. ‘Normative Ethics: Back to the Future’. In The Future for Philosophy, ed. B. Leiter, pp. 246–64 (Oxford: Clarendon Press, 2004). Hurka, Thomas. Perfectionism (New York: Oxford University Press, 1993). Hurka, Thomas. ‘Rights and Capital Punishment’, Dialogue 21 (1982): 647–60. Hurka, Thomas. ‘Two Kinds of Organic Unity’, Journal of Ethics 2 (1998): 283–304. Hurka, Thomas., ed. Underivative Duty: British Moral Philosophers from Sidgwick to Ewing (Oxford: Clarendon Press, 2011). Hurka, Thomas. ‘Value and Population Size’, Ethics 93 (1983): 496–507. Hurka, Thomas. Virtue, Vice, and Value (New York: Oxford University Press, 2001). Hurka, Thomas. ‘Virtuous Acts, Virtuous Dispositions’, Analysis 66 (2006): 69–76. Hurka, Thomas and Schubert, Esther. ‘Permissions to Do Less Than the Best: A Moving Band’, Oxford Studies in Normative Ethics 2 (2012): 1–27. Huxley, Julian. Evolutionary Ethics (Oxford: Oxford University Press, 1943). Irwin, Terence. The Development of Ethics: A Historical and Critical Study, 3 vols. (Oxford: Clarendon Press, 2007–9). Irwin, Terence. ‘Eminent Victorians and Greek Ethics: Sidgwick, Green, and Aristotle’. In Essays on Henry Sidgwick, ed. B. Schultz, pp. 279–310 (Cambridge: Cambridge University Press, 1992).

288

bibliography

Irwin, Terence. ‘Mistakes About Good: Prichard, Carritt, and Aristotle’. In Underivative Duty, ed. T. Hurka, pp. 106–25 (Oxford: Clarendon Press, 2011). Jackson, Frank. ‘Decision-theoretic Consequentialism and the Nearest and Dearest Objection’, Ethics 101 (1991): 461–82. Jackson, Frank and Pettit, Philip. ‘A Problem for Expressivism’, Analysis 58 (1998): 239–51. Jones, E.E. Constance. ‘Mr. Moore on Hedonism’, International Journal of Ethics 16 (1906): 429–64. Joseph, H.W.B. Some Problems in Ethics (Oxford: Clarendon Press, 1931). Kagan, Shelly. ‘The Additive Fallacy’, Ethics 99 (1988): 5–31. Kagan, Shelly. The Geometry of Desert (New York: Oxford University Press, 2012). Kagan, Shelly. The Limits of Morality (Oxford: Clarendon Press, 1989). Kagan, Shelly. ‘The Limits of Well-Being’, Social Philosophy and Policy 9/2 (1992): 169–89. Kamm, F.M. Morality/Morality, Vol. II: Rights, Duties, and Status (New York: Oxford University Press, 1996). Kant, Immanuel. Critique of Practical Reason. In Critique of Practical Reason and Other Writings, ed. Lewis White Beck, pp. 118–260 (Chicago: University of Chicago Press, 1949). Kant, Immanuel. Critique of Practical Reason and Other Writings in Moral Philosophy, ed. L.W. Beck (Chicago: University of Chicago Press, 1949). Kant, Immanuel. The Doctrine of Virtue: Part II of The Metaphysic of Morals, trans. Mary J. Gregor (Philadelphia: University of Pennsylvania Press, 1964). Kant, Immanuel. Groundwork of the Metaphysics of Morals, trans. M. Gregor (Cambridge: Cambridge University Press, 1997). Kant, Immanuel. ‘On a Supposed Right to Lie From Altruistic Motives’. In Critique of Practical Reason and Other Writings, ed. L.W. Beck, pp. 346–50 (Chicago: University of Chicago Press, 1949). Kelly, Thomas. ‘The Epistemic Significance of Disagreement’. In Oxford Studies in Epistemology, vol. 1, eds J. Hawthorne and T. Gendler-Szabo, pp. 167–96 (Oxford: Oxford University Press, 2005). Keynes, John Maynard. ‘My Early Beliefs’. In The Bloomsbury Group: A Collection of Memoirs, Commentary and Criticism, ed. S.P. Rosenbaum, pp. 48–64 (London: Croom Helm, 1975). Klagge, James C. ‘An Alleged Difficulty Concerning Moral Properties’, Mind 93 (1984): 370–80. Korsgaard, Christine M. ‘From Duty and for the Sake of the Noble: Kant and Aristotle on Morally Good Action’. In Aristotle, Kant, and the Stoics: Rethinking Happiness and Duty, eds S. Engstrom and J. Whiting, pp. 203–36 (Cambridge: Cambridge University Press, 1966). Korsgaard, Christine M. ‘Kant’s Formula of Universal Law’, Pacific Philosophical Quarterly 66 (1985): 24–47. Korsgaard, Christine M. ‘Two Distinctions in Goodness’, Philosophical Review 92 (1983): 169–95. Kraut, Richard. ‘The Rationality of Prudence’, Philosophical Review 81 (1972): 351–9. Kraut, Richard. What Is Good and Why: The Ethics of Well-Being (Cambridge, MA: Harvard University Press, 2007). Kymlicka, Will. Contemporary Political Philosophy (Oxford: Clarendon Press, 1990). Lacey, Nicola. A Life of H.L.A. Hart: The Nightmare and the Noble Dream (Oxford: Oxford University Press, 2004). Laird, John. A Study in Moral Theory (London: George Allen & Unwin, 1926).

bibliography

289

Langford, C.H. ‘Moore’s Notion of Analysis’. In The Philosophy of G.E. Moore, ed. P.A. Schilpp, pp. 319–42 (Chicago: Northwestern University Press, 1942). de Lazari-Radek, Katarzyna and Singer, Peter. ‘The Objectivity of Ethics and the Unity of Practical Reason’, Ethics 123 (2012): 9–31. Lemos, Noah M. Intrinsic Value: Concept and Warrant (Cambridge: Cambridge University Press, 1994). Levy, Paul. Moore: G.E. Moore and the Cambridge Apostles (New York: Holt, Rinehart and Winston, 1979). Locke, John. An Essay Concerning Human Understanding, ed. J.W. Yolton (London: Dent, 1977). Mabbott, J.D. ‘Punishment’, Mind 48 (1939): 152–67. MacIntyre, Alasdair. After Virtue: A Study in Moral Theory (London: Duckworth, 1981). Mackenzie, J.S. Review of G.E. Moore, Principia Ethica, International Journal of Ethics 14 (1904): 377–82. Mackie, J.L. Ethics: Inventing Right and Wrong (Harmondsworth: Penguin, 1977). Mackie, J.L. ‘A Refutation of Morals’, Australasian Journal of Psychology and Philosophy 24 (1946): 77–90. Mackie, J.L. ‘Sidgwick’s Pessimism’, Philosophical Quarterly 26 (1976): 317–28. Mahon, James Edwin. ‘Emotivism and Internalism: Ayer and Stevenson’, Studies in the History of Ethics (2005), online at . Malcolm, Norman. Ludwig Wittgenstein: A Memoir (London: Oxford University Press, 1958). Matheson, P.E. The Life of Hastings Rashdall, D.D. (London: Oxford University Press, 1928). Mayerfeld, Jamie. Suffering and Moral Responsibility (New York: Oxford University Press, 1999). McCloskey, H.J. ‘An Examination of Restricted Utilitarianism’, Philosophical Review 66 (1957): 466–85. McDowell, John. ‘Non-Cognitivism and Rule-Following’. In Wittgenstein: To Follow a Rule, eds S.H. Holtzman and C.M. Leich, pp. 141–62 (London: Routledge & Kegan Paul, 1981). McDowell, John. ‘Virtue and Reason’, The Monist 62 (1979): 331–50. McKerlie, Dennis. ‘McTaggart on Love’. In Underivative Duty, ed. T. Hurka, pp. 66–86 (Oxford: Clarendon Press, 2011). McTaggart, John McTaggart Ellis. ‘The Ethics of Henry Sidgwick’, Quarterly Review 205 (1906): 398–409. McTaggart, John McTaggart Ellis. ‘Hegel’s Theory of Punishment’, International Journal of Ethics 6 (1896): 479–502. McTaggart, John McTaggart Ellis. ‘The Individualism of Value’, International Journal of Ethics 18 (1908): 433–45. McTaggart, John McTaggart Ellis. The Nature of Existence, 2 vols. (Cambridge: Cambridge University Press, 1921 and 1927). McTaggart, John McTaggart Ellis. Some Dogmas of Religion (London: E. Arnold, 1906). McTaggart, John McTaggart Ellis. Studies in Hegelian Cosmology (Cambridge: Cambridge University Press, 1901). McTaggart, John McTaggart Ellis. ‘The Unreality of Time’, Mind 17 (1908): 457–74. Metz, Rudolf. ‘Recent Trends in Ethical Theory’, Philosophy 14 (1939): 299–312. Meyerhoff, Hans. ‘Emotive and Existentialist Theories of Ethics’, Journal of Philosophy 48 (1951): 769–83. Moore, G.E. ‘An Autobiography’. In The Philosophy of G.E. Moore, ed. P.A. Schilpp, pp. 1–39 (Chicago: Northwestern University Press, 1942).

290

bibliography

Moore, G.E. ‘The Conception of Intrinsic Value’. In Moore, Philosophical Studies, pp. 253–75 (London: Kegan Paul, Trench, Trubner, 1922). Moore, G.E. ‘Death of Dr. McTaggart’, Mind 34 (1925): 269–71. Moore, G.E. The Elements of Ethics, ed. Tom Regan (Philadelphia: Temple University Press, 1991); contains the text of lectures given in 1898. Moore, G.E. Ethics (London: Williams and Norgate, 1912); reprinted Oxford University Press, 1965. Moore, G.E. ‘External and Internal Relations’, Proceedings of the Aristotelian Society 20 (1919–20): 40–62; reprinted in Philosophical Studies, 276–309. Moore, G.E. ‘Is Goodness a Quality?’, Proceedings of the Aristotelian Society, Supplementary Volume 11 (1932): 116–31. Moore, G.E. ‘Mr. McTaggart’s Ethics’, International Journal of Ethics 13 (1903): 341–70. Moore, G.E. ‘The Nature of Judgment’, Mind 8 (1899): 176–93. Moore, G.E. ‘The Nature of Moral Philosophy’. In Moore, Philosophical Studies, pp. 310–39 (London: Kegan Paul, Trench, Trubner, 1922). Moore, G.E. Philosophical Studies (London: Kegan Paul, Trench, Trubner, 1922). Moore, G.E. ‘Preface to the Second Edition’. In Principia Ethica: Revised Edition, ed. Thomas Baldwin, pp. 1–27 (Cambridge: Cambridge University Press, 1993). Moore, G.E. Principia Ethica (Cambridge: Cambridge University Press, 1903); Principia Ethica: Revised Edition, ed. T. Baldwin (Cambridge: Cambridge University Press, 1993). Moore, G.E. ‘A Reply to My Critics’. In The Philosophy of G.E. Moore, ed. Schilpp, pp. 533–687 (Chicago: Northwestern University Press, 1942). Moore, G.E. Review of Franz Brentano, The Origin of the Knowledge of Right and Wrong, International Journal of Ethics 14 (1903): 115–23. Moore, G.E. Review of Hastings Rashdall, The Theory of Good and Evil, Hibbert Journal 6 (1907–8): 446–51. Moore, G.E. Some Main Problems of Philosophy (London: George Allen & Unwin, 1953); contains the text of lectures given in 1910–11. Moore, G.E. ‘Wittgenstein’s Lectures in 1930–33’, Mind 63 (1954): 1–15, 289–316; Mind 64 (1955): 1–27. Morris, Herbert. ‘Persons and Punishment’, The Monist 52 (1968): 475–501. Nagel, Thomas. ‘The Fragmentation of Value’. In Mortal Questions, pp. 128–41 (Cambridge: Cambridge University Press, 1979). Nagel, Thomas. ‘The Limits of Objectivity’. In The Tanner Lectures on Human Values, vol. 1, ed. S.R. McMurrin (Salt Lake City: University of Utah Press, 1980), 77–139. Nagel, Thomas. The Possibility of Altruism (Oxford: Clarendon Press, 1970). Nagel, Thomas. The View From Nowhere (New York: Oxford University Press, 1986). Nell (O’Neill), Onora. Acting on Principle: An Essay on Kantian Ethics (New York: Columbia University Press, 1975). Newman, John Henry. Certain Difficulties Felt by Anglicans in Catholic Teaching, 2 vols. (London: Longmans, 1900). Norcross, Alastair. ‘The Scalar Approach to Utilitarianism’. In H.R. West, ed., Blackwell Guide to Mill’s Utilitarianism, pp. 217–32 (Oxford: Blackwell, 2006). Nozick, Robert. Anarchy, State, and Utopia (New York: Basic Books, 1974). Nozick, Robert. The Examined Life (New York: Simon & Schuster, 1989).

bibliography

291

Nozick, Robert. Philosophical Investigations (Cambridge, MA: Harvard University Press, 1981). Nussbaum, Martha C. ‘Aristotelian Social Democracy’. In Liberalism and the Good, eds R. B. Douglas, G.M. Mara, and H. Richardson, pp. 203–52 (New York: Routledge, 1990). Nussbaum, Martha C. The Fragility of Goodness: Luck and Ethics in Greek Tragedy and Philosophy (Cambridge: Cambridge University Press, 1986). Ogden, C.K. and Richards, I.A. The Meaning of Meaning: A Study of the Influence of Language upon Thought and of the Science of Symbolism (London: Kegan Paul, Trench, 1923). Olson, Jonas and Timmons, Mark. ‘A.C. Ewing’s First and Second Thoughts in Metaethics’. In Underivative Duty, ed. T. Hurka, pp. 183–211 (Oxford: Clarendon Press, 2011). Owens, David. ‘A Simple Theory of Promising’, Philosophical Review 115 (2006): 51–77. Parfit, Derek. Equality or Priority?, The Lindley Lecture (Lawrence, KS: University of Kansas Press, 1995). Parfit, Derek. On What Matters, 2 vols. (Oxford: Oxford University Press, 2011). Parfit, Derek. Reasons and Persons (Oxford: Clarendon Press, 1984). Phillips, David. Sidgwickian Ethics (New York: Oxford University Press, 2011). Pickard-Cambridge, W.A. ‘Two Problems About Duty (I.), (II.), and (III.)’, Mind 41 (1932): 72–96, 145–72, 311–40. Pigden, Charles R. ‘Desiring to Desire: Russell, Lewis, and G.E. Moore’. In Themes from G.E. Moore: New Essays in Epistemology and Ethics, eds S. Nuccetelli and G. Seay, pp. 244–60 (Oxford: Oxford University Press, 2007). Pigden, Charles R. ‘Geach on Good’, Philosophical Quarterly 40 (1990): 129–54. Plato. Euthyphro, trans. L. Cooper. In The Collected Dialogues of Plato, eds E. Hamilton and H. Cairns, pp. 169–85 (Princeton, NJ: Bollinger, 1961). Portmore, Douglas W. Commonsense Consequentialism: Wherein Morality Meets Rationality (New York: Oxford University Press, 2011). Price, H.H. Critical Notice of H.A. Prichard, Knowledge and Perception, Mind 60 (1951): 103–21. Price, H.H. Critical Notice of W.D. Ross, The Right and the Good, Mind 40 (1931): 341–54. Price, H.H. ‘Harold Arthur Prichard’, Proceedings of the British Academy 33 (1947): 331–50. Price, Richard ‘A Review of the Principal Questions in Morals’. In British Moralists, vol. 2, ed. D.D. Raphael, pp. 131–98 (Oxford: Clarendon Press, 1969). Prichard, H.A. ‘Acting, Willing, Desiring’. In Moral Writings, pp. 272–81 (Oxford: Clarendon Press, 2002). Prichard, H.A. ‘A Conflict of Duties’. In Moral Writings, pp. 77–83 (Oxford: Clarendon Press, 2002). Prichard, H.A. Correspondence, The Prichard Papers, Bodleian Library, University of Oxford. Prichard, H.A. ‘Does Moral Philosophy Rest on a Mistake?’, Mind 21 (1912): 21–37; reprinted in Moral Writings, 7–20. Prichard, H.A. ‘Duty and Ignorance of Fact’, Proceedings of the British Academy 18 (1932): 67–92; reprinted in Moral Writings, 84–101. Prichard, H.A. ‘Duty and Interest’, Inaugural Lecture, White’s Professor of Moral Philosophy (Oxford: Clarendon Press, 1929); reprinted in Moral Writings, 21–49. Prichard, H.A. ‘Green: Political Obligation’. In Moral Writings, pp. 226–52 (Oxford: Clarendon Press, 2002). Prichard, H.A. ‘Kant’s Fundamental Principles of the Metaphysic of Morals’. In Moral Writings, 50–76 (Oxford: Clarendon Press, 2002).

292

bibliography

Prichard, H.A. Kant’s Theory of Knowledge (Oxford: Clarendon Press, 1909). Prichard, H.A. Knowledge and Perception: Essays and Lectures, ed. W.D. Ross (Oxford: Clarendon Press, 1950). Prichard, H.A. ‘Manuscript on Morals’. In Moral Writings, pp. 114–62 (Oxford: Clarendon Press, 2002). Prichard, H.A. ‘The Meaning of ‘agathon in the Ethics of Aristotle’, Philosophy 37 (1935): 27–39; reprinted in Moral Writings, 102–13. Prichard, H.A. ‘Moral Obligation’. In Moral Writings, pp. 163–225 (Oxford: Clarendon Press, 2002). Prichard, H.A. Moral Obligation: Essays and Lectures, ed. W.D. Ross (Oxford: Clarendon Press, 1949). Prichard, H.A. Moral Writings, ed. J. MacAdam (Oxford: Clarendon Press, 2002). Prichard, H.A. ‘The Obligation to Keep a Promise’. In Moral Writings, pp. 257–65 (Oxford: Clarendon Press, 2002). Prichard, H.A. ‘What is the Basis of Moral Obligation?’. In Moral Writings, pp. 1–6 (Oxford: Clarendon Press, 2002). Prior, Arthur N. Logic and the Basis of Ethics (Oxford: Clarendon Press, 1949). Quinn, Warren. ‘The Right to Threaten and the Right to Punish’, Philosophy and Public Affairs 14 (1985): 327–73. Rabinowicz, Wlodek and Ronnow-Rasmussen, Toni. ‘The Strike of the Demon: On Fitting Pro-Attitudes and Value’, Ethics 114 (2004). Raphael, D.D., ed. British Moralists 1650–1800, 2 vols. (Oxford: Clarendon Press, 1969). Raphael, D.D. ‘Edgar Frederick Carritt’, Proceedings of the British Academy 51 (1965): 439–53. Rashdall, Hastings. ‘The Commensurability of All Values’, Mind 11 (1902): 145–61. Rashdall, Hastings. Is Conscience an Emotion? (Boston: Houghton Mifflin, 1914). Rashdall, Hastings. Ethics (London: T.C. and E.C. Jack, 1913). Rashdall, Hastings. ‘The Limits of Casuistry’, International Journal of Ethics 4 (1894): 459–80. Rashdall, Hastings. ‘The Philosophical Theory of Property’. In Property; Its Duties and Rights: Historically, Philosophically and Religiously Regarded, eds C. Gore and L.T. Hobhouse (London: Macmillan, 1913), 35–64. Rashdall, Hastings. ‘Professor Sidgwick on the Ethics of Religious Conformity: A Reply’, International Journal of Ethics 7 (1897): 137–68. Rashdall, Hastings. ‘Professor Sidgwick’s Utilitarianism’, Mind 10 (1885): 200–26. Rashdall, Hastings. The Theory of Good and Evil, 2 vols. (London: Oxford University Press, 1907; 2nd ed. 1924). Rashdall, Hastings. ‘The Theory of Punishment’, International Journal of Ethics 2 (1891): 20–31. Rawls, John. A Theory of Justice (Cambridge, MA: Harvard University Press, 1971). Rawls, John. ‘Two Concepts of Rules’, Philosophical Review 64 (1955): 3–32. Regan, Donald H. Utilitarianism and Co-operation (Oxford: Clarendon Press, 1980). Regan, Donald H. ‘Why Am I My Brother’s Keeper?’. In Reason and Value, eds R.J. Wallace, P. Pettit, S. Scheffler, and M. Smith, pp. 202–30 (Oxford: Oxford University Press, 2004). Regan, Tom. Bloomsbury’s Prophet: G.E. Moore and the Development of His Moral Philosophy (Philadelphia: Temple University Press, 1986). Ritchie, David G. Review of Henry Sidgwick, The Elements of Politics, International Journal of Ethics 2 (1892): 254–7.

bibliography

293

Robinson, Richard. ‘The Emotive Theory of Ethics’, Proceedings of the Aristotelian Society, Supplementary Volume 22 (1948): 79–106. Robinson, Richard. Review of W.D. Ross, The Right and the Good, International Journal of Ethics 41 (1931): 343–51. Ross, W.D. (Sir David). Aristotle (London: Methuen, 1923). Ross, W.D. (Sir David). ‘The Basis of Objective Judgments in Ethics’, International Journal of Ethics 37 (1927): 113–27. Ross, W.D. (Sir David). ‘The Coherence Theory of Goodness’, Proceedings of the Aristotelian Society, Supplementary Volume 10 (1931): 61–70. Ross, W.D. (Sir David). ‘The Ethics of Punishment’, Journal of Philosophical Studies (later Philosophy) 4 (1929): 205–11. Ross, W.D. (Sir David). Foundations of Ethics: The Gifford Lectures Delivered in the University of Aberdeen 1935–6 (Oxford: Clarendon Press, 1939). Ross, W.D. (Sir David). ‘Is There a Moral End?’, Proceedings of the Aristotelian Society, Supplementary Volume 8 (1928): 90–8. Ross, W.D. (Sir David). Kant’s Ethical Theory: A Commentary on the Grundlegung zur Metaphysik der Sitten (Oxford: Clarendon Press, 1954). Ross, W.D. (Sir David). ‘The Nature of Morally Good Action’, Proceedings of the Aristotelian Society 29 (1928–29): 251–74. Ross, W.D. (Sir David). Preface to H.A. Prichard, Knowledge and Perception: Essays and Lectures, ed. W.D. Ross (Oxford: Clarendon Press, 1950). Ross, W.D. (Sir David). The Right and the Good (Oxford: Clarendon Press, 1930). Rubin, Michael. ‘Is Goodness a Homeostatic Property Cluster?’, Ethics 118 (2008): 496–528. Russell, Bertrand. The Autobiography of Bertrand Russell, 3 vols. (London: George Allen & Unwin, 1966–9). Russell, Bertrand. ‘The Elements of Ethics’. In Philosophical Essays, pp. 13–59 (London: George Allen & Unwin, 1966) [1st ed. 1910]. Russell, Bertrand. ‘Is Ethics a Branch of Empirical Psychology?’. In Russell on Ethics, ed. C. Pigden, pp. 71–8 (London: Routledge, 1999). Russell, Bertrand. ‘My Mental Development’. In The Philosophy of Bertrand Russell, ed. P. A. Schilpp, pp. 3–20 (Evanston, IL: Northwestern University Press, 1944). Russell, Bertrand. Review of G.E. Moore, Principia Ethica, Independent Review 2 (1904): 328–33. Russell, Bertrand. Russell on Ethics: Selections From the Writings of Bertrand Russell, ed. C. Pigden (London: Routledge, 1999). Sachs, David. ‘A Fallacy in Plato’s Republic’, Philosophical Review 72 (1963): 141–58. Salt, Henry S. ‘The Rights of Animals’. International Journal of Ethics 10 (1900): 206–22. Scanlon, T.M. Moral Dimensions: Permissibility, Meaning, Blame (Cambridge, MA: Harvard University Press, 2008). Scanlon, T.M. What We Owe to Each Other (Cambridge, MA: Harvard University Press, 1998). Scheffler, Samuel. The Rejection of Consequentialism: A Philosophical Investigation of the Considerations Underlying Rival Moral Conceptions (Oxford: Clarendon Press, 1982). Scheffler, Samuel. ‘Relationships and Responsibilities’, Philosophy and Public Affairs 26 (1997): 189–209. Schilpp, P.A., ed. The Philosophy of Bertrand Russell (Evanston, IL: Northwestern University Press, 1944).

294

bibliography

Schilpp, P.A., ed. The Philosophy of Brand Blanshard (La Salle, IL: Open Court, 1980). Schilpp, P.A., ed. The Philosophy of C.D. Broad (New York: Tudor Publishing, 1959). Schilpp, P.A., ed. The Philosophy of G.E. Moore (Chicago: Northwestern University Press, 1942). Schneewind, J.B. Sidgwick’s Ethics and Victorian Moral Philosophy (Oxford: Clarendon Press, 1977). Schultz, Bart., ed. Essays on Henry Sidgwick (Cambridge: Cambridge University Press, 1992). Schultz, Bart. Henry Sidgwick: Eye of the Universe (Cambridge: Cambridge University Press, 2004). Schultz, Bart. ‘Introduction’. In Essays on Henry Sidgwick, ed. B. Schultz, pp. 1–61 (Cambridge: Cambridge University Press, 1992). Searle, John. ‘Prima Facie Obligations’. In Practical Reasoning, ed. J. Raz, pp. 81–90 (Oxford: Oxford University Press, 1978). Sen, Amartya K. Collective Choice and Social Welfare (San Francisco: Holden-Day, 1970). Sen, Amartya K. ‘Equality of What?’. In Tanner Lectures on Human Values, vol. 1, ed. S.M McMurrin, pp. 195–220 (Salt Lake City: University of Utah Press, 1980). Shafer-Landau, Russ. Moral Realism: A Defence (Oxford: Clarendon Press, 2003). Shaver, Robert. ‘The Birth of Deontology’. In Underivative Duty, ed. T. Hurka, pp. 126–45 (Oxford: Clarendon Press, 2011). Shaver, Robert. Rational Egoism (Cambridge: Cambridge University Press, 1998). Shaver, Robert. ‘Sidgwick on Moral Motivation’, Philosophers’ Imprint 6 (2006): 1–14. Shaver, Robert. ‘Sidgwick’s Axioms and Consequentialism’, Philosophical Review 123 (2014): 173–204. Shaver, Robert. ‘Sidgwick’s False Friends’, Ethics 107 (1997): 314–20. Shaver, Robert. ‘Sidgwick’s Minimal Metaethics’, Utilitas 12 (2000): 261–77. Sher, George. Desert (Princeton, NJ: Princeton University Press, 1987). Shiffrin, Seana Valentine. ‘Promising, Intimate Relationships, and Conventionalism’, Philosophical Review 117 (2008): 481–524. Sidgwick, Henry. Critical Notice of David G. Ritchie, Natural Rights, Mind 4 (1895): 384–8. Sidgwick, Henry. Critical Notice of F.H. Bradley, Ethical Studies, Mind o.s. 1 (1876): 545–9, reprinted in Essays on Ethics and Method, 185–9. Sidgwick, Henry. Critical Notice of Thomas Fowler, Progressive Morality, Mind o.s. 10 (1885): 266–71, reprinted in Essays on Ethics and Method, 259–63. Sidgwick, Henry. The Elements of Politics (London: Macmillan, 1891). Sidgwick, Henry. Essays on Ethics and Method, ed. M.G. Singer (Oxford: Oxford University Press, 2000). Sidgwick, Henry. ‘The Establishment of Ethical First Principles’, Mind o.s. 4 (1879): 106–11, reprinted in Essays on Ethics and Method, 29–34. Sidgwick, Henry. The Ethics of Conformity and Subscription (London: Williams and Norgate, 1870). Sidgwick, Henry. ‘The Ethics of Religious Conformity’, International Journal of Ethics 6 (1896): 273–90. Sidgwick, Henry. ‘Further on the Criteria of Truth and Error’. In Lectures on the Philosophy of Kant and Other Philosophical Lectures and Essays, ed. J. Ward (London: Macmillan, 1905), reprinted in Essays on Ethics and Method, 166–70.

bibliography

295

Sidgwick, Henry. ‘Green’s Ethics’, Mind o.s. 9 (1884): 169–84, reprinted in Essays on Ethics and Method, 243–58. Sidgwick, Henry. ‘Hedonism and Ultimate Good’, Mind o.s. 2 (1877): 27–38, reprinted in Essays on Ethics and Method, 89–98. Sidgwick, Henry. Henry Sidgwick: A Memoir, eds Arthur Sidgwick and Eleanor Mildred Sidgwick (London: Macmillan, 1906). Sidgwick, Henry. ‘Idiopsychological Ethics’, Mind o.s.12 (1887): 31–44, reprinted in Essays on Ethics and Method, 264–75. Sidgwick, Henry. ‘Is the Distinction Between “Is” and “Ought” Ultimate and Irreducible?’, Proceedings of the Aristotelian Society 2/1 (1891–2): 88–92, reprinted in Essays on Ethics and Method, 59–62. Sidgwick, Henry. Lectures on the Ethics of T.H. Green, H. Spencer, and J. Martineau, ed. E.E. Constance Jones (London: Macmillan, 1902). Sidgwick, Henry. The Methods of Ethics (London: Macmillan, 1874; 2nd ed. 1877; 3rd ed. 1884; 4th ed. 1890; 5th ed. 1893; 6th ed. 1901; 7th ed., 1907). Sidgwick, Henry. ‘Mr. Barratt on “The Suppression of Egoism” ’, Mind o.s. 2 (1877): 411–12, reprinted in Essays on Ethics and Method, 27–8. Sidgwick, Henry. Outlines of the History of Ethics for English Readers (London: Macmillan, 1886). Sidgwick, Henry. Practical Ethics: A Collection of Addresses and Essays (London: Swan Sonnenschein, 1898). Sidgwick, Henry. ‘Professor Calderwood on Intuitionism in Morals’, Mind o.s. 1 (1876): 563–6, reprinted in Essays on Ethics and Method, 23–8. Sidgwick, Henry. Reply to Bradley, ‘Mr. Sidgwick on Ethical Studies’, Mind o.s. 2 (1877): 125–6, reprinted in Essays on Ethics and Method, 193–4. Sidgwick, Henry. ‘Some Fundamental Ethical Controversies’, Mind o.s. 14 (1889): 473–87, reprinted in Essays on Ethics and Method, 35–46. Sidgwick, Henry. ‘The Theory of Evolution in its Application to Practice’, Mind o.s. 1 (1876): 52–67, reprinted in Essays on Ethics and Method, 10–22. Sidgwick, Henry. ‘Utilitarianism’. In Essays on Ethics and Method, 3–9. Sidgwick, Henry. ‘The Verification of Beliefs’, Contemporary Review 17 (1871): 582–90, reprinted in Essays on Ethics and Method, 121–8. Sidorsky, David. ‘The Uses of the Philosophy of G.E. Moore in the Works of E.M. Forster’, New Literary History 38 (2007): 245–71. Singer, Peter. ‘Sidgwick and Reflective Equilibrium’, The Monist 58 (1974): 490–517. Skelton, Anthony. ‘Henry Sidgwick’s Moral Epistemology’, Journal of the History of Philosophy 48 (2010): 491–520. Skelton, Anthony. ‘Ideal Utilitarianism: Rashdall and Moore’. In Underivative Duty, ed. T. Hurka, pp. 45–65 (Oxford: Clarendon Press, 2011). Skelton, Anthony. ‘Sidgwick’s Philosophical Intuitions’, Etica & Politica/Ethics and Politics 10 (2008): 185–209. Slote, Michael. Common-Sense Morality and Consequentialism (London: Routledge & Kegan Paul, 1985). Slote, Michael. ‘The Virtue in Self-Interest’, Social Philosophy and Policy 14/1 (1997): 264–85.

296

bibliography

Smith, Michael. ‘Consequentialism and the Nearest and Dearest Objection’. In Minds, Ethics, and Conditionals: Themes from the Philosophy of Frank Jackson, ed. I. Ravenscroft, pp. 237–66 (Oxford: Clarendon Press, 2009). Smith, Michael. The Moral Problem (Oxford: Blackwell, 1994). Smith, Michael. ‘Neutral and Relative Value after Moore’, Ethics 113 (2003): 576–98. Smuts, Aaron. ‘The Feels Good Theory of Pleasure’, Philosophical Studies 155 (2011): 241–65. Stevenson, Charles L. ‘The Emotive Meaning of Ethical Terms’, Mind 46 (1937): 14–31. Stevenson, Charles L. ‘Moore’s Arguments against Certain Forms of Ethical Naturalism’. In The Philosophy of G.E. Moore, ed. P.A. Schilpp, pp. 71–90 (Chicago: Northwestern University Press, 1942). Stocker, Michael. ‘Agent and Other: Against Ethical Universalism’, Australasian Journal of Philosophy 54 (1976): 206–20. Stocks, J.L. ‘The Golden Mean’, The Monist 41 (1931): 161–79. Stocks, J.L.. ‘Morality and Purpose’, Hibbert Journal 26 (1927–28): 54–68. Stratton-Lake, Philip, ed. Ethical Intuitionism: Re-evaluations (Oxford: Clarendon Press, 2002). Stratton-Lake, Philip. ‘Introduction’. In Ethical Intuitionism, ed. P. Stratton-Lake, pp. 1–28 (Oxford: Clarendon Press, 2002). Stratton-Lake, Philip. ‘Pleasure and Reflection in Ross’s Intuitionism’. In Ethical Intuitionism, ed. P. Stratton-Lake, pp. 113–36 (Oxford: Clarendon Press, 2002). Strawson, P.F. ‘Ethical Intuitionism’, Philosophy 24 (1949): 23–33. Street, Sharon. ‘A Darwinian Dilemma for Realist Theories of Value’, Philosophical Studies 127 (2006): 109–66. Sturgeon, Nicholas L. ‘Ethical Intuitionism and Ethical Naturalism’. In Ethical Intuitionism, ed. P. Stratton-Lake, pp. 184–211 (Oxford: Clarendon Press, 2002). Suits, Bernard. The Grasshopper: Games, Life, and Utopia (Toronto: University of Toronto Press, 1978). Sumner, L.W. Welfare, Happiness, and Ethics (Oxford: Clarendon Press, 1996). Svavarsdottir, Sigrun. ‘Objective Values: Does Metaethics Rest on a Mistake?’. In Objectivity in Law and Morals, ed. B. Leiter, pp. 144–83 (Cambridge: Cambridge University Press, 2001). Sverdlik, Steven. ‘Motive and Rightness’, Ethics 106 (1996): 327–49. Sverdlik, Steven. Motive and Rightness (Oxford: Oxford University Press, 2011). Sverdlik, Steven. ‘The Origins of “The Objection” ’, History of Philosophy Quarterly 29 (2012): 79–101. Taylor, C.C.W. Aristotle: Nicomachean Ethics Books II–IV (Oxford: Clarendon Press, 2006). Thomson, Judith Jarvis. ‘On Some Ways in Which a Thing Can Be Good’, Social Philosophy and Policy 9/2 (1992): 96–117. Thomson, Judith Jarvis. ‘The Right and the Good’, Journal of Philosophy 94 (1997): 273–98. Thomson, Judith Jarvis. ‘Self-Defense’, Philosophy and Public Affairs 20 (1991): 283–310. Urmson, J.O. ‘A Defence of Intuitionism’, Proceedings of the Aristotelian Society 75 (1974–75): 111–19. Urmson, J.O. The Emotive Theory of Ethics (London: Hutchinson University Library, 1968). Vlastos, Gregory. ‘The Individual as Object of Love in Plato’. In Platonic Studies (Princeton, NJ: Princeton University Press, 1973). Waddington, C.H. ‘The Relations Between Science and Ethics’, Nature 148 (1941): 270–4.

bibliography

297

Wallace, Ray J., Pettit, Philip, Scheffler, Samuel, and Smith, Michael, eds. Reason and Value: Themes from the Moral Philosophy of Joseph Raz (Oxford: Oxford University Press, 2004). Warnock, G.J. Contemporary Moral Philosophy (London: Macmillan, 1970). Warnock, Mary. Ethics Since 1900 (London: Oxford University Press, 1960). Way, Jonathan. ‘Defending the Wide-Scope Approach to Instrumental Reason’, Philosophical Studies 147 (2010): 213–33. Welchman, Jennifer. ‘G.E. Moore and the Revolution in Ethics: A Reappraisal’, History of Philosophy Quarterly 6 (1989): 317–29. Whewell, William. The Elements of Morality, including Polity, 4th ed. (Cambridge: Deighton, Bell, 1864) [1st ed. 1845]. White, Nicholas P. ‘The Attractive and the Imperative: Sidgwick’s View of Greek Ethics’. In Essays on Henry Sidgwick, ed. B. Schultz, pp. 311–30 (Cambridge: Cambridge University Press, 1992). Whitely, C.H. ‘On Duties’, Proceedings of the Aristotelian Society 53 (1952–53): 97–104. Whitely, C.H. ‘Mr. Warnock on Ordinary Language’, Mind 68 (1959): 396–8. Wiggins, David. ‘Deliberation and Practical Reason’, Proceedings of the Aristotelian Society 76 (1975/76): 29–51. Wiggins, David. ‘The Right and the Good and W.D. Ross’s Criticism of Consequentialism’, Utilitas 10 (1998): 261–80. Wilde, Norman. Review of G.E. Moore, Principia Ethica, Journal of Philosophy, Psychology and Scientific Method 2 (1905): 581–3. Williams, Bernard. ‘A Critique of Utilitarianism’. In J.J.C. Smart and B. Williams, Utilitarianism: For and Against, pp. 77–150 (Cambridge: Cambridge University Press, 1973). Williams, Bernard. ‘Ethical Consistency’. In Problems of the Self, pp. 166–86 (Cambridge: Cambridge University Press, 1973). Williams, Bernard. Ethics and the Limits of Philosophy (Cambridge, MA: Harvard University Press, 1985). Williams, Bernard. Morality: An Introduction to Ethics (New York: Harper & Row, 1972). Williams, Bernard. ‘The Point of View of the Universe: Sidgwick and the Ambitions of Ethics’. In Making Sense of Humanity, pp. 153–71 (Cambridge: Cambridge University Press, 1995). Williams, Bernard. ‘Utilitarianism and Self-Indulgence’. In Moral Luck, pp. 40–53 (Cambridge: Cambridge University Press, 1981). Woolf, Leonard. Sowing (London: Hogarth, 1960). Zimmerman, Michael J. ‘The Good and the Right’, Utilitas 19 (2007): 326–53. Zimmerman, Michael J. ‘In Defense of the Concept of Intrinsic Value’, Canadian Journal of Philosophy 29 (1999): 389–410.

Index achievement, value of 68, 209–11 aesthetic appreciation, value of 122, 142–4, 199, 202–6, 227, 228–9, 243, 276 aesthetics 16–17, 203–5 agent-neutrality, see ‘good’, agent-neutral; ‘ought’, agent-neutral agent-relativity, see ‘good’, agent-relative; ‘ought’, agent-relative analysis, conceptual 94; paradox of 99–100 ancient ethics 259–67 animals, duties concerning 89, 214–15 Anscombe, G.E.M. 167, 188n28 anti-theory 139–41, 146, 266 Apostles 6, 7, 9, 12–13, 14–15, 51n7, 229 Aristotle 2, 13, 18, 40, 71, 96, 119, 131–2, 134, 138, 205, 207, 210, 223, 234, 259–67 Audi, Robert 110, 117n14, 118n16, 119n18 Austin, J.L. 4, 14, 40–1, 262 autonomy of ethics 1, 86, 93–101, 103, 109, 128–30, 131, 134, 272, 279 more important to school than moral realism 106–7 Ayer, A.J. 3, 4, 8, 14, 18, 20, 103–6 Baldwin, Thomas 122n23 beauty error theory of 34, 87–8 value of 122, 199, 204, 220, 227 ‘beauty’, analysed using ‘good’ 23, 34, 51, 203, 229 beneficence, duty of 31, 132, 148, 151, 153–4, 161–3, 165, 167, 178–9, 184–5, 186–7, 189–92, 215, 265, 270–1, 277 benevolence, axiom of 50, 89, 120–1, 158–64, 239, 241 benevolence, virtue of 48, 148, 217, 219, 265; see also beneficence Bentham, Jeremy 96, 115, 117, 132, 166, 177, 214, 244–6, 271 Berlin, Isaiah 5 biology as ground of moral truth 129, 147 Blanshard, Brand 17, 60, 105n49, 229n21, 260n2 Bosanquet, Bernard 3, 15, 139–41, 146 Bradley, F.H. 3, 8, 11, 12, 17, 98n33, 129, 136–7, 139–41, 146, 166, 195–6, 197n7, 200–1, 207, 252 Braithwaite, R.B. 16 Brentano, Franz 6, 52–5, 139n15, 204, 218 British moralists 267–72

Broad, C.D. on analysis, paradox of 99n35 on autonomy of ethics 97 on beneficence 277 on biology as ground of moral truth 129 on Bradley 8 on Butler 268, 271 on common-sense morality 154, 238, 241 on conscientiousness 219, 222–3 on consequentialism 172, 174–5 on deontology 154, 165 on desert 130, 233 on dualism of the practical reason 136 on egoism, normative 241 on egoism, psychological 267–8 on ‘fitting’ 23, 56, 63, 74, 77 on free will, relation to morality 130 on good, measurement of 145 on good, total vs. average 212 on ‘good’ 23, 34, 54 on Green 8–9 on Hobbes 268n23 on Hume 102, 269–70 on hypothetical imperative 30 on intuition 111–12 on Kant 268, 273–5, 277–8 on knowledge, value of 206 life of 19–20 on logical positivism 103 on love 231–2 on McTaggart 7, 12, 13, 145, 213–14, 231–2 on metaphysics, as ground of moral truth 130 on Mill 197 on monism vs. pluralism 42, 74, 142 on moral concepts, deontic 30 on moral explanations 136, 148, 232, 277 on moral motivation 98, 105 on moral theory 138, 142, 145 on moral truths as underivative 128 on ‘moral’ ought 26–7, 30n10 on non-cognitivism 103, 105–6 on objective vs. subjective duty 173 on ordinary language 40, 41n38 on organic unities 130, 200, 211, 214 on ‘ought’, agent-relative 148, 238, 241 on ‘ought’ as involving ‘fitting’ 23, 56, 74, 77 on ‘ought’ and motives 46 on ‘ought’ as unanalysable 97 on ‘ought’, wide vs. narrow 45, 63

300

INDEX

Broad, C.D. (cont.) on permissions, agent-favouring 238, 241 on pleasure, duty to pursue your own 26–7, 238 on pleasure, nature of 194–5 on pleasure, value of 197–8, 200–2, 213–14, 226 on Price 269, 272 on Prichard 9, 103n45 on prima facie duty 9, 74, 77 on promise-keeping 275 on proportionality 225, 232 on religion as ground of moral truth 129 on repugnant conclusion 213–14 on Ross 9 on self-evidence 109, 111–12, 116, 241, 272 on self-referential altruism 148–9, 192 on Sidgwick 7, 8, 53, 98n33, 136, 154, 164, 177, 200 on subjective naturalism 102, 269–70 on supervenience 92 on utilitarianism 142 on variety, value of 214 on virtue 25, 201, 216, 219, 221, 224–6, 273 Butler, Joseph 4, 52, 132, 134, 162, 241, 259, 267–8, 270–2, 273, 279 Carritt, E.F. on aesthetic appreciation 204–5 on aesthetics 16–17, 116n13, 137n12, 203n19, 204–5 on Aristotle 40, 263–4 on beauty, value of 199, 205 on beneficence 277 on common-sense morality 40, 116 on conscientiousness 33, 219, 242 on consequentialism 165–7, 169–71 on conservatism, moral 117 on deontology 165–7, 169–71, 182, 185 on desert 165, 233, 248, 251, 254 on equality 165, 248, 251 on error theory 34, 87–8 on freedom, duty concerning 167 on good, measurement of 145 on ‘good’ 34, 35, 263 on ‘good for’ 35, 263 on Hobbes 16, 268n23 on intuition 111, 122–6, 165–6 on justice, distributive 165, 233, 248, 251 on Kant 274–7, 279 on knowledge, value of 205–7, 209 life of 16–17 on metaphysics as ground of moral truth 129–30 on monism vs. pluralism 42, 142, 181–2, 248, 254

on moral concepts, deontic 30–1 on moral explanations 126, 137, 170 on moral theory 137–8, 142, 145 on moral truths as underivative 128 on ‘moral’ good/ought 26–7, 33 on non-cognitivism 103–4 on non-maleficence 166–7 on ordinary language 40 on ‘ought’, agent-neutral/agent-relative 46, 169, 182, 242 on Plato 40, 263 on pleasure, duty to pursue your own 26–7, 238 on pleasure, value of 145, 197–8, 226, 238 on Price 269 on Prichard 8, 13 on prima facie duty 71, 72, 137, 167, 182–3 on promise-keeping 165–7 on punishment 31, 166, 254 on Rashdall 8 on realism, general/moral 87–8 on religion as ground of moral truth 129 on rights, moral 31 on self-evidence 88, 111, 116 on self-referential altruism 192–3 on supererogation 180 on time-relativity 20 on utilitarianism 166 on virtue 24, 33, 201, 219, 222, 226, 238, 242 on ‘why be moral?’ 27 casuistry, see anti-theory categorical imperatives 26, 28–30, 107, 134, 138, 242, 263, 270, 274–9 Clarke, Samuel 162, 241, 268–9, 271–2 coherentism 107, 116, 120–1, 125–7 common-sense morality 7, 24, 26, 40, 52, 71, 78–9, 116, 120–1, 124, 126, 131–2, 135, 147–8, 151, 153–4, 161–2, 165–6, 168, 170–1, 176–80, 186, 193, 200, 202, 216, 210, 218, 222, 237–8, 241, 246, 260, 266, 270–1 compensation, duty of 153–4, 178, 183–4, 187 conscientiousness 33, 47–8, 148, 219, 222–3, 225, 242, 272–4 consequentialism 7–9, 42, 70, 136, 139, 150, 153–4, 172–8, 179–83, 185, 188–9, 215, 245–6, 247, 251, 252–5, 271, 278 act- 7, 78–9, 172, 175, 178 critique of 125–6, 157–8, 165–9, 180, 182, 185–6, 192, 271, 277 defence of 50–2, 70–1, 112, 120, 142, 150, 153–4, 158–64, 165 ideal 3, 7n12, 9, 11, 42, 56–8, 137, 142–3, 169–71, 173–4 indirect 7, 136–7, 168, 175–8, 255, 260 nature of 1, 170–1, 174–5

INDEX

as reducing deontology 56–8, 142–3, 169–71, 174 scalar 76 conservatism, moral 116–17, 176–8, 271–2 Cook Wilson, John 7, 8, 87, 88 courage 23–4, 261, 265–6, 273 Cox, H.H. 92n18 Croce, Benedetto 17 Cudworth, Ralph 268–9, 271 Cumberland, Richard 268, 270–1 definition, see analysis, conceptual deontic concepts, see duty proper; moral concepts, deontic; ‘ought’; prima facie duty deontology 3, 8, 46, 56–8, 89, 142, 178–93, 215, 239, 255, 264–5 absolute vs. moderate 70–1, 78, 152–4, 164, 185, 255, 274 critique of 70–1, 112, 120, 135, 142–3, 150–8, 162, 271–2 defence of 125–6, 165–71 nature of 1, 71, 136, 152–4, 181–3, 188–9 elaborations of Ross’s 186–93 neo-Thomist 188–9, 191, 278 as reduced to consequentialism 56–8, 142–3, 169–71, 174 scalar 76–8 desert, economic 232, 247–51 desert, moral 35, 130, 142, 165, 174, 185, 233–6, 244, 248, 250, 256–7, 267 and proportionality 192, 234–6, 244, 250–1, 256–7 desert, retributive 11, 68, 115, 129, 130, 147–9, 233–5, 252, 254–8 doing/allowing distinction 185–6, 187–8 Donagan, Alan 156, 164, 178n11 double effect, doctrine of 48–9, 185–6, 187–8 dualism of the practical reason 26, 35, 42, 132, 136, 159–60, 162, 164, 239 ‘duty’ 30–1; see also duty proper; moral concepts, deontic; ‘ought’; prima facie duty duty proper 27, 31–2, 44, 69–78, 125n25, 152–3, 162–4, 181, 183–4, 187, 224, 239–41, 260, 272 economic distribution 244–51 egoism, normative 26, 36–7, 57, 131–4, 159–60, 162–4, 239–41, 260 egoism, psychological 131–2, 134, 220–1, 263, 267–8, 272 emotivism, see non-cognitivism entitlement theory 246–7 equality 147, 165, 244–6, 251 error theory of beauty 34, 87–8 of morals 87–9, 107n55

301

Ewing, A.C. on aesthetic appreciation 203 on analysis, paradox of 99–100 on ancient ethics 261–4 on animals, duties concerning 215 on anti-theory 140 on autonomy of ethics 60n17, 97, 99–100 on beauty, value of 199 on biology as ground of moral truth 129 on coherentism 61–2, 116, 126–7 on common-sense morality 40, 116 on conscientiousness 222–3, 225 on consequentialism, defence of 142, 150 on consequentialism, ideal 56–8, 142–3, 169–71, 173–4 on consequentialism as reducing deontology 56–8, 169–70 on desert 129, 236, 255–6 on error theory 88 on ‘fitting’ 23, 34n20, 56, 63, 84–5 on free will, relation to morality 130 on good, measurement of 145 on ‘good’, agent-relative 55, 57–8 on ‘good’ as analysed using ‘fitting’ 23, 40, 54–64 on ‘good’, attributive 34 on ‘good’, intrinsic 54, 59, 67 on intuition 92–3, 108, 109, 111–12, 114–16, 118–19, 122–5, 171 on justice, distributive 235 on Kant 172n3, 273, 275n32, 277–9 on knowledge, value of 55, 206, 208 life of 20–1 on logical positivism 21, 103 on lying 277 on metaphysics as ground of moral truth 129–30 on monism vs. pluralism 142–3, 255–6 on Moore 102n32, 105 on moral concepts, deontic 23, 30–1, 84 on moral concepts, deontic vs. evaluative 48–9 on moral concepts, thick vs. thin 24 on moral explanations 60–2, 277–8 on moral motivation 98, 106 on moral theory 137, 139, 140, 142–3, 145 on ‘moral’ good/ought 23, 26–7, 30n10, 55–6, 84 on non-cognitivism 103, 106 on non-maleficence 166 on non-naturalism 56, 88, 90–1, 92–3, 97, 99–100, 106, 108 on objective vs. subjective duty 83–5, 173 on ordinary language 40 on organic unities 59, 67, 141, 211–12 on ‘ought’, agent-neutral 57 on ‘ought’ as involving ‘fitting’ 23, 30, 56, 63, 84

302

INDEX

Ewing, A.C. (cont.) on ‘ought’ and motives 48 on ‘ought’ as unanalysable 97 on pleasure 26–7, 55, 181, 197, 224n15, 226 on praise and blame 23, 83–4 on prima facie duty 31, 56–7, 84, 92, 127 on proportionality 54–5, 225, 236 on punishment 31, 129, 166, 254–6 on realism, moral 88, 90–1, 92–3 on reasons, normative 27, 31, 106 on relativity, argument from 92–3 on religion as ground of moral truth 129 on Ross 71, 181 on self-evidence 92–3, 111–12, 115–16, 119, 157 on self-referential altruism 169–70, 193 on Strawson 73 on subjective naturalism 102–3 on supererogation 180 on variety, value of 212 on virtue 5, 25, 54–5, 210–11, 219, 222–3, 224–6, 238, 256, 261, 273 on ‘why be moral?’ 27 Field, G.C. 99n34, 105n50, 264 Finnis, John 188 ‘fitting’ 23, 34n20, 54–6, 68, 74–5, 77–8, 234, 267 analysed using ‘ought’ 62–4, 77, 85 as unanalysable 23, 56, 63, 84–5; see also ‘good’, analysed using ‘fitting’; ‘ought’, involves ‘fitting’ Forster, E.M. 230 Frankena, William K. 95n23 freedom, duty concerning 167, 177n9, 190–1, 246–7, 251 free will, relation to morality 130 Frege-Geach objection 104–5 Geach, P.T. 7n13, 39–41, 104n48 good aggregation of 211–14 conflict of different persons’ 243–4, 260–1 duty to promote your own 237–44, 276 measurement of 7, 9, 139, 144–6, 155, 211, 244 total vs. average 157–8, 161, 185n25, 212–14 ‘good’ agent-neutral 36–7, 39, 46, 57, 229–30 agent-relative 36–7, 55, 57–8, 170, 225, 230, 232, 239–40, 242 analysed using ‘fitting’ 23, 40, 54–64, 68–9, 77, 170 analysed using ‘ought’ 22, 36–7, 40, 44, 50, 52–4, 59–60, 94, 102 attributive 26, 34, 39–40

instrumental/contributive 59, 67, 69, 78, 206, 218 intrinsic 9, 22–3, 25–6, 33, 35, 38, 39–40, 41, 42, 51, 54, 59, 62–3, 65–9, 78, 91, 137, 141, 179, 218, 224, 246, 263, 278 as property of states of affairs 45, 55–6 as unanalysable 22–3, 38, 40, 46, 51, 54, 56, 57, 59–62, 64, 69, 93–7, 230, 242, 263, 269 vs. ‘ought’, see moral concepts, deontic vs. evaluative ‘good for’ 26, 34–9, 42, 57, 179, 233n26, 263 gratitude 74, 77, 125, 142–3, 151, 156, 170, 178–9, 182, 187, 192–3 Green, T.H. 3, 8–10, 31, 96, 129, 134, 137, 195–6, 197n7, 200–1, 214, 217, 243, 253, 269 Grice, G.R. 20–1 happiness 198 Hare, R.M. 3, 41, 104–5 harm, duty not to, see non-maleficence, duty of harm principle 177n9 Hart, H.L.A 5, 189–90, 255 Hayward, F.H. 152, 177n10, 200n12, 219n6, 243n9 hedonism, evaluative 198–202; see also pleasure, value of Hegel, G.W.F. 6, 7, 12, 129, 157, 205, 253–4 Heidegger, Martin 157 Hill, Jr, Thomas E. 277 Hobbes, Thomas 2, 13, 267, 268n24 Hume, David 13, 87, 102, 105, 131, 132–3, 216n1, 259, 267–71, 273 Hutcheson, Frances 267–9, 271 hypothetical imperative 29–30, 53, 134, 263, 274, 276 Idealism 3–4, 7–8, 11–12, 15, 17, 87, 129, 136–7, 139–1, 144, 205, 272 impartiality, see ‘good’, agent-neutral; ‘ought’, agent-neutral intuition 38, 88–9, 92–3, 107, 108–27, 240, 269 about explanation 125–6, 133–4, 167–8, 200, 262 of abstract vs. particular truths 119–27, 135, 165–8 as certain vs. fallible 112, 114–15, 117–18, 127, 157, 163–4, 171 and inference 118–19 see also self-evidence intuitionism, moral 1–2, 42, 86, 108–27 isolation, method of 35, 65–6, 68, 122, 140, 273 Jones, E.E. Constance 38, 66n3, 199n11 Joseph, H.W.B. 3–5, 9, 11, 17, 46, 130, 142, 150, 167, 169, 174, 198, 199

INDEX

judgement-externalism vs. -internalism, see moral motivation justice, axiom of 91, 120–1, 158–9 justice, distributive 25, 147, 151–2, 165, 167, 178–9, 189, 192, 235, 236, 244–51 Kant, Immanuel 2, 13, 28, 30, 35, 67, 70, 130, 134, 162, 166, 220, 222, 237, 272–9 Keynes, John Maynard 7, 10, 15, 203, 211, 229 knowledge nature of 66, 80, 117, 124, 199, 206, 209 value of 33, 55, 57, 59, 66, 68, 131, 142, 199–200, 205–9, 213, 226, 242, 267 Korsgaard, Christine M. 69n6 Laird, John 3–4, 52, 129, 140, 150, 206, 211 Langford, C.H. 99 libertarianism, see entitlement theory Locke, John 97, 113n10, 246–8, 267 logical positivism 4, 21, 103 love, personal nature of 228–32 value of 7, 122, 142, 144, 202, 213, 228–32 lying, duty against 70, 106–7, 137, 143, 147, 154, 173, 178, 188–9, 275–8 Mabbott, J.D. 167, 257 McCloskey, H.J. 167 McDowell, John 25 MacIntyre, Alasdair 107 McKerlie, Dennis 145n28, 213 Mackie, J.L. 87–8, 88–90, 92, 110 McTaggart, John McTaggart Ellis on autonomy of ethics 7n12, 96–7 on beauty, value of 199 on Bradley 8, 12, 200–1 on common-sense morality 241 on consequentialism 7n12, 42, 137 on desert 130n4, 234, 254 on egoism, normative 241 on free will, relation to morality 130 on good, measurement of 144–5 on ‘good’ 67, 69, 97 on Green 200–1 Idealism of 7–8, 11 on knowledge, value of 209, 213 life of 12–13 on love 213, 230–2 on metaphysics as ground of moral truth 129 on monism vs. pluralism 42 on moral explanations, inherent vs. external 147–8 on moral theory 137, 144 on ‘moral’ good 33 on non-naturalism 7 on organic unities 67 on other Idealists 12

303

on perfectionism 7 on pleasure, value of 144, 161, 194, 198, 200, 213–14, 245 on punishment 253–4, 256 on repugnant conclusion 213–14 on Sidgwick 164 on time-relativity 161 on utilitarianism 137 on virtue 25, 147–8, 213, 219 malice 49, 201 218, 224, 226, 227, 234 Marx, Karl 17, 210 maximizing 1, 23, 50–1, 70, 76, 78–9, 141, 172, 179–80, 182, 215, 245 Meinong, Alexius 41 metaphysics as ground of moral truth 129–31, 210, 240–1 Metz, Rudolf 179n14 Mill, John Stuart 94, 96, 115, 117, 132, 177, 197, 200, 214, 269, 271 monism vs. pluralism about duty 42–3, 69–70, 74, 141–3, 145–6, 151, 153–6, 164, 167–8, 178–93, 254–8 about good 42–3, 95–6, 141–3, 199, 248 Moore, G.E. on aesthetic appreciation 122, 142, 144, 202–4, 206, 227, 228–9 on aesthetics 203 on analysis, conceptual 94, 99 on anti-theory 140–1 on autonomy of ethics 93–7, 102 on beauty, value of 122, 199, 204, 220, 227 on ‘beauty’, analysed using ‘good’ 23, 34, 51, 203, 229 on Bentham 115 on biology as ground of moral truth 129 on Brentano 6 on coherentism 122 on common-sense morality 7, 176–7, 200 on conscientiousness 219, 222 on consequentialism, act- 7, 174–5 on consequentialism, defence of 50–2, 70–1, 150, 165 on consequentialism, ideal 42, 137, 173–4 on consequentialism, indirect 7, 137, 175–8 and conservatism, moral 176–8 on deontology 71, 152 on desert 142, 148, 174, 233–4, 236, 254 on dualism of the practical reason 136 on egoism, normative 36, 57, 239–41 on egoism, psychological 267 on free will, relation to morality 130 on good, measurement of 144 on ‘good’, agent-neutral 57, 229–30, 239–40 on ‘good’, instrumental/contributive 67, 69, 204, 206 on ‘good’, intrinsic 9, 22–3, 38, 42, 51, 54, 65–9, 91–2, 122, 137, 141

304

INDEX

Moore, G.E. (cont.) on ‘good’ as unanalysable 22, 38, 40, 46, 54, 57, 69, 93–7, 230, 239 on ‘good for’ 34–6, 57 ignores prima facie duty 71 on intuition 38, 108–9, 115–16, 118–19, 122, 126, 200 on isolation, method of 65–6, 68, 122 on Kant 272–4 on knowledge, value of 206 life of 14–16 on love 122, 142, 144, 202, 228–30, 232 on McTaggart 7, 13 on maximizing 23, 50–1 on metaphysics as ground of moral truth 129–30 on Mill 94, 115, 197 on monism vs. pluralism 42, 95–6, 142 on moral concepts, deontic 30 on moral explanations 119, 136, 137, 147–8, 200 on moral motivation 97 on moral theory 137, 138, 140–1, 142, 144–5 on moral truths as underivative 128 on ‘moral’ good/ought 27, 33 on ‘my good’ 34–6, 57 on naturalistic fallacy 93–6; see also autonomy of ethics on non-cognitivism 105–6 on non-naturalism, 86–7, 90–2, 93–7, 99 on objective vs. subjective duty 79, 84–5 on open-question argument 51, 93–7, 99, 101, 102 on organic unities 42, 65–6, 140–1, 148–9, 174, 200, 204, 211, 233–4, 254 on ‘ought’, agent-neutral, 57, 239–40, 242 on ‘ought’, analysed using ‘good’ 22–3, 50–2 on ‘ought’ and motives 174n4 on ‘ought’ as unanalysable 23, 51, 94 on perfectionism 7, 137, 242 on Plato 267 on pleasure 115, 122, 194, 197–200, 202, 213, 227, 237, 245 on praise and blame 79–80 on punishment 148, 254 on Rashdall 9 on realism, general/moral 86–7, 90–2, 272 on relativity, argument from 92–3 on religion as ground of moral truth 129 on self-evidence 51, 109, 115–16, 118, 122, 221, 239, 241 on self-referential altruism 177 on Sidgwick 7, 36, 57, 96, 115, 136, 200 on subjective naturalism 101–2, 105 on supervenience 91–2 on virtue, duty to promote 173–4 on virtue, forms of 222–4, 229

on virtue, nature of 23–4, 147–8, 218–19 on virtue, value of 204, 218–19, 221, 227 on ‘why be moral?’ 27 moral concepts deontic 23, 25–33, 69–78, 84; see also ‘duty’; ‘obligation’; ‘ought’; ‘right’ deontic vs. evaluative 44–9, 170–1, 274 evaluative 33–43, 65–9; see also ‘good’ minimalism about 22–3, 25–6, 33, 41–3, 133 thick vs. thin 23–5 moral explanations 2–3, 24, 60–2, 119, 125–6, 128, 133–4, 135–8, 141, 147–9, 167–8, 170, 188–90, 200, 232, 239, 243–4, 261–4, 276, 277–8, 279 inherent vs. external 2–3, 147–9 moral motivation externalism about 32, 97–8, 138 internalism about 89–90, 97–8, 105–6 moral realism, see realism, moral moral theory 1, 135–49, 156, 165, 171, 260 degrees of 141–6; see also monism vs. pluralism discovery and explanatory functions of 135–8 rejection of, see anti-theory ‘salad dressing’ view of 43 moral truths fully vs. partly determinate 135, 146, 154–7, 199 as underivative 1, 27, 32, 128–34, 143, 149, 178, 180–1, 188n30, 190, 193, 258, 260, 279 ‘moral’ good, vs. ‘non-moral’ 26, 33, 55 ought, vs. ‘rational’ or ‘prudential’ 25–8, 42, 133–4 More, Henry 268–9 motive of duty, see conscientiousness ‘my good’ 34–9, 57 naturalism, subjective, see subjective naturalism naturalistic fallacy 93–6; see also autonomy of ethics Newman, John Henry 226 non-cognitivism 3, 5, 42, 103–7 irrelevant to normative ethics 106–7, 127 non-maleficence, duty of 151, 154, 157, 161, 166–7, 178, 182–4, 185–6, 187–8, 189, 191 non-naturalism 1–2, 3, 5–6, 7, 32, 42, 56, 86–101, 108, 110, 127, 203, 269, 272 modest vs. ontological 90–1 see also autonomy of ethics; realism, moral Nozick, Robert 126, 202, 208, 246–7, 250 ‘obligation’ 23, 30–1, 84; see also moral concepts, deontic objective vs. subjective duty 78–85, 170–5, 186, 225, 244, 249

INDEX

open-question argument 51, 93–101, 102, 195–6, 203 ordinary language 4, 39–41 organic unities, principle of 42, 140–1, 174, 200, 254, 273, 200, 273 and aesthetic appreciation 204 and deontology 185–6 and desert 130, 148, 174, 233–4, 254 holistic vs. variability interpretation 65–9, 140–1, 186, 233, 254 in single life 211–12, 214 ‘ought’ agent-neutral 37, 50, 57, 89, 147–8, 158–61, 165, 177, 180–1, 237–8, 241, 242 agent-relative 46, 56–7, 168–71, 182, 189, 192–3, 238, 241 all things considered, see duty proper involves ‘fitting’ 23, 30, 56, 63, 74, 77, 84, 267, 272 analysed using ‘good’ 23, 50–2 concerns only acts of will 81–2 ‘moral’, vs. ‘rational’ or ‘prudential’, see ‘moral’ and motives 46–9 not derivable from ‘is’, see autonomy of ethics objective vs. subjective, see objective vs. subjective duty other things equal; see prima facie duty as property of persons 46, 82 as unanalysable 22–3, 51, 64, 96–7 vs. ‘good’, see moral concepts, deontic vs. evaluative wide vs. narrow 45, 50, 53, 63 ‘ “ought” implies “can” ’ 32, 44–5, 47–9, 53, 63–4, 77, 85, 170–1, 183–4 pain, see pleasure Paley, William 271 Parfit, Derek 90–1, 100, 146n30, 161, 201, 213 partiality; see self-referential altruism Peirce, Charles Sanders 139n15 perfectionism 7, 9, 136–7, 194, 203–11, 211–12, 216–32, 241–4, 244–6, 253, 261, 262–3, 267; see also consequentialism, ideal permissions, moral 161, 165, 179–81, 180–81, 190 agent-favouring 161, 179–81, 190, 231–9 agent-sacrificing 165, 181, 190, 237 Phillips, David 240–1 Pickard-Cambridge, W.A. 3, 156 Plato 16, 19, 33, 40, 96, 128–9, 134, 205, 207, 210, 259–64, 267 pleasure comparative value of 42, 145, 197–8, 206, 213–14, 224, 226–7 degrees of value in 144–6, 155, 158, 196–8, 213, 245 duty to promote your own 26–7, 181, 237–8

305

malicious 201, 224, 226, 227, 234 no duty to promote your own 26, 28–9, 37, 57, 178, 181, 237–8 nature of 194–6, 218 value of 33, 42, 54, 55, 57, 63, 66, 115, 120, 122, 131, 136–7, 141, 142, 143, 161, 198–202, 214, 220, 224, 228, 232, 233, 235, 237, 238–9, 242, 243, 263, 267 praise and blame 23, 79–80, 83–4, 130, 180 Price, H.H. 14, 70n9, 125, 207 Price, Richard 268–9, 271–2 Prichard, H.A. on ancient ethics 261–4 on Aristotle 134, 261–4 on autonomy of ethics 97, 272 on beneficence 165, 179 on Butler 134, 271 on Clarke 269 on common-sense morality 40, 126, 165 on conscientiousness 33, 47–8, 219, 274 on consequentialism, critique of 125–6, 165, 167–8, 277 on Cudworth 269 on deontology, defence of 125–6, 165, 167–8, 171 on deontology, nature of 8, 181–2 on deontology, scalar 76, 78 on dualism of the practical reason 136 on duty proper 125n25, 70, 75–7 on egoism, normative 132–4, 263–4, 272 on egoism, psychological 132–3, 262–3, 272 on equality 165 on ‘good’ 23, 34, 35, 67, 69 on ‘good for’ 35–6 on Green 134 on Hutcheson 269 on hypothetical imperative 29–30 on intuition 92–3, 111, 117–18 on intuition, about explanation 125–6, 133–4, 167–8, 262 on intuition, of abstract vs. particular truths 123–4, 165, 167–8 on justice, distributive 165 on Kant 47n3, 138, 272, 274–5, 277–8 on knowledge, nature of 80, 117, 206 on knowledge, value of 143n24, 206 life of 12–13 on monism vs. pluralism 42, 142–3, 167–8, 181–2 on Moore 9 on moral concepts, deontic 30 on moral concepts, deontic vs. evaluative 45–9, 170, 274 on moral explanations 125–6, 133–4, 136, 138, 149, 167–8, 261–4, 277 on moral motivation 98 on moral theory 137–8, 142, 143, 145n27 on moral truths as underivative 27, 128

306

INDEX

Prichard, H.A. (cont.) on ‘moral’ good/ought 26–7, 33 on objective vs. subjective duty 80–2, 85 on ordinary language 40 on ‘ought’, agent-relative 46, 181–2 on ‘ought’ as concerning only acts of will 81–2 on ‘ought’ and motives 46–9 on ‘ought’ as property of persons 46, 82 on ‘ought’ as unanalysable 23, 97 on ‘ “ought” implies “can” ’ 45, 47–9, 50, 53 on permissions, agent-sacrificing 165 on Plato 134, 262–4, 271 on pleasure 26, 28, 143n24, 237–8 on prima facie duty 8, 70, 72–3, 75–8, 145n27, 181–3 on promise-keeping 126, 134, 149, 275 on Rashdall 8 on realism, general 8, 17, 46, 87, 272n30 on relativity, argument from 92–3 on rights, moral 31 on self-evidence 92–3, 111, 118, 124, 133 on self-interest as ground of moral truth 132–4, 262–4, 271 on Sidgwick 134, 136 on virtue 28, 33, 143, 165, 182, 216, 219, 222, 242 on ‘why be moral?’ 27, 132–4 prima facie duty 8, 9, 31, 44, 56–7, 69–78, 84, 127, 142–3, 145, 148, 156, 167, 168–9, 178–83, 193, 224, 257, 266 analysed using duty proper 33, 72–3, 75–8 conflicts of 62, 70–3, 124–5, 143, 151, 153, 154–6, 183–6, 187, 239 ignored by previous writers 70–1, 78, 152–4, 162–4, 266 as involving fittingness 33, 73–8, 84 knowledge of 92, 111, 123–4, 137, 168 Prior, A.N. 96n30 promise-keeping, duty of 46, 56–7, 77, 125–6, 134, 147, 149, 151, 156–7, 165–9, 170–1, 178–9, 182, 184, 186–7, 188, 189–90, 192, 275 property, rights to 191–2, 246–7, 251, 257 proportionality 54–5, 192, 212, 225, 227, 230, 232, 234, 235–6, 238–9, 242, 244, 250–1, 256–7 prudence, axiom of 89, 120–1, 158–60, 162, 240 Psychical Research, Society for 10, 19 punishment, criminal 31, 115, 140, 147–8, 166–7, 184, 185, 191, 233–6, 252–8, 260 punishment, divine 11, 129, 136, 234 queerness, argument from 88–91, 110 Rashdall, Hastings on achievement 210–11 on aesthetic appreciation 203

on ancient ethics 262–4 on animals, duties concerning 214–15 on anti-theory 140 on Aristotle 264 on autonomy of ethics 96–7 on benevolence, axiom of 50 on Bentham 96, 244–6 on biology as ground of moral truth 129 on Bradley 11, 140, 200–1 on Butler 268 on categorical imperative 274 on conscientiousness 222–3 on consequentialism 150, 245–6, 251 on consequentialism, defence of 50, 71, 150 on consequentialism, ideal 9, 42, 173–4 on conservatism, moral 116–17 on Cudworth, 269 on deontology 71, 151–2 on desert 11, 115, 129, 234–6, 248–50, 252 on economic distribution 244–51 on egoism, psychological 220–1, 267–8 on entitlement theory 247 on equality 244–5, 251 on freedom, duty concerning 251 on free will, relation to morality 130 on good, conflict of different people’s 243 on good, measurement of 9, 144, 146 on ‘good’ 46, 50, 66, 96–7, 269 on Green 9, 200–1, 214, 253 on happiness 198 on Hume 270 Idealism of 11 ignores prima facie duty 71 on intuition 93, 115–17 on Kant 273–4, 275n32, 276–8 on knowledge, value of 66, 131, 142, 206, 208 life of 11–12 on lying 276 on maximizing 245 on metaphysics as ground of moral truth 129–30, 131 on Mill 96, 197 on monism vs. pluralism 42, 142 on Moore 9 on moral concepts, deontic vs. evaluative 44–5 on moral explanations 137, 147–8 on moral motivation 98 on moral theory 137, 139, 140, 142, 144, 146 on moral truths, fully vs. partly determinate 146 on moral truths as underivative 128 on ‘moral’ good 33 on ‘my good’ 35, 37 on non-naturalism 86–7, 91, 96–7 on objective vs. subjective duty 79 on open-question argument 97 on organic unities 66, 141, 200, 212, 234n28

INDEX

on ‘ought’ 50, 242 on perfectionism 9, 137, 214–15, 242, 243, 245–6, 253 on permissions, agent-sacrificing 237 on pleasure, value of 66, 131, 142, 144, 194, 197, 200–1, 206, 214–15, 220, 226–7, 235 on Price 269 on property, rights to 247, 251 on proportionality 212, 225, 234, 235 on punishment 11, 115, 129, 140, 234, 252–3, 256 on realism, moral 87, 91 on relativity, argument from 93 on religion as ground of moral truth 11, 129 on self-evidence 93, 111, 115 on Sidgwick 9, 35, 151–2, 200 on subjective naturalism 270 on variety, value of 212 on virtue, duty to promote 173–4 on virtue, forms of 222–3, 273 on virtue, nature of 24–5, 46, 147–8, 217–9, 270 on virtue, value of 66, 131, 137, 142, 174, 210, 212, 214, 216–21, 224–5, 226–7, 244–5, 252 Rawls, John 2n1, 131, 147, 149, 167, 248–9, 255 realism, general 8, 17, 46, 87, 146, 272 realism, moral 86–93, 146 less important to school than autonomy of ethics 106–7 reasons, normative 22, 27, 31–3, 36, 61–3, 90, 106, 135–6, 174, 180 reflective equilibrium, see coherentism Reid, Thomas 268, 271–2 relativity, argument from 92–3 religion as ground of moral truth 128–9 reparation, see compensation, duty of repugnant conclusion 213–14 ‘right’ 30; see also moral concepts, deontic rights, moral 31–3, 189–92, 246–7, 257–8 Ritchie, David G. 177 Robinson, Richard 88, 107n55 Ross, W.D. on aesthetic appreciation 143, 203, 276 on analysis, paradox of 99–100 on ancient ethics 262–4 on animals, duties concerning 214–15 on Aristotle 18, 260n3, 264, 266–7 on autonomy of ethics 97, 99 on beauty, value of 199 on beneficence 31, 178–9, 184–5, 186–7, 189–92, 264 on Broad 9 on Butler 268, 271–2 on Clarke 272 on common-sense morality 40, 116, 126, 135, 180 on compensation 71–3, 178, 187

307

on conscientiousness 33, 47–8, 219, 223, 272–4 on consequentialism, critique of 125–6, 165–9, 169–71, 192, 271, 277 on conservatism, moral 117 on deontology, absolute vs. moderate 70–1, 185 on deontology, defence of 125–6, 165–71 deontology, elaborations of his 186–93 on deontology, nature of 179–80, 182–3 on desert 130, 165, 174, 185, 233–4, 236, 244, 256–7, 267 on duty proper 69–75, 78, 183–4 on equality 165, 244 on error theory 34, 87–8 on free will, relation to morality 130 on Frege-Geach objection 104–5 on good, duty to promote your own 241–2 on good, measurement of 144–6 on good, total vs. average 157n6, 185n25, 212–13 on ‘good’, agent-neutral/agent-relative 37, 57, 242 on ‘good’ analysed using ‘ought’ 50, 54, 57, 102, 242 on ‘good’, attributive 34, 39n34 on ‘good’, instrumental/contributive 67, 69 on ‘good’, intrinsic 9, 41, 66, 224 on ‘good’ as property of states of affairs 45 on ‘good’ as unanalysable 23, 46, 54, 57, 59–60, 69, 242 on gratitude 178, 187, 192 on hypothetical imperative 29–30, 45n2 on intuition 93, 111, 117–18, 118–19 on intuition about explanation 125–6, 167–8, 276 on intuition of abstract vs. particular truths 123–5, 126, 144, 165–8 on justice, distributive 165, 178–9, 189, 192, 236, 244 on Kant 47n3, 273–9 on knowledge, nature of 80, 117, 124, 206 on knowledge, value of 57, 59, 131, 202, 206–9, 226, 242, 267 life of 17–19 on logical positivism 103 on love 143, 228, 232 on lying 137, 143, 147, 178, 188, 276, 275 on maximizing 179–80, 182 on metaphysics as ground of moral truth 131 on monism vs. pluralism 42–3, 69–70, 142–3, 167–8, 178–9, 181–6, 188–9, 193, 256–8 on Moore 9 on moral concepts, deontic 30–1, 45n2 on moral concepts, deontic vs. evaluative 45–9, 170, 274

308

INDEX

Ross, W.D. (cont.) on moral concepts, thick vs. thin 24 on moral explanations 125–6, 134, 138, 147–8, 167–8, 170, 188–90, 264, 276, 277 on moral motivation 32, 98 on moral theory 135, 137–8, 142, 143, 144–5, 146, 171 on moral truths, fully vs. partly determinate 146, 155 on moral truths as underivative 128, 178 on ‘moral’ good/ought 26–7, 33 on non-cognitivism 103–5 on non-maleficence 166–7, 178, 182–3, 187–8, 189 on non-naturalism 87–8, 91, 92–3, 97, 99 on objective vs. subjective duty 80–3, 85 on open-question argument 51, 99 on ordinary language 40, 41 on organic unities 185, 233, 273 on ‘ought’, agent-neutral/agent-relative 46, 56–7, 168–9, 182–3, 242 on ‘ought’ as concerning only acts of will 81 on ‘ought’ as involving ‘fitting’ 74, 77, 267, 272 on ‘ought’ and motives 46–9, 188 on ‘ought’ as property of persons 46 on ‘ought’ as unanalysable 23, 97 on ‘ought’, wide vs. narrow 45 on ‘ “ought” implies “can” ’ 45, 47–9, 50, 53 on pleasure, comparative value of 144–5, 206, 215, 224, 226–7 on pleasure, malicious 224 on pleasure, nature of 194–5 on pleasure, no duty to promote your own 26, 28, 37, 57, 178, 181, 237–8 on pleasure, value of 54, 57, 63, 131, 142, 143, 197, 201–2, 220, 224, 228, 238, 242, 267 on Price 272 on Prichard 8, 9, 14, 18 on prima facie duty 9, 69–75, 77–8, 142–3, 148, 156, 168–9, 178–83, 193, 224, 257 on prima facie duty as analysed using duty proper 72–3, 77–8 on prima facie duty, conflicts of 70–3, 124–5, 153, 183–6, 187 on prima facie duty as involving ‘fitting’ 73–5, 77–8 on prima facie duty, knowledge of 111, 123–4, 168 on promise-keeping 125–6, 156–7, 165–6, 168, 178, 182, 184, 186–7, 188, 189–90, 192, 275 on property, rights to 191–2, 257 on proportionality 192, 225, 234, 236, 244, 256–7

on punishment 31, 166–7, 184, 236, 252, 254, 256–8 on realism, general/moral 87–8, 91, 92–3, 146 on relativity, argument from 93 on religion, as ground of moral truth 129 on rights, moral 31, 190, 257–8 on self-evidence 88, 93, 110, 111, 116, 118–19, 124 on self-improvement, duty of 72, 178, 241–2, 276 on self-interest as ground of moral truth 134, 264 on self-referential altruism 165, 192–3 on subjective naturalism 102 on supererogation 180 on supervenience 92 on utilitarianism 125–6, 166, 167–8, 258 on virtue, comparative value of 28, 144–5, 224, 226–7, 228 on virtue, duty to promote 26, 28, 72, 178, 238, 242 on virtue, forms of 219, 221–3, 238, 273 on virtue, nature of 25, 46, 147–8, 219, 238, 270 on virtue, value of 54, 57, 59, 125, 131, 143, 201–2, 209–10, 216, 219, 221, 224–5, 228, 238–9, 242, 267 Russell, Bertrand 3, 9, 12, 95 on free will, relation to morality 130 on ‘good’, intrinsic 66 on Moore 15, 51 on moral concepts, deontic vs. evaluative 44–5 on moral theory 138 on ‘my good’ 35–6 on non-naturalism 87 on objective vs. subjective duty 79, 83, 85, 173 on open-question argument 51 on ‘ought’ as unanalysable 51 on relativity, argument from 92–3 Salt, Henry S. 214–15 Searle, John 73 self-evidence 51, 88, 109–22, 124, 133, 221, 239, 240–1, 260, 272 conditions for 92–3, 111, 112–17, 119–20, 150–8, 161–4, 165, 168, 171, 202 see also intuition self-improvement, duty of 72, 178, 241, 276 self-indulgence, moral 227, 264–6 self-interest as ground of moral truth 131–4, 147, 262–6 self-referential altruism 37, 148–9, 165, 169–70, 177, 192–3, 229–30, 232

INDEX

Shaftesbury, Lord 268–71 Shaver, Robert 112–13, 159–60, 163, 170 Sidgwick, Henry on aesthetic appreciation 199 on analysis, conceptual 99n35 on ancient ethics 259–67 on animals, duties concerning 89, 214 on anti-theory 140 on Aristotle 259–60, 265–6 on autonomy of ethics 96, 131 on beauty, value of 199 on beneficence 132, 151, 153–4, 161–2, 265, 270–1, 277 on benevolence, axiom of 89, 120–1, 158–64, 239, 241 on Bentham 96, 132, 214, 244, 271 on biology as ground of moral truth 129 on Bradley 8, 140, 200–1 on Butler 52, 132, 162, 241, 267–8, 271–2 on Clarke 162, 241, 269, 271 on coherentism 120–1 on common-sense morality 7, 24, 26, 52, 71, 120–1, 131–2, 135, 147, 153–4, 161–2, 168, 176–8, 200, 216, 218, 222, 237, 260, 266, 270–1 on compensation 153 on consequentialism, act- 7, 78–9, 172, 175 on consequentialism, defence of 112, 120, 150, 153–4, 158–64 on consequentialism, indirect 7, 168, 175–8, 260 and conservatism, moral 176–8, 260 on Cudworth 271 on Cumberland 270–1 on deontology, critique of 120, 135, 150–8, 162, 164, 271–2 on deontology, nature of 136, 152–4 on desert 130, 234, 248–50, 252 on double effect, doctrine of 188 on dualism of the practical reason 26, 42, 121, 132, 136, 159–60, 162, 164, 200n9, 239 on duty proper 44, 152, 162–4, 239, 260 on economic distribution 151, 244–51 on egoism, normative 26, 36–7, 57, 131–2, 159–60, 162–4, 239–41 on egoism, psychological 132, 267–8 on entitlement theory 246–7 on equality 244 on error theory 88–9 on free will, relation to morality 130 on good, conflict of different persons’ 243, 260–1, 270–1 on good, duty to promote your own 26, 36–7, 132, 237, 239–42 on good, measurement of 7, 9, 144–6, 155, 196–7, 199, 244 on good, total vs. average 157, 212

309

on ‘good’, agent-neutral/agent-relative 36–7, 57–8, 239–40 on ‘good’, intrinsic 33, 66–7, 218 on ‘good’ as analysed using ‘ought’ 22, 36–7, 40, 44, 52–4 on ‘good for’ 34–8, 57 on gratitude 151, 187 on Green 8, 10–11, 200–1 on happiness 198 on Hume 270–1 on Hutcheson 269–71 on hypothetical imperative 30 ignores prima facie duty 70, 152–4, 162–4 on intuition 38, 108–19, 119–22, 126, 240, 269 on justice, axiom of 120–1, 158–9 on justice, distributive 151, 244, 246–51 on Kant 162, 276–8 on knowledge, value of 66, 199–200 life of 9–11 on love 228, 232 on lying 154, 173 on McTaggart 6 on maximizing 78–9, 141 on metaphysics as ground of moral truth 129, 130–1, 240–1 on ‘method of ethics’ 26, 135–6 on Mill 177n9, 132, 197, 214, 269, 271 on monism vs. pluralism 42, 141–2, 145–6, 151, 153–6, 164, 199 on Moore 15 on moral concepts, deontic 30 on moral concepts, deontic vs. evaluative 44–6, 170 on moral concepts, thick vs. thin 24 on moral explanations 24, 119, 135, 147–8 on moral motivation 97–8 on moral theory 135–6, 138–9, 140, 141–3, 144, 145, 146, 260 on moral truths, fully vs. partly determinate 146, 154–6 on moral truths as underivative 128 on ‘moral’ good/ought 26–8, 33 on More 269 on ‘my good’ 34–8, 57 on non-maleficence, 151, 154 on non-naturalism 7, 86–7, 88–9, 90, 96, 108, 131, 196 on objective vs. subjective duty 78–9, 82, 84 on open-question argument 96 on ‘ought’, agent-neutral 50, 89, 147–8, 158–64, 177, 241 on ‘ought’ and motives 46–9 on ‘ought’ as unanalysable 22, 96 on ‘ought’, wide vs. narrow 45, 50, 53, 63 on perfectionism 136, 214, 242, 243, 279 on Plato 259–61

310

INDEX

Sidgwick, Henry (cont.) on pleasure, duty to promote your own 36–7, 132, 237, 239–41 on pleasure, nature of 195–7, 218 on pleasure, value of 42, 115, 120, 141, 144, 146, 155, 158, 196–7, 199–201, 214 on praise and blame 130 on Price 269, 272 on promise-keeping 151, 186 on property, rights to 246–7 on prudence, axiom of 89, 120–1, 158–60, 162, 240 on punishment 252, 260 on Rashdall 9 on realism, general/moral 86–7, 88–9, 90, 196 on reasons, normative 22, 31, 135–6 on Reid 272 on religion, as ground of moral truth 128–9 on rights, moral 31 on self-evidence 109–17, 119–21, 150–8, 161–4, 168, 240–1, 260, 272 on self-interest as ground of moral truth 132 on self-referential altruism 177 on Shaftesbury 269–71 on Socrates 260 on Stewart 272 on Stoics 259–60 on subjective naturalism 101–2 on supervenience 26, 91, 158–9 on time-neutrality 158–62, 178 as unfair in argument 155–6, 157–8, 160–4, 243n8 on utilitarianism 26, 42, 78–9, 120–1, 132, 135, 147–8, 153–4, 163–4, 168, 176–8, 212, 239, 244, 246, 251, 252, 260, 269, 270–1 on virtue, duty to promote 136, 242, 279 on virtue, forms of 222–3 on virtue, nature of 24, 216, 259, 270 on virtue, value of 136, 216–18, 220–1, 259 on ‘why be moral?’ 27 Smith, Adam 268–9 Socrates 260 Spencer, Herbert 31, 96, 155, 195 Stevenson, Charles L. 3, 101–2, 103, 105n50 Stewart, Dugald 268, 271–2 Stocks, J.L. 3, 71n12, 264n10 Stoics 217, 222, 259–60, 267 Strachey, Lytton 15 Strawson, P.F. 73 Street, Sharon 110

subjective naturalism 101–3, 105, 270 supererogation 180 supervenience 23, 26, 91–2, 109, 124, 158–9, 265 Sverdlik, Steven 49 synthetic a priori truth 21, 23, 51–2, 64, 92, 103, 109–10, 150 theft 166, 191, 257 time-neutrality 158–62, 178, 201 time-relativity 57, 161, 170–71, 201 tragic moral conflicts 183–4, 239 trolley problem 124, 154, 157, 182, 184–5, 188–9 universalizability; see supervenience Urmson, J.O. 261 utilitarianism 3, 26, 36, 42, 78–9, 120–1, 125–6, 132, 135, 137, 142, 145–6, 147–8, 153–4, 163–4, 166, 167–8, 173, 176–8, 197, 212, 239, 244, 246, 251, 252, 255, 258, 260, 269, 270–1; see also consequentialism variety, value of 211–12, 214 vice, see virtue virtue comparative value of 28, 33, 42, 144–5, 224, 226–7, 228, 243, 261 degrees of value in 54–5, 174, 204, 224–5, 230, 232 duty to promote 26, 28, 72, 136, 143, 165, 173–4, 178, 238, 242, 243, 279 forms of 219, 221–4, 229, 238, 273 nature of 23–5, 33, 46, 147–8, 216–19, 227, 233, 238–9, 259, 270 and proportionality 54–5, 225, 227, 230, 232, 238–9, 242 value of 54, 55, 57, 59, 66, 125, 131, 132, 136, 137, 142, 143, 149, 182, 195, 199, 201–2, 207, 209, 210, 211, 212, 213, 214, 216–27, 228, 232, 233–6, 238–9, 242–3, 244–5, 252, 256, 263, 267 Warnock, G.J. 92n18 welfare, well-being 34, 38–9; see also ‘good for’ Whewell, William 150, 156, 162, 186 Whitehead, Alfred North 157 ‘why be moral?’ 27–8, 132–4, 261, 269 Williams, Bernard 37n25, 164, 175–6, 178, 183–4, 239, 266 Wittgenstein, Ludwig 4, 5, 16, 20, 58–9 Woolf, Leonard 16