London and New York
First published 1996
by Routledge 11 New Fetter Lane, London EC4P 4EE
Simultaneously published in the USA and Canada
by Routledge 29 West 35th Street, New York, NY 10001
Routledge is an International Thomson Publishing company.
? 1996 Geoffrey Scarre
Typeset in Times by Ponting-Green Publishing Services, Chesham, Bucks Printed and bound in Great Britain by Mackays of Chatham PLC, Chatham, Kent All rights reserved. No part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording or in any information storage or retrieval system, without permission in writing from the publishers.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library. Library of Congress Cataloging in Publication Data
1. Includes bibliographical references and index. 2. Ethics. 3. Utilitarianism. I. Title. BJ1012.S336 1995
ISBN 0-415-09527-1 (hbk)
ISBN 0-415-12197-3 (pbk)
I Introduction: The Character of the Theory 1
II Four Ancient Moralists 27
1 Mo Tzŭ 27
2 Jesus 33
3 Aristotle 37
4 Epicurus 39
III Utilitarianism and Enlightenment 48
1 Chastellux and Helvétius 50
2 Hutcheson 53
3 Hume 57
4 Priestley and Paley 60
5 Godwin 66
6 Bentham 72
IV John Stuart Mill 82
1 Early years 82
2 James Mill 85
3 The importance of character 87
4 Higher and lower pleasures 90
5 The „proof of utility‟ 96
6 Utility and justice 101
V Some Later Developments 106
1 Intuitional utilitarianism: Sidgwick 106
2 Ideal Utilitarianism: Moore and Rashdall 114
3 Rule-utilitarianism 122
VI Happiness and Other Ends 133
1 Preference and happiness 133
2 Dominant-and inclusive-end conceptions of happiness 137
3 Problems about multiple ends 141
4 Two contrasting responses 144
VII Maximisation, Fairness and Respect for Persons 152
1 Is utilitarian justice just? 152
2 Panem et circenses 155
3 „Whoever debases others is debasing himself‟ 158
4 But should the consequences count? 162
5 Limitations of the self-respect argument 166
6 Archangels, proles and the natural man 172
VIII Utilitarianism and Personality 182
1 Does utilitarian morality demand too much? 182
2 The hard line: utilitarians should be saints 186
3 A softer line: utilitarians may be human 187
4 Maximisation and alienation 193
5 Non-alienating direct utilitarianism 199
A few words about structure are in order. Following an introductory chapter setting out the basic features of modern utilitarianism, four historical chapters provide a critical survey of some of the most important utilitarian thinkers from the fifth century BC to the present. There is no sharp transition from the historical to the non-historical parts of the book; the final section of Chapter V, on rule-utilitarianism, appraises a debate within utilitarianism which began over half a century ago and is still going strong. The last three chapters are purely analytical and focus on what I believe to be the three most important areas of concern for contemporary utilitarian moral theorists: the definition of a philosophically viable concept of utility; the justification of utilitarian ideas about justice and fair treatment; and the defence of utilitarianism against the charge that it is too demanding a moral doctrine, requiring of individual agents a readiness for self-sacrifice that is possible only for moral saints. Special emphasis has been laid, in the historical sections, on the contributions made by older writers to clear thinking in these problem areas.
I am very grateful to Jonathan Wolff and to an anonymous reader for the publisher for extremely helpful comments on an earlier draft of this work. I have learned much also from spirited discussion of many of its themes with colleagues and students in the University of Durham. All have forced me to clarify the exposition and sharpen the argument at many points. It is hoped that the errors it contains will be fewer and less egregious for their assistance. I should also like to express my gratitude to my copy-editor, Marguerite Nesling, for exemplary editing of the text.
After some hesitation, I have decided to follow the traditional practice of using masculine word-forms throughout in gender-neutral contexts. It is difficult to incorporate „she or he‟ and „his or her‟ into a text while retaining tolerable prose, and the switch to feminine forms constitutes merely a kind of stylistic „out of the frying pan into the fire‟. The only really
satisfactory solution would be to introduce some new gender-neutral terms into the language: a task to be attempted by some bolder writer of the future.
Ancestors of two sections of the text have appeared in the pages of Utilitas: „Epicurus as a
forerunner of utilitarianism‟ was a longer version of Chapter II, section (4), while „Utilitarianism and self-respect‟ corresponds to the first half of Chapter VII; fuller references are given in the Bibliography.
Durham, June 1995
everything of importance.
This book is concerned both with the history of utilitarianism as an ethical theory and with the lively contemporary debate over its forms, problems and prospects. It will become clear as I proceed, but it it is only fair to declare at the outset, that my own sympathies run broadly in favour of the utilitarian point of view, though I am not convinced that a wholly satisfactory form of the theory is yet available. Perhaps the best that can
currently be said about utilitarianism is that it is a very bad form of moral philosophy, but that all the others are so much worse. At any rate, I am far from believing, as some philosophers profess to do, that the days of utilitarianism are numbered. Reports of the death, or imminent demise, of utilitarianism have been greatly exaggerated. „Utilitarianism is destroyed,‟ claimed John Plamenatz in 1949, „and no part of it left standing‟ (Plamenatz 1966:145). Yet there was
still enough life left in the corpse a quarter of a century later for Bernard Williams to hope that „The day cannot be too far off in which we hear no more of it‟ (Williams 1973: 150). Twenty years further on we hear as much about utilitarianism as we ever did, and the flow of new writings on the subject is unceasing. For a viewpoint often accused of simple-mindedness, utilitarianism has demonstrated a striking ability to go on stimulating forceful and subtle argument on both sides of the case. If the utilitarian theory is wrong, proving it to be so is
taking a remarkable amount of intellectual effort.
John Stuart Mill believed that the appeal of utilitarianism, to a certain cast of mind, was perennial. In every age of philosophy, he thought, one of its schools had been utilitarian (J.S. Mill 1838:87). But in every era too, the doctrine had provoked virulent opposition: from a very early period, the theory of utility had by some been designated as „utterly mean and grovelling‟ and „worthy only of swine‟ (J.S. Mill 1861:210). Utilitarianism is to the present
day the moral philosophy par excellence which people love to hate, though recent debate has
mostly been couched in more temperate language than in the days of Mill. It should be noted, however, that utilitarianism is not so much a single theory of morals as a family of theories, of markedly differing sophistication and plausibility. Few contemporary utilitarians would now defend without large quali-fications the views of a Godwin or a Bentham, and the failure of the more primitive versions of the doctrine does not entail that all varieties of it must fail. But the existence of several forms of utilitarianism gives rise to a problem of definition: it is not easy to say in few words exactly what the essence of utilitarianism is.
Consider, for instance, Mill‟s thumb-nail characterisation of it at the beginning of
The creed which accepts as the foundation of morals, Utility, or the Greatest Happiness Principle, holds that actions are right in proportion as they tend to promote happiness, wrong
as they tend to produce the reverse of happiness. By happiness is intended pleasure, and the absence of pain; by unhappiness, pain, and the privation of pleasure (210).
This is inadequate even as a statement of Mill‟s own theory, though it is a fair representation of Bentham‟s. Mill, as we shall see later, thought that intellectual and moral activities like reading poetry or being kind to one‟s neighbour contribute much more fully to our happiness
than the merely physical pleasures of sex or eating caviare. Moreover in some of his writings he moved away from a pleasure-pain definition of happiness towards a more Aristotelian conception of the happy subject as one who concentrates on developing the excellence of his character. Later utilitarians have also devoted careful attention to the analysis of happiness, and while many believe that it involves more than the attainment of pleasant experiences and the avoidance of painful ones, specific accounts of it vary. Some, particularly those influenced by developments in economics and decision theory, have jettisoned altogether the idea of happiness as the proper goal of moral action. In the tradition of utilitarian welfare economics, utility has generally been explained, as Alfred Marshall expressed it, as „correlative to Desire or Want‟ (Marshall 1920: 92). On this approach, the maximal satisfaction of human preferences (alternatively, of interests or desires) is considered to be the central concern of a utilitarian theory. A few writers again have opted for a broader construal of utility, to embrace not merely human satisfactions but also intrinsic goods like knowledge and beauty—a
position sometimes referred to as „Ideal Utilitarianism‟ (though it is arguable that the term
„utility‟ is being stretched here beyond tolerable limits). Attempts to sum up the doctrine of utility in a few sentences, then, are likely to fail; „utilitarianism‟ names a large, and loosely connected, group of theories.
For Mill, the label „utilitarian‟ „supplies a want in the language, and offers, in many cases, a convenient mode of avoiding tiresome circum-locution‟ (J.S. Mill 1861:210n.). But he
acknowledged, as subsequent writers have done, that the name adds little to the theory‟s
attractions. The word „utilitarian‟ presents to many people an uninviting picture of an earth-
bound and limited philosophy which cheerfully sacrifices more inspiring ideals on the altar of the useful. Bentham, who coined the term, took his cue from Hume‟s advocacy of „utility‟ in the Enquiry Concerning the Principles of Morals as the justifying ground of both the social 1virtues and the rules of equity and justice. Hume had claimed that no principles of action or
habits of mind were praiseworthy unless they conduced to the „happiness and welfare‟, either of the individual or society:
In common life, we may observe, that the circumstance of utility is always appealed to; nor is it supposed, that a greater eulogy can
be given to any man, than to display his usefulness to the public, and enumerate the services, which he has performed to mankind and society (Hume 1751:212).
Bentham adopted the term „utility‟, in the Fragment on Government of 1776, to denominate
the „tendency‟ of actions to promote „the common end of all of them‟, namely „Happiness‟
(Bentham 1776:237). According to the Oxford English Dictionary, the neologism „utilitarian‟
occurs for the first time in a letter of Bentham‟s of 1781, where a certain clergyman is
described as „a very worthy creature…a naturalist, a chemist, a physician‟—and „a utilitarian‟.
Despite his invention of the name „utilitarianism‟, however, Bentham was not the inventor of the doctrine, and did not claim to be. But the almost casual way in which the name entered the philosophical (and later, the popular) vocabulary suggests that he gave little thought to its suitability for the position he championed. By the early nineteenth century the label was too well established to be dropped, though Mill was not the only writer to regret the mis-understandings it so frequently caused: utility, he noted, was often thought to be opposed not merely to the „agreeable and ornamental‟ but, ironically, even to the pleasant (J.S. Mill 1861:209). A major purpose of Mill‟s 1861 manifesto for Fraser’s Magazine was thus to
counter the prevalent impression (recently reinfored by Dickens‟s satirical novel Hard Times
(1854)) that the utilitarian theory was dour, dry and informed by the lowest estimate of human 2 possibilities.
Although resistant to any simple definition, utilitarianism is not quite beyond characterisation in general terms. The theories which compose the family are linked by some common ideas and structural features which roughly set utilitarian positions off from others. There are five points of „family resemblance‟ of particular importance, though not every version of the doctrine exhibits them all. For the most part (exceptions will be pointed out as we proceed), utilitarian theories have been welfarist, consequentialist, aggregative and maximising;
historically, too, many have been universalist. Each of these notions requires some careful
1 Welfarist The „utility‟ of which utilitarians speak is most commonly identified with the
welfare (the faring well) of human beings, though some utilitarians (following Bentham)
extend their concern to the well-being of animals too. (Roughly, any subjects capable of having a better or a worse quality of life—which includes even the lowliest creatures capable
of sentience—may be thought of as fit objects of utilitarian
consideration.) Views of utility, or welfare, however, vary widely. The eighteenth-century utilitarians tended to construe welfare, as Bentham did, in terms of pleasure and pain: a life was held to be going well to the extent that it contained a balance of pleasurable sensations over painful ones. On the whole, these early writers disregarded a long tradition going back to Aristotle whereby pleasure was to be distinguished from happiness. For Aristotle, happiness
(eudaimonia) consisted in worthwhile activity (specifically, activity in accordance with human excellences of mind and character), rather than in pleasant amusements; happiness, moreover, was more properly ascribed to a life as a whole— a life exhibiting a certain
structured and successful directedness on suitable ends—than to individual phases of it.
Pleasures were, rather, experiences of the moment, and while Aristotle did not, like the Cynic Antisthenes, consider that pleasure was a thing best avoided, he firmly rejected the identification of the happy existence with the life of pleasure; to prefer the latter to the former was to prefer a life more „suitable to beasts‟—and „we call neither ox nor horse nor any of the other animals happy‟ (Aristotle, Nicomachean Ethics, 1095b, 1099b). Bentham, in contrast,
notoriously declared that the words „utility‟, „benefit‟, „advantage‟, „pleasure‟, „good‟, and „happiness‟ all came to the same thing (Bentham 1789:1-2).
Mill‟s subtler position, which we shall investigate more fully in later chapters, owes more to Aristotle in its definition of happiness than it does to Bentham; but Mill was in verbal agreement with Bentham that morally right actions were those which promoted happiness. Yet Mill‟s theory raises a difficulty which does not affect the less sophisticated utilitarianism
of his predecessor. Mill believed that the happy person is someone who pays close attention to his self-development, who works hard to hone his talents, refine his tastes and increase his sympathies— in short, a person who desires „for its own sake, the conformity of his own character to his standard of excellence‟ (J.S. Mill 1838:95). But this reference to a „standard of excellence‟ raises the question whether happiness is still being treated as the sole ultimate criterion of right action. Mill plausibly says that the development of personal excellence is a major source of happiness; but it is hard to see how achieving excellence could have this effect unless it were seen as something valuable in its own right; and if it were valuable for its
own sake, it would presumably be worth pursuing independently of its propensity to create happiness. Mill usually denied a causal model of the relations between happiness and whatever makes us happy, holding that happiness is not
so much the effect of happiness-making activities as a condition constituted by them. But even
if improving one‟s character is (in Mill‟s terminology) a „part of happiness‟ (J.S. Mill 1861: ch. 4), it seems not to be a part of it in the way that (say) listening to jazz or eating cream buns may be parts. These activities are valued just because they help us to be happy; whereas the pursuit of personal excellence would seem capable of making us happy only because it is valued. We can postpone until later a fuller scrutiny of these issues, but it is worth noting now that there is a question how purely utilitarian a theory is, once it allows that other things besides welfare have basic value.
Some contemporary utilitarians, in a tradition which might reasonably be described as Millian, affirm that the holding of certain objective facts about a person‟s condition is more salient than the history of his subjective states to judgements about the satisfactoriness of his life. David Brink, for instance, has proposed that utilitarians should adopt an „objective‟ theory of well-being whereby
a valuable life consists in the possession of certain character traits, the exercise of certain capacities, and the development of certain relations with others and to the world, and…the value of such a life is independent of the pleasure it contains and whether or not this sort of life is desired or would be desired in some preferred epistemic state (Brink 1989:221). Brink thinks such a theory called for because he disbelieves that any amount of purely pleasing experience could be sufficient to make a life worth living. A brain in a vat might be fed an unbroken sequence of pleasures, without having an existence which anyone would envy. We want to be certain sorts of people, and to do certain kinds of thing, rather than
merely have illusory subjective experiences (however pleasant) of being those persons and of doing those things (223-4; cf. Nozick 1974: 42). Brink concludes that a plausible theory of well-being „counts reflective pursuit and realization of agents‟ reasonable projects and certain personal and social relationships as the primary components of valuable lives‟ (231). (This
idea of the good life corresponds quite closely to what Elizabeth Telfer has called „eudaimonistic happiness‟, defined as the state of someone who is „truly fortunate‟ or „truly well-off‟ according to some objective standard of worthwhile life (Telfer 1980:37).)
Objectivist views of welfare are controversial. They imply that even a rational, well-informed person‟s evaluation of the satisfactoriness of his life could in principle be wrong. (Even rational, well-informed people can make mistakes about objective facts.) John Harsanyi has argued in
favour of a more subjectivist account of individual utility on the ground that an impartial sympathetic observer „cannot determine what is “good” for different persons with reference to
any pre-existing moral standard, but only with reference to the preferences of these persons themselves‟. To the question how do we want to be treated by other people, the „only answer
necessarily true, because tautologically true, …is: “I want to be treated in accordance with my
own wants”‟ (Harsanyi 1976: 31; cf. Harsanyi 1977:27). Yet not all a person‟s wants are sensible ones, and Harsanyi recognises that „benevolence cannot require us to satisfy people‟s foolish wants and preferences in the same way as their sensible wants and preferences‟.
Rejecting any objectivist criterion of the distinction between rational and foolish wants, he proposes to represent the difference „without reference to any standard outside of the own attitudes of the persons concerned‟, by identifying the „true preferences‟ which a person would manifest „on due reflection and in possession of all relevant information (including information on the pleasures and pains resulting from alternative courses of action)‟ (1976:31-
Objectivists may complain, however, that the identification of a person‟s „true‟ preferences will often call for some very difficult counterfactual reflection. (How do we determine what Jim, who is feckless and silly, would have chosen had he been sensible? And in what sense anyway would any inferred preference be a „true‟ one if it were one which Jim would never actually have had?) Moreover, even the preferences which were, for a given person, the most rational which could realistically be expected from him may not seem to wiser judges to be for
things which were truly in his interests (he may, for instance, be a child, or a mentally defective person). T.M. Scanlon has consequently suggested that in practice „the criteria of well being that we actually employ in making moral judgements are objective‟, and that our
appraisals of a person‟s well-being could be reasonable even where they „conflicted with the preferences of the individual in question, not only as he believes they are but even as they would be if rendered consistent, corrected for factual errors, etc.‟ (Scanlon 1975:658; cf. Scanlon 1993). Nevertheless objectivists need not deny the importance of subjective preferences to well-being; in Scanlon‟s view, „A high objective value may be attached to providing those conditions which are necessary to allow individuals to develop their own preferences and interests and to make these felt in the determination of social policy‟ (1975:658).
One reason why preference-satisfaction accounts of utility have been popular is that they appear nicely liberal in their refusal to make more than limited assumptions a priori about the character of reasonable goals:
utility is identified with the satisfaction of whatever goals people happen to have, provided that they attain some threshold degree of rationality (admittedly there is room for dispute as to where this threshold lies and, in regard to some preferences, on which side of the line they fall). Sen expresses such a conception when he explains utility „to stand for a person‟s conception of his own well-being‟ (Sen 1979:463). No question-begging claims are made
about the nature of happiness or the good life; no attempt is made to lay down what human beings should want for themselves or for others; and one man‟s meat is conceded to be
another man‟s poison. (This is related to the doctrine, popular with economists, of the „supremacy of the consumer‟.) While the preferences of different persons can conflict, and a
calculation needs to be done to determine how to maximise preference-satisfaction overall, no (minimally rational) preferences are deemed to be intrinsically less suitable than any others to enter the calculation. Harsanyi explains why it is better to judge what is good or bad for persons in terms of their preferences rather than their desires or wants: a person can have
conflicting wants and desires, and these can therefore guide his behaviour or those of other people who wish to benefit him only if he decides which wants or preferences represent his considered preferences (Harsanyi 1976:32). To be measurable these preferences must of course be revealed by „outward phenomena‟, such as, in the easiest case, the price which an individual is willing to pay for their satisfaction (cf. Marshall 1920:92). It is a standard assumption of the utilitarian doctrine that utility comes in greater and lesser amounts. The right action or policy is normally that one of the available alternatives which is likely to realise the largest quantity of utility. Many utilitarians, from Hutcheson and Bentham to Brandt, have accordingly been tantalised by the prospect of a „calculus of utility‟—a
method of reducing moral decisions to mathematical calculation. For outcomes to be evaluated mathematically, it is necessary, first of all, to have a way of calibrating individual utility. (It is also necessary to define a consistent notion of social utility which takes account of the fact of conflicts among the preferences of individuals; but of this more later.) The trouble is that the philosophical sophistication of a concept of utility and its susceptibility to quantitative treatment are generally in inverse proportion to one another. It is not easy to see, for instance, how we could calibrate utility as represented in Brink‟s rich account of the
„valuable life‟, or even in Sen‟s more subjectivist sense of „a person‟s conception of his own well-being‟ (cf. Mirlees 1982:65).
Nevertheless, following the pioneering work in game theory of J. von Neumann and O. Morgenstern, progress has been made in providing a
mathematical theory of utility capable at least of application in such a limited domain as welfare economics. While the idea of a calculus of utilities able to generate solutions to all moral problems is likely to remain a philosopher‟s pipe-dream, mathematical techniques have
been used with some success in appraising the relative utility of economic outcomes. Von Neumann and Morgenstern explained how an individual can compare the utilities not only of single events, but of combinations of events with stated probabilities. Suppose that a person prefers, in regard to three events A, B and C, event A over event B, but also event C over event
A. Let α be a real number in the interval 0 to 1, such that the desirability of A is exactly equal
to that of the combined event consisting of a chance of probability 1-α for B and the chance of
probability α for C. Von Neumann and Morgenstern‟s ingenious suggestion is that a can now
be employed to provide a numerical estimate for the ratio of the preference of A over B to that
of C over B. This germ of theory is the starting point for an elaborate mathematical treatment of utility by axiomatic methods (Von Neumann and Morgenstern 1953:17-19).
Economic utilitarians have laid much stress on the fact that measur-ability of utility is attainable provided that some standard of comparison can be determined. One way of doing this is as follows. An individual arbitrarily fixes two very similar outcomes, A and B, as his
standard of comparison, assigning zero utility to A and unit utility to B. The utility difference
between two different outcomes C and D can then be taken to be unity if he is indifferent
between the combinations (A, D) and (B, C) (cf. Mirlees 1982:65). Such methods presuppose,
however, that a person is able to determine clear orders of preference among alternative
outcomes, a condition which is not always fulfilled in practice. One source of difficulty here is that outcomes to be compared may differ in more than one parameter. A person would have little difficulty in deciding between a trip to London plus ?100 spending money and a trip to London with only ?95 spending money. But he might find it much harder to choose between a trip to Paris with ?75 worth of spending money and a trip to London with ?100 (this example simplified from Vickrey 1960:520-1).
We shall return in Chapter VI to the question of the relative merits of „happiness‟ versus „preference-satisfaction‟ theories of utility. The case will be argued there that a satisfactory utilitarian ethics demands a richer, more expansive theory of human well-being—of what is
required for a life to go well—than can readily be supplied on a preference-satisfaction
account. For the moment, however, the differences between classical
and welfare-economic-inspired approaches are less important than the common emphasis in their theories of value on the living of satisfying lives—in short, their shared concern with
welfare, in a broad sense of that term.
2 Consequentialist Utilitarianism is a consequentialist (or, to use an older term, a
„teleological‟) doctrine in the sense that it maintains that the proper response to its values is to promote them (cf. Pettit 1993a: 231). Act consequentialism (the simplest form) maintains that an action is right if it can reasonably be expected to result in a state of affairs at least as good as the alternative states of affairs that would have resulted from alternative feasible acts (cf. Sen 1979:464). Rule consequentialism is a little more complicated: it holds that actions are right if they conform to rules whose general observance could reasonably be expected to result in a set of states of affairs at least as good as the sets of states of affairs that would have resulted from the adoption of alternative feasible rules. (Rule-utilitarianism is open to a number of interpretations, as we shall see in Chapter V.) For all forms of consequentialism the good is logically prior to the right, as no criterion of right action can be laid down until a conception of the good is specified.
Utilitarian theories agree that the good is utility, though they differ in their accounts of what utility is; they also differ, as we shall see, as to whether it is the total utility, or the average utility of individuals, that we should be seeking to maximise. But non-utilitarian forms of consequentialism are conceivable as well. For instance, egoism is a species of consequentialism which evaluates outcomes according to their propensity to enhance the agent‟s own welfare. So too is counter-utilitarianism (or „philosophical sadism‟), which
affirms that right actions are the ones which diminish human welfare. Other possible consequentialist views hold that actions are right if, and only if, they promote the advantage of a certain god or king or social grouping. But the only form of consequentialism other than utilitarianism which will be considered in this book is the ethical theory of Epicurus, which while strictly a variety of egoism is in important respects a precursor of utilitarian ideas. It is easy to misrepresent the character of consequentialism. Thus a recent, and sympathetic, writer has described utilitarianism as „a member of the family of moral doctrines which judge
actions neither by their motives nor their intrinsic qualities, but by their consequences‟ (Brittan 1990:76). This gives the misleading impression that utilitarians take no account of motives in the appraisal of actions. Consequentialists reject a view like Kant‟s, that only
actions done from a sense of duty and respect