Towards Reunion In Ethics 3030124096, 9783030124090

Citation preview

Philosophical Studies Series

Jan Österberg

Towards Reunion in Ethics

Philosophical Studies Series Volume 138

Editor-in-Chief Mariarosaria Taddeo, Oxford Internet Institute, Digital Ethics Lab, University of Oxford, Oxford, UK Executive Editorial Board Patrick Allo, Vrije Universiteit Brussel, Brussel, Belgium Massimo Durante, Università degli Studi di Torino, Torino, Italy Phyllis Illari, University College London, London, UK Shannon Vallor, Santa Clara University, Santa Clara, CA, USA Board of Consulting Editors Lynne Baker, Department of Philosophy, University of Massachusetts, Amherst, USA Stewart Cohen, Arizona State University, Tempe, AZ, USA Radu Bogdan, Dept. Philosophy, Tulane University, New Orleans, LA, USA Marian David, Karl-Franzens-Universität, Graz, Austria John Fischer, University of California, Riverside, Riverside, CA, USA Keith Lehrer, University of Arizona, Tucson, AZ, USA Denise Meyerson, Macquarie University, Sydney, NSW, Australia Francois Recanati, Ecole Normale Supérieure, Institut Jean Nicod, Paris, France Mark Sainsbury, University of Texas at Austin, Austin, TX, USA Barry Smith, State University of New York at Buffalo, Buffalo, NY, USA Nicholas Smith, Department of Philosophy, Lewis and Clark College, Portland, OR, USA Linda Zagzebski, Department of Philosophy, University of Oklahoma, Norman, OK, USA

More information about this series at http://www.springer.com/series/6459

Jan Österberg

Towards Reunion in Ethics

Jan Österberg Department of Philosophy University of Uppsala Uppsala, Sweden

ISSN 0921-8599     ISSN 2542-8349 (electronic) Philosophical Studies Series ISBN 978-3-030-12409-0    ISBN 978-3-030-12410-6 (eBook) https://doi.org/10.1007/978-3-030-12410-6 Library of Congress Control Number: 2019936167 © Springer Nature Switzerland AG 2019 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Editorial Note

This is a posthumous publication. At the time of his death in July 2011, Jan Österberg had nearly completed the manuscript. A final section, 7.6, called “Envoi,” remained unfinished. This uncompleted section has been appended at the end of the preceding Sect. 7.5. In other respects, the editorial changes made prior to publication are only minor. They mainly consist in the correction of incomplete or faulty cross-­references and typographical errors. Uppsala, May 2018 Erik Carlson and Rysiek Sliwinski

v

Preface

A person living quite alone, like Robinson Crusoe before he met Friday, does not need a moral code; in order to survive and lead an acceptable life, he needs a prudential code, a set of rules for the guidance of his private life. But people who do not live alone but interact with each other need a moral code, a set of rules for the guidance of their common life; otherwise, their lives will be, in Hobbes’s words, “solitary, poor, nasty, brutish, and short.” The question is: What moral code is the most reasonable one? For almost three centuries, the main answers to the question have been those provided by consequentialism and deontological ethics. A lot of eminent philosophers on each side have tried to persuade their opponents to come over to their side, but in vain. The present essay is an attempt to decide the point at issue in this Homeric struggle. Two convictions have formed the shape of my essay. One is the conviction that consequentialism and deontological ethics really are the main competitors in the moral field. (To some extent I try to defend this conviction in Chap. 1.) The other conviction is that neither of the moralities, which have survived for so long, can be wholly wrong; the true morality must be some compromise between consequentialism and deontological ethics (what compromise I try to spell out in the last chapter). The task that I have set myself is, of course, a rather hopeless one. The ground that I tread on is everywhere mined with deep philosophical problems. In order to do justice to these, I should have devoted at least a chapter to each of them, instead of, as now, a few pages. But sometimes it may be defensible to launch an attack without caring too much about the supply lines, or so I hope. My essay has a feature which distinguishes it from at least many other works in moral philosophy: it is (to some extent) occupied with the psychological, sociological, and historical background of the moral theories discussed. Thus, for example, I set forth my conviction as to the origins of consequentialism, discuss what psychological features of human nature have determined the shape of deontological ethics, and point out the gradual development of common-sense morality. The reason is that moral theories are variously embedded in social and psychological contexts and vii

viii

Preface

are better understood if the contexts are made explicit. (To some extent, the same is true of scientific theories. Thus, witness the recent trend among philosophers of science to describe the historical background of scientific theories.) To forestall a possible misunderstanding, I want to stress that, in my opinion, this contextualization has no relativistic implications whatever. Henry Sidgwick thought that the prima facie reasonable moral theories are those of intuitionism, utilitarianism, and ethical egoism. He further thought that intuitionism (roughly, deontological ethics) had to give way to utilitarianism (roughly, consequentialism), together with ethical egoism the two main competitors in the moral field. In an earlier book, Self and Others, I argued that ethical egoism is not a plausible moral theory. In the present book, I will argue that deontological ethics does not have to give way to consequentialism. Once more, I thus find myself having written a Sidgwickean book. This testifies either to my own narrowness of mind or to the enduring value (interest) of Sidgwick’s view on ethics. Uppsala, Sweden

Jan Österberg

Contents

1 The Conflict������������������������������������������������������������������������������������������������    1 1.1 Dramatis Personae������������������������������������������������������������������������������    1 1.2 Consequentialism vs. Deontology—The Received View ������������������    4 1.3 The Received View Questioned����������������������������������������������������������    9 1.4 Other Moral Theories�������������������������������������������������������������������������   13 1.5 Conceptions of Morality ��������������������������������������������������������������������   22 1.6 A Short Preview����������������������������������������������������������������������������������   25 2 Deontological Ethics: Exposition��������������������������������������������������������������   27 2.1 Common-Sense Morality��������������������������������������������������������������������   27 2.2 Common-Sense Morality and Contractualism������������������������������������   32 2.3 Collective Versus Individualistic Codes����������������������������������������������   37 2.4 Deontological Features in Common-Sense Morality ������������������������   40 2.5 Common-Sense-Moral Features in Deontological Ethics������������������   53 2.6 Common-Sense Morality and Deontological Ethics��������������������������   58 3 Deontological Ethics: Assessment������������������������������������������������������������   61 3.1 Can the Deontological Features Be Justified?������������������������������������   61 3.2 Common-Sense Morality and Its Critics��������������������������������������������   63 3.3 Is Common-Sense Morality Self-Defeating?��������������������������������������   77 3.4 Is Common-Sense Morality Irrational?����������������������������������������������   79 3.5 The Limits of Common-Sense Morality��������������������������������������������   87 3.6 Précis of Chapters 1–3������������������������������������������������������������������������   91 4 Consequentialism: Exposition������������������������������������������������������������������   93 4.1 Forms of Consequentialism����������������������������������������������������������������   94 4.2 The Interpretation of Objective Consequentialism ����������������������������   99 4.3 The Interpretation of Subjective Consequentialism����������������������������  104 4.4 The Golden Rule��������������������������������������������������������������������������������  108

ix

x

Contents

5 Consequentialism: Assessment ����������������������������������������������������������������  121 5.1 Some Objections��������������������������������������������������������������������������������  122 5.2 The Arguments from Being Too Demanding��������������������������������������  133 5.3 The Argument from Horrendous Actions ������������������������������������������  136 5.4 Is Consequentialism Self-Defeating? ������������������������������������������������  143 5.5 The Dependence of Consequentialism on Deontological Ethics��������  148 5.6 Précis of Chapters 1–5������������������������������������������������������������������������  154 6 Possible Solutions ��������������������������������������������������������������������������������������  155 6.1 Eliminations and Reconciliations ������������������������������������������������������  156 6.2 Consequentialist Attempts at Reconciliation��������������������������������������  159 6.3 Two Attempts at Partial Reconciliation����������������������������������������������  166 6.4 Rule Consequentialism ����������������������������������������������������������������������  171 6.5 Justifying Ultimate Principles: Some Traditional Proposals��������������������������������������������������������������������������������������������  174 6.6 Justifying Ultimate Principles: Some Non-Traditional Proposals��������������������������������������������������������������������������������������������  182 6.7 Concluding Remarks��������������������������������������������������������������������������  188 7 My Solution������������������������������������������������������������������������������������������������  189 7.1 Contract Theories��������������������������������������������������������������������������������  190 7.2 The Autonomy Condition ������������������������������������������������������������������  192 7.3 Parfit on Contractualism and Rule Consequentialism������������������������  199 7.4 The Present Standing of C* and D* ��������������������������������������������������  203 7.5 How Should We Act?��������������������������������������������������������������������������  204 Appendix: Hodgson’s Argument��������������������������������������������������������������������  213 Bibliography ����������������������������������������������������������������������������������������������������  249

Chapter 1

The Conflict

1.1  Dramatis Personae A fundamental, not to say the fundamental, conflict in normative ethics is the conflict between consequentialism (C) and deontological ethics (D), that is, (very roughly) between the view that you should do what has the best consequences and the view that there are actions which, for certain stated reasons, it is wrong to do even if they have the best consequences.1 An example that vividly brings out the contrast between the two views is the following one, taken from Gilbert Harman. (I shall call it “the doctor’s choice”.) You have five patients in the hospital who are dying, each in need of a separate organ. One needs a kidney, another a lung, a third a heart, and so forth. You can save all five if you take a single healthy person and remove his heart, lungs, kidneys, and so forth, to distribute to these five patients. Just such a healthy person is in room 306. He is in the hospital for routine tests. Having seen his test results, you know that he is perfectly healthy and of the right tissue compatibility. If you do nothing, he will survive without incident; the other patients will die, however. The other five patients can be saved only if the person in Room 306 is cut up and his organs distributed. In that case, there would be one dead but five saved.2

In the doctor’s choice—given some not too far-fetched assumptions, such as that you and your assistant nurse can keep the whole thing secret, that both of you are so convinced about the moral rightness of the action that you don’t suffer any emotional disturbance or pangs of bad conscience later on, that each of the five patients and the healthy man would be about equally missed by his near and dear, etc.—C says that you should cut up the person in room 306 and distribute his organs to the five patients. Doing this, we assume, has better consequences than not doing it: it is better that five persons survive and one dies than that one person survives and five 1  Obviously, then, this does not mean that a moral theory is either consequentialist or deontological. For a short review of other moral theories, see Sect. 1.4 below. 2  Gilbert Harman, The Nature of Morality: An Introduction to Ethics (New York: Oxford University Press, 1977), p. 3 f.

© Springer Nature Switzerland AG 2019 J. Österberg, Towards Reunion in Ethics, Philosophical Studies Series 138, https://doi.org/10.1007/978-3-030-12410-6_1

1

2

1  The Conflict

die. Any deontological theory, on the other hand, would say that you should not do that: it is wrong to kill an innocent person in order to prolong the lives of other people. Confronted with the doctor’s choice you may think that C is obviously mistaken, that an acceptable moral theory must be some version of D. This is, I believe, how many people would spontaneously react. But reflection should give us pause. The core idea of C—that one should do what has the best consequences, that one should minimize evil and maximize good—seems hard to resist. For, as Samuel Scheffler points out, this idea seems to be one of the main elements of our conception of practical rationality. Anyone who resists consequentialism seems committed to the claim that morality tells us to do less good than we are in a position to do, and to prevent less evil than we are in a position to prevent. And this does not sound nearly as plausible.3

How our intuitions vary and now support one theory, now the other, depending on the details of the case, can be illustrated by the following example, proposed by the eighteenth century consequentialist William Godwin. Godwin asks us to consider the following situation: In a burning building two people are trapped, Fénelon, the famous author, and his valet. It is possible to save only one of them. It is evident, Godwin says, that you ought to save Fénelon, not his valet. For, if saved, Fénelon will continue to instruct and entertain many people, so the consequences of saving him are better than those of saving his valet. I think that, given only this information, many people would agree with Godwin’s verdict that one should save Fénelon. (This verdict is supported, not only by C, but by many deontological theories as well.) But the agreement would probably disappear when Godwin adds to his description the fact that the valet is your brother or your father, claiming that this does not change the verdict. The life of Fenelon would still be more valuable than that of the valet; and justice, pure, unadulterated justice, would still have preferred that which was most valuable. Justice would have taught me to save the life of Fenelon at the expense of the other. What magic is there in the pronoun ‘my,’ that should justify us in over-turning the decisions of impartial truth?4

In this situation consequentialism prescribes that, since saving Fénelon has better consequences than either saving the valet or saving neither person, this is what you should do. Many people, I think, feel the weightiness of the consequentialist consideration, at least to some extent. At the same time they suspect that, were they in the rescuer’s position, they would probably save the other person, their father or their brother, thinking that it would be morally permissible to do so. To defend their 3  Samuel Scheffler, “Introduction” to Scheffler (ed.), Consequentialism and Its Critics (Oxford: Oxford University Press, 1988), p. 1. Cf. G. E. Moore, Ethics (London: Oxford University Press, 1912), p. 76f. 4  William Godwin, Enquiry Concerning Political Justice (Dublin: Luke White, 1793); quoted from D. H. Monro (ed.), A Guide to the British Moralists (London: Collins, 1972), p. 189. It should be noticed that Godwin uses the term “justice” as “a general appellation for all moral duty” (ibid., p. 187).

1.1  Dramatis Personae

3

(hypothetical) choice they would cite principles such as, “Everyone has a special obligation to his family that overrides the general obligation to do what has the best consequences”. And most deontological theories would support this choice, citing principles of a similar kind. The question is, as Godwin points out, how on earth is this partiality to near and dear and other similar deontological restrictions to be morally justified? I think that, upon reflection, many people feel themselves divided when confronted with cases where C and deontological theories conflict with each other: C seems to have a strong case, but it is not able to silence the strongly held intuitions that oppose it. Small wonder then that the intrapersonal conflict has turned into an interpersonal debate between moral philosophers. The debate between consequentialists and deontologists has been carried on since the eighteenth century and is as lively today as it was in the days of Bentham and Godwin, Price and Butler. Textbooks in normative ethics are usually structured round this debate, and a never ending flood of books and articles defend and criticize these positions. Deontologists are accused by their opponents for, among other things, rule-fetishism and for not providing any justification for their diverse rules. Against C it is objected, among other things, that it is too demanding and sometimes, as in the doctor’s choice, prescribes grossly immoral actions. I think that what engenders this conflict in normative ethics is the conflict between common-sense morality (CSM), on one hand, and (what may be called) universalistic rationality, on the other. Let me explain the latter expression: Practical rationality tells the agent to try to bring about the best possible outcome. The individualistic conception of (practical) rationality takes the best outcome to be the best outcome for the agent, whereas the universalistic conception takes it to be the impersonally best outcome. As Samuel Scheffler says, this conception seems hard to resist. For given only the innocent-sounding assumption that good is morally preferable to evil, it seems to embody the principle that we should maximize the desirable and minimize the undesirable, and that principle seems to be one of the main elements of our conception of practical rationality.5

Different deontological theories can be seen as attempts to defend CSM against the claims of universalistic rationality, and different versions of C as attempts to work out the most acceptable version of universalistic rationality. I spoke about “deontological theories”. Now many deontologists, it is true, do not offer fully worked-out theories; they are content to criticize different aspects of C. But common to deontologists—evincing their dependence on common-sense morality—is their reference to our moral intuitions, to what we unreflectively consider right and wrong, as well as their acceptance of certain crucial features of CSM, later to be surveyed. And common to consequentialists is their conviction that it is contrary to reason not to (attempt to) bring about the best outcome. At the most fundamental level, then, the conflict is not a conflict between two kinds of philosophical moral theories, but between a conception of rationality and positive morality, the moral  Samuel Scheffler, op. cit., p. 1.

5

4

1  The Conflict

code actually accepted by people, shaping their ways of conceiving themselves and other people and guiding their actions, often not wholly consciously. In the next section of this chapter, I shall briefly characterize deontological and consequentialist theories. In Sect. 1.3, I shall discuss certain attempts to question this characterization. Then in Sect. 1.4, I shall give a brief overview of other moral theories, arguing that those that are neither deontological nor consequentialist are either implausible or best seen as complements to these. In Sect. 1.5, I shall introduce some conceptions of morality that are taken for granted in the following. In Sect. 1.6, finally, I shall give a preview of the rest of the book.

1.2  Consequentialism vs. Deontology—The Received View How are C (consequentialist ethics) and D (deontology), the key concepts of my account, to be understood? Since, in my terminology, C is a species of teleological ethics, I shall first discuss the distinction between D and teleological ethics and then state what distinguishes C from other forms of teleological ethics. When deciding how to use these technical terms, it is, of course, desirable not to create confusion by unnecessarily departing from established usage. The problem is that there is no firmly established usage in this area; moral philosophers use the terms somewhat differently, sometimes, no doubt, due to their taking different positions on meta-­ ethical and other philosophical questions.6 For me, the most important objective when defining these terms is to bring out as clearly as possible (what I take to be) the essential nature of the actual conflict between C and D while presupposing as little as possible of controversial ideas. The term “deontology” was introduced (and contrasted with “utilitarianism”) by William Whewell.7 But the opposition between “deontological” and “teleological” ethics was introduced by C. D. Broad in his discussion of the connection between concepts of obligation (‘notions like “right”, “ought”, “duty”, etc.’) and concepts of value (‘concepts like “goodness”, “merit”, etc.’).8 Some moral theories, Broad observes, claim that there is some connation between the two kinds of concepts. 6  For a rather exhaustive survey of the usage of “deontological”, see Shelly Kagan, Normative Ethics (Boulder, Col.: Westview Press, 1998), p. 73 f. 7  In his preface to James Mackintosh, A Dissertation on the Progress of Ethical Philosophy, chiefly during the Seventeenth and Eighteenth Centuries (Edinburgh: Adam and Charles Black, 1836). (I owe this information to Knud Haakonssen’s article on Mackintosh in W. J. Mander and Alan P. F. Sell (eds.), Dictionary of Nineteenth-Century British Philosophers, 2 vols. (Bristol: Thoemmes Press, 2002), Vol. 2: 715–19.) Bentham uses the term “deontology” in the title of one of his works. But, as is evident from the rest of the title, the term here means something quite different, viz. the Science of Morality. 8  C. D. Broad, Five Types of Ethical Theory (London: Routledge & Kegan Paul, 1930), p. 277. (In her article “Teleological Ethics” in Routledge Encyclopedia of Philosophy (London and New York: Routledge, 1998), Christine Korsgaard says that the terms originate with J. H. Muirhead, Rule and

1.2  Consequentialism vs. Deontology—The Received View

5

Now such theories might take the following forms. […] The concepts of obligation are fundamental and the concepts of value are definable in terms of them. Thus it might be held that the notion of fittingness is fundamental, and that “X is intrinsically good” means that it is fitting for every rational being to desire X. Such theories might be called Deontological. […] The concepts of value are fundamental and the concepts of obligation are definable in terms of them. Such theories may be called Teleological. E.g., it might be held that “X is a right action” means that X is likely to produce at least as good consequences as any action open to the agent at the time.9

A prominent feature of this characterization of deontological and teleological ethics is that the connection between the concepts of obligation and the concepts of value is taken to be definitional. Since nowadays hardly any proponent of a theory recognized as deontological or teleological makes such a strong assumption, the characterization unnecessarily narrows down the set of theories to be discussed.10 However, the idea that, according to deontological theories, concepts of obligation, whereas, according to teleological theories, concepts of value are more fundamental, seems plausible. The question is how the connection between the two kinds of concepts should be understood, if not as definitional. One such possibility is mentioned by Broad. Neither concept might be definable in terms of the other, but there might be synthetic and necessary connexions between them. Many people who would deny that the proposition “I ought to do X” means that X will probably have the best consequences of all actions open to me at the time, would yet hold it to be self-evident that I ought to do the action which will probably have the best consequences of those open to me at the time.11

The proposal, then, is that, according to both deontological and teleological theories, what they state are synthetic a priori claims. Now this is probably true of, for example, the consequentialist theory defended by G.  E. Moore in his Ethics, but most proponents of deontological and teleological theories make no such strong claims. Some, like, for example, J. J. C. Smart, are non-cognitivists, denying that moral principles have truth values, and many, both cognitivists and non-cognitivists, are critical of the very notion of synthetic a priori truths. These critics may, of course, be wrong, but even if we assume that there are a priori truths, those who End in Morals (London: Oxford University Press, 1932). But this is wrong: Broad’s book antedates Muirhead’s by 2 years.) 9  Broad, op. cit., p. 278. 10  One of the few moral philosophers who take the connection between the concepts of obligation and the concepts of value to be definitional is G.  E. Moore in Principia Ethica (Cambridge: Cambridge University Press, 1903). However, in his later Ethics, this claim is dropped. To be true, some philosophers discussing deontological and teleological theories in a general way follow Broad in characterizing the connection in terms of “define”. (See, e.g., John Rawls, A Theory of Justice (Cambridge, Mass.: Harvard University Press, 1971, p. 24 f.) In such cases, however, one gets the impression that these philosophers use “define” in some rather loose sense. 11  Broad, op. cit., p. 278. Broad’s characterization of the theories might give the impression that he is talking about meta-ethical theories. But what Broad says in the following (se, e.g., my next quote from him) makes it evident that the theories are (also) meant to be normative.

6

1  The Conflict

deny that there are moral truths of this kind seem to have a strong case: moral principles do not seem to have the same kind of uncontradictability as have those ­propositions usually offered as examples of synthetic a priori truths, propositions expressed by sentences such as, “We see with our eyes” or, “Red is a colour”. Whereas such propositions are hardly seriously disputed by anyone who understands their meaning, any interesting deontological and teleological principle is denied by at least some people whose understanding of its meaning does not seem to be any less good than that of its adherents. (The conviction that a moral principle is self-evident is no proof of its truth, especially not if other people hold other moral principles, incompatible with it, to be equally self-evident). We seem to be left with the following idea. Moral principles (of the kind that interests us now) state that an action is right (wrong, ought to be done) in virtue of some other, internal or external, property (or properties) that the action has, for example, being the fulfilment of a promise or having better consequences than any alternative action; having this other property is what makes the action right, etc.12 The distinction between teleological and deontological ethics may then be brought out as follows: A principle is teleological if, and only if, according to it, what makes an action right, etc. is (either the total or only some of) the (intrinsic) value of its consequences—usually as compared with the (intrinsic) value of the consequences of alternative actions.13 A principle is deontological if, and only if, according to it, some natural (empirical) or supernatural property of an action, such as being the fulfilment of a promise, or being commanded by God, is what makes the action right, etc.14 A moral theory, I stipulate, is teleological if, and only if, all its principles are teleological; if some or all of its principles are deontological, the theory is deontological.15 According to my above definition of “teleological principle”, either the total or only some of the value (positive and negative) of the consequences of an action is  A non-cognitivist may object that the value of something is not a property of it. But nothing substantial hinges on my use of the term; it is used only for ease of exposition. 13  David Brink claims that “[t]eleological and nonteleological theories can be distinguished along traditional lines if it holds rightness to consist in maximal goodness” (Moral Realism and the Foundations of Ethics (Cambridge: Cambridge University Press, 1989), p. 215). (Brink says that this restriction “is sufficient to distinguish between teleological and non-teleological theories” (ibid.); he does not, however, claim that it is necessary, and, as far as I can see, it is not. So it is not clear why he wants to restrict teleological theories the way he suggests.) The drawback with Brink’s suggestion is that, according to it, teleological and non-teleological theories will not, contrary to what he says, “be distinguished along traditional lines”. Thus, for example, non-comparative and satisficing versions of consequentialism as well as ethical egoism, traditionally classified as teleological theories, will—for no good reason—be classified as non-teleological (deontological) theories. (Different versions of consequentialism will be discerned and discussed in Sect. 4.1.) 14  It might be objected that, according to this definition, also virtue-ethical principles that characterize right action in terms of what the virtuous person would do (see the next section) are deontological. But this property is not, I take it, an empirical property, since “virtuous” is a (thick) value term. 15  I do not assume that deontological and teleological theories together exhaust the field of moral theories. (In Sect. 1.4 below, I discuss some theories which fall outside this division.) This assumption is, however, made—wrongly, in my opinion—by, e.g., John Rawls, op. cit., p. 24 f. 12

1.2  Consequentialism vs. Deontology—The Received View

7

relevant to its normative status. A teleological principle of the first kind I shall call a consequentialist principle. Thus, for example, utilitarianism is a consequentialist theory, whereas, for example, ethical egoism is not.16 There is a wealth of consequentialist theories, since goodness of outcomes can be (and has been) specified in many different ways. One is in terms of well-being, a notion which itself can be further specified in several different ways: pleasure, happiness, desire-satisfaction, etc., each of which, in turn, can be further specified. I shall use the term “utilitarianism” for the version of C that takes goodness of outcomes to be well-being. Utilitarianism is the most discussed and, consequently, the most developed form of C. I shall therefore often use this form for illustrative purposes, but it should be kept in mind that what I say about it is meant to apply to C in general. In the case of D there are not, as in the case of teleological ethics, any subdivisions based on formal features. I said above that a principle is deontological if, and only if, it claims that what makes an action right, etc. is something other than the value of its consequences, for example, its being the fulfilment of a promise. And a theory is deontological if, and only if, at least one of its principles is deontological. Depending on which properties are taken as right-making, it is possible to devise a great many deontological theories. But, as we will see, the deontological theories

 Some authors use “teleological theory” for what I call “consequentialist theory”, or use both terms in the latter sense. See, e.g., David O. Brink, Moral Realism and the Foundations of Ethics, pp.  213–16; Robert Olson, “Teleological Ethics”, in Paul Edwards (ed.), The Encyclopedia of Philosophy (New York: Macmillan, 1967); John Rawls, op. cit., p. 24 f. In my opinion, the distinction between consequentialist theories (in my sense) and other forms of teleological theories (in my sense) is an important one, and the distinction should be made terminologically explicit. And the terms “teleological” and “consequentialist” serve that purpose quite well. A rather deviant use of “consequentialism” is the one made by David Gauthier, who uses the term to include all kinds of theories concerned with utility maximizing, whether general or individual. See his “Coordination”, Dialogue, 14 (1975): 195–221; repr. in his Moral Dealing: Contract, Ethics, and Reason (Ithaca and London: Cornell University Press, 1990): 274–97. The term “consequentialism” was introduced in moral philosophy by Elizabeth Anscombe in a sense quite similar to the one used here. The main differences is that she takes the principles of consequentialism as definitions and restricts its range of application to comparative and maximizing versions (see Sect. 4.1 below) of what I call consequentialism. According to consequentialism, she says, “‘the right action’ means the one which produces the best possible consequences” (G. E. M. Anscombe, “Modern Moral Philosophy”, originally published in Philosophy, 33 (1958): 1–19. I quote from the reprint in Roger Crisp & Michael Slote (eds.), Virtue Ethics (Oxford: Oxford University Press, 1997): 26–44; p. 34.) In Thomas Mautner (ed.), Dictionary of Philosophy, 2nd ed. (Harmondsworth: Penguin, 1997), p. 107, it is said that Anscombe meant by her term “the view that an agent is equally responsible for the intended consequences of an act and its unintended but foreseen consequences”. What she actually says, however, is that Sidgwick was the first who held this view and that “this move on the part of Sidgwick explains [my emphasis] the difference between old-fashioned Utilitarianism and that consequentialism, as I name it, which marks him and every English academic philosopher since him” (Anscombe, op. cit., p. 37). For much more elaborate characterizations of consequentialism, see Erik Carlson, Consequentialism Reconsidered (Dordrecht: Kluwer, 1995) and Krister Bykvist, “Normative Supervenience and Consequentialism”, Utilitas, 15 (2003): 27–49.

16

8

1  The Conflict

that actually have been propounded are quite similar to each other. (This, I shall argue, is due to their being heavily influenced by common-sense morality.) Deontological and teleological theories do not, of course, have to be restricted to conduct. In addition, or alternatively, they may be brought to apply to things such as motives, character traits, institutions, rules, laws, and policies.17 But such applications fall outside the controversy which I intend to discuss in this book. It is also true, as, among others, Marcia Baron has pointed out, that the proper response to a value is not necessarily that of promoting it; other possible responses are, for example, those of expressing, appreciating, or honouring it.18 Taking account also of these responses, both deontological and teleological theories could be given a wider scope. Since, however, the controversy being discussed in this book confines itself to the promoting response to value, so shall I. It should be noticed that (what is called) “rule consequentialism” and, in general, (so-called) rule-teleological theories—according to which (roughly) an action is right if, and only if, it conforms to a rule justified in terms of the value of the outcome of its general acceptance or observance—are, in my terminology, versions, not of teleological, but of deontological ethics. I think this is as it should be, since there is, as far as I can see, no common genus for act-teleological and (so-called) rule-teleological theories that excludes deontological theories.19 (Often, to be sure, rule-teleological theories are classified as teleological. But there are notable exceptions. Thus Robert Olson classifies rule utilitarianism as a version of deontological ethics.20 And Richard Brandt views rule-teleological theories as “a group of compromise theories, which form a bridge” between deontological and teleological theories.21 But, according to Brandt’s own definition, they should be classified as deontological. For deontological (or “formalist” theories, as Brandt prefers to call them) are said to “have in common the view that one’s moral obligations are dependent on some other things about one’s situation in addition to (or perhaps, instead of) the intrinsic value of the actual or expectable results of one’s act”.22 And this characterization certainly covers rule-teleological theories.)

17  For proposals to extend the scope of moral theories to include such things, see, e.g., David Brink, op. cit.; John Rawls, op. cit., and Bernard Williams, “A Critique of Utilitarianism”, in J. J. C. Smart and Bernard Williams, Utilitarianism: For and Against (London: Cambridge University Press, 1973). 18  Marcia Baron, “Kantian Ethics”, in Marcia Baron, Philip Pettit, and Michael Slote, Three Methods of Ethics: A Debate (Oxford: Blackwell, 1997): 3–91, esp. 21–32. This multi-response view is somewhat in vogue at present. 19  I discuss rule consequentialism in Sect. 6.4. 20  Robert Olson, “Deontological Ethics”, in Paul Edwards (ed.), op. cit. 21  Richard Brandt, Ethical Theory: The Problems of Normative and Critical Ethics (Englewood Cliffs, N. J.: Prentice-Hall, 1959), p. 355. 22  Ibid., p. 354.

1.3  The Received View Questioned

9

1.3  The Received View Questioned The above, rather traditional, characterization of the difference between consequentialism and teleological ethics in general, on one hand, and deontological ethics, on the other, may be—indeed, has been—questioned in several ways. Thus one might point out that, just as something is right, etc. in virtue of something else, so something has value in virtue of something else. Ultimately, therefore, what makes an action right, etc. according to a teleological principle is not the value of its consequences but, just as in the case of a deontological principle, some empirical property (or properties). Hence, it might be argued, the above characterization of the distinction between teleological and deontological principles is amiss. I think, however, that the characterization is appropriate. According to a teleological principle, what ultimately makes an action right, etc. is not just that it possesses some empirical property but that the property is a value-making property. This is not true of deontological principles. A deontologist might, of course, say that the fact that an action is, for example, the keeping of a promise, makes the action good, and the goodness of the action, in turn, makes it right. But—and this is the crucial point—the deontologist does not have to make this detour via the value of the action (and, as far as I know, he almost never does). He is able to state what he means without using axiological terms, whereas the teleologist is not. Suppose, however, that the deontologist insists that he has to make the axiological detour, claiming, for example, that promise-keeping is right because it is good. Then I would say that he talks about moral goodness (whereas the teleologist talks about non-moral goodness); that is, the deontologist is claiming that the action is good in virtue of its deontic moral properties.23 For what else could make an action, abstracted from how it affects the world, good but its deontic moral properties?24 If, therefore, a deontologist claims that an action, such as the keeping of a promise, is good, he certainly means that it is good because of it deontic moral properties, in other words, because it is right. So what I made the deontologist say boils down to  I know that many moral philosophers bristle at the notion of moral goodness, thinking that “good” is here assumed to have a special “moral” meaning. I want to assure them that I do not asume that. Nor do I assume that what makes something morally good is always its having a certain deontic moral property. But it might be the case that this is what ultimately makes something morally good. Thus, what makes a person morally good is presumably his having certain morally good properties (i.e., virtues). But what, in turn, makes these properties morally good is presumably their making the person act rightly. Kant seems to be of the same opinion. “The concept of [moral] good and evil must not be determined before the moral law (of which it seems as if it must be the foundation), but only after it and by means of it.” (Immanuel Kant, Kritik der praktischen Vernunft; quoted from Kant’s Critique of Practical Reason and Other Works on the Theory of Ethics, tr. T. K. Abbott (London: Longmans, Green, 1873), p. 154. (For a good discussion of the distinction between moral and non-moral value, see William Frankena, Ethics (Englewwod Cliffs, N. J.: prentice-Hall, 2nd Ed., 1973), p. 62.) 24  Sometimes, of course, an action is considered good in virtue of its aesthetic properties. But, I take it, aesthetic properties are not what the deontologist would adduce as good-making in this context. 23

10

1  The Conflict

the statement that the action is right because it is right; its alleged goodness is just an epiphenomenon and does no real work. But maybe one could agree with what I have just said and still deny that there is any fundamental difference between teleological and deontological ethics. One way of doing this is to argue that moralities of one type are extensionally equivalent to moralities of the other type. Such a proposal has been made by Sven Danielsson with respect to consequentialism and deontological ethics. Danielson puts forward the following theses: It seems that every consequentialist morality can be transformed into an extensionally equivalent deontological morality simply by replacing statements about value with statements about their empirical grounds. It seems that every deontological morality can be transformed into an extensionally equivalent consequentialist morality by replacing statements about duty with suitable statements about value.25

I am inclined to agree with the first but not with the second of these theses. But to adequately deal with this intricate issue would require a lengthy digression.26 Fortunately, however, this is not necessary. In Sect. 2.6, I will characterize D in a way that is not vulnerable to the above objection, and in Sect. 4.1, I will do the same for C. I mentioned above utilitarianism as an example of a consequentialist theory. This is also the received view. But the received view has been questioned by Will Kymlicka.27 There are two possible interpretations of utilitarianism, Kymlicka says. One is unquestionably consequentialist, claiming that “maximizing the good is primary”.28 But Kymlicka thinks that this interpretation is not a genuine moral principle. For it is entirely unclear why maximizing utility, as our direct goal, should be considered a moral duty. Whom is it a duty to? Morality, in our everyday view at least, is a matter of interpersonal obligations—the obligations we owe to each other. But to whom do we owe the duty of maximizing utility?29

(I will discuss this objection in Sect. 5.1.) According to the other, the deontological interpretation, favoured by Kymlicka, utilitarianism is a procedure for aggregating individual interests and desires, a procedure for making social choices, specifying which trade-offs are acceptable. It is a moral theory because it purports to treat people as equals, with equal concern and respect. It does so by counting everyone for one, and no one for more than one.30

 Sven Danielsson, “Konsekvensetikens gränser” (The Limits of Consequentialism), in his Filosofiska utredningar (n.p.: Thales, 1988), p. 78 (my translation). 26  For a good discussion of the issue, see Erik Carlson, op. cit., pp. 40–43. 27  Will Kymlicka, “Rawls on Teleology and Deontology”, Philosophy & Public Affairs, 17 (1988): 173–90. 28  Ibid., p. 178. Thus Kymlicka restricts utilitarianism to comparative versions of it. 29  Ibid., p. 180 30  Ibid., p. 176. Thus Kymlicka restricts comparative versions of utilitarianism to desire-satisfaction versions. 25

1.3  The Received View Questioned

11

According to this version of utilitarianism, maximizing the good is not primary. “Maximization occurs, but as a by-product of a decision procedure that is intended to aggregate people’s preferences fairly.”31 Now, is this really a deontological version of utilitarianism, based on the principle that one should treat people with equal concern and respect? I do not think so. It is true, of course, that, like all other versions of utilitarianism, the present version counts everyone for one, and no one for more than one, in the sense of disregarding whose desire a desire is, taking the value of the satisfaction of a desire to depend only on its strength.32 But this is certainly not what is meant by treating people “with equal concern and respect”. To see this, consider the following example: You can act in either of two ways. If you act in one way, either person A’s desire will be satisfied or person B’s equally strong desire will be satisfied. If you act in the other way, A’s desire will be satisfied, but B’s will not.33

Counting both A and B for one and neither for more than one, you may act in either way, although acting in the second way will favour A and disfavour B, whereas acting in the first way will not favour any of them. Thus it is evident that acting in the second way will not treat B fairly; it is not a way of showing equal concern for B. So counting everyone for one, and no one for more than one, is not a way of treating people with equal concern and respect. The suggested interpretation of utilitarianism is therefore not a genuine deontological theory “intended to aggregate people’s preferences fairly”. And I cannot think of any interpretation of utilitarianism which is also a genuine deontological theory. In fact, the Benthamite principle of counting everyone for one, and no one for more than one, does not do any work at all in a (maximizing) utilitarian theory. This is convincingly argued by James Griffin: It is not a separate principle over and above the principle of utility. Mill’s account of its status, it seems to me, is right. It is, he says, simply part of what is involved in applying the principle of utility. If one’s aim is to maximize utility, whose utility it is is irrelevant. One is barred from being a respecter of persons by being a respecter of utilities alone.34

I have characterized C (and teleological theories in general) in terms of “consequences”, but, as Erik Carlson points out, it is better to state C in terms of “outcomes” instead of “consequences”, understanding the relevant notion of the outcome of an action counterfactually, that is, as “the possible world that would be actual if  Ibid., p. 177. Kymlicka says that “it is the concern with equal consideration that clearly underlies Bentham’s argument [in An Introduction to the Principles of Morals and Legislation] and is explicitly affirmed by recent utilitarians such as John Harsanyi and James Griffin” (ibid.). If my criticism of Kymlicka is correct, this is either false or misleading. 32  This is true of the version of desire-satisfaction utilitarianism discussed by Kymlicka. There are, however, other versions which in different ways restrict the range of desires worthy of being satisfied. 33  The example is a version of an example given by James Griffin in Well-Being: Its Meaning, Measurement, and Moral Importance (Oxford: Oxford University Press, 1986), p. 208. 34  Ibid., p. 168. Griffin here refers to Mill’s Utilitarianism, Ch. V, note to Par. 36. (Griffin (roughly) equates utility with desire-satisfaction.) 31

12

1  The Conflict

the action were performed”.35 The reason for talking about outcomes instead of consequences is that it better accords with what C is generally taken to claim. Consider, for example, the case of someone’s harming someone else. If the agent could abstain from harming the other person, and the consequences of this alternative was neither good nor bad, then (most versions of) C would say that the act of harming was wrong. But it cannot be said to be wrong because of its consequences being bad. For the total value of the consequences of this act of harming—what the act of harming caused—is, we assume, positive: the harm is not among the consequences of an act of harming as the concept of “harming” is normally understood. But the total outcome of this act of harming, we assume, has negative value, and the world where the action is performed is worse than the world that would be actual if it were not performed. So the act of harming is wrong. Thus, substituting “outcome” for “consequences” in the formulation of C makes its verdicts in such cases as the above one better correspond to what, intuitively, we expect the theory to say. Again, consider the case of someone’s not throwing a lifebuoy to a drowning man, although he could easily do that, thus saving the man. Many people would say that inaction cannot be the cause of anything; hence the death of the drowning man would not be a consequence of the other person’s not throwing the lifebuoy. Yet consequentialism, we may assume, considers this death normatively relevant—as something bad or, maybe, as something good—and it can express this conviction by talking about outcomes instead of consequences. In yet another kind of cases it is controversial, that, according to C, one should only consider the value of the consequences when assessing whether actions are right or wrong. These are the cases where A makes B do an action a by, for example, persuading him to do it. Consequentialists will certainly assess A’s act of persuading B as right or wrong depending, among other things, on what A persuaded B to do, viz. the outcome of a. But some people, the libertarians, subscribe to a conception of free will according to which A does not cause B to do a; hence the outcome of a is not a consequence of what A does. There are, then, two reasons why C, despite its name, is better formulated in terms of outcomes than in terms of consequences. One, illustrated by the first example, is that, according to almost any description of what an agent did, the action itself, not (only) its consequences, might have intrinsic value. To avoid this and only deal with evaluatively neutral actions one could, it might be suggested, have recourse to (what is sometimes called) basic actions. A basic action is something which one simply does, not something which one does by doing something else; basic actions are usually identified with (voluntary) bodily movements.36 The problem with this suggestion is that even such actions might have intrinsic value: the scratching of my finger may be the satisfaction of a preference and thus have value according to the

 Erik Carlson, op. cit., p. 10.  See Arthur Danto, “Basic Actions”, American Philosophical Quarterly, 2 (1965): 141–8, where the term was introduced.

35 36

1.4  Other Moral Theories

13

version of consequentialism that takes preference satisfaction to constitute what is intrinsically good.37 The other reason, illustrated by the two other cases, is that “consequence” is a “causal term”: a is a consequence of b if, and only if, b is a cause of a. But, as we have seen, consequentialism considers also certain events which are not causally related to an action as relevant for the normative status of the action. It is, of course possible to redefine “consequence” in such a way that it has the same extension as “outcome”. But I think it is simpler and less confusing to use the latter term instead.

1.4  Other Moral Theories It might be objected that the conflict between C and D is not the fundamental conflict in normative ethics. These theories are not the only plausible moral positions; there are other moral theories which are at least as plausible as these. One of them, it might be suggested, is the version of moral particularism according to which there are no genuine moral principles or rules: what is a morally relevant feature in one situation may be quite irrelevant in another situation, and there is no (detectable) way of specifying when a feature is relevant and when it is not.38 But particularism does not fall outside the opposition between C and D. It is usually considered a form of deontological ethics, albeit not of rule- but of act-­ deontological ethics.39 The same is also true of much existentialist ethics and of one version of so-called situation ethics, according to which what should be done in a situation is determined by God’s immediate command.40 (The view that moral theories such as these are not deontological is possibly due to the widespread use of “rule ethics” as a synonym for “deontological ethics”.) According to another version of situation ethics, set forth by Joseph Fletcher, what ought to be done in a situation is determined by what features in the situation

 This is so according to one version of the preference satisfaction theory of intrinsic value. According to another version, what has intrinsic value is not the object the occurrence of which satisfies a preference, but the state of affairs that a preference is satisfied. For a discussion of which version is the most plausible one, see Wlodek Rabinowicz and Jan Österberg, “Value Based on Preferences: On Two Interpretations of Preference Utilitarianism”, Economics and Philosophy, 12 (1996): 1–27. 38  There are more or less radical versions of particularism. A rather radical version is defended by Jonathan Dancy in Moral Reasons (Oxford: Blackwell, 1993). 39  But the received view of particularism might be mistaken. It has, quite convincingly, been argued that “[t]he consequentialist could perfectly well be a particularist about value. Hence, given a particular [sic] axiological view, there need be no conflict between particularism and consequentialism”. (Jonas Olson and Frans Svenson, “A Particular Consequentialism”, Utilitas, 15 (2003): 194–205; p. 205. (Particularism will be further discussed in Sect. 3.2.) 40  This kind of situation ethics is propounded by, e.g., Reinhold Niebuhr in The Nature and Destiny of Man (New York: Charles Scribner’s Sons, 1949). 37

14

1  The Conflict

the injunction “to love your neighbour as yourself” directs attention to.41 This theory, Fletcher thinks, is extensionally equivalent to act-utilitarianism. The same interpretation has been given, quite generally, of agapism, the Christian ethics of love.42 But some theologians reject this interpretation claiming that agapism is neither a deontological nor a teleological theory.43 Since, however, it is never spelled out what the principle of agapism, thus understood, states, this cannot be said to be a plausible moral position. Some moral theories are stated in terms of rights, and, depending on how rights are conceived and what rights we are taken to have, there are several kinds of right-­ based moralities. But to every right, in the sense of a claim that a person has, there is a duty for other people to do or omit certain kinds of action.44 So right-based moralities are deontological, or, to take a more cautious line, to every right-based morality there is a deontological morality that is extensionally equivalent to the former. So the assumption of rights does not engender another fundamental conflict in ethics. A moral position that, although it is a teleological theory, falls outside the opposition between consequentialism (as here understood) and deontological ethics is ethical egoism. There are several forms of ethical egoism. Strong act egoism, the most interesting one, states that an action is right if, and only if, there is no alternative action open to the agent which has (would have) a better outcome for him. Is any form of ethical egoism a plausible alternative to C or D? The answer is No, but I am not going to justify this negative verdict here, my excuse being that I have done so at length elsewhere.45 I shall confine myself to rendering an argument against strong act egoism given by James Rachels. Rachels reports what he says actually happened in a small town in South Georgia some years ago. A poor, black woman visited the town’s only doctor, who discovered that the woman suffered from malnutrition. The problem was that the woman did not have enough money to buy food for herself and her several small children. She worked, whenever she could, as a cleaning-woman in the homes of the better-off people in town, but she was able to earn only a few dollars in this way. All this was known to the doctor. After spending no more than five minutes with the

 See Joseph Fletcher, Situation Ethics (Philadelphia, PA: The Westminster Press, 1966). Fletcher is not very clear, so it might be the case that he intends the injunction “to love your neighbour as thyself” to furnish, not the ratio essendi, but the ratio cognoscendi of right action. 42  See, e.g., J. S. Mill, Utilitarianism, ed. Roger Crisp (Oxford: Oxford University Press, 1998), Ch. 2, Par. 18, p.  18; and A.  C. Garnett, Ethics: A Critical Introduction (New York: Ronald Press, 1960), passim. As for Mill’s view, see also Sect. 4.4 above. 43  I am indebted here to William Frankena, op. cit., p. 56 ff. (It should be added that most versions of Christian ethics are deontological theories resting on “We should obey God” or some similar principle. 44  Exceptions to this are (what may be called) “political” rights, e.g., the one stated in Article 23 in United Nation’s Declaration of Human Rights: “Everyone has the right to work, to free choice of employment, to just and favourable conditions of work and to protection against unemployment.” 45  In Jan Österberg, Self and Others:A Study of Ethical Egoism (Dordrecht: Kluwer, 1988). 41

1.4  Other Moral Theories

15

woman, and having done absolutely nothing for her, the doctor told her that the charge would be twenty-five dollars. The woman only had twelve dollars—this was, literally, all the money she had in the world—so the doctor took that.46

The argument that Rachels then gives is of the modus tollens form: (i) If ethical egoism is correct, then the doctor did the right thing. (ii) The doctor did not do the right thing. (iii) Therefore ethical egoism is not correct. To rebut the argument you have to rebut one or both of the premises. The first premise, Rachels assures us, is— regrettably—true; the 12 dollars improved the doctor’s financial position to some extent, and this was the only noticeable effect that the actions had on his interests. “There were no harmful reactions toward him from the black community, or from any other quarter, and none were expected. The doctor himself had no guilty conscience, for he never gave it another thought.”47 As for the second premise I think it is evident that the doctor did not do the right thing. Since, furthermore, the two premises together entail the conclusion, and the argument does not beg the question at issue, I have a good reason to reject this strong version of act egoism. And so have all people—hopefully the vast majority— who accept the second premise. At least strong act egoism is not, in the considered opinion of most people, a serious competitor to consequentialist or deontological ethics. What about self-actualization, or self-realization, theories? The starting point of such theories is a certain conception of what human beings are, and if this conception is selected with some care, the theory need not have wildly implausible implications. But, it seems, self-actualization theories are best classified as consequentialist theories: the paramount intrinsic value is the actualization of selves; this is the value we should all promote. (The concern with one’s own self, stressed by self-­ actualization theories, is to be explained by the contingent fact that we can only marginally influence the actualization of selves other than our own.) Alternatively, self-actualization theories could be classified as forms of ethical egoism. In any case, it seems, they are to be classified as teleological theories. But this is something that is denied by Allen Wood, at least as far as Hegel’s version of “self-­ actualizationism” is concerned: “Hegel’s self-actualization theory represents a distinctive type of ethical theory, different from both deontological and teleological theories.”48 The reason Wood gives for this claim is that, according to Hegelian theory, the ground of all value and normative force is a certain ideal of the self. The starting point of Hegelian ethics, he says, is the conception of a certain self or identity to be exercised or actualized, to be embodied and expressed in action. The theory selects the actions to be performed and the ends to be  James Rachels, “Two Arguments against Ethical Egoism”, Philosophia, 4 (1974): 297–314; p. 308 f. 47  Ibid., p. 309. 48  Allen W. Wood, Hegel’s Ethical Thought (Cambridge: Cambridge University Press, 1990), p. 31. (This version of self-actualizationism resembles a version of virtue ethics (to be discussed below) and therefore falls victim to the same kind of objection.) 46

16

1  The Conflict pursued because they are the actions and ends of that kind of self. In such a theory, laws and commandments owe their force to the fact that they turn out to be principles which the right sort of self would follow. Ends owe their desirability to the fact that they turn out to be the ends which that sort of self would pursue.49

This theory is not plausible, since it gives the wrong kind of ground for actions being right and ends being good. Surely the ground why it is pro tanto wrong for A to torture B is some fact about B, viz. that he experiences agonizing pain, not some fact about A. To paraphrase the lesson of Plato’s Euthyphro: Given that an ideal self would follow the right principles and pursue the good ends, principles are not right nor are ends good because the self would follow the principles and pursue the ends; rather the self is ideal because it would follow the right principles and pursue the good ends. Although the ideal self may provide the ratio cognoscendi of right and wrong, good and bad, it does not provide its ratio essendi. An ethical theory that several philosophers during the last few decades have heralded as a major competitor to C and D is virtue ethics. (I will therefore discuss this theory at some length.) The revival of virtue ethics is often thought to begin with Elizabeth Anscombe’s paper “Modern Moral Philosophy” in 1958. Here Anscombe argues that the notion of moral obligation or moral ought, common to both teleological and deontological ethics, is a quasi-juridical notion presupposing God as law-giver, which does not make sense outside a divine law conception of morality. For secular moral philosophy, the only viable alternative is to return to and build upon the virtue ethics of Plato and Aristotle. And this was what several moral philosophers, for instance, Philippa Foot and Alasdair MacIntyre, began to do. I do not want to disparage the importance of virtues for morality. Virtues, that is, good traits of character, are certainly instrumentally good in motivating us to do what we (think we) ought to do. Probably they also have an “epistemic” value. For it is likely that the virtuous person is more perceptive, as well as more sensitive to the strengths of competing moral claims, than is the non-virtuous person. He may therefore be better at comparing the strengths of conflicting moral claims and at “perceiving” their resultant. But, in order to be a genuine competitor to deontological and teleological (consequentialist) ethics, virtue ethics has to provide “virtue-­ related” criteria of right and wrong.50 Some virtue ethicists think that this can be done. According to one proposal, whether an action is right or wrong is determined by what the virtuous person would do. If a virtuous person who is familiar with the situation would act in a certain way, this makes the action right (wrong).51 It is a  Ibid.  It has been objected that particularist virtue ethicists, such as John McDowell, will reject the proposal that they have to provide criteria of right and wrong: since a particularist denies the existence of moral principles, he cannot, it is said acknowledge the existence of such criteria. But I think he can: the criteria proposed by virtue ethicists (see below) are as available to the particularist as they are to the non-particularist. A particularist has no special reason to repudiate, e.g., the proposal, “An action is right if, and only if, it is what a virtuous person would do”, even if he claims that the virtuous person himself does not pick out the right action by means of a general criterion. 51  It is sometimes claimed that virtue ethics can dispense with deontic notions altogether, using epithets such as “just” and “cowardly” instead of “right” and “wrong”. I agree. But this is possible 49 50

1.4  Other Moral Theories

17

contentious issue whether Aristotle accepted this claim, but some modern virtue ethicists do. Thus, for example, Rosalind Hursthouse says that, just as the fundamental principle of act utilitarianism is, “An action is right iff it promotes the best consequences”, the fundamental principle of virtue ethics runs, “An action is right iff it is what a virtuous agent would characteristically (i.e. acting in character) do in the circumstances”.52 There is a problem with attempts that, like this one, try to spell out the criterion of rightness in terms of behaviour. Suppose that in a situation there are, according to virtuous people, several morally permissible actions. Then there is not one action that a virtuous agent would characteristically do. Still, I take it, the virtue ethicist would want to claim that any of these actions is right. I will therefore henceforth state the claim in terms of what the virtuous person would consider right. (If there is more than one action in the situation that virtuous people find permissible, then they would consider it right to do any of these.)53 There is another reason for substituting “what a virtuous agent would consider right” for “what a virtuous agent would do”, viz. that there are (kinds of) actions that a virtuous agent would never do. Suppose, for example, that Jones has wronged Smith. A virtuous person might consider it right for Jones to apologize to Smith for that, although this is not a kind of action that he himself would ever do—for the simple reason that a virtuous agent would never do anything that he had to apologize for.54 It is, of course, conceivable that a person who is not virtuous could regard the verdict of the virtuous person as the criterion of right action. But could the virtuous person himself do that? I do not think so. When confronted with a choice situation the virtuous agent, like anybody else, has to look for morally relevant features in the situation in order to come up with a decision about what to do. (What else could he do if he is a responsible agent?) When he comes up with a verdict to the effect that in this situation that action is to be performed, his verdict is based on (what he takes to be) the morally relevant features in the situation. So the virtuous person himself cannot consistently believe that it is his verdicts that makes actions right or wrong.55 only because the former notions too are action-guiding, doing the same job as the latter. I therefore see no reason why, when applied to actions, they should not be taken as (thick) deontic notions entailing the thin notions of “right” and “wrong”. 52  Rosalind Hursthouse, On Virtue Ethics (Oxford: Oxford University Press, 1999), p. 28. 53  There is nothing circular in letting what is right depend on what a virtuous agent would consider right. And, as far as I can see, Hursthouse’s proposal, too, ultimately relies on what a virtuous agent considers right. For, I take it, a virtuous agent would not do anything unless he considered it right; the virtuous agent is not an automaton. 54  There are other cases where, due to actual people being less than fully virtuous, an account in terms of what a virtuous person would consider right works, whereas one in terms of what he would do does not; for an account of these cases, see Frans Svensson, Some Basic Issues in NeoAristotelian Virtue Ethics (Uppsala: n. p., 2006), p. 48 ff. 55  This right-making claim has analogues, not only in the case of self-actualizationism (see above), but also in the philosophy of law and in aesthetics. Oliver Holmes’s prediction theory of law says that “[t]he prophecies of what the courts will do in fact, and nothing more pretentious, are what I

18

1  The Conflict

And, although (as I said) someone who is not virtuous might conceivably regard the verdict of the virtuous person as the criterion of right and wrong action, we in fact do not—and I see no reason why we should. It is evident that what makes, say, an act of torture (at least prima facie) wrong is the victim’s suffering, not what anybody else thinks or does. To (once more) paraphrase the lesson of Plato’s Euthyphro: Given that virtuous people know and do what is right, an action is not right because they know and do it; rather they know and do it because (they think) it is right. A third argument against the view that the verdict of the virtuous person is what makes a right action right starts from the epistemic priority of right action over good traits of character. In order to determine what virtues people should have, we must take cognizance of what a person with such and such a trait of character would characteristically think or do. This is not only a normative claim, saying that it would be preposterous to do otherwise. It is also a conceptual claim: we would not, I believe, call a trait of character a virtue if, in our opinion, it did not more often than not give rise to actions that we thought laudable and right. Our conception of a virtuous person thus presupposes that we have an independent grasp of what a right action amounts to. The former cannot therefore be the criterion of the latter.56 Similar objections can be raised against the version of the right-making claim set forth by Plato in the Republic. In Book Four of this dialogue Plato makes Socrates state the view that just (meaning, roughly, right) actions are those that produce and maintain the justice of the soul, that is, the harmonious functioning of its three “parts”. Justice, Socrates says, is a principle of this kind; its real concern is not with external actions, but with a man’s inward self, his true concern and interest. The just man […] will be ready for action of any kind, whether it concerns his personal or financial welfare, whether it is political or private; and he will reckon and call any of these actions just and honourable if it contributes to and helps to maintain this disposition of mind. […] Similarly, he will call unjust any action destructive of this disposition […].57 mean by the law” (Oliver Wendell Holmes, Jr., “The Path of the Law”, Harvard Law Review, 10 (1896): 457–78, p. 458; quoted from H. L. A. Hart, The Concept of Law (Oxford: Oxford University Press, 1961), p. 8). And the institutional theory of art, in somewhat different versions propounded by Arthur Danto and George Dickie, states that “works of art are art as the result of the position or place they occupy within an established practice, namely the art world” (George Dickie, “The New Institutional Theory of Art”, in Rudolf Haller (ed.), Proceedings of the Eighth International Wittgenstein Symposium, Part I (Vienna: Holder-Pichler-Temsky, 1984): 57–64; quoted from the reprint in George Dickie et  al. (eds.), Aesthetics: A Critical Anthology, 2nd ed. (New York: St. Martin’s Press, 1977): 196–205; p. 196. My criticism of the right-making claim applies, mutatis mutandis, to these theories too. 56  These objections also apply to other virtue-ethical accounts of what makes an action right. It applies to Michael Slote’s “agent-based” account, set forth in Moral from Motives (Oxford: Oxford University Press, 2001); to Julia Annas’s “developmental” account in “Being Virtuous and Doing the Right Thing”, Proceedings and Addresses of the American Philosophical Association, 78 (2004): 61–75; and to Valerie Tiberius’s “reason-based” account, according to which “[a]n action A is right for S in circumstances C iff it is the action in accordance with the reasons that would guide the action of a completely virtuous person acting in C” (“How to Think About Virtue and Right”, Philosophical Papers, 35 (2006): 247–65; p. 248.) 57  The Republic 443e; quoted from Desmond Lee’s translation (Harmondsworth: Penguin, 1987), p. 161 f.

1.4  Other Moral Theories

19

Let us call “P-just” those actions that the Platonically just man would consider just (as well as perform) and “O-just” those actions which we would ordinarily call just. It is very important for a central claim of the Republic—viz. that the just man, as ordinarily conceived, is more happy than the unjust man—that P-just and O-just actions coincide; and this means that (all and only) O-just actions contribute to and help to maintain the just disposition of the agent. But even if we grant this controversial claim, we should not grant, as Socrates and Plato seem to do, that it is because of this that they are just. Even if, say, helping a person in need strengthens my own just disposition, this is certainly not what makes it just. My objections to virtue ethics have been questioned by Frans Svensson.58 According to Svensson, defenders of (non-Platonic) virtue ethics can say in response that we are actually just dealing with two different ways of referring to the same property here. Hence, to say that a certain action is what a virtuous agent characteristically would do in the circumstances may in fact be taken as a right-making claim, though we could also refer to what makes the relevant action right in terms of, for instance, that it would constitute the fulfilment of a promise, or to help someone in need.59

I think that this is wrong: the two properties are not identical. For ease of exposition, let “A” be short for “the property of being what a virtuous agent would characteristically do (or, according to my suggestion: consider right) in the circumstances”, and let “B” denote the set of right-making properties as normally characterized (such as being the fulfilment of a promise or helping someone in need), at least one member of which is instantiated by any action that instantiates A.60 The suggested proposal then says: (S) The property A is the same property as the property B. Now, what started the present discussion was the virtue ethicist’s claim that A is a right-making property. Whether this claim is acceptable or not depends on how the virtuous agent is characterized. As far as I can see, for the claim to be acceptable the virtuous agent must be assumed to be, among other things, an ideal (moral) observer in the following sense: he must be perfectly cognizant of, and responsive to, every morally relevant fact in the situation. If the virtuous agent was not perfectly cognizant of the morally relevant facts in a situation, he might come up with the wrong answer concerning what was right to do in the situation. So only by being an ideal observer is it possible for the virtuous agent to always know what actions are right. (This means, of course, that it is very unlikely that any actually existing person could measure up to this ideal.)  See Frans Svensson, op. cit. My objections first appeared in “The Virtues of Virtue Ethics”, in Rysiek Sliwinski (ed.), Philosophical crumbs (Uppsala: Uppsala Philosophical Studies, 1999), and it is this text that is the target of Svensson’s criticism. 59  Svensson, op. cit., p. 34. What Svensson explicitly criticizes is my first argument against nonPlatonic virtue ethics, but his criticism is relevant to the second argument too. 60  B is as large, complex, and fine-grained as your favourite moral theory requires. (You might even be a moral particularist.) The members of B might be referred to by statements of the form: in such and such circumstances, property P is a (pro tanto) right-making property. 58

20

1  The Conflict

But the morally relevant facts that the virtuous agent is cognizant of, and responsive to, are just the properties that belong to B. Thus A is formulated in terms of the virtuous agent, and the virtuous agent, in turn, has to be (partly) defined in terms of B. This means that A is ultimately (partly) defined in terms of B, whereas, obviously, B is not defined in terms of A. So (S) is false: although A and B are necessarily coextensive, A is not the same property as (any of the properties belonging to) B. For, I take it, if a property P is defined in terms of a property Q, but not vice versa, then P is not the same property as Q.61 (Thus it differs from genuine identities, such as water  =  H2O: neither “water” nor H2O is (even partly) definable in terms of the other.) If it were the case that “A” and “B” are, as it is claimed, “two ways of dealing with the same property”, this would indeed have far-reaching consequences. Take any deontological or consequentialist theory T, and let “B” denote the set of right-­ making properties according to T. Svenssons’s proposal then says that we can also refer to B by means of “what a virtuous agent would characteristically do (or, according to my proposal: consider right) in the circumstances”, that is, by means of “A”. But, we have been told, A gives the distinctive virtue-ethical account of right-making properties. This means that T counts as a virtue-ethical theory. Thus, according to the present proposal, every deontological and consequentialist theory is in fact a virtue-ethical theory, though, of course, what the theories take as virtues differ. The commonly assumed distinction between virtue ethics and other moral theories hence disappears. What is left is a distinction between those virtue-ethical theories that are also deontological and those that are also consequentialist. I think that this consequence comes close to a reductio ad absurdum of the thesis that A = B. According to my third objection, our conception of right action is prior to our conception of a virtuous agent: we would not consider a trait of character a virtue if we did not think that it, at least more often than not, gave rise to right actions. To this Svensson replies that this does not make the present right-making account viciously circular. For, he says, (i) we commonly take the virtues to include, for instance, courage, generosity, justice, and temperance. And (ii) “it seems as if we actually can say a few things about what the possession and practice of these virtues involve, without immediately referring to right action”.62 These remarks are, no doubt, true. But, as far as I can see, this does not show that courage, generosity, etc. would have been promoted to the rank of virtues if we did not think that they usually gave rise to (independently characterized) right actions. The best way to defend the virtue ethicist’s right-making claim that I can think of would be, starting from the epistemic claim, to argue along the following lines: In  This asymmetry explains why it sounds all right to say that a virtuous agent, acting in character, would do an action because it is, say, the fulfilment of a promise, whereas it sounds odd to say that the action is the fulfilment of a promise because the virtuous person, acting in character, would do it. 62  Svensson, op. cit., p. 36. 61

1.4  Other Moral Theories

21

morally problematic situations we have to take into consideration reasons that pull in different directions; for each alternative action, there are reasons both for and against its performance. The reasons, moreover, often seem to be incomparable with each other. To decide what is, on balance, best to do is therefore a matter of judgement. The virtuous person is, ex hypothesi, the person with the best judgement; because of his moral upbringing and his experience of similar situations he is maximally sensitive to what is morally relevant in the situation. So, if the virtuous person thinks that one of the available actions in the situation is right, this is the best possible evidence for the action’s being right that we can get. Now, in order for an action to be fully right, the defence continues, the action must be both objectively and subjectively right.63 Thus, it is not enough that the action in question is a best response to the morally relevant features of the situation; the agent must also believe that it is the best response, and he must, furthermore, have good reasons for believing that.64 In morally problematic situations only the action truly believed to be right by the agent with the best judgement, that is, by the virtuous person, satisfies all three conditions. The agent’s virtue is therefore necessary for the fulfilment of the third condition and, thus, for the action’s being fully right.65 This defence of the right-making claim is, however, of rather limited interest. Firstly, it is elitist: only virtuous people are able to act rightly. Perhaps also some non-virtuous people who have an opportunity to consult a virtuous person before they act are able to act rightly. But how many of us has this opportunity? Moreover, the claim that a fully right action has to fulfil all three conditions is, at most, true when “action” is taken in a wide sense, not in the narrow sense used by deontologists and consequentialists. In the narrow sense, the agent’s state of mind is not part of the action; so for an action to be right only the first condition has to be satisfied. Only if it could be shown that the agent’s virtue makes an action in the narrow sense right would the right-making claim of virtue ethics be substantiated in any interesting way, that is, in a way that would make it a competitor to deontological and consequentialists ethics. One may speculate why virtue ethicists so eagerly try to mould their theory into a theory of right action. Why are they not satisfied with working out a theory of the  Perhaps this is what Aristotle claims in the Nicomachean Ethics, Bk. 2.  My use of “subjectively right” here slightly deviates from certain other uses of it. According to a rather common conception, an action is subjectively right if, and only if, it would have been objectively right if the facts of the case had been as the agent thought they were. See, e.g., Richard Brandt, op. cit., p. 365. (Brandt’s proposal is, however, stated in terms of “duty”, not of “right”.) 65  This view is somewhat akin to the view that, in order to act fully rightly, the agent’s motive must be of the right kind. An example is given by Derek Parfit. Parfit considers a man who saves a child from drowning because he hopes to get a reward. Parfit comments: “It might be claimed that, when this man saves this drowning child, what he is doing is not wrong, but his doing of it is. Kant suggests a similar distinction when he claims that, to fulfil some duties of virtue, we must not only act rightly, but also with the right motive.” (On What Matters (Draft of 28 April 2008, Derek Parfit’s Homepage), p. 227.) 63 64

22

1  The Conflict

good person? Perhaps because in their opinion a theory of right action is more important or more prestigious than a theory of good character. In my opinion this is wrong; each kind of theory is an indispensable part of morality. I am wholly at one with James Rachels that theories of ethics that emphasize only right action will never provide a completely satisfactory account of the moral life. For that, we need a theory that emphasizes personal qualities such as friendship, love, and loyalty—in other words, a theory of the virtues.66

This ends my brief and rather dogmatic survey of moral theories that compete, or seem to compete, with deontological and consequentialist theories. (An adequate treatment of the theories would have made the book at least twice as long.) Some of these theories are (not very interesting) versions of either C or D. But three of the theories mentioned—viz. the Hegelian version of self-actualizationism, ethical egoism, and virtue ethics—are neither. Since I claim that C and D are the most plausible moral theories, I took the opportunity to criticize these rival theories. Needless to say, in neither case will my criticism convince all supporters of the theory in question. (There is no short way with dissenters.) Besides, my survey is possibly not complete: there may be theories not mentioned here that some find plausible. This means that what I will say in the rest of this book is primarily addressed to those who find (some version of) either C or D the most plausible moral theory. Assuming that some version of these theories is the most plausible (correct, true) moral theory, my purpose is to find out which one it is. But before I end this introductory chapter I will add a few words about what conceptions of morality are relevant to and/or presupposed in the following chapters.

1.5  Conceptions of Morality There are several different conceptions of morality.67 Some of them can be combined, whereas others are opposed to each other. Most of the conceptions are of no interest in the present essay: what I say here neither presupposes nor excludes them. But four conceptions are of some interest. One of them, I will claim, is presupposed by C, and another by D; a third is presupposed by me, and its rival, the fourth conception, is therefore rejected. I will therefore devote a section to briefly introducing them. One way of conceiving of morality starts from the individual’s concern: What kind of life is it rational to lead? What is it rational to do? Depending on different conceptions of rationality, some of these answers consider the agent’s own benefit to be of paramount importance, whereas, according to others, his treatment of other  James Rachels, The Elements of Moral Philosophy, 3rd Ed. (Boston: McGraw-Hill, 1999), p. 189. 67  A good survey is provided by the articles in G. Wallace and A. D. M. Walker (eds.), The Definition of Morality (London: Methuen, 1970). 66

1.5  Conceptions of Morality

23

people takes pride of place. To the former belong, for example, the ethical theories of most ancient philosophers, as well as most modern versions of ethical egoism.68 To the latter belong, for example, such diverse theories as those of Kant, Richard Hare, and David Gauthier.69 I shall call this conception of morality the rationalistic conception. According to its main alternative, (what may be called) the social conception of morality, morality takes its point of departure from society’s, not the individual’s, concern; morality is seen as the cement of the society, dealing with, in Hobbes’s words, “those qualities of mankind that concern their living together in Peace and Unity”.70 Other spokesmen for the social conception include most social contract theorists, as well as, for example, Bernard Gert and P. F. Strawson.71 Notice that, like the rationalistic conception, the social conception is not restricted to positive morality; proposed moralities, no less than actual ones, may concern themselves with the question how people should live together in peace and unity.72 The opposition between these competing conceptions of morality is the topic of a thorough and penetrating discussion by W.  D. Falk.73 Ordinary language, Falk thinks, backs both conceptions, but the social one more than the rationalistic one. Our concept of morality vacillates between the Greek and the Christian traditions. We associate ‘moral’ with ‘social’ commitment, and the ‘morally good man’ with the ‘selfless man’. But we also speak of man as a ‘moral agent’, of his ‘moral freedom’ and ‘moral powers’; and here we refer to his whole capacity of self-direction by good and sufficient reasons. […] There are contemporary moralists who call ‘moral’ any ‘authentic’ commitment of a self-governing person, whether its grounds are social or personal. What justifies them is the broader use of the term which is also part of our language and tradition.74

 For an example of the first kind, se Aristotle, the Nicomachean Ethics; for an example of the second kind, see Jesse Kalin, “In Defense of Egoism”, in David Gauthier (ed.), Morality and Rational Self-Interest (Englewood Cliffs, N. J.: Prentice-Hall, 1970): 64–87. 69  See, e.g., Immanuel Kant, Grundlegung zur Metaphysik der Sitten; Richard Hare, Freedom and Reason (Oxford: Oxford University Press, 1963); David Gauthier, Morals by Agreement (Oxford: Oxford University Press, 1986). 70  Thomas Hobbes, Leviathan, Ch. XI, Par. 1; in C. B. MacPherson’s ed. (Harmondsworth: Penguin, 1968), p. 160. 71  See, e.g., T. M. Scanlon, What We Owe to Each Other (Cambridge, MA and London, England: Harvard University Press, 1998; Bernard Gert, Morality: Its Nature and Justification (New York: Oxford University Press, 2005 (1988); P. F. Strawson, “Social Morality and Individual Ideal”, in Wallace and Walker (eds.), op. cit.: 98–118. 72  The distinction between the rationalistic and the social conceptions of morality is quite similar to Hegel’s distinction between Moralität and (the higher form of) Sittlichkeit. See G. W. F. Hegel’, Elements of the Philosophy of Right, tr. H. B. Nisbet, ed. Allen W. Wood (Cambridge: Cambridge University Press, 1991), Pts. two and three. (This is a translation of Hegel’s Grundlinien der Philosophie des Rechts.) 73  W. D. Falk, “Morality, Self, and Others”, in Ought, Reasons, and Morality: The Collected Papers of W. D. Falk (Ithaca: Cornell University Press, 1986): 198–231. (Falk’s terms for the conceptions are different from mine.) 74  Ibid., p. 200 f. 68

24

1  The Conflict

I think that the rationalistic conception underlies (and supports) C, whereas the social conception underlies (and supports) D.75 Given the universalistic conception of rationality accepted by C (see Sect. 1.1 above), it follows that “good and sufficient reasons” prescribe that one acts so as to maximize the impersonally good. C is thus only indirectly social: one should act rationally, and rationality requires that one considers the interests of other people. D, on the other hand, is directly social: one should consider the interests of other people, period. And this is certainly not astonishing if, as I will argue in Chap. 2, D has its source in common-sense morality, and CSM is the paradigmatic expression of the social conception of morality. The connections between, on one hand, C and the rationalistic conception and, on the other, D and the social conception may explain the formal (structural) differences between the two kinds of moralities (see Sect. 1.2 above). Given that C is meant to give an answer to the question, What is it rational to do?, it is small wonder that the theory is formulated in terms of value and not in terms of the underlying value-making properties. For, according to the standard conception of practical rationality, rational action is value-maximizing action. And, given that deontological ethics is supposed to answer the question, What should we do in order to live together in peace and unity?, it is to be expected that the answer is given, not in terms of lofty talk of maximization of value, but in terms of concrete injunctions: Do not lie! Do not kill! Keep your promises! (and these differences between the two kinds of moralities would exist even if Danielsson’s equivalence thesis, mentioned in Sect. 1.2, were true.) It might even be the case that some form of C is the best theory given one conception of morality, and some form of D is the best given the other conception. The crucial question then becomes, Which conception is the superior one? Perhaps the answer is that neither conception can be shown to be superior to the other, and that we are therefore faced with a stalemate. Other possibilities are that, although neither conception is superior to the other, either is valid within certain confines, or that they can be lexically ordered. These are question that I will return to in the following, especially in Chap. 7. Now for a different issue. It is sometimes discussed whether the “true” morality, the one we should all conform to, is something we discover or something we invent; is the moral law more like a natural law or a law in a legal system. I think that different bones of contention lurk behind the opposition characterized in these terms, but the most interesting interpretation of the opposition is perhaps the following: According to the discovery conception of morality, the true morality, if it has to be justified at all, is justified exclusively by means of non-normative facts, whereas, according to the invention conception, it must be justified, at least partly, by normative facts. Adherents of the discovery conception try to justify their favourite morality by claiming, for example, that its requirements are “objective”, that is, wholly determined by the nature of things, or that they are commanded by God. The

 There is also an intimate connection between these conceptions of morality and the two kinds of moral codes to be discussed in Sect. 2.3.

75

1.6  A Short Preview

25

e­ ighteenth century philosopher Samuel Clarke wants to have it both of these ways. There are, he says, eternal and necessary relations between things, and these eternal and necessary differences of things make it fit and reasonable for creatures so to act; they cause it to be their duty, or lay an obligation upon them, so to do; even separate from the consideration of these rules being the positive will or command of God76;

Justifications advanced by adherents of the invention conception claim, for example, that general adherence to the proposed morality offers the best life for all, or fulfils certain requirements of fairness, or is uniquely rational. Thus Kurt Baier offers the following Hobbesian justification of (defensible) moralities: Moralities are systems of principles whose acceptance by everyone as overriding the dictates of self-interest is in the interest of everyone alike, though following the rules of a morality is not of course identical with following self-interest.77

I opt for the invention conception. To my mind, the discovery conception does not make sense. Just as most philosophers accept Hume’s Law, epitomized in the slogan, “You cannot derive an ‘ought’ from an ‘is’”, so, I think, one should accept the related slogan, “You cannot (completely) justify an ‘ought’ by an ‘is’”. The reason is in both cases the same: the invention conception expresses the conviction that morality (cum rationality) is autonomous, that it is not wholly dependent on what is the case. This conviction is, I think, part and parcel of our concept of morality.

1.6  A Short Preview The rest of the book consists of six chapters and an appendix. The first two chapters, Chaps. 2 and 3, are devoted to D, especially to what I take to be its source, viz. CSM. In Chap. 2, I characterize this morality, and in Chap. 3, I discuss actual and possible objections to it. In Chaps. 4 and 5, I do the same with respect to C. My discussion in Chaps. 2, 3, 4 and 5 will bring the two theories somewhat closer to each other. Is it then possible to wholly solve the conflict, either by defensibly rejecting one of the theories, or by somehow reconciling them? In Chap. 6, I discuss both actual and possible attempts to this end; the result is negative. In Chap. 7, finally, I set forth my own solution of the conflict, a contractual theory that tries to do justice to both theories. The appendix defends a criticism of C made by D. H. Hodgson, against those who have criticized him. Since my solution of the conflict presupposes that Hodgson is right, it is important for me that my defence is successful.

 Samuel Clarke, “A Discourse Concerning the Unchangeable Obligations of Natural Religion, and the Truth and Certainty of the Christian Revelation”, 2nd series of the Boyle lectures, delivered at St. Paul’s in 1705; quoted from J. L. Mackie, Hume’s Moral Theory (London: Routledge and Kegan Paul, 1987 (1980)), p. 15. 77  Kurt Baier, The Moral Point of View (Ithaca, New York: Cornell University Press, 1958), p. 314. 76

Chapter 2

Deontological Ethics: Exposition

The topic of this and the following chapter is deontological ethics (D). This chapter is concerned with characterizing it, and the next with assessing it. More specifically, in this chapter I argue that the different deontological theories take their point of departure from common-sense morality (CSM) and are as many attempts to elaborate, systematize, and justify that morality. As the characteristic features of deontological theories are inherited from CSM, I focus on the nature of the latter: what are its distinguishing features, and how are they to be explained? In the first section, I give a general characterization of CSM and try to identify its core, the basic norms common to its different variants. In the second section, I argue that the core is (in a certain sense) contractualist, and in the third section, I argue that it is a collectivistic code. Both of these factors are of crucial importance for the explanation of the more specific features characterizing (the core of) CSM and, hence, the specific deontological theories built on it. In the fourth section, I identify and discuss seven such specific deontological features and try to explain why CSM exhibits them. In the fifth section I consider three representative deontological theories and show that they too manifest these features. In the sixth and last section, I state my final characterizations of (the core of) CSM and of D.

2.1  Common-Sense Morality Every society has a morality, a common collection of norms which are impressed on its members from their early childhood.1 On the basis of a thorough examination of anthropological evidence, Richard Brandt concludes that “no societies have been 1  This is partly an empirical, partly a conceptual claim. If people living in a certain area did not have a common set of norms regulating their dealings with each other, there would almost certainly be so much strife and dissension that we would not call this congregation of people a society.

© Springer Nature Switzerland AG 2019 J. Österberg, Towards Reunion in Ethics, Philosophical Studies Series 138, https://doi.org/10.1007/978-3-030-12410-6_2

27

28

2  Deontological Ethics: Exposition

reported that are without some authoritative rules for behavior in some sorts of recurrent social situations; nor have societies been reported in which criticism and praise of conduct do not occur”.2 And, Brandt continues, the kinds of conduct regulated in different societies are roughly the same, since there are probably everywhere ethical prohibitions and requirements relevant to the major areas where clash of interest seems inevitable in social life: sex and marriage, the family, property and income, the security of life and limb and reputation, contract and promises, the care of the aged or sick and needy, and so on.3

This set of prohibitions and requirements I shall call “common-sense morality” (“CSM”).4 (Other names, having slightly different connotations that are often used for this phenomenon include “social morality”, “conventional morality”, “ordinary morality”, and “positive morality”. I will sometimes use the name “social morality”, in order to stress the social function of CSM.) CSM is essentially different from everything else that goes under the name of morality. To be sure, like the moralities set forth by moral philosophers, CSM is something which can (partly) be stated in propositions and be compared with what is stated or implied by those moralities. But it is also—like custom (etiquette), law, and religion—a social institution with specific functions, causally and conceptually related to other social institutions and, together with these, in various ways shaping the minds of the members of society. To quote William Frankena: As first encountered by the individual, at any rate, it is an instrument of society as a whole for the guidance of individuals and smaller groups. It makes demands on individuals that are, initially at least, external to them. Even if the individuals become spokesmen of these demands […] the demands are still not merely theirs nor directed only at themselves. If they come to disagree with the demands, then […] they must still do so from the moral point of view that has somehow been inculcated into them.5

That CSM is a social morality, a morality directed to all members of society, puts some restrictions on both its form and its content. (In Sect. 2.4 I shall try to show how the functions of CSM can explain its characteristic features.) Now, even if every society has a common morality, it does not follow that there is a morality common to all societies. And available evidence shows that there is not. As anthropologists and historians have shown, the norms of CSM to a certain extent vary, both geographically and historically. Many of these differences can be Anthropologists inform us that in certain so-called “primitive” societies there is no distinction between morality, custom (etiquette), and law; hence in such societies there is no concept corresponding to our concept “morality”. Yet, I would contend, they have a norm system that encompasses norms that we count as moral, and in that sense they certainly have a morality. 2  Richard Brandt, Ethical Theory, p. 87 f. 3  Ibid., p. 92. 4  According to Derek Parfit (Reasons and Persons (Oxford: Oxford University Press, 1984), p. 40), the term was coined by Henry Sidgwick. 5  William Frankena, Ethics, p. 6.

2.1  Common-Sense Morality

29

explained as due to different conditions of living or to different factual beliefs ­concerning man and world, but there seems to be some disagreement in basic principles too. There are societies where a “master” is thought to have an absolute right to treat his slaves in any way he chooses, including arbitrarily beating and killing them. Similarly, there may be no limitations on what a husband can do to his wife, or a father to his young children. Infanticide is considered acceptable in some societies. […] A person who is able successfully to cheat outsiders may be treated as an admirable person. Similarly for someone who is able to harm and kill outsiders.6

But the differences between social moralities should not overshadow the fact that there is a considerable amount of overlap. As Richard Brandt points out, there is no group that thinks it right to kill an adult, healthy member of the society who has committed no crime and whose death is not required by the welfare of the group. There is no group in which marriage or sexual intercourse is approved between members of the immediate family […] There is no society in which kindliness, sympathy, hospitality, or regard for others and their rights is disapproved.7

(I shall return to this set of common norms in Sect. 2.2.) The norms belonging to the CSM of a society are usually seen as very important by members of the society. People tend to view them as overriding other norms— norms of custom and etiquette, even of law— and to react with indignation and blame when the norms are violated by other people, and with feelings of guilt and remorse when they have violated them themselves.8 It is sometimes suggested that the norms of CSM are defeasible. Since they sometimes conflict per accidens, that is, since the situation sometimes is such that you have to disobey at least one of them, they cannot, it is said, be absolute. Thus, for example, it is claimed, what the norm forbidding lying really says is that lying is wrong unless, by not lying, you violate another norm which, in the case in question, takes precedence over it. But this, I think, is not a correct description of the way (at least many) people view cases of moral conflict. They consider each norm absolutely valid and hold that clashes between norms represent genuine moral dilemmas, so that the agent acts morally wrong (to some extent) whatever he does. Hitherto I have spoken of CSM as consisting primarily of norms concerning what actions are right and wrong, and secondarily of moral reactions to the honouring and, especially, the not honouring of the norms. But this is, of course, an oversimplified picture. To some extent CSM also consists of second-order norms of

 Gilbert Harman in Gilbert Harman and Judith Jarvis Thomson, Moral Relativism and Moral Objectivity (Oxford: Blackwell, 1996), p. 9. For a vastly detailed account of the variety in social moralities—an account that amply testifies to Harman’s verdict—see Edward Westermarck, The Origin and Development of the Moral Ideas, 2 vols. (London: Macmillan, 1908), passim. 7  Brandt, op. cit., p. 95. 8  The first observation is perhaps only true of societies where these norm systems are clearly distinguished from each other. 6

30

2  Deontological Ethics: Exposition

priority for deciding what to do when two or more first-order norms conflict (per accidens). Moreover, it consists of views concerning what traits of character are good and bad; there is, in other words, a common-sense ethics of virtue. A person well brought up therefore possesses virtues which—if, in addition, he has judgement and some experience—helps him to interpret and supplement the norms of CSM, that is, to see more to their spirit than to their letter, and, therefore, to be able to adjudicate conflicts between them. Both norms of action and virtues are probably to a great extent inculcated by means of examples; thus every society has its stock of moral heroes and villains, real and fictional, whose deeds are set forth as good and bad examples. (The importance of moral education by means of narratives— whether oral, written, or filmic—is probably considerable.) These moral paradigms not only illustrate what actions are morally right or wrong, good or bad; they also make vivid the virtue ethics of CSM, its ideals of how a person should be. CSM is thus a rather sophisticated fabric. A successful moral upbringing will impart to a person a set of moral dispositions to act which are much more flexible and fine-grained than any based merely on the simple rules commonly cited as making up CSM.  Criticizing Richard Hare’s characterization of CSM as furnishing people with quite rigid norms of action—in contradistinction to reflective “critical thinking”, whose task it is to decide which norms people should accept—T.  M. Scanlon claims that such common-sense moral principles as ‘it is wrong to kill’ are not simple self-contained rules but ways of referring to the same moral ideals whose complexities are explored in ‘critical thinking’. Qualifications having to do with intent, justifications, excuses, and so on, while not always explicit in any formulation of the principle, are part of the idea referred to (though their exact boundaries are never totally clear).9

Being initiated in the morality of one’s society is in certain respects analogous to being initiated in one’s first language. In both cases a person’s initiation starts already when he is a child, and the initiation process works more through setting good examples and by criticism of wrong behaviour than by citing rules out of context. The rules that are cited are usually rather simple and rigid, but with growing experience the child learns how to extend and qualify them in order to cope with situations never before met with. Since much of this process is not fully conscious, the rather complex and subtle rules which have come to be internalized by most adults are implicit and not easily statable by them. Like linguistic competence, moral competence amounts to a kind of tacit knowledge. (In the following, however, I shall disregard the complex nature of CSM and concentrate on its norms of action, and I will use rough and ready formulations of them. The necessary qualifications will have to be supplied by the reader.) It is an astonishing fact that CSM is so little discussed by moral philosophers, and that, when discussed, it is often done so in a rather cursory and/or condescending manner. There are exceptions, of course; thus, for example, both Aristotle and Sidgwick have given sympathetic accounts of the CSMs of their respective societies, 9  T. M. Scanlon, “Levels of Moral Thinking”, in Douglas Seanor and N. Fotion (eds.), Hare and Critics: Essays on Moral Thinking (Oxford: Oxford University Press, 1988), p. 134.

2.1  Common-Sense Morality

31

and Kant regarded himself as a defender of CSM. But on the whole, moral philosophers, especially in later times, tend to ignore or be critical of CSM. The explanation of this is, no doubt, bound up with the whole historical development from the Renaissance onwards. All since then, philosophers, like other intellectuals, have increasingly come to look upon themselves as members of the enlightened vanguard, leading their fellow-beings to the promised land of freedom and material abundance, where the shackles of tradition-governed society are broken asunder and supplanted by the scientific outlook. And the pretensions of the intellectual vanguard have not been excessive: during this era, Western society in all its aspects has been immensely transformed, to a great extent due to scientific and technological developments. Among these changes are conspicuous changes in many areas of CSM. (Consider, for example, the changes in sexual morality that have taken place during the last centuries in many Western countries.) These changes have not happened simultaneously in all strata of society. They have been more rapid in the bigger cities and among intellectuals, thus creating a cultural lag between the “enlightened” few and the more tradition-bound majority. Identifying society’s morality with that of the majority, the intellectual vanguard will naturally look upon social morality as something old-fashioned and narrow-minded, full of irrational beliefs and conservative prejudices. A characteristic expression of this attitude is found in the description of the vanguard ethos of the famous Bloomsbury group, given by Lord Keynes, himself one of its leading members: “[W]e claimed the right to judge every case on its own merits […] repudiat[ing] entirely customary morals, convention and traditional wisdom.”10 There are other circumstances that tend to reinforce this attitude to CSM. Firstly, for many centuries—to some extent even today—morality was intimately bound up with religion and the established churches. Therefore the secularization process, characteristic of the modern era and first taking place among intellectuals, brought with it a critical attitude not only to established religion but also to established morality. Secondly, the contacts with foreign cultures that accompanied the increasing domination by Western Europe over other continents brought with it awareness that different societies had different moralities. To the extent that the moralities of foreign cultures were not considered as representing “lower” levels of development—and this was emphatically denied by, for example, many leading anthropologists—this insight has probably contributed, in the eyes of many Europeans, to a loss of prestige for the morality of their own society.11 Thirdly, a central tenet of the ideology of modern times is the demand that one should be critical of conventional ideas, think for oneself, and not accept anything on authority. It is therefore not surprising that many people pride themselves of living up to this ideal in, among  John Maynard Keynes, “My Early Beliefs”, in his Two Memoirs (New York: Augustus M. Kelley, 1949), p. 97; quoted from Tom Regan, Bloomsbury’s Prophet: G. E. Moore and the Development of His Moral Philosophy (Philadelphia: Temple University Press, 1986), p. 247. 11  The insight, due to modern historiography, of how different the moralities of past societies have been no doubt, had a similar impact. And so, I think, has the critical attitude to established morality taken by Marxism and psychoanalysis. 10

32

2  Deontological Ethics: Exposition

other things, moral matters, not realizing the extent to which their opinions are still due to the moulding influence of CSM. I have met with quite a few people, even moral philosophers, who confess their allegiance to some lofty moral theory, though their practice clearly shows their adherence to the morality they imbibed with their mother’s milk. It is not surprising, then, that CSM has often been criticized, the criticism ranging from wholesale condemnation to a questioning of specific norms. (I will discuss the objections to CSM in Sects. 3.2 and 3.3.) And in the case of many societies we certainly find norms contained in its CSM that are clearly objectionable: norms that prescribe (or permit) actions which are cruel, or unjust, or simply pointless.12 (As already mentioned, some of these norms can, no doubt, be ascribed to mistaken factual beliefs or to special conditions of living, but probably not all can.)

2.2  Common-Sense Morality and Contractualism Even if the morality of many societies contains norms which are offensive, every social morality also contains a set of norms that can hardly be found objectionable. As Richard Brandt (see Sect. 2.1) and others affirm, in every society actions such as lying, cheating, stealing, injuring, killing, etc. are forbidden—at least with respect to members of one’s own society. In addition to these norms enjoining us not to harm other people (at least not members of one’s own society), there exist in every society norms that enjoin us to benefit other people (at least members of one’s own society) when the cost of doing this is small, and to benefit our near and dear. In the following, “CSM*” will refer to the set of norms belonging to these three categories. Of these norms, those belonging to the first category are by far the most important: as I said above, were not these norms honoured by most people on most occasions, there would not exist a society, only a group of people living in a Hobbesian state of nature.13 General non-compliance with the norms belonging to the second category does not have similar disastrous consequences. But we would all be worse off if, for example, people did not warn us of imminent dangers, did not throw a lifebuoy to us when we had fallen into the water, or did not show us the way when we had lost it and asked for help. Norms belonging to the third category are different: for each of us it does not matter very much whether people in general benefit their near and dear; what is important is that people in our near and dear groups benefit each other and can be expected to do that. Otherwise we will be confronted with situations such as this:  See the many articles on the CSMs of different societies, both past and contemporary, under the heading of “Ethics and Morality”, in James Hastings (ed.), Encyclopedia of Religion and Ethics, 13 vols. (Edinburgh: T. and T. Clark; New York: Charles Scribner’s sons, 1908–26). 13  For a vivid illustration of this, see Colin Turnbull, The Mountain People (New York: Simon and Schuster, 1972). 12

2.2  Common-Sense Morality and Contractualism

33

Your corn is ripe to-day; mine will be so to-morrow. ’Tis profitable for us both, that I shou’d labour with you to-day, and that you shou’d aid me to-morrow. I have no kindness for you, and know you have as little for me. I will not, therefore, take any pains upon your account, and shou’d I labour with you upon my own account, in expectation of a return, I know I shou’d be disappointed, and that I shou’d in vain depend upon your gratitude. Here then I leave you to labour alone: You treat me in the same manner. The seasons change; and both of us lose our harvests for want of mutual confidence and security.14

This suggests that these norms, common to the CSMs of different societies, are (what I shall call) “solutions” to “n-person (non-iterated) prisoners’ dilemma situations” (PD-situations). Let me first define “PD-situation” and then “solution to a PD-situation”.15 A situation is a PD-situation if, and only if:   (i). there is a set of people S such that each member M of S has two actions at his disposal, a and a′ (usually a′ = −a)16; (ii). for each M, the value for him of the outcome of his doing a is greater than that of the outcome of his doing a′—whatever other members of S do; but (iii). for each M, the value for him of the outcome of each M doing a is less than that of the outcome of each M doing a′. A norm N is a solution to a PD-situation for a set of people S, if and only if: N prescribes, for each member M of S, that M should perform the action such that the value for him of the outcome of each M doing this kind of action is greater than that of the outcome of each M doing its alternative. An obvious explanation of this most important part of existing morality is then the insight that, on the collective level, “it pays”: we should all be worse off if we did not live in, what Hobbes called, “the state of society”, that is, if (at least most) people we meet with did not comply with the norms of CSM.  Life would then indeed be “solitary, poor, nasty, brutish, and short”.17 That this is so is evident from even a cursory inspection of the norms in question. It is obvious that if people did not generally comply with, for example, the norm not to lie to each other whenever it suited their purposes, all would be worse off than if, as is actually the case, the norm is generally complied with. Still, anyone would be better off if he did not comply with the norm whenever it suited his purpose. This is, of course, most obvious in cases when other people do not comply either. Therefore much hinges on the shaping of “moral” preferences by internalizing the norms among the members of society. (This probably requires some rudimentary awareness that the facts are as  David Hume, A Treatise of Human Nature, Bk. III, Pt. II, Sec. V; ed. David Fate Norton and Mary Fate Norton (Oxford: Oxford University Press, 2004 (2000)), p. 334. 15  Prisoners’ dilemmas—like lawn-crossers’ and consequentialists’ dilemmas, both of which will be introduced in subsequent chapters—are not, of course, dilemmas in the strict sense. 16  For the sake of simplicity, the term “action” here includes omissions. 17  This is obvious in the case of the norms belonging to the first category. It is perhaps a slight exaggeration in the case of the norms belonging to the second and third categories. In the following, I will mostly dwell upon the norms belonging to the first category. 14

34

2  Deontological Ethics: Exposition

mentioned above. And this also seems to be the case: if people did not have (at least) a tacit understanding that they (collectively) gained by general compliance with the norms in question, it would be difficult to motivate them to do their part; and false beliefs in this matter would not be long-lived. Not even morality can fool all people all the time.)18 The claim that CSM* is a solution to a PD-situation entails the claim that (the most important part of) existing morality is, or is as if it were, the outcome of a contract, or an agreement, which serves the interests of everyone who is party to it. I shall call this claim “descriptive moral contractualism”.19 (This factual claim should be distinguished from normative claims to the effect that a morality is justified if, and only if, it is, or is as if it were, the outcome of a contract agreed upon under such and such conditions. Such claims, too, are sometimes called “contractualism” (or “contractarianism”). They will be discussed later on in Chap. 7.) As I will now show, (descriptive) moral contractualism has a long history. The, to my knowledge, first statement that CSM* is a solution to a PD-situation, together with the first explicit statement of moral contractualism, is found in Plato’s Republic— in a passage where Glaucon reports “what Thrasymachos and hundreds of others” have dinned into his ears.20 What they say is that it is according to nature a good thing to inflict wrong or injury, and a bad thing to suffer it, but that the disadvantages of suffering it exceed the advantages of inflicting it; after a taste of both, therefore, men decide that, as they can’t evade the one and achieve the other, it will pay to make a compact with each other by which they forgo both. They accordingly proceed to make laws and mutual agreements, and what the law lays down they call lawful and right. This is the origin and nature of justice. It lies between what is most desirable, to do wrong and avoid punishment, and what is most undesirable, to suffer wrong without being able to get redress.21

(This is part of the picture of morality that Socrates is invited to refute. But does he attempt to refute this particular part of it? He certainly tries to give a better justification of social morality, but that does not mean that the Thrasymachean account of its origin is considered wrong. Socrates also argues that “real” justice concerns the structure of man’s soul, but that does not mean that the above explanation of social morality is (taken to be) mistaken.) In the account given by Glaucon, the fact that social upbringing involves the internalization of moral norms, and the fact that the internalization of moral norms transforms the pattern of people’s preferences, are not recognized. It is the same in Hobbes’s account in Leviathan; even in the state of society Hobbesian man is an  Something that bears witness to, at least, a dim awareness of the pertinent facts is the common reaction to moral misdemeanours, “What if everybody did that!” 19  This claim is sometimes intertwined with and not clearly distinguished from (what may be called) social contractualism, the claim that political authority is, and can be justified as, the outcome of an express or tacit contract, viz. the time-honoured social contract. 20  The Republic, 358c; quoted from H. D. P. Lee’s translation of The Republic, p. 45. 21  Ibid., 358e–359a; Lee’s translation, p. 45. Notice that the Greek term dikaiosune, here translated as “justice”, actually “has a less legal and more moral meaning than ‘justice’; it is in fact the most general Greek word for morality, both as a personal quality and as issuing in right action” (Lee in the “Introduction” to his translation, p. 7, n. 1). 18

2.2  Common-Sense Morality and Contractualism

35

unregenerate creature propelled by “a perpetuall and restlesse desire of Power after power, that ceaseth onely in Death”.22 Only with Rousseau’s Du contrat social are these facts explicitly mentioned: The passage from the state of nature to the civil state produces a very remarkable change in man, by substituting justice for instinct in his conduct, and giving his actions the morality they had formerly lacked. […] We might […] add, to what man acquires in the civil state, moral liberty, which alone makes him truly master of himself; for the mere impulse of appetite is slavery, while obedience to a law which we prescribe to ourselves is liberty.23

In Glaucon’s statement of moral contractualism, as in several other statements of it, it is said that the morality of a society is the outcome of an agreement that has actually once taken place. Like the similar suggestion with regard to the social contract, taken literally, this is, to put it mildly, an absurd suggestion. But, in all probability, it is a way of dramatizing and making vivid the indubitable fact that existing morality is as if its norms were the outcome of an actual contract.24 An explanation of this fact is given by Hume in his account of the “circumstances of justice”.25 If, Hume argues, (i) we had either lived in material superabundance or (ii) in extreme scarcity; or if (iii) human beings were either totally selfless and altruistic or (iv) totally selfish and malevolent; or if (v) human beings had no need of each other or (vi) were of greatly different strength, both mentally and physically, then “justice” (morality) would be of no use and would never had arisen among us. But, as things stand, the opposite circumstances obtain: we neither live in superabundance or extreme scarcity; we are neither totally selfless nor totally selfish; and we all need each other in many ways and are of roughly equal strength. Being placed in such circumstances, the utility of morality will become evident to people, and, starting with the family,  Thomas Hobbes, Leviathan, Part I, Ch. XI; quoted from C. B. Macpherson’s ed., p. 161.  Quoted from Jean-Jacques Rousseau, The Social Contract and Discourses, tr. G.  D. H.  Cole (London: Everyman, 1973), p. 177 f. The “law which we prescribe to ourselves” spoken of in the last sentence of the quoted passage is (the outcome of) that much discussed phenomenon which Rousseau calls “the general will” (“la volonté générale”). For a convincing interpretation of this concept that takes it to embody (what I call) a solution to a PD-situation, see W. G. Runciman and Amartya Sen, “Games, Justice and the General Will”, Mind, 74 (1965): 554–62. The connection between the general will and central elements in Kant’s moral philosophy is adumbrated by Cole in the “Introduction” to his translation: The idea of the General Will is indeed essentially ethical: it is a principle of moral conduct applied to political behaviour. Ethically, it is one and the same as Kant’s conception of moral rationality. Kant, in effect, took it from Rousseau and applied it to the entire realm of conduct. The justification for this extension is to be found in Rousseau’s own attitude; for he protested more than once against attempts to treat moral and political philosophy apart, as distinct studies, and asserted their absolute unity. (Ibid., p. xxxvi) 24  As a matter of fact, Hobbes explicitly denies that the state of nature actually once prevailed “all over the world”. See Leviathan, p. 187. 25  What Hume terms “justice” is, roughly, the norms that are included in CSM, although Hume stresses the rules concerning property. “‘Justice’ is, in effect, Hume’s term for the virtue of following the rules and institutions that society creates to thwart the elements of the wolf and the serpent in our nature” (Tom Beauchamp in “Editor’s Introduction” to David Hume, An Enquiry concerning the Principles of Morals (Oxford: Oxford University Press, 1998), p. 35.) 22 23

36

2  Deontological Ethics: Exposition

the increasing interdependence between people will cause the boundaries of the moral community to gradually expand. For, says Hume, suppose the conjunction of the sexes to be established in nature, a family immediately arises; and particular rules being found requisite for its subsistence, these are immediately embraced; though without comprehending the rest of mankind within their prescriptions. Suppose, that several families unite together into one society, which is totally disjoined from all others, the rules, which preserve peace and order, enlarge themselves to the utmost extent of that society; but becoming then entirely useless, lose their force when carried one step further. But again, suppose that several distinct societies maintain a kind of intercourse for mutual convenience and advantage, the boundaries of justice still grow larger, in proportion to the largeness of men’s views and the force of their mutual connexions. History, experience, reason sufficiently instruct us in this natural progress of human sentiments, and in the gradual enlargement of our regards to justice, in proportion as we become acquainted with the extensive utility of that virtue.26

There are several other forces, Hume points out, that contribute to the development of moral attitudes, viz. sympathy, that is, a natural tendency to share other people’s feelings, and the efforts of politicians and private educators to inculcate such attitudes in children. Perhaps Hume would also accept, as a complementary account of the forces lying behind the “natural progress of human sentiments”, the following suggestion by J. L. Mackie: For the reasons given [that is, those given by Hume for the utility of “justice”], moral sentiments which “annex the idea of virtue to justice” will enable social groups in which they take root to flourish. Consequently the ordinary evolutionary pressures, the differential survival of groups in which such sentiments are stronger, either as inherited psychological tendencies or as socially maintained traditions, will help to explain why such sentiments become strong and widespread.27

But be that as it may. In order to account for the contractual nature of social morality, “the sixth circumstance of justice”, viz. that human beings are of roughly equal strength, both mentally and physically, is especially important. This circumstance is also stressed by Hobbes: Nature hath made men so equall, in the faculties of body and mind; as that though there bee found one man sometimes manifestly stronger in body, or of quicker mind then another; yet when all is reckoned together, the difference between man, and man, is not so considerable, as that one man can thereupon claim to himselfe any benefit, to which another may not pretend, as well as he. For as to the strength of body, the weakest has strength enough to kill the strongest, either by secret machination, or by confederacy with others, that are in the same danger with himselfe. And as to the faculties of the mind, [… ] I find yet a greater equality amongst men, than that of strength.28

Before bringing this section to an end, I should add that not everybody thinks that the case for moral contractualism is sufficiently established. Thus the claim is

 David Hume, op. cit., Sec. 3, Par. 21; p. 89.  J. L. Mackie, Ethics: Inventing Right and Wrong (Harmondsworth: Penguin, 1977), p. 113. 28  Thomas Hobbes, op. cit., p. 183. There is a similar account in H. L. A. Hart, The Concept of Law, p. 190 f. 26 27

2.3 Collective Versus Individualistic Codes

37

criticized by Mary Midgley as resting on a false picture of human psychology.29 People, she says, “are not so prudent or consistent as this account would imply”.30 They are not constantly acting from egoistic motives; human beings “have natural social dispositions”, and those “who do make an effort to behave decently plainly are often moved by a quite different set of motives, arising directly out of consideration for the claims of others”.31 But Midgley’s criticism misses the mark. As should be evident by now, contractualists need not, and most of them do not, assume that human beings are egoists, let alone prudent and consistent ones. Hobbes did, but, as I mentioned above, Rousseau did not. And Hume thought that, in order to account for the existence of social morality, psychological egoism had to be rejected: It is sufficient for our present purpose, if it be allowed, what surely, without the greatest absurdity, cannot be disputed, that there is some benevolence, however small, infused into our bosom; some spark of friendship for human kind; some particle of the dove, kneaded into our frame, along with the elements of the wolf and the serpent.32

2.3  Collective Versus Individualistic Codes On most views, it has been said, “morality is essentially a collective code—an answer to the question, ‘How should we all act?’”33 In my view, whatever is the case with “morality”, CSM* is basically a collective, not an individualistic code. Consider a CSM* norm to the effect that people shall do (actions of the kind) A. That CSM* is a collective code means that the form of the prescription is: (1) It shall be the case that each member of society does A, not: (2) Each member of society shall do A. That CSM* is to be understood in the first, the collective sense, follows from its being of a contractual nature. (As far as I can see, being a contractualist code is not only a sufficient, but also a necessary condition for being a collective code.) If CSM* were to be understood in the second, the individualistic sense, it would prescribe that any member of society ought to conform to CSM* even when interacting with members who do not conform, at least not with respect to him. But this strong principle would run contrary to the contractual nature of CSM*, and it is not, I take

 Mary Midgley, “The Origin of Ethics”, in Peter Singer (ed.), A Companion to Ethics (Oxford: Blackwell, 1993): 3–13. 30  Ibid., p. 4. 31  Ibid., pp. 9 and 5, respectively. 32  David Hume, op. cit., Sec. 9, Par. 4; p. 147. 33  Derek Parfit, Reasons and Persons, p. 106. Perhaps, however, Parfit means by “collective code” something different from what I mean. 29

38

2  Deontological Ethics: Exposition

it, how, on reflection, CSM* is understood: it does not claim that we have an obligation to return good for evil.34 To meet this objection it might be suggested that CSM* is to be understood as prescribing: (3) Each member of society shall do A with respect to those who conform with respect to him. But if this interpretation of CSM* were correct, it would be an intolerably weak system of norms. For then it would be completely satisfied either if all conformed or if none did. But, surely, the very point of CSM* is to guarantee a state of affairs that approximates to the former rather than to the latter state. And it sounds preposterous to claim that CSM* would be completely satisfied in the latter state, that is to say, in the state of nature. The import of a CSM* norm is thus similar to that of a military order given to a group of soldiers. When the commanding officer shouts “forward” to his men, he does not only mean that each soldier should attack if the other soldiers attack. (This is to demand too little.) Nor does he mean that each soldier should attack whether or not the others attack. (This is to demand needless bravery.) What he means is that they should all attack together. But, it might be objected, CSM* is an individualistic code, for certainly its norms are directed to individual members of society, such as you and me. The latter claim should be readily admitted, but it is compatible with CSM* being primarily a collective code, since a collective code conditionally generates an individualistic code. For, I take it, it is a conceptual truth that if all members of a group ought to comply with a norm, then each member of the group ought to comply if the other members of the group comply. So living in the state of society, that is, in a state where other members of society comply with CSM*, each person ought to comply. (CSM* certainly encompasses (2) but this is not its fundamental, but a derivative, principle.) Using some terminology from linguistics, we may say that principles of the kind (1)—which almost never manifest themselves in people’s consciousness and have no direct influence on people’s thought and conduct— belong to the deep structure of CSM*, whereas principles of the kind (2)—which are not seldom consciously entertained— belong to its surface structure. But, it may be further objected, this means that no one ought to comply with CSM*. For it is probably never the case that all members of a society comply with its morality. And if there is less than universal compliance, the collective code does not generate a corresponding set of individualistic norms. The answer to the objection is that, not only ought each member of society to comply with CSM* if all other members comply; also any subgroup of that group ought to comply if it is better for each of its members that all in the group comply than that none does. So any person  In the case of some norms, e.g., those concerning lying and promise-keeping, partial conformity to the norm would often be pointless since, without general conformity, people would not trust each other to abstain from lying and promise-breaking. (See my discussion in Sect. 5.5 and the Appendix.)

34

2.3 Collective Versus Individualistic Codes

39

ought to comply with respect to members of such a subgroup of which he is a member if these comply with respect to him. And in the state of society almost everyone lives most of his life as a member of complying subgroups, and should accordingly (often) comply. Many people will, no doubt, take exception to the very idea of a collective moral code, claiming that morality is inherently an individualistic business. I think such a reaction is partly due to the all-embracing ideology of contemporary society, to its emphasis on personal autonomy.35 In the moral sphere, this individualistic ideology takes the form of a conviction that the source of morality, of moral values and principles, the creator of the very criteria of moral evaluation, is the individual: he becomes the supreme arbiter of moral (and, by implication, other) values, the final moral authority in the most fundamental sense.36

If one thinks that it is up to each individual to decide what morality he shall conform to, the conception of morality as primarily a collective code must seem clearly mistaken. For if everyone else in a group of which I am a member conforms to a “contractually justified” code, then, on the collective conception of morality, I have no choice but to conform too. Moreover, as I said, CSM* does not reveal itself as a collective code, at least not in reasonably stable and peaceful societies. The child is not taught to conform to the norms of CSM* on condition that those with whom he interacts conform too. Since it is taken for granted that this condition is largely fulfilled, the child is taught to conform, period. (Sometimes the child is even told to conform “even if other children will not conform”. This pious fraud is certainly justified (in a reasonable stable and peaceful society), since it confers an added degree of sanctity upon the norms: they become more strongly internalized by the individual, thus helping him to resist temptations to break them.)37 Another objection to the idea of a collective code that I have come across says that the implied notion of a collective agent is incomprehensible. My guess is that this reaction rests on a misunderstanding of the very notion of a collective: one imagines a collective as a kind of mysterious super-individual. And, of course, it is not easy to understand what it would mean for such an entity to exist and to act. But a collective is not any kind of individual—it is a set of individuals. And that a collective act simply means that its members act. Similarly, that the collective shall (ought to) act means only that its members, taken together, shall (ought to) act.  For a survey of the many kinds of individualism characteristic of contemporary Western society—political, economic, religious, etc.—see Steven Lukes, Individualism (Oxford: Blackwell, 1973). 36  Ibid., p. 101. 37  Consider the following parallel: In Sweden, people are taught to drive a car (ride a bike, etc.) on the right-hand side of the road. They are not, of course, told the collective rule: It shall be the case that each person keeps to the right. This would not give the individual much guidance. Nevertheless, this is the fundamental rule, which, together with facts about people’s conformity to the rule, justifies the corresponding individualistic rule: You shall keep to the right. It was the collective rule that was the object of a referendum in Sweden in 1955. 35

40

2  Deontological Ethics: Exposition

Now, what kind of statement does the claim that CSM* is basically a collective, not an individualistic, code express—a (disguised) normative statement, a factual, or a conceptual one? Well, I take it to be an interpretive statement, stating how CSM* should be (or is best) understood. (To some extent, then, it is a normative, or an axiological, claim.) This claim, in turn, is based on two premises. Firstly, that this is how, on reflection, CSM* is understood by us (a factual claim). And, secondly, how CSM* is understood furnishes the ground for how it should be understood. (This is, I think, partly a conceptual, partly a normative, claim.) As for the first premise, I think that, upon reflection, most people hold that a person should conform to the norms of CSM* if, and only if, other people with whom he interacts conform too. The “only if” part of this condition contains the gist of Hume’s fourth circumstance of justice: Should a man fall into “the society of ruffians”, Hume says, he can have no other expedient than to arm himself, to whomever the sword he seized, or the buckler, may belong: To make provision of all means of defence and security: And his particular regard to justice being no longer of USE to his own safety or that of others, he must consult the dictates of self-preservation alone, without concern for those who no longer merit his care and attention.38

I think that most people would agree with Hume on this matter. I also think that most people consider the state of society to be preferable to that of nature, that is, prefer general conformity to CSM* to general non-conformity. And, taken together, holding these two convictions amounts to subscribing to the claim that CSM* is basically a collective code. The second premise states, I think, a reasonable principle of interpretation of codes such as CSM*. It would be otherwise if CSM* were commands issued by some moral legislator, for example, God. In such a case, it could be argued, an interpretation of the nature of CSM* would have to refer to the intentions of the “norm-­ giver(s)”, just as the interpretation of positive law has to refer to the intentions of the law-giver. But, since there are no (justifiable) moral norm-givers, how a norm should be understood has to refer to how those whose norm it is understand it, just as the interpretation of a linguistic rule should refer to how those whose rule it is understand it. (This claim, it should be noted, has nothing to do with the moral realism–non-realism issue.)

2.4  Deontological Features in Common-Sense Morality One of the claims that I try to defend in this work is that CSM* is both the origin and the touchstone of the deontological theories put forward by moral theorists. This claim is corroborated by the fact that CSM* contains several deontological features, that is, that it considers morally relevant certain properties of action the 38

 David Hume, op. cit., p. 86.

2.4  Deontological Features in Common-Sense Morality

41

relevance of which cannot be explained by reference to the value of the outcomes of actions. These deontological features, common to CSM* and deontological moral theories, have often been criticized by philosophers of a consequentialist bent for being irrational and without justification. The question whether they can be justified will be touched upon in the next chapter, leaving the final verdict on it to Chap. 7. In this section I shall point out the most important of the deontological features and, for each feature, offer an explanation of why CSM* contains it.39 (How do I know that the features I mention really belong to CSM*? Well, I rely on my own intuitions and on my experience of what other people say and do. But, of course, if other people do not agree with me, I am wrong. So in each case I implicitly appeal to my readers’ opinions: does not this feature belong to the morality that you imbibed with your mother’s milk?) (i) Constraints. Recall the doctor’s choice, introduced in Sect. 1.1. Given that the outcome of cutting up the healthy person in room 306 and distributing his organs to the five needy patients is optimal, consequentialism tells the doctor to do that. But CSM* (and every deontological theory) says that this is morally forbidden: even if the outcome of killing an innocent person is, as in this case, (non-­morally) better than not killing him, we are not allowed to do it. More generally, as I pointed out in Sect. 2.2, CSM* recognizes constraints against harming other people in various ways: we must not, for example, deceive, injure, maim, torture, or kill people even if this is the only way to bring about the best results overall. The strength of these constraints probably varies, at least to some extent, from person to person, from society to society, and from one time to another. In general, no constraint is considered absolute but of a pro tanto character and whether it may be overridden depends on the degree and kind of harm involved, as well as on the amount of good at stake. Thus hardly anyone would agree with Kant that the ban on deceiving other people is so strong that one must not even lie to a presumptive murderer in search of his victim.40 But other constraints—for example, the constraint against killing an innocent person is nearly absolute. Why does CSM* recognize constraints against harming other people? The (simple) reason is, of course, that people prefer such a morality to its opposite, a morality that permitted other people to harm them—even if the infliction of harm was the only way to bring about the best outcome overall. People are so constituted that they would live in constant fear if they knew that (most) other people did not recognize restrictions against inflicting harm. (Of course, some people may sometimes be willing to suffer harm for the sake of the common good, but they usually want to be asked for permission beforehand.)

 In the following, I will sometimes talk as if I thought that other people who have commented on CSM have had CSM* in mind. Actually, what I think is that they have had at least something similar in mind, and I trust that my simplified way of talking will not be misleading. 40  Immanuel Kant, “On a Supposed Right to Lie from Altruistic Motives”, in Lewis White Beck (ed. and tr.), Immanuel Kant: Critique of Practical Reason and Other Writings in Moral Philosophy, reprint (New York: Garland, 1976): 346–50. 39

42

2  Deontological Ethics: Exposition

(ii) Options. Since CSM* is addressed to all (grown-up and psychologically sane) members of society, it is important that it is not too demanding: a system of norms that goes against the grain will not be complied with and will fall into disrespect. This fact, together with the fact that, for obvious reasons, moral education has to begin early in life, also explains why the norms of CSM* are comparatively few and not too complex. More importantly, it also explains why CSM* contains agentfavouring permissions or, as they are sometimes called, options: to a large extent you are permitted to fulfil your own projects and, in general, do what you prefer, even if this is not best from an impersonal point of view. (You are allowed to take “moral holidays”.) In many cases, doing what is (believed to be) impersonally best is optional and regarded as acts of supererogation. (iii) Agent-Relativity. According to CSM* there is, as I said, a list of wrong actions (constraints), but, notice, one should not act so as to minimize the performance of wrong actions; rather each person should see to it that he himself does not commit any wrong action.41 Suppose, for example, that if you do not tell a lie on a certain occasion, then, as a consequence, several other people will do wrong actions, for example, tell lies. That is, according to CSM*, not your problem. Your moral task is to see to it that you yourself do no tell lies. I will call this feature of CSM*, which is part and parcel of the deontological outlook, agent-relativity.42 (My guess is that it is CSM*’s agent-relative conception of morality that, at least partly, motivates those non-religious conscientious objectors who refuse to defend their country under any circumstances.) Agent-relativity is explicitly defended by several philosophers of a deontological bent. Thus Bernard Williams claims that “each of us is specially responsible for what he does, rather than for what other people do”.43 The claim is made in his discussion of two fictitious examples. One of them, which I shall call Jim’s choice, runs as follows: On a botanical expedition in South America Jim one day arrives at a small town, where he finds a group of 20 Indians tied against the wall. An army patrol is just about to kill them, in order to remind the other villagers that the government does not like the recent acts of protest among them and will tolerate no more of it. To celebrate the arrival of a stranger, however, the captain in command offers Jim the privilege of killing one of the Indians. If Jim accepts, then as a special mark of the occasion, the other Indians will be left off. Of course, if Jim refuses, then there is no special occasion, and Pedro here will do what he was about to do when Jim arrived, and will kill all of them. […] The men against the wall and the other villagers understand the situation, and are obviously begging him [Jim] to accept. What should he do?44

 For the sake of brevity I use the term “action” in the present context so as to include also omissions. 42  Unlike CSM*, the norms of consequentialism are agent-neutral. 43  Bernard Williams, “A Critique of Utilitarianism”, p. 99. 44  Ibid., p. 98 f. 41

2.4  Deontological Features in Common-Sense Morality

43

Williams notes that the utilitarian answer is that Jim should accept the offer. And he himself thinks that “the utilitarian is probably right in this case”.45 This is interesting, since it shows that the commitment to agent-relativity need not be total: if what other people will do unless you perform a wrong action is something very bad, you are permitted to do the (prima facie) wrong action. This is probably the view of CSM. In the other example invented by Williams, George, an unemployed chemist, is offered a job in a laboratory which pursues research into chemical and biological warfare. Here Williams’s answer is strictly deontological: George should not accept the job, although “if George refuses the job, it will certainly go to a contemporary of George’s who […] is likely to push along the research with greater zeal than George would”.46 Here, obviously, the alternative is not considered so bad that it outweighs the duty to keep one’s hands clean. The agent-relativity of CSM* and other deontological moralities has been criticized by consequentialist philosophers as irrational. Why, they ask, is it considered obligatory for an agent not to do a certain action if doing it would result in a better state of affairs, a state of fewer wrong actions being done? It has also been claimed that a person who subscribes to an agent-relative morality, and acts accordingly, is guilty of “moral self-indulgence”, with having a too great concern for his own moral integrity.47 But, as Bernard Williams points out, this kind of reflexive concern can, but does not have to, arise with respect to any moral motivation whatever, even a consequentialist one.48 You may act, for example, from the direct, first-order motives of honesty or generosity, but, of course, you may also act from the reflexive, second-­ order motives of concern for your own honesty and generosity. And the same is true of the adherent of an agent-relative morality: he may, of course, act out of concern for his moral integrity, but he may also be directly motivated by the conviction that the action in question is wrong. Whether agent-relativity can be justified will be discussed in the next section. But, whether or not it can, it is not difficult to explain why CSM* is agent-relative. If social morality is to work, its norms must be strongly internalized by the members of society. In order for people to resist temptations, the norms must be deeply held and combined with strong emotional reactions towards transgressing them. If the norms are to keep their required sanctity, it is not psychologically possible, or at least very difficult, for people to look at the norms in such a way that they are able to transgress them whenever the common good requires it. So even if in some situations the agent-relativity of the norms of CSM* leads to more transgressions of

 Ibid., p. 117.  Ibid., p. 98. Both of Williams’s examples are also examples of CSM*’s above mentioned recognition of constraints against harming people. Perhaps they might also be seen as illustrating the acts and omissions doctrine, to be discussed later in this section. 47  For such criticism, see Jonathan Glover, “It Makes No Difference Whether or Not I do It”, Proceedings of the Aristotelian Society, Supp. Vol. 49 (1975): 171–90. 48  Bernard Williams, “Utilitarianism and Moral Self-Indulgence”, in his Moral Luck (Cambridge: Cambridge University Press, 1981). 45 46

44

2  Deontological Ethics: Exposition

them than if they had been agent-neutral, it most likely leads to less transgressions on the whole—given that most people have internalized the norms. (iv) Special Relationships. A deontological feature of CSM* is its being, in C. D. Broad’s words, self-referentially altruistic—that is, that, according to it, we have special obligations to our near and dear. Whereas agent-favouring permissions allow us to do what is, from an interpersonal point of view, not for the best, selfreferential altruism may require us to do that in certain cases (see below). For, as Broad points out, CSM* does not hold that any of us has an equally strong obligation to benefit everyone whom he can affect by his actions. According to it, each of us has specially strong obligations to benefit certain persons and groups of persons who stand in certain special relations to himself. And these special relations to himself are the ultimate and sufficient ground of these specially urgent obligations. Each person may be regarded as a centre of a number of concentric circles. The persons and groups to whom he has the most urgent obligations may be regarded as forming the innermost circle. Then comes a circle of persons and groups to whom his obligations are moderately urgent. Finally there is the outermost circle of persons (and animals) to whom he has only the obligation of “common humanity.”49

Thus I should do things for members of my family, my friends, my neighbours, my countrymen, etc. that I am not obliged to do for other people. It was this doctrine that was challenged by the consequentialist William Godwin when (in the example rendered in Sect. 1.1) he claimed that you ought to save the famous Fénelon rather than your father or your brother if you cannot save both. (“What magic is there in the pronoun ‘my,’ that should justify us in overturning the decision of impartial truth?”) The explanation of the fact that CSM* holds the doctrine of self-referential altruism is, of course, that we are united with strong bonds of affection and common interest to certain other people; a morality that required that we should disregard these bonds and act in a universalistic spirit, treating all people on an equal basis, would go against the grain and would probably not be complied with by most people. This is often stressed by critics of consequentialism. Thus the eighteenth century philosopher Thomas Green writes: We cannot change our nature. By a law of that nature we proceed from personal affection to general regard: from the love of offspring, of kindred, of neighbours, and acquaintance, to that of our district, our community, our country and our kind. In this order our affections are diffused; and in this order, by the constitution of our being, they weaken as they spread.50

 C. D. Broad, “Certain Features in G. E. Moore’s Ethical Doctrines”, in Paul Schilpp (ed.), The Philosophy of G.  E. Moore (Chicago and Evanston, Ill.: Northwestern University Press, 1942): 43–67; p. 54 f. 50  Thomas Green, An Examination of the Leading Principle in the New System of Morals, 1798; quoted from D. H. Monro (ed.), A Guide to the British Moralists, p. 199. 49

2.4  Deontological Features in Common-Sense Morality

45

(There is, in turn, an evolutionistic (a socio-biological) explanation of these bonds and of our bent for binding ourselves with them—roughly that kin altruism and reciprocal altruism have survival value.51) The requirements of self-referential altruism make up a great deal of the requirements arising from special relationships. But there are also other requirements coming from such relationships. Thus, for example, CSM* prescribes that we should, if possible, benefit those who have benefited us, and that we should try to compensate those whom we have harmed. And, most important, we should keep our promises, even implicit ones. (v) The Acts and Omissions Doctrine. Like deontological theories, CSM* sees the distinction between acts and omissions, between doing and allowing, as morally relevant and of paramount importance. According to this view, it is, for example, morally wrong to do something with the result that a person dies, but it is not morally wrong—or, at least, not wrong to the same extent—to omit an action with the same result. Thus, for you to (intentionally) send poisoned food to a starving person picked out at random and thereby kill him is considered wrong, but not sending him food, with the result that he starves to death, is, if at all considered wrong, considered much less so. (It is assumed, of course, that there is no relation of kinship, friendship, etc. between you and the starving person.) On the whole you comply with social morality by not performing certain kinds of action. I think we all know sufficiently well what the distinction between acts and omissions amounts to.52 In order to render the acts and omissions doctrine a little more precise, however, we may use the notions of benefit and harm. To benefit someone, I stipulate, is to make him better off than he would otherwise have been, and to harm someone is to make him worse off than he would otherwise have been. The doctrine can then be taken to state that it is wrong (intentionally) to harm a person, but not wrong, or at least less wrong, (intentionally) not to benefit him. (This is a rough and not wholly accurate characterization, but it will do for the present purpose.) It should be noted that not benefiting a person covers two cases: (i) leaving him on the same welfare level as before and (ii) leaving him on a worse level than before. An example of the first case is: not giving someone an unexpected, wanted, but not needed gift; an example of the second case is: not warning someone of a threat. (Notice that this is a case of not benefiting, not a case of harming. For even if, because B did not warn him of some threat, A is harmed, he is not harmed by B.) It is understandable why the distinction between acts and omissions is considered morally relevant by CSM*: to comply with the norm not to harm another p­ erson  The, to my knowledge, best treatment of the socio-biological claims concerning CSM* is still the one contained in Peter Singer, The Expanding Circle: Ethics and Sociobiology (Oxford: Oxford University Press, 1981). 52  Excellent analyses of the distinction are given in Jonathan Bennett, The Act Itself (Oxford: Oxford University Press, 1995) and Alan Donagan, The Theory of Morality (Chicago: University of Chicago Press, 1977). 51

46

2  Deontological Ethics: Exposition

does not normally require too much self-sacrifice, but it is a very demanding task to benefit all those in need whom we are able to help, given the amount of need in the world. To comply with a norm to benefit all those whom we are able to help from dying an untimely death or living wretched lives is, for almost anyone of us, beyond what is psychologically possible. If, somehow, we should come to accept that the distinction between acts and omissions is morally irrelevant, it is hardly probable that we would come to believe that omissions are just as wrong as the “corresponding” acts are thought to be. Since our willingness to self-sacrifice is rather limited, we would not be able to live with this conviction. It is therefore far more likely that, should the relevance of the distinction somehow come to be eroded, we would come to hold these acts to be no more wrong than we now hold the corresponding omissions to be. That social morality, the morality actually in use, takes the acts and omissions distinction to be morally relevant is, therefore, what should be expected. This is also to be expected if, as I claim, CSM* subscribes to a contractual conception of morality: morality is a set of norms general compliance with which makes everyone better off. By and large, a ban on actions that harm people is in everyone’s interest, whereas a demand to benefit people is not. Therefore harming people is morally forbidden, whereas not benefiting them is not. For, however unequal a society is, anyone runs a risk of being harmed by other members of society. Wealth and power, although they make the infliction of certain kinds of harm on their possessors more unlikely, are no guarantee against that: even the King or the President is vulnerable to violence and runs the risk of being killed. Indeed, the well-off may run a greater risk than the not well-off of being victims of certain kinds of violence, such as theft, burglary, armed robbery, and blackmail, since it is more profitable to commit such actions against them. Almost everyone, therefore, wants there being generally accepted norms against doing harm. On the other hand, there being generally accepted norms for benefiting other people, for doing good to them, is not in everyone’s interest and are therefore not included in CSM*. There are, however, some exceptions to this, some kinds of benefiting actions which are considered obligatory by CSM*. (This was pointed out in Sect. 2.2.) But in these cases general acceptance of the norms prescribing these kinds of actions evidently is in everyone’s interest. There is, first, the case of promising. CSM* says that you ought to do as you, explicitly or implicitly, have promised: to be back at a certain time, to read my paper, to pay back the money I lent you, etc. (Moreover, the institution of promising is itself of a contractual nature. A demand that people fulfil their contracts is obviously in everybody’s interest, and if, as I claim, CSM* harbours a contractual conception of morality, it is to be expected that it contains a norm to the effect that one should honour one’s contracts.) In order to explain a second kind of cases where benefiting others are obligatory according to CSM*, I will first dispel a possible misconception. The reason why a general policy of benefiting other people is not in everyone’s interest is, it may be thought, that in many cases people do not stand in “reciprocal” relations to each other: many people who need the help of other people are not able to reciprocate their potential benefactors. But a contractual conception of morality need not be based on direct reciprocity. Acts and omissions, it should be noticed, are strategically different. Although it is usually in my interest not to harm you only on condition that

2.4  Deontological Features in Common-Sense Morality

47

you are disposed not to harm me, it may be in my interest to benefit you on condition that, not you, but others benefit me when I need help. This explains why, according to CSM*, we ought to benefit other people in certain kinds of situations if it is not too costly. If, for example, I have fallen into the water and will drown unless you save me, and you can save me without putting your life at stake, CSM* requires you to save me. It is certainly not very probable that I will ever be able to do the same for you, should you get into similar trouble, but other people probably will. Since we all run the risk of getting into emergency situations, everyone is—at least, probably— better off if there is a norm, generally complied with, which enjoins helping other people in such situations. But it is not in everyone’s interest to adopt norms of mutual help in many other situations. If, for example, I am diseased and do not have enough money to buy the medicine that I need to recover my health, while you are rich and could easily help me, it may not be in your interest to help me. Therefore you would probably not be better off if there were a norm, generally complied with, which required helping other people in such situations. (It is assumed that we are neither relatives nor friends.) For the probability that you would ever need such help is very low.53 This, I think, explains why CSM* does not prescribe that we benefit people in such situations. Generally, acts of benevolence are considered by social morality, not as acts of duty, but as acts of supererogation.54 Kant thought otherwise. In the Grundlegung he illustrates the working of the categorical imperative by means of four examples. In the fourth example we meet with a man who is himself flourishing, but he sees others who have to struggle with great hardships (and whom he could easily help); and he thinks ‘What does it matter to me? Let everyone be as happy as Heaven wills or as he can make himself; I won’t deprive him of anything; I won’t even envy him; only I have no wish to contribute anything to his well-being or to his support in distress!’55

But, says Kant, the categorical imperative states that the rich man has a duty to help those in distress. The duty is, in Kant’s terminology, not a “perfect” but an “imperfect” duty; there is, therefore, no contradiction in conceiving a world where this duty is generally neglected, but it is impossible to will that such a world existed. For a will which decided in this way would be in conflict with itself, since many a situation might arise in which the man needed love and sympathy from others, and in which, by such a law of nature sprung from his own will, he would rob himself of all hope of the help he wants for himself.56

 In some societies, the probability that you might need help in such cases is higher, and in these societies I may accordingly be required to help you. I take up this issue later in the present subsection. 54  The above explanation of the acts and omissions doctrine also explains the traditional view of positive and negative rights: the existence of negative rights is usually taken for granted, but the existence of positive rights is often called in question and, if acknowledged, not considered equally stringent. 55  Immanuel Kant, Grundlegung, Paton’s tr., p. 86. 56  Ibid. 53

48

2  Deontological Ethics: Exposition

Is Kant right on this point? Would the man in this example prefer a world where people generally helped each other to a world where they did not, or would he consider the second world at least as good as the first? I think that if he is rich and estimates the risk of ever becoming poor as very small, nothing speaks against the latter possibility. Like the rest of us he certainly needs love and sympathy from others, but not from everyone. It is enough if he gets it from his family and his friends, and why shouldn’t he? It is a well-attested fact that warm and affectionate relations within a small circle of friends and relatives can happily co-exist with total callousness to strangers. It would be otherwise in a society of both economic equality and scarcity, where (nearly) everyone runs the risk of suddenly being pressed below the level of subsistence. It is likely that in such a society (nearly) everyone would prefer that people in distress are helped rather than that they are not—the help presumably being effectuated by the combined efforts of sufficiently many other people—and in such a society a norm of helping each other would probably be adopted.57 Such societies are rare, however, since societies both tend to raise their standard of living and to become socially stratified, but within a society there may exist groups where a principle of mutual aid is adopted. (This is said to have been the case, for example, among the early Christians.) (To avoid misunderstanding, I want to stress that I do not claim that people generally lack compassion and sympathy with other people—only that it is rather limited. Nor do I claim that people never voluntary help (for them) unknown people in distress—the existence of organizations such as Amnesty International proves the opposite—only that such help is not very common and is not commonly regarded as acts of duty but as acts of supererogation.) The above explanation of why CSM* subscribes to the acts and omissions doctrine refers essentially to the contractual nature of CSM*. There is, however, another possible explanation of this fact, an explanation that refers to the structural difference between acts and omissions and between benefiting and harming. This explanation points out that you cannot do all the good things each of which, however, it is possible for you to do. You cannot, for example, provide food for all who will be starving to death in the near future, each of whom you are able to help. This means that none of these potential recipients has any legitimate claims on us. On the other hand, it is often possible for you to abstain from harming anyone. In this respect benefiting/not benefiting differs from harming/not harming. Suppose, for example, that you can either benefit A or B, but not both. Then neither A nor B has any legitimate claim on you: A has not been wronged if you choose to benefit B, and the same goes for B if, instead, you choose to benefit A. Moreover,  An example of this is given in Fridtjof Nansen’s description of the Greenlanders: “A hard life has taught the Eskimo that even if he is a skilful hunter and can, as a rule, manage to hold his own well enough, there may come times when, without the help of his fellows, he would have to succumb. It is better, therefore, for him to help in his turn.” (Fridtjof Nansen, The First Crossing of Greenland, 2 vols. (London: Longman, Green & Co., 1890), Vol. 1, p. 304 f.; quoted from Edward Westermarck, op. cit., Vol. 1, p. 560.

57

2.4  Deontological Features in Common-Sense Morality

49

it should be noted, neither A nor B has been wronged even if you choose to benefit none of them. Even if it is also in such cases considered better to benefit someone rather than none, it is understandable why in many cases CSM does not consider this as obligatory, as something you shall do, but as supererogatory, as something it would be good of you to do.58 It is otherwise with harming people. There are exceptions, of course, but normally we are able to abstain from harming other people. This means that, normally, when you harm a person, you wrong that person; you disregard his legitimate claim on you. In such cases, your harming someone is therefore severely held against you. By contrast, in those rare cases where, whatever you do, you will harm some person, your harming him is not similarly held against you. For in such cases you do not disregard anyone’s legitimate claim. Thus, given that wronging people, that is, disregarding their legitimate interests, is considered central to moral wrong-doing, it is understandable why CSM* treats (morally good) acts and omissions differently.59 (This explanation does not, of course, rule out the former explanation.) (vi) The Self/Other Asymmetry. A more puzzling deontological feature of CSM*, first discussed at length by Michael Slote, is a self/other asymmetry regarding what we are permitted to do.60 CSM* permits us to do (or omit) things to ourselves that we are not permitted to do (or to omit) to other people. For example, according to CSM*, it is morally permissible—though, perhaps, stupid or irrational—to sacrifice one’s own greater benefit to the (much) lesser benefit of another person. Slote writes: In the absence of some special relation or obligation to that other, common-sense might concede it was irrational, stupid or gratuitous to do so, but surely not that it was morally wrong. Similarly, in a situation where no one else is concerned (or even, if you will, where no one else exists) if I ignore an opportunity to enjoy a pleasure or do not bother to avoid a pain, then (other things being equal) I do wrong by consequentialist standards, but, again, not by ordinary standards.61

 Cf. my characterization of CSM* at the end of Sect. 2.6.  CSM and, more generally, D side with the negative version of the golden rule (see Sect. 4.4), whereas C sides with the positive version. 60  Michael Slote, Common-sense Morality and Consequentialism (London: Routledge & Kegan Paul, 1985). The asymmetry had not gone totally unnoticed before Slote put his finger on it. Both Henry Sidgwick (The Methods of Ethics 7th ed. (London: Macmillan, 1907), p. 431 f.) and W. D. Ross (Foundations of Ethics (London: Oxford University Press, 1939), pp. 72 ff., 272 ff.) comment on it. Sidgwick thinks that the asymmetry is, all things considered, compatible with utilitarianism, and Ross, who thinks that it is not, finds the asymmetry intuitively acceptable and therefore unproblematic. 61  Slote, op. cit., p. 10. It should be noted that some deontologists, e.g., Kant, think that, according to CSM, we have moral obligations to ourselves. This may be due to a lingering Christian influence: we are created by God, are his property, and therefore cannot deal with ourselves as we please. Consider how Shakespeare makes Hamlet burst out: “O, […] that the Everlasting had not fix’d/His canon ’gainst self-slaughter!” (Hamlet, Act 5, Scene 2, lines 129–32. 58 59

50

2  Deontological Ethics: Exposition

The agent-sacrificing permissions, as Slote calls them, thus conflict with the restrictions imposed by constraints (see Subsection (i) above). Consider the doctor’s choice. Although CSM* forbids the doctor to cut up the healthy person and distribute his organs to the five people in need of organ transplants, it does not forbid the doctor to cut up himself (or, more realistically, to have himself cut up) and then have his organs similarly distributed. The asymmetry, as Slote points out, is also reflected in language: benefiting oneself is not praised as benevolence, nor is suicide stigmatized as murder, or rough treatment of oneself called injustice.62 The agent-sacrificing permissions of CSM*, Slote remarks, are not only incompatible with the impersonal standpoint adopted by consequentialism; it is also, at least on the face of it, difficult to reconcile the self/other asymmetry of CSM*, with its self-referential altruism. On one hand, Slote says, CSM* encourages the idea that strength of obligations weakens as one gets further from the agent, but on the other hand, and in seeming opposition to the first idea, it assumes that there is no moral obligation whatever (except indirectly) for the agent to benefit himself or concern himself with his own projects. Once one leaves the agent himself behind [,] the agent’s obligations vary in proportion to his reasons for concern, but where he has greatest reason for concern in the natural course of things, he has no direct obligation whatever.63

Slote discusses three possible ways of explaining/justifying the self/other asymmetry. The first starts from a suggestion by G. J. Warnock. According to Warnock, the object of CSM* is to improve general well-being by countervailing our limited sympathies.64 Perhaps, then, the self/other asymmetry could be regarded as contributing to this object. But, as Slote points out, sometimes the agent can improve general well-being only by improving his own situation, but this is what the self/other asymmetry permits him not to do. Of course, in such situations the permission is in accord with (or, at least, not contrary to) the policy of countervailing our limited sympathies, but why should CSM enjoin us to follow that policy except as a means of improving general well-being? Another possible way of explaining the asymmetry considered by Slote focuses on the lack of consent involved in harming (or not benefiting) others but not in harming (or not benefiting) ourselves. If I intentionally harm myself, I consent to it, but if (whether intentionally or not) I harm another person, he presumably does not consent to it (and if he does, there seems to be nothing wrong with my action). But, Slote argues, the difference with respect to consent in the two cases does not explain the asymmetry. If someone irrationally asks me to harm or kill him, it will presumably be irrational and wrong of me to kill him, more wrong at any rate than if I irrationally choose to kill myself; yet the consent seems to be equal in the two cases.65

 But language is not always a reliable indicator; suicide is called “Selbstmord” in German and “självmord” in Swedish, both words meaning self-murder. 63  Ibid., op. cit., p.14. 64  G. J. Warnock, The Object of Morality (London: Methuen, 1976). 65  Slote, op. cit., p. 20. Slote gives two other examples, but they add nothing essential. 62

2.4  Deontological Features in Common-Sense Morality

51

It may be objected that if I choose to kill myself, I do not really consent to my doing it. “Consent” means “acquiescence to or acceptance of something done or planned by another”.66 So neither the term “consent” nor its opposite is applicable to what an agent does to himself. It is therefore not true that in Slote’s example “the consent seems to be equal in the two cases”. But Slote could admit this verbal point and instead use the phrase “what someone irrationally wants you to do to him”. He could then argue in the following way: There are situations such that, if someone else irrationally wants you to kill him, it is, according to CSM*, morally wrong for you to comply with his want. But, ceteris paribus, it is not morally wrong, or anyway less wrong, for you to comply with your irrational want to kill yourself. This is just considered stupid or irrational. So there is, according to CSM*, a self/other asymmetry concerning what we are permitted to do, which cannot be explained by (the distinction that is really intended when one argues in terms of) the consent/lack of consent distinction. In a later work, Slote seems to claim that he has found the correct CSM explanation of the asymmetry. But shouldn’t we have obligations to advance and not to damage our own prospects, our own happiness? Common sense (and again you can check for yourselves) has an answer ready for this question, and it is that it makes no sense to suppose there is an obligation to do things we are already inclined to do and can naturally be expected to do. Since we naturally and expectably do care for our natural interests, there can’t—there is no moral need for—an obligation to do so.67

But this explanation, Slote says, is incoherent with CSM being self-referentially altruistic. For, Slote argues, we are more inclined to help our near and dear than people we do not know. But in this case, CSM holds that we have a stronger obligation to help those whom we are already more inclined to help. “And that is the very opposite of what one should expect given the above rationale for the absence of an obligation to pursue and advance our own well-being.”68 I do not think that there is any incoherence involved here. For I do not think that Slote has found the correct explanation of the asymmetry. How, then, is the self/other asymmetry to be explained? I think that the answer is the following: CSM is, as another common name of it indicates, a social morality; its task is to regulate what we do to other people in society; what we do to ourselves lies outside morality—to the extent that our actions do not affect other people too. Small wonder, then, that we are morally permitted to do things to ourselves that we are not permitted to do to others. There is, however, no great risk that we will ignore our own interests since nearly all of us have a natural propensity to care for them. If, nevertheless, under the influence of moral or other considerations, we are tempted to make a considerable sacrifice of our interests, another normative principle, viz. prudence, comes into play, telling us that it would be irrational to do the contemplated action and that charity begins at home.  Collins Dictionary of the English Language (Glasgow: Collins, 1985); my emphasis.  Michael Slote, “Virtue Ethics”, in Marcia Baron, Philip Pettit, and Michael Slote, Three Methods of Ethics: A Debate (Oxford: Blackwell, 1997), p. 181. 68  Ibid., p. 182. 66 67

52

2  Deontological Ethics: Exposition

That CSM is a social morality—concerned with legislating what we may and may not do to other people—can, in turn, be explained by its contractual nature. We have an interest in the inculcation of a system of norms which gives a considerable protection against being harmed by other people, but we have no interest in the erection of an otherwise similar system that protects us against ourselves. In fact, we have an interest in not being thus protected.69 At least most people think they are sufficiently motivated and competent to take care of themselves and are willing not to morally stigmatize other people for acting against their own interests if other people are willing to follow suit with respect to them. And many people want themselves and, especially, other people to be able to do acts of moral supererogation— that is, sacrificing their own interests to those of other people—without being morally censured for that. Thus, unlike Michael Slote, I think that the self/other asymmetry of CSM can be explained. Whether it also can be justified depends on whether the contractual conception of morality inherent in CSM can be justified. This question will be discussed in Chap. 7. (vii) Truth-Telling. CSM* attaches great importance to truth-telling, that is, to not deceive people by lying to them. Like other requirements of CSM*, (pace Kant) the requirement not to lie has exceptions of different kinds. There are “white lies”; thus, for example, you are permitted to lie in order to mislead a prospective murderer concerning the whereabouts of his intended victim. But the built-in exceptions aside, lying is very strictly censured by CSM*. And this is not astonishing. If people could not trust each other to generally speak the truth, there would be no mutually beneficial co-operation and, hence, no society. (More about this in Sect. 7.5.) This concludes my survey of the deontological features in CSM*. But before I leave the topic, I will briefly show how these features are linked to the three categories of norms characteristic of CSM* that I mentioned in Sect. 2.2. Constraints simply describes the first category of the CSM* norms, those that forbid us to harm other people. Options follow from the fact that the range of benefits required of us is very limited: norms in the second category prescribe that we benefit other people only when the cost of doing this is small, and norms in the third category that we benefit our near and dear. So, given, the CSM* norms, we have ample time to pursue our private interests. Agent-relativity follows simply from the fact that the CSM* norms are agent-relative. And special relationships is just an amplification of the third category. The self/other asymmetry follows from the fact that no CSM* norm forbids the agent to harm himself. The acts and omissions doctrine says that it is wrong to harm a person, but not wrong, or at least less wrong, not to benefit him. And this is a direct consequence of what the norms of the first and the second categories state. Truth-­ telling, finally, is one of the norms in the first category. It is evident, then, that, with one exception, the CSM* norms manifest the deontological features. The exception is that special relationships is more detailed than 69

 This explanation is tentatively suggested by Shelly Kagan in Normative Ethics, p. 250 f.

2.5  Common-Sense-Moral Features in Deontological Ethics

53

the third category. (But this could easily be remedied.) Are the CSM* norms also the only set of norms that manifest the deontological features? I think so. It may be thought that a set of norms could manifest the deontological features, options included, although what was here optional was, not benefiting other people, but not harming them. But such a set of norms would not then manifest the acts and omission doctrine. So I think that the deontological features in CSM are very intimately linked to the three categories of norms characteristic of CSM.

2.5  Common-Sense-Moral Features in Deontological Ethics It is a reasonable assumption that moral speculation begins at home, in other words, that all moral theories have their origin in, and take their inspiration from, CSM. Where else could speculation start from, and how else could the theories get whatever plausibility they have? (Cf. the quotation from Frankena in Sect. 2.1 above.) Some moral theories, such as C and ethical egoism, take hold of more peripheral parts of CSM, while other moral theories, such as virtue ethics and, especially, D, fasten on its more central parts.70 (Conversely, moral theories have, no doubt, to some extent influenced CSM. Several of the differences between Victorian CSM, as found in Sidgwick’s account of it in The Methods of Ethics, and the CSM of Aristotle’s time, found in the Nicomachean Ethics, are probably due to the workings of the morality of Christianity on the former.) CSM* is, no doubt, the source of the deontological theories concocted by moral theorists. A study of the history of moral philosophy bears ample witness to the fact that deontological theories often have been propounded as direct responses to teleological (consequentialist) theories which explicitly or implicitly challenge CSM*.71 Moreover, deontological theories and CSM* to a large extent overlap: nearly all deontological theories set forth by moral philosophers contain norms that coincide with CSM*. In addition, the theories contain systematizations and justifications of the norms, justifications that ultimately appeal to our intuitions, that is, to the ­intuitions of CSM*. I think that the best explanation of this is that CSM* is the source of deontological theories, what gives them their moral appeal. Moreover, several deontologists explicitly admit that their purpose is not to change or add to CSM* but rather to defend it. Even Kant does not claim to do any It is not, perhaps, strictly true that ethical egoism has its origin in common-sense morality. Though common sense, no doubt, harbours the conviction that one ought (or, at least, is permitted) to fulfil one’s own interest, the conviction is not, perhaps, (properly) regarded as an unmistakably moral one. 71  To verify that this is true with respect to English-speaking philosophy from Hobbes onwards, the reader should consult the following anthologies: L.  A. Selby-Bigge, British Moralists (Oxford: Oxford University Press, 1897); D.  D. Raphael, British Moralists 1650–1800, 2 vols. (Oxford: Oxford University Press, 1969; D. H. Monro, op. cit. 70

54

2  Deontological Ethics: Exposition

thing else. Having stated (the first formulation of) the categorical imperative at the end of the first chapter of the Grundlegung, he adds: In studying the moral knowledge of ordinary human reason we have now arrived at its first principle. This principle it admittedly does not conceive thus abstractly in its universal form; but it does always have it actually before its eyes and does use it as a norm of judgement.72

The rest of the Grundlegung is devoted to a defence of this “moral knowledge of ordinary human reason”. And among modern deontologists, W. D. Ross, who is in many ways opposed to Kantian moral philosophy, is at one with Kant that CSM* is the starting point as well as the test of moral theories. In his opinion, what we are apt to describe as ‘what we think’ about moral questions […] forms the standard by reference to which the truth of any moral theory has to be tested, instead of having itself to be tested by reference to any theory. […] We have no more direct way of access to the facts about rightness and goodness and about what things are right or good, than by thinking about them; the moral convictions of thoughtful and well-educated people are the data of ethics just as sense-perceptions are the data of a natural science.73

In order to further justify my claim that CSM* is the source of the deontological theories worked out by moral philosophers, I shall consider three representative deontological theories, starting with that of Ross, and show that they exhibit (at least) most of the features characteristic of CSM*.74 (To show that all or most bona fide deontological theories exhibit these features would be tedious as well as pointless: we would not, I think, consider a theory deontological if it lacked most of the characteristics in question.75 But this is, of course, compatible with there actually being no genuine deontological theories. So giving some evidence for there being such theories may be appropriate.) (i) Ross’s Theory. Ross thinks that there are several possible characteristics of an action that are morally relevant. (He lists six such characteristics, but admits that the list may not be final.) Each characteristic gives rise to (what Ross calls) a prima facie duty, and, together, the prima facie duties determine (in a loose sense) what an agent should do in the situation in which he finds himself. How is this done? Here is Ross’s explanation:

 Immanuel Kant, Grundlegung zur Metaphysik der Sitten, 404; the quotation is from H. J. Paton’s translation, The Moral Law (London: Hutchinson, 1948), p. 69. 73  W. D. Ross, The Right and the Good (London: Oxford University Press, 1930), p. 40 f. 74  In the previous section, I called these features “deontological”. And it might seem to be a trivial task to show that deontological theories manifest deontological features. But it is not trivial that deontological features manifest most of the features characteristic of CSM*, and this is what I will try to show. Notice also that the fact that a theory does not manifest a deontological feature does not, of course, mean that it does not acknowledge it; the feature may just be taken for granted. 75  According to David McNaughton and Piers Rawling, the basic distinguishing features of deontology standardly fall under three rubrics, viz. constraints, options, and duties of special relationships. See their “Deontology” in David Copp (ed.), The Oxford Handbook of Ethical Theory (New York: Oxford University Press, 2006): 424–58. 72

2.5  Common-Sense-Moral Features in Deontological Ethics

55

When I am in a situation, as perhaps I always am, in which more than one of these prima facie duties is incumbent on me, what I have to do is to study the situation as fully as I can until I form the considered opinion (it is never more) that in the circumstances one of them is more incumbent than any other; then I am bound to think that to do this prima facie duty is my duty sans phrase in the situation.76

Ross’s moral theory recognizes most of the features characteristic of CSM*. One exception is that Ross does not recognize options, since there is, he says, a duty of “beneficence”.77 Having established this, he continues: [E]ven when we have come to recognize the duty of beneficence, it appears to me that the duty of non-maleficence is recognized as a distinct one, and as prima facie more binding. We should not in general consider it as justifiable to kill one person in order to keep another alive, or to steal from one in order to give to another.78

Thus morality according to Ross embraces both constraints and the acts and omissions doctrine: you should not harm people, and harming people is more wrong than not benefiting them. Among the duties recognized by Ross are those of “fidelity” (promise-keeping and truth-telling), “reparation” (compensating those whom we have wronged), and “gratitude” (benefiting those who have benefited us).79 So truth-telling and duties of special relationships make up an important part of morality according to Ross. And in the Foundations of Ethics, Ross wholeheartedly endorses the self/other asymmetry. But in his earlier The Right and the Good he is more cautious: “while we clearly recognize a duty to produce pleasure for others, it is by no means so clear that we recognize a duty to produce pleasure for ourselves”.80 Ross does not address the question whether his moral theory is agent-relative. But since agent-relativity is the default position, his silence on this question may perhaps be taken as tacit consent. It may be thought, however, that the fact that the theory includes the duty of beneficence makes it non-relative. Suppose, for example, the objection says, that, unless A tells a lie, B and C will tell lies, and that the outcome of A’s lying (hence B’s and C’s not lying) contains much more (non-moral) good than the outcome of A’s not lying (hence B’s and C’s lying). Then Ross’s theory would presumably tell A to lie. Thus, the objection concludes, according to Ross’s theory, it might be the case that someone should do a wrong action in order that two wrongs actions not be done. And this is obviously to reject the principle of agent-relativity. I think that in this case Ross’s theory would tell A to lie. (At any rate, since, according to Ross, both the duty of truth-telling and the duty of beneficence are (not lexically ordered) prima facie duties, there must be some, actual or possible, situation where the latter duty takes precedence over the former.) But the reason is not  Ibid., p. 19.  Ibid., p. 21; cf. p. 39. 78  Ibid., p. 22. (Here, I take it, “prima facie” means intuitively.) 79  Ibid., p. 21. Ross puts truth-telling and promise-keeping together since he considers the former as resting on an implicit promise. 80  Ibid., p. 24. 76 77

56

2  Deontological Ethics: Exposition

that, by doing a wrong action, A would minimize the performance of wrong actions; the reason is that (i), by doing the wrong action, A would (indirectly) produce more good than by not doing it, and (ii) in this case the prima facie duty of beneficence takes precedence over the prima facie duty of not lying. If the outcome of A’s lying (hence B’s and C’s not lying) did not contain (much) more good than the outcome of A’s not lying (hence B’s and C’s lying), there is no reason to think that Ross’s theory would tell A to lie. There is therefore no reason to think that the theory is not agent-relative. (ii) Alan Donagan’s Theory. Another deontological theory is the one propounded by Alan Donagan.81 Donagan considers his theory to be an exposition of what he variously calls “the Hebrew-Christian moral tradition”, “the traditional morality of the Western world”, or just “common morality” or “traditional morality”. This morality, we are told, acknowledges both constraints and options: [E]vil is not to be done that good may come of it. The evil that is not to be done is the violation of the prohibitory precepts—the precepts of perfect duty; and the good for which it is not to be done, even though common morality requires that it be promoted, is the well-being of oneself and others as human.82

And: [E]xcept for special institutional duties, for example those of a parent, to promote the well-­ being of others at the cost of one’s life or fundamental well-being would be supererogatory.83

Traditional morality, Donagan emphasizes, is agent-relative. As for possible bad consequences of acting according to such a morality, Donagan is sanguine: “the nature of traditional morality is such that observing it cannot, except by unpredictable accidents, have calamitous consequences.”84 Commenting on Jim’s choice, Donagan advises Jim not to accept the offer of killing one of the Indians. Jim, Donagan suggests, should reason like this: Well, the whole thing may be a particularly nasty hoax—the gun he [the commanding officer]’ll give me may be loaded with blanks. Or he may intend to murder me. Or he may think to make me an accomplice: if I should kill one prisoner, he could kill the others with impunity; for my testimony would incriminate myself as well.85

Donagan does not acknowledge any self/other asymmetry since he thinks that the fundamental principle of common morality is “to respect every human being,

 Alan Donagan, op. cit.  Ibid., p. 155. (A perfect duty “is simply a duty not to do, or not to omit, an action of a certain kind: not to murder, [n]ot to lie, not to omit to pay a debt” (ibid., p. 154). This quotation also illustrates Donagan’s adherence to the acts and omissions doctrine. 83  Ibid., p. 86. 84  Ibid., p. 206. 85  Ibid., p. 208. 81 82

2.5  Common-Sense-Moral Features in Deontological Ethics

57

oneself or any other, as a rational creature”.86 But he emphasizes the importance of promise-keeping and truth-telling.87 The latter, Donagan thinks, is especially important since it rests on the respect due to other people as rational creatures. (iii) Charles Fried’s Theory. The last theory to be considered is the one set forth by Charles Fried.88 Unlike Ross and Donagan, Fried expressly calls the theory he defends “deontological”. With one exception, viz. the self/other asymmetry, the theory displays all the features characteristic of CSM* mentioned in the previous section. (However, as we shall see, Fried does not fully subscribe to the acts and omissions doctrine.) Thus the theory recognizes constraints: “When we say that one must not grievously harm an innocent person, that one must not lie, these are categorical prohibitions in the sense that (within limits) no amount of good can justify them.”89 Likewise the theory admits of options. For, says Fried, categorical norms do not take up the totality of moral space. After having avoided wrong and done one’s duty, an infinity of choices is left to be made. Indeed it is consequentialist systems like utilitarianism which are oppressive in the totality of the claim they make on moral agents.90

As we have seen, lying is categorically forbidden. So is also the breaking of promises. Actually, Fried thinks that lying is a kind of promise-breaking, the reason being that if I assert that p, I thereby assure you that p is true; lying is therefore a breach of trust.91 As for the existence of duties of special relationships Fried is quite outspoken: Now, I believe that to condemn any preference for particular persons as morally invalid, to require us to treat all human beings with strict partiality, would leave the very concept of human welfare with a remarkable poor content: there would be no room for concern based just on the fact of relationships with others.92

Fried also subscribes to agent-relativity, taking lying as an example: Since it is not consequences or states of the world as such that are proscribed, it is irrelevant that we might be causing, risking causing, or failing to avoid the same consequences (such as false beliefs) as tend to be caused by the forbidden action. Since we are not enjoined to minimize false beliefs at all costs but only not to lie at all costs, we can honour the prohibition and still have ample room left over for other things.93

As I said, Fried does not fully subscribe to the acts and omissions doctrine, although at one place he says that his theory “distinguishes between the evil which a moral  Ibid., p. 66. As I said in Sect. 2.4, such convictions may be due to Christian influence. And, as I mentioned above, Donagan considers his theory to be an exposition of the Hebrew-Christian moral tradition. 87  Ibid., p. 92 ff. and p. 88 ff., respectively. 88  Charles Fried, Right and Wrong (Cambridge, Mass.: Harvard University Press, 1978). 89  Ibid., p. 11. 90  Ibid., p. 13. 91  Ibid., p. 67 f. 92  Ibid., p. 36. Among special relationships, kinship and friendship are mentioned (ibid., p. 173). 93  Ibid., p. 17. 86

58

2  Deontological Ethics: Exposition

agent does and the evils which he allows to occur”.94 The reason why he does not fully subscribe to the doctrine is, Fried says, that “the notion of doing as opposed to allowing to happen […] is vague and requires further specification”.95 This specification, Fried tells us, is provided by the distinction between intended and merely foreseen consequences of one’s actions. And, he continues, this distinction is formalized in the so-called law of double effect, by which certain bad results—for example, the death of an innocent person, the sinful act of another—may not be “directly” willed, that is, chosen either as one’s ultimate goal or as the means to one[’s] goal, though they may be tolerated in appropriate circumstances as the foreseen concomitants of one’s chosen means or ends.96

The doctrine of double effect is certainly a deontological principle.97 But it is not, I think, a component of CSM. That is why it does not figure in the list of features characteristic of CSM* propounded in the previous section. (I discuss the doctrine of double effect, as well as the acts and omissions doctrine, in Sect. 5.3 below.)

2.6  Common-Sense Morality and Deontological Ethics In Sect. 2.2, I said that in the following “CSM*” will refer to a morality consisting of the three categories of norms that I listed there. I also said that these are the norms that are common to the CSMs of different societies. I cannot, of course, prove that this empirical claim is true: it might therefore be the case that the CSM of some society or societies does not contain all of these norms. If this were the case, I would stick to the first characterization and accordingly weaken the empirical claim. So, to repeat, “CSM*” will henceforth refer to a morality that contains norms which: (i) forbid harming other people (at least members of one’s own society); (ii) enjoin benefiting other people (at least members of one’s own society) when the cost to the benefactor is small—especially if the gain to the beneficiary is comparatively great98; (iii) enjoin benefiting our near and dear.99

 Ibid., p. 1.  Ibid., p. 21. 96  Ibid., p. 21 f. The doctrine (law, principle) of double effect is not, as Fried claims, a specification of the acts and omissions doctrine. If, for example, I intentionally omit an action, I may be acting wrongly according to the former doctrine but not according to the latter. At other places, e.g., at p. 52, Fried implicitly recognizes that the doctrines do not come to the same thing. 97  The doctrine of double effect is often traced back to Aquinas. But, according to Jonathan Bennett, this attribution is not correct: “The history of the principle of double effect, indeed, is clear only back into the nineteenth century.” (Op. cit., p. 200.) Among modern defenders of the doctrine are G. E. M. Anscombe, in “Modern Moral Philosophy”: 26–42, and Thomas Nagel, in The View from Nowhere (Oxford: Oxford University Press, 1986). The doctrine is scathingly criticized by Bennett (op. cit., p. 196 ff.) 98  More costly benefits are normally considered supererogatory, that is, praiseworthy but not obligatory. 99  In Sect. 2.4, i showed that this set of norms manifest the deontological features described there. 94 95

2.6  Common-Sense Morality and Deontological Ethics

59

I think that this stipulative definition captures pretty well what moral philosophers usually mean by “common-sense morality”. In Sect. 1.2, I characterized D in terms of what properties, according to deontological principles, make an action right, etc. Since this way of characterizing D has been objected to, I said that I would later propose a characterization that is not vulnerable to the objection. The alternative characterization runs as follows: A morality is deontological if, and only if, it (i) contains the norms of CSM*, or norms which are close to those of CSM*, and (ii) exhibits at least most of the deontological features characteristic of CSM*. This definition is, admittedly, somewhat vague—due to the phrase “close to”. It certainly includes CSM* itself, as well as the moral theories described in the previous section, and it excludes all forms of consequentialism. But it may be thought that there are moral theories which—since it cannot non-arbitrarily be decided whether its norms are, or are not, “close to” CSM*—are neither included nor excluded. I admit that this is possible in theory. I think, however, that, due to the intimate connection between the norms and the deontological features of CSM*, this will not happen: if a theory exhibits most of CSM*’s deontological features, then its norms are sufficiently close to CSM*. As we saw in the previous section, although different versions of D have much in common, they also differ from each other to some extent. Since I cannot possibly deal with all of them, I must make a choice. For the time being, I choose CSM*, the deontological morality par excellence, to represent D. (Later, in Sect. 3.5, I will reject CSM* in favour of a version of D that I call D*.) CSM*, as I have repeatedly said, is the point of departure of the other deontological moralities, the centre around which the other versions group. It is also, to a large extent, incorporated into the other versions. So in many ways it is suited to represent this kind of morality.

Chapter 3

Deontological Ethics: Assessment

In Chap. 1, I claimed that CSM is the archetype and source of D. In Chap. 2, I tried to characterize and explain this central version of D, especially what I called its “deontological features”; I also tried to identify the core of CSM, which I baptized CSM*. In this chapter, I shall begin the task of critically evaluating CSM* (a task to be continued in Chap. 7). In the first section, I discuss whether the deontological features can be justified. My answer is that they can—provided that “morality” is collective (see Sect. 2.3 above). (Whether it is will be discussed in Chap. 7.) In Sect. 3.2, I take up some traditional objections to CSM*, as well as a non-traditional one provided by Shelly Kagan. In Sects. 3.3 and 3.4, I discuss two objections which threaten CSM* with being, respectively, self-defeating and irrational, one objection emanating from Derek Parfit and the other from J. J. C. Smart. In the fifth section, I suggest one way in which CSM should be more limited, and one way in which it should be less limited, than what the contractual nature of CSM* allows. Since the result of the latter change is a moral code that is somewhat different from CSM*, I think it should have a name of its own. I dub it “D*”. (In the rest of the book it is D* that will confront C*, the most defensible version of C.) I conclude the chapter with a summary of what has gone before.

3.1  Can the Deontological Features Be Justified? In Sect. 2.4, I tried to show why CSM* has the several deontological features mentioned there. (I will not repeat the details.) The underlying reason, I claimed, is the contractual nature of CSM*: (nearly) every one of us prefers our common morality to have these features rather than not having them. Thus we want to have a morality that can be strongly internalized and be complied with. This explains options, agent-­ relativity, special relationships, and the acts and omissions doctrine. We further want morality to tell us what to do and not to do to other people, but not to take us to task © Springer Nature Switzerland AG 2019 J. Österberg, Towards Reunion in Ethics, Philosophical Studies Series 138, https://doi.org/10.1007/978-3-030-12410-6_3

61

62

3  Deontological Ethics: Assessment

for what we do and not do to ourselves. This explains the self/other asymmetry. We also want people to be brought up according to a morality that proscribes harming other people. This explains constraints. And we certainly want our morality to stress the importance of truth-telling—the last feature on the list—since this is a condition of mutually beneficial co-operation and, hence, of the very existence of society. Can a morality having these features really be justified? That our morality forbids harming other people and acknowledges the importance of truth-telling hardly needs further justification. But what about the other features? Don’t they show the hall-mark of a compromise with human frailty, with our deep-seated unwillingness to impartially pursue the good? It is understandable that biased and not very benevolent people prefer to regulate their lives by a system of norms that mirror their moral shortcomings, but this, it may be thought, cannot justify the norms. I think that the deontological features of CSM* can be justified. We need a normative system, a social morality general compliance with which enables us to live in peaceful and fruitful co-operation. This is also what nearly everyone wants. In order for social morality to do its work, it must be complied with by (almost) everyone. But then it cannot be too demanding. If it is to be complied with by most people, it must take the nature of Homo sapiens into account. It cannot therefore require that people do all the good they can (vide options and the acts and omissions doctrine), that they do not care more for their near and dear than for strangers (vide special relationships), or that they are not free to do to themselves what they are forbidden to do to other people (vide the self/other asymmetry). And, given that people are the way they are, a morality that required that people should sometimes do what is normally taken to be wrong would in all probability be less effectively internalized by the members of society (vide agent-relativity). The justification of the deontological features in question thus starts with the desirability of having a common morality enabling people to successfully co-­ operate. It then argues that, people being the way they are, their morality must have those features in order to do its job. In this way social morality, both in its function and its content, resembles law. I fully agree with P. F. Strawson when he says that the nature of social morality cannot be properly understood without some consideration of its relationship to law. It is not merely that the spheres of morality and law are largely overlapping, or that their demands often coincide. It is also that in the way law functions to give cohesiveness to the most important of social groupings we may find a coarse model of the way in which systems of moral demand function to give cohesiveness to social groupings in general.1

But the resemblance to a legal system is just what some people find wrong with social morality. Having internalized social morality, people thereby have come to take the moral point of view, which involves caring for other people. And, if people matter, critics may ask, why ought we not to benefit them, not only abstain from

1  P. F. Strawson, “Social Morality and Individual Ideal”, Philosophy, 36 (1961): 1–17; quoted from the reprint in G. Wallace and A. D. M Walker (eds.), The Definition of Morality (London: Methuen, 1970): 98–118; p. 117.

3.2  Common-Sense Morality and Its Critics

63

harming them? Especially options and the acts and omissions doctrine may, on reflection, seem contrary to reason. To this a defender of social morality will point out that this morality says that it is a good thing to benefit other people who are badly off. But, with the exception of emergency cases (see Sect. 2.2), such actions are supererogatory, not actions that you are required to do, in the sense that you act morally wrong and are blameworthy if you do not do them. And the reason is, to repeat, that such a requirement would be too strong; you cannot expect people to comply with it. Consequentialism is more exacting in this respect. C tells you to do much more in the way of benefiting other people; for example, in certain cases you even ought to benefit people who are not badly off. But the question is whether this demand can be justified. Another question is whether the “ought” used by C really is different in meaning from the term “a good thing” used by CSM with respect to supererogatory actions. (I will discuss these questions in Chap. 4.) CSM* does not, of course, forbid anyone to adopt moral ideals of a strict kind; you are free to aspire to the status of a saint or a hero, or whatever. But, CSM* will insist, the ideal in question should not conflict with CSM*, and it should not be advocated as a moral requirement pertaining to everyone.

3.2  Common-Sense Morality and Its Critics In comparison with C, CSM*, as well as D in general, is not much critically discussed, and most criticism is made in passing and is seldom very detailed. One possible explanation of this is, as I remarked in Sect. 2.1, that those who are critical of CSM* find it obviously wrong and think it is a waste of time and effort to go into the matter in detail.2 It is also possible that those who reject CSM* abstain from criticism because they find CSM* rather indeterminate—it is not always easy to tell what it prescribes in controversial cases—and refrain from taking it to task since they are uncertain whether they have a case. (But would not this alleged fact itself be the ground of an objection?) A third possible explanation why CSM* is not very much questioned is that we are all brought up with it. Therefore, even when as adults we want to critically assess it, it is difficult to break the spell and reject the moral intuitions that we imbibed early in our lives. Nevertheless, CSM* has been the object of some criticism. In this and the following sections, I shall discuss (what in my opinion are) the most important of the objections that have been raised.3 2  I want to repeat something that I said in Sect. 2.4: Sometimes when I let other people refer to CSM, I talk as if I thought that they have CSM* in mind. But this is only a simplified way of expressing myself. What I mean is that what they say or think about CSM applies to CSM*. 3  In a way, my discussion of objections raised against CSM has already begun and will continue in Chap. 7. In Sect. 2.4, I tried to explain the presence of the deontological features in CSM as ultimately due to CSM being a contractualist morality. I also said that the presence of these, often criticized, features is justified if, and only if, the “true” morality is contractualist, a question that is the topic of Chap. 7.

64

3  Deontological Ethics: Assessment

CSM* has been criticized from different points of view. Thus, for example, when amoralists or immoralists, such as Thrasymachos (in Plato’s Republic) or Nietzsche, criticize “morality”, they often have CSM* in view, although their criticism applies to many other moralities too. In the following I will only discuss those objections against CSM* that do not equally affect consequentialism. Not surprisingly these have most often been raised by people of a consequentialist bent. (i) Different CSMs contradict each other. It is not only the case that CSM varies from one society to another; the CSMs of different societies sometimes contradict each other: what is right according to the CSM of one society is sometimes wrong according to the CSM of some other society. Thus, for example, killing handicapped infants was permitted in ancient Greece and Rome but is considered wrong in most societies of today. And the slave system has been allowed in many societies all over the world but is now almost universally considered morally wrong. It is not only that we consider these practices wrong in our societies, we think they are wrong everywhere. Such examples, critics claim, show that the deliverances of CSM always might be, and in some cases actually are, invalid.4 (For, if according to the CSM of society S it is wrong to do A, but according to the CSM of society S′ it is not wrong to do A, then at least one CSM is so far mistaken.) If it were objected that the CSMs of (many) other societies are, or might be, mistaken, but certainly not our own, the obvious reply would be that such a smug and conceited reaction is just what is to be expected: this is what the great majority in all societies have always believed of their own morality—due to the efficacy of their moral upbringing. Sometimes coupled with the argument from contradiction is an objection to the effect that CSM is partial to certain social groups, perhaps to different groups in different societies. The most elaborate version of this claim is, of course, the Marxist one: social morality belongs to “ideology” and therefore always promotes the interests of the ruling class, those controlling the means of production. Thus Marx writes that in the capitalist era morality, along with law and religion, are with respect to the proletarian “so many bourgeois prejudices, behind which lurk in ambush just as many bourgeois interests”.5 An adequate assessment of this claim would involve me in a long digression. Luckily, there is no need for that, since, according to Marx, C (utilitarianism) is in the same boat. In one of his diatribes against utilitarianism, Marx says: The apparent stupidity of merging all the manifold relationships of people in the one relation of usefulness, this apparently metaphysical abstraction arises from the fact that, in modern bourgeois society, all relations are subordinated in practice to the one abstract monetary-commercial relation.6 4  See, e.g., Gilbert Harman’s discussion in his and Judith Jarvis Thomson’s, Moral Relativism and Moral Objectivity, Ch. 1. 5  Karl Marx, The Communist Manifesto, quoted from Saxe Commins and Robert Linscott (eds.), The World’s Great Thinkers, Vol. 3: Man and the State; tr. Samuel Moore (New York: Random House, 1947), p. 499. 6  Karl Marx, The German Ideology, quoted from David McLellan (ed.), Karl Marx: Selected Writings (Oxford: Oxford University Press, 1977), p. 185. (It is not likely that Marx would have been more lenient toward other forms of C than the Benthamite one which is his primary target in the quoted passage.)

3.2  Common-Sense Morality and Its Critics

65

Let us return to the argument from contradiction. I think that the argument is to some extent valid. The mere circumstance that CSM varies from one society to another does not, of course, pose any problem. As has often been pointed out, this variability can often be accounted for by the fact that people in different societies live under different material conditions and/or have different (empirical, religious, etc.) convictions; even if people belonging to different societies have the same basic moral beliefs, it is only to be expected that in many cases their derived moral beliefs differ. Thus it is a well-known fact that certain tribes consider it morally permitted to leave their elderly members to die when moving away to other dwelling-places, whereas such treatment of elders is considered morally wrong in our society. But the tribes in question live under conditions quite different from ours; in their case, bringing elders with them would slow down the pace of their movement, thus causing general starvation and perhaps the extinction of the tribe. An illustration of the second possible account is furnished by a tale told by Herodotus. The Persian king Darius asked members of a certain tribe who ate their dead and members of another tribe who burned their dead what they thought about the practice of the other tribe. As might have been expected, they considered it morally wrong. But, no doubt, they all agreed that one should show respect to the dead and that their own practice was the way of doing this. It is therefore likely that their moral disagreement concerning the proper treatment of the dead was due to different religious convictions.7 But not all such differences between moral codes can be similarly explained, much less justified. Consider, for example, the different opinions concerning the moral legitimacy of the slave system. According to the CSMs of some societies, the slave system is morally wrong, according to those of other societies, it is not. It is difficult to see how this difference could always be explained by the putative fact that members of the two kinds of societies live under different material conditions and/or have different (relevant) non-moral beliefs. And, even if it could be thus explained, it is still more difficult to see how both convictions as to the moral legitimacy of the slave system could be justified. For, as I said above, the convictions in question are not always of a relativistic kind: according to the CSM of, for example, our society, the slave system is not only wrong for our society, but for every society. CSM is thus sometimes invalid. This is, of course, more easily recognized when the CSMs of different societies contradict each other. But, as far as we know, CSM might be (what may be called) systematically invalid: every society might, for example, harbour norms that are partial to those in power. It is, however, difficult to see that this is true of CSM*. The norms belonging to CSM* are such that their general observance by those to whom they are directed is in the interest of each one. CSM* may, of course, be criticized for not requiring more than this, but that is another question.

7  Herodotus, Histories, 3:38; The Loeb Classical Library, ed. and tr. A.  D. Godley, Vol. 2 (Cambridge, Mass.: Heineman, 1963 (1921)), p. 51.

66

3  Deontological Ethics: Assessment

(ii) CSM lacks ultimate principles. A conspicuous feature of CSM is its heterogeneity and its concrete character: CSM issues a set of rather concrete norms which are seemingly unconnected with each other. According to CSM it is, for example, wrong to tell lies, and it is also wrong to steal, but what is the connection between these norms? On what ground, it may be asked, are both kinds of actions forbidden? The concrete character of the norms militates against their being themselves ultimate, and their heterogeneity militates against their being derivable from some underlying principle. (Nor does CSM itself point to any ultimate principles.) But, the objection goes, without any plausible ultimate principle or principles, the norms of CSM appear arbitrary and gratuitous. The defender of CSM might answer that there exist deontological moralities worked out by moral philosophers which are intended to remedy this shortcoming of CSM. To this the critic might reply that these attempts do not succeed: the heterogeneity is perhaps lessened but it is still there. Thus, in W. D. Ross’s system the (prima facie) obligation to keep promises is not further justified but taken as an ultimate principle. But, as J. J. C Smart says, “the obligation to keep promises seems to be too artificial, to smack too much of human social conventions, to do duty as an ultimate principle”.8 Admittedly, the critic continues, there is one moral system that attempts to give CSM an ultimate foundation, viz. that of Kant. But it is, to say the least, a controversial issue whether Kant really succeeds. Even some of his sympathisers have their doubts about that.9 A moral particularist (see Sect. 1.4) need not worry about CSM’s lack of ultimate principles. There are no moral principles, he contends, only several prima facie right- and wrong-making features which cannot be subsumed under any more general principles. A radical proponent of moral particularism goes further, claiming that there are indefinitely many right- and wrong-making features and/or that such features are context-dependent: what is a right-making feature in one context might be a neutral or even a wrong-making feature in another context.10 It is not always clear whether the particularist’s claim is to be taken as a normative claim concerning the shape of true morality, or (also) as a descriptive claim concerning the shape of CSM.11 Taken in the latter way—and it is this interpretation that is of interest in the present connection—I think the claim is false, at least as far as CSM* is concerned. In Sect. 2.2, I argued that norms belonging to CSM* can be understood in a contractualist way, as solutions to PD-situations. If this is correct, then CSM* contains after all an implicit ultimate principle, a principle which can be formulated as follows:

8  J.  J. C.  Smart, “An Outline of a System of Utilitarian Ethics”, in J.  J. C.  Smart and Bernard Williams, Utilitarianism: For and Against, p. 6. 9  See, e.g., Onora Nell, Acting on Principle: An Essay on Kantian Ethics (New York and London: Columbia University Press, 1975). 10  See, e.g., Jonathan Dancy, Moral Reasons. 11  I interpreted particularism as a normative claim in Sect. 1.4.

3.2  Common-Sense Morality and Its Critics

67

An action is prima facie wrong if, and only if, it violates a norm which (i) is a solution to a PD-situation and (ii) is (generally) complied with by those who are (would be) affected by the action; otherwise it is prima facie right. This is not to deny that there is some truth in moral particularism. According to CSM*, a feature might be right-making in one context and wrong-making in another. Although lying is usually considered wrong, it is sometimes considered right: pace Kant, you should lie to the man who plans to murder your neighbour when he asks you whether the latter is at home.12 Likewise, CSM* holds harming other people to be in general wrong, but there are exceptions even to this; thus CSM* most probably considers it permissible, perhaps even obligatory, for Jim to kill the Indian (see Sect. 2.4). But I think that all such exceptions from the general case can be explained by the contractual nature of CSM*. (Whether this amounts to a justification is the topic of Chap. 7.) (iii) CSM is an alienated morality. At least to most people morality appears as a set of requirements to which the individual must conform. Not seldom the requirements command him to do or omit what he strongly wants. A person who conforms to morality will therefore often frustrate his deepest desires; if, however, he yields to temptation, he will afterwards experience strong feelings of guilt and anxiety. In either case he will be unhappy. Thus the moral institution adds to the natural calamities that befall humankind. The above is a, slightly exaggerated, statement of a criticism of CSM that has been raised by several “enlightened” intellectuals of different persuasions during the last 200 years or so. (Often the criticism has been specifically directed to the alleged repressive nature of sexual morality.) An elaborated version of this criticism was expressed by Freud in his Civilization and Its Discontents.13 The target of Freud’s criticism is what he calls the “super-ego”, a term roughly synonymous with “conscience”. According to Freud, the super-ego is formed by the child’s internalization of the commands issued by his parents. As the child comes to identify with his parents, he experiences the commands as having their source within himself. Since the commands of the super-ego often conflict with his natural desires, the child feels anger when these have to be frustrated, and since the anger cannot be directed against the respected commands, it is directed to the sinful desires. If, however, the desires win out, or are even imagined to do that, the child feels the guilt of bad conscience, a trace of the (explicit or implicit) threat with which his parents accompanied the issuing of the commands. The child is father of the man, so this is how the super-ego confronts the adult person too. The basic fault with traditional (common-sense) morality, it might accordingly be thought, is its alienated and, therefore, mysterious character. Just because its  Kant, “On a Supposed Right to Lie from Altruistic Motives”, in Immanuel Kant: Critique of Practical Reason and Other Writings in Moral Philosophy, ed. and tr. Lewis White Beck (New York: Garland, 1976 (1949)). 13  First published in 1930, it is included in Vol. XXI of The Standard Edition of the Complete Psychological Works of Sigmund Freud (London: The Hogarth Press, 1953–64). 12

68

3  Deontological Ethics: Assessment

workings are hidden from and unknown for us, we are in the grip of disastrous, sometimes mentally crippling, feelings stemming from the super-ego. The hidden workings of our morality, it might further be claimed, also militate against our ideals of being rational and having self-knowledge. Isn’t it obvious that traditional CSM should be abolished and replaced by a rational and non-alienated morality? However, Freud himself thought otherwise: the alienated morality of the super-­ ego is the price we have to pay in order to live peacefully together in society, and, all in all, it is worth the price. Freud’s view of human beings is no less pessimistic than that of Hobbes. Throughout history, he claims, we find that human beings have been very aggressive towards each other. This aggressiveness, Freud thought, is constitutional, and there is therefore no hope that social changes would eradicate what Freud called the “death instinct”. To a certain extent, however, the inborn aggressiveness can be checked by a moral upbringing of a traditional kind, and, all things considered, the negative effects of moral upbringing are in Freud’s view preferable to the social chaos that would otherwise obtain. Some followers of Freud have been less pessimistic than Freud himself concerning the possibility of combining a non-alienated morality with a tolerably peaceful society.14 A contemporary philosopher who thinks this is possible is Richard Norman. Norman thinks that such a non-alienated morality would still require great sacrifices from the individual. But, however great the sacrifices, one feels oneself affirmed in them. They are not undertaken in the name of a Kantian duty which stands opposed to one’s inclinations; rather, they coincide with one’s deepest inclinations. One may, for example, be willing to sacrifice even one’s life for a person one loves, but the sacrifice is made because one’s love for the other is an integral part of one’s own life and one’s own self; in that sense it is not imposed by duty, but it is a form of self-affirmation. Similar things could be said about not only devotion to other individuals but also devotion to a social group or a political cause.15

I think that Norman presents a too rosy view concerning the possibility of doing without an alienated social morality. I think that Freud’s pessimistic outlook is more realistic. But be that as it may. Even if our CSM was reformed into something less alienated, the norms belonging to CSM* would certainly still be part of it. For, as I argued in Sect. 2.2, if the norms of CSM* are not generally complied with, there is no society. So the present objection, whether or not it succeeds with respect to certain CSM norms, does not strike against CSM*. (iv) CSM is unwittingly consequentialist. Perhaps the most interesting objection to CSM is a long and detailed one made by Shelly Kagan.16 True morality, Kagan claims, requires that you promote the good. CSM—or, as Kagan calls it, ordinary morality—does not, of course, accept this thesis; but, Kagan claims, ordinary moral See, e.g., Wilhelm Reich, The Sexual Revolution: Toward a Self-Governing Character Structure (New York: Farrar, Straus and Giroux, 1969); Erich Fromm, Man for Himself: An Inquiry into the Psychology of Ethics (New York: Rinehart, 1947). 15  Richard Norman, The Moral Philosophers: An Introduction to Ethics (Oxford: Oxford University Press, 1983), p. 221. 16  Shelly Kagan, The Limits of Morality (Oxford: Oxford University Press, 1989). 14

3.2  Common-Sense Morality and Its Critics

69

ity holds a weaker thesis, viz. that there is a standing pro tanto reason to promote the good.17 What prevent this reason from taking the form of a requirement to promote the good are two things. Firstly, ordinary morality includes agent-centred options permitting you not to promote the good in cases where promoting the good would be too “costly”, that is, would involve a significant sacrifice of your interests. Secondly, ordinary morality recognizes the existence of agent-centred constraints against performing certain kinds of actions, for example, harming other people even when performing such actions would promote the good.18 Now, Kagan continues, suppose that an adequate defence of options by appeal to cost could be provided: you are then permitted to allow harm in cases where not allowing it would be too costly. But, it seems, by the same reason you would also be permitted to do harm. But this would cause problems for ordinary morality. That is, it seems that it is not merely options to allow harm that would be supported by the appeal to cost, but also options to do harm. Yet although ordinary morality insists on the existence of options to allow harm, it rejects options to do harm. Only the minimalist believes in the existence of options to do harm for the sake of promoting one’s own interests.19

Kagan devotes much space to investigating whether it is possible to defend both the existence of an option to allow harm and the existence of a constraint against doing harm. The upshot is negative: it is difficult to defend the existence of either options or constraints, and it seems quite impossible to defend the existence of both. This, Kagan thinks, speaks in favour of C: “As I see it, we are indeed under a general moral requirement to promote the good. Ordinary morality may tell us otherwise; but ordinary morality cannot be defended.”20 Is Kagan right? Cannot the existence of both options and constraints be defended? In Sect. 2.4, I tried to show that the fact that CSM views the distinction between acts and omissions, between doing and allowing, as morally relevant can be given a contractualist explanation.21 For, as I said, “[b]y and large, a ban on performing actions that harm people is in everyone’s interest, whereas a demand to benefit other people is not.” That explains why, according to CSM, there is a constraint against harming other people, whereas there is an option, but no requirement, to benefit them. The primary moral distinction is then between what is and what is not in the interest of (almost) everyone. Now the world being what it is, it happens to be the case that the general inculcation and maintenance of dispositions not to commit certain kinds of actions belong to the former category, while the general inculcation and maintenance of dispositions to commit certain kinds of actions do not. Since an  The term “pro tanto reason” is intended to refer to, and better describe, what W. D. Ross calls “prima facie duty” (ibid., p. 17, n. 13). 18  “Options” and “constraints” were introduced in Sect. 2.4 above. 19  Shelly Kagan, “Précis of The Limits of Morality”, Philosophy and Phenomenological Research, 51 (1991), 897–901; p. 899. (Minimalists include, among others, egoists and nihilists.) 20  Ibid., p. 901. 21  I also mentioned another, complementary, explanation. 17

70

3  Deontological Ethics: Assessment

effective way of inculcating and maintaining behavioural dispositions is by means of moral education, it is not astonishing that generally the norms of CSM forbid many kinds of commissions, but not many kinds of omissions. If, as I claim, the primary moral distinction according to CSM is the one between what is and what is not in the interest of (almost) everyone, then a number of Kagan’s objections lose their force. Obviously, it does not matter, for example, that it is, as Kagan says, extremely difficult to come up with an intuitively adequate specification of the constraint: various formulations seem inevitably to either condemn too much or to excuse too much. And modifications intended to make for a better intuitive fit often seem ad hoc from the perspective of plausible motivations for the underlying constraint.22

The reason why, according to CSM, there is a constraint against behaving in a certain way is (I claim) that it is against the interest of (almost) everyone. It is not astonishing that there is no perfect fit between this property and some other property of the actions forbidden by CSM. (It is not even astonishing that some omissions are forbidden, for example, not saving a person from drowning when this can be done with little effort and no risk for oneself.) Nor does it matter that we sometimes intuitively judge a piece of behaviour as an action, but judge what seems to be a relevantly similar case as an omission. This happens not seldom. As Kagan points out, if parents starve their children to death, many of us want to describe the reaction of the parents as killing—a case of doing harm. Yet if the same people were to pass by a starving beggar, knowing that he will soon die unless he eats, their reaction would typically be described as allowing to die. Given the structural similarities, it is difficult to see how both judgments can be maintained. For in both cases the reaction is failure to provide food, knowing that death will be the outcome.23

What Kagan here reminds us of is that CSM is, in the terminology of Sect. 2.4, self-referentially altruistic: according to CSM, people have a strong obligation to benefit their own children, especially to help them survive. (This fact, socio-­ biologists tell us, is ultimately to be explained by means of the “will to survive” that belongs to our genes.24) Probably the best way of fulfilling this obligation is that people tend to their own offspring, both because they are normally strongly predisposed to do that anyway, and (pace Plato) because any other arrangement would be too burdensome to administer and probably not successful. Thus starving one’s children to death is to ignore (what is taken to be) an obligation, whereas not helping beggars from starving to death is not. It is therefore only  Ibid., p. 900. Kagan discusses these issues in Limits, Chs. 3 and 4.  Ibid., p. 102. Kagan’s use of the term “reaction” is, he admits, “a little idiosyncratic: I use it to refer to the agent’s response in a situation, even if that response amounts to doing ‘nothing at all’” (ibid., p. 47, n.). 24  “So genes that lead parents to take care of their children are, other things being equal, more likely to survive than genes that lead parents to abandon their children.” (Peter Singer, The Expanding Circle, p. 13.) 22 23

3.2  Common-Sense Morality and Its Critics

71

to be expected that CSM would condemn the former “reaction” more strongly than the latter, and that we use “killing” in the former case, but “allowing to die” in the latter. For even if “killing” usually denotes an action and “allowing to die” an omission, it is not always so. Since, generally, an act of killing is considered wrong, whereas omitting to keep alive most often is not considered wrong, or considered less wrong, we find it appropriate to use the former expression for those omissions to keep alive that we consider wrongful. To call a reaction “killing” is to condemn it in a way that calling it “allowing to die” is not. It is thus not astonishing that we are inclined to stigmatize failure to provide food for one’s children as “killing”, whereas the same omission with respect to a stranger is just “allowing to die”.25 Nor is it, as Kagan thinks, a problem for CSM that some kinds of actions are forbidden although refraining from performing them is very costly for the agent. Kagan holds that the appeal to “cost” provides the most plausible defence of options that CSM can give. (The basic cost is the agent’s sacrifice of his interest, but derivatively other things, such as the sacrifice of time and money, may be counted as costs.) As was pointed out in Sect. 2.4, it is very costly to benefit all those whom we are able to benefit. Also, with some exceptions, benefiting other people is not required by CSM: there is no general requirement to benefit but in most cases a moral option whether to benefit or not. It is easily thought then that in each case of benefiting other people the cost of doing so is what, according to CSM, justifies its being optional. And this is what Kagan thinks. But, as he points out, this attempt at justifying options does not work: it is sometimes very costly also to comply with CSM’s requirements. I stand to gain one million dollars by killing my uncle [Albert], and if morality is going to close off that act to me, then it exacts a tremendous cost. If high cost is indeed ground for denying the existence of a requirement, then we seem to have such a case here—and it is permissible to kill Albert. But this is quite unacceptable to the moderate.26

But Kagan has misunderstood the rationale of CSM options. Why is it not, according to CSM, optional, hence permissible, for me to kill my uncle even if not killing him is very costly? As should be clear from what was said in Chap. 2, it is because general acceptance of a norm permitting people to kill each other, whether unrestrictedly or only when (they think) they stand to gain by it, is too costly for everyone and therefore against everyone’s interest. That, for example, “I” will gain by killing my uncle is not compensated for by the loss that I will incur if others are allowed to kill me. It is because people sometimes would gain by killing other people, but would not gain by a general practice of killing each other whenever it is (considered) profitable, that CSM strongly forbids us to kill each other. As for ben If, on the other hand, a doctor disconnects the life-support machine from a patient instead of meaninglessly prolonging his life, this is, strictly speaking, an act of killing, not of allowing to die. But if we agree that this is the right thing to do—for to be killed in such circumstances is arguably in nearly all cases in the patient’s interest—we choose the less harsh description and call it “allowing to die”. 26  Kagan, Limits, p. 22. By a “moderate” Kagan understands the defender of ordinary morality, i.e., CSM. 25

72

3  Deontological Ethics: Assessment

efiting people, it is not, as Kagan seems to think, the idea that each benefit is costly for the benefactor that grounds the option whether to benefit or not; some benefits involve very little cost, and some no cost at all. It is compliance with a norm to benefit all those whom we are able to benefit that is costly, indeed too costly. This, together with the fact that not everyone gains from a general practice of benefiting, explains why CSM contains a norm of not killing other people, but not a norm to benefit them. As I mentioned above, Kagan thinks that CSM (ordinary morality) holds that there is a standing pro tanto reason to “promote the good”. Most of the time Kagan uses this vague non-comparative characterization, but several passages make it clear that he intends the corresponding comparative consequentialist notion. Thus he writes that, according to ordinary morality, “[e]everyone has a standing reason to promote the objectively best outcome”.27 As I mentioned above, what prevents this reason from becoming a requirement of ordinary morality, Kagan thinks, is that ordinary morality contains constraints and options.28 If, therefore, the moderate would accept Kagan’s criticism of ordinary morality, he would become a consequentialist: the reasonable moderate is unwittingly a consequentialist. Does ordinary morality really hold, as Kagan thinks, that there is a pro tanto reason to promote the good? I do not think so. Suppose, for example, that you intend to rob a person, correctly believing that his assets would give you more welfare than it would deprive him of. I strongly doubt that, according to ordinary morality, this gives you a reason to rob the person, a reason which is, however, outweighed by the reason stemming from the constraint against robbery. According to ordinary morality, it gives you no reason at all. Nor do I think that the moderate holds that the reason to promote the good turns into a requirement in cases where an action that would promote the good neither violates a constraint nor goes against the agent’s interest—thus not grounding an option. Suppose that I sincerely want to give my neighbour’s son a present that I know he would enjoy. Suppose further that I correctly believe that this is the best thing I could do with my spare money, that it is not against my interest, and that there is no other reason for not doing it. Is it then, according to ordinary morality, a moral requirement that I give the boy the present? Does the moderate think that I act morally wrong if, for some reason, I do not? I am fairly confident that the answer is No. There is, according to ordinary morality, no requirement to perform an action just for the sake of the greater good. Kagan bases his opposite view on a very slender base. Suppose, he says, that a child is drowning in front of me and that I can save his life by throwing a life preserver. In that case, Kagan points out, ordinary morality requires me to save the  Limits, p. 61.  There is sometimes, Kagan thinks, a requirement to promote the good even according to ordinary morality. The rejection of a general requirement to promote the good does not mean, Kagan says, “that the moderate believes that there is never any case at all in which an agent is required to perform some act for the sake of greater good. But such cases tend to be rather modest and limited.” (“Précis”, p. 897.) I question this claim below.

27 28

3.2  Common-Sense Morality and Its Critics

73

child, even at the price of some discomfort to myself.29 From this and a few similar examples Kagan draws the conclusion that ordinary morality holds that there is a pro tanto reason—and in some cases even a requirement—to promote the good. The best explanation of the moderate’s position is, I think, this: the opportunity to promote the good (say, by saving the life of the drowning child) generates a reason to perform the act. And this reason—in these cases—is sufficiently forceful to outweigh any opposing reasons, or other countervailing considerations. That is, in these cases at least, the reason is a morally decisive one, and thus grounds a moral requirement.30

I think that the contractualist conception of morality, which I ascribe to ordinary morality, gives a better explanation of the moderate’s position: it explains the requirement to save the child’s life, and there is, as I claim, independent support for ordinary morality’s being contractualist. Occam’s razor thus cuts away the assumption that ordinary morality holds that there is a pro tanto reason to promote the good, a reason that in some cases inexplicably turns into a requirement. For all that Kagan says, then, the reasonable moderate does not have to surrender; he is not a consequentialist malgré lui. (v) CSM subscribes to moral luck. According to CSM, the objection points out, people are not responsible, hence not blameworthy, for what happens due to circumstances over which they have no control. But, critics say, CSM also supports the opposite principle—CSM subscribes to moral luck.31 Hence, CSM is both inconsistent and subscribes to the highly implausible and morally repugnant principle of moral luck. A representative version of (what may be called) the argument from moral luck is offered by Michael Slote: A person driving along a lonely country road and paying too much attention to the scenery might swerve into the oncoming traffic lane, incur no accident, and blame herself very little if at all for her inattention or negligence. And our own attitude, thinking of such a case on its own, would normally be quite similar. However, if we imagine the same scenario except that a car in fact is coming in the opposite direction, with the result that an accident occurs and the other driver or passenger in her own car is killed, the negligent driver will very much blame herself. And our own inclination, say, as observers would be very similar. Yet the difference between the two cases is, from the standpoint of the agent, a matter of luck or accident, and our common-sense moral intuitions (check this in yourselves!) find it implausible and morally repugnant to believe that differences in blameworthiness and other serious moral differences between people should be a matter of luck or accident, beyond anyone’s control or advance knowledge. So the moral judgments we make and intuitions we have about cases of negligence are in fact inconsistent […].32  Limits, p. 3 f.  “Précis”, p. 898. 31  See the discussion in Sect. 4.2 whether C subscribes to the version of moral luck according to which one is blameworthy for one’s action having (not having) a certain property if one could not personally know that. 32  Michael Slote, “Virtue Ethics”, p. 181. The contemporary interest in moral luck started with two articles: Bernard Williams, “Moral Luck”, P. A. S., Supp. Vol. L (1976): 115–35; repr. in his Moral Luck: Philosophical Papers 1973–1980 (Cambridge: Cambridge University Press, 1981); and Thomas Nagel, “Moral Luck”, P. A. S., Supp. Vol. L (1976): 137–55; repr. in his Mortal Questions (Cambridge: Cambridge University Press, 1979). 29 30

74

3  Deontological Ethics: Assessment

I think (i) that Slote’s driver really is much more blameworthy in the second than in the first scenario (where, maybe, she is not at all blameworthy), (ii) that this is also CSM’s opinion, (iii) but that it does not conflict with any other opinion held by CSM. Slote agrees with (ii), but not with (i) and (iii). Let us first consider what can be said in defence of (i). The two scenarios contrasted by Slote may be schematically described thus: (1) A (a) drives inattentively, (b) therefore swerves into the oncoming lane, but (c) meets no other car and incurs no accident. (2) A (a) drives inattentively, (b) therefore swerves into the oncoming lane, and (c) collides with another car, whose driver is killed. Comparing (1) and (2) with each other, one is led to believe that whether or not there is an accident depends on luck, on circumstances external to and not controlled by A. And one is therefore inclined to find A not much more blameworthy in the second than in the first scenario. But let us contrast (2) with the following scenario: (3) A (a) drives attentively, (b) therefore does not swerve into the oncoming lane, and (c) incurs no accident. Comparing (2) and (3) with each other, one is led to believe that whether or not there is an accident depends on circumstances internal to and controlled by A. And one is inclined to find A much more blameworthy in the second than in the first scenario. Whether we find A very little, if at all, blameworthy or very much blameworthy in the second scenario thus depends on which alternative scenario it is contrasted with. This is an instance of a general feature of the pragmatics of causal attributions: Suppose we find that in situation S an event E occurs, and that there are n necessary conditions: C1, C2, …, Cn for the occurrence of E. We then compare S in our imagination with a situation S′ where one of the necessary conditions, say C1 is lacking, and where, therefore, E does not occur either. More or less spontaneously, we then dub C1 the cause of E and hold it to be responsible for the occurrence of E. If a human action (activity) is among the necessary conditions for the occurrence of E, and E is something we consider bad, we consider the agent blameworthy if, and only if, his action is the necessary condition that the comparison has selected as the cause. In Slote’s example, the comparison between scenarios (1) and (2) selects the external factor (the occurrence of the oncoming car) as the cause of the accident. The comparison between scenarios (2) and (3) selects the internal factor (A’s inattentiveness) as the cause of the accident, and therefore considers A blameworthy. Either comparison yields part of the truth: both the external and the internal factor are responsible for, and causes of, the accident. Since A is partly responsible for the accident, we hold him blameworthy in scenario (2). We do not hold him blameworthy in scenario (1) since, as a matter of fact, nothing bad happens there. And this, I contend, is a quite reasonable position.

3.2  Common-Sense Morality and Its Critics

75

Let us now consider what can be said in defence of (iii), that is, my claim that holding A much more blameworthy in the second than in the first scenario does not conflict with CSM. Slote claims that holding A responsible for the accident, hence extra blameworthy, runs counter to the following CSM principle: People are not responsible, hence not blameworthy, for what happens due to circumstances over which they have no control. But this formulation may be given a strong or a weak interpretation, depending on whether “due to” is interpreted as partly due to or as wholly due to.33 I think that CSM subscribes to the weak version of the principle: People are not responsible for what happens wholly due to circumstances over which they have no control. But CSM certainly does not subscribe to the strong version of the principle: People are not responsible for what happens partly due to circumstances over which they have no control. So, although holding A responsible for the accident in the second scenario runs counter to the (implausible) strong version of the principle, it does not run counter to the (plausible) weak version. The kind of moral luck that I have hitherto discussed, resultant luck, is but one of four kinds of moral luck that have been distinguished.34 The other three are circumstantial luck, constitutive luck, and causal luck. Do any of these pose a problem for CSM? Circumstantial luck is luck with respect to the circumstances in which one finds oneself. Consider A, a German who became a Nazi and then committed several horrendous acts during the Nazi era. Had he, as he once planned, moved to Argentina in 1929, he would (we postulate) have lived an impeccable life. Holding A responsible for having committed horrendous acts is therefore a case of a person not having circumstantial (moral) luck. I think that circumstantial luck can be handled in the same way as resultant luck was handled above. Let “P” stand for that set of mental properties which, together with his German circumstances, explains why A became a Nazi and committed the horrendous acts.35 The two scenarios that we are invited to compare with each other may then be schematically rendered thus: (1′) (a) A has P, (b) goes to Argentina, and (c) does not become a Nazi. (2′) (a) A has P, (b) remains in Germany, and (c) becomes a Nazi. Comparing (1′) and (2′) with each other, one is spontaneously led to believe that whether or not A became a Nazi and committed the horrendous acts depended on the circumstances in which he happened to find himself. And one is therefore, on reflection, inclined to find A about equally blameworthy in the two scenarios. But let us compare (2′) with the following scenario: (3′) (a) A lacks P, (b) remains in Germany, and (c) does not become a Nazi.

 Notice that the strong interpretation of the principle contains the weak interpretation of “due to”, viz. partly due to, whereas the weak interpretation of the principle contains the strong interpretation of “due to”, viz. wholly due to. 34  See Thomas Nagel, op. cit. 35  For the sake of simplicity, I assume that there is only one set of mental properties that satisfies the condition. 33

76

3  Deontological Ethics: Assessment

Comparing (2′) and (3′) with each other, one is spontaneously led to believe that whether or not A became a Nazi depended on himself, not on external circumstances. And one is therefore, on reflection, inclined to find A much more blameworthy in the second scenario. From here the argument is strictly parallel to the above argument to the effect that resultant luck does not pose a problem for CSM. And the conclusion is that nor does circumstantial luck pose a problem for CSM. But, someone might object, whether A has P or not is not something that is under his control. So whether, while remaining in Germany, A became a Nazi or not depends, after all, on external circumstances for which he is not responsible. The objection thus deals with the third kind of luck distinguished by Nagel, constitutive luck: we have no control over the hereditary and environmental factors that have made us who we are and thus determine how we act. Still we are responsible for who we are and how we act. Does constitutive luck pose a problem for CSM? Well, in the first place it is uncertain whether CSM itself has any opinion concerning this metaphysical issue. But adherents of CSM—and to some extent we are all adherents of CSM— certainly have. It turns out, however, that people embrace contradictory opinions on this issue. On one hand, people spontaneously believe in agent-causality, that is to say, believe that we are always able to act against the impact of our hereditary and environmental influences, even if this sometimes requires a considerable effort. So, given the spontaneous view, who we are and what we do is taken to be under our control. In other words, there is no constitutive luck. But, on the other hand, on reflection many of us become convinced that the picture of our agency assumed by the spontaneous view is a myth: there are no independent selves with the power to counteract the combined impact of hereditary and environmental influences.36 All empirical evidence supports the view that we are what these influences have made us to be. But this does not support constitutive luck. For the reflective view holds that, since we do not have control over these influences, we are not responsible for how they have shaped us. So, again, there is no constitutive luck. Thus, whether the spontaneous or the reflective view is attributed to CSM, constitutive luck does not pose a problem. The fourth kind of moral luck distinguished by Nagel is causal luck: our actions are caused by external events over which we have no control.37 Nor does causal luck pose a problem for CSM. The same reason that explains why constitutive luck does not pose a problem for CSM explains why causal luck does not do that either.

36  As for the spontaneous and the reflective views, cf. the discussion in Sect. 6.6 of the internal and the external perspectives. 37  Nagel has been criticized for bringing in causal luck as a separate category. This category is redundant, it has been objected, for any case covered by it is already covered by circumstantial or constitutive luck.

3.3  Is Common-Sense Morality Self-Defeating?

77

I thus conclude that CSM does not subscribe to any implausible and repugnant kind of moral luck, inconsistent with some principle that it holds concerning responsibility and blameworthiness.

3.3  Is Common-Sense Morality Self-Defeating? CSM* is collectively self-defeating. In Sect. 2.2, I claimed that since CSM* is of a contractualist nature, its norms furnish solutions to PD-situations. According to Derek Parfit, however, there are certain PD-situations that CSM* itself creates but fails to solve. In other words, CSM* is, Parfit claims, sometimes self-defeating.38 More exactly, the claim is that CSM* sometimes is “directly collectively self-­ defeating”. Here is Parfit’s explanation of this notion: A moral theory gives each agent certain substantial aims, things he ought to try to achieve. “[W]e successfully obey a theory T when each succeeds in doing what, of the acts available, best achieves his T-given aims.”39 T is directly collectively self-defeating “when it is certain that, if all rather than none successfully obey T, we will thereby cause the T-given aims of each to be worse achieved”.40 Parfit presents some (schematic) examples to illustrate his claim, the first of which runs: We cannot communicate. But each could either (1) save his own child from some lesser harm or (2) save the other’s child from another somewhat greater harm […]. Since we cannot communicate, neither’s choice will affect the other’s. If we believe we ought to give priority to our own children, we must believe that each should do (1) rather than (2). Each would thus ensure that, whatever the other does, his own child will be harmed less. But if both do (1) rather than (2) both our children will be harmed more.41

As I said in Sect. 2.4, CSM* claims that we have special obligations to our near and dear: CSM* is self-referentially altruistic. This means that “we ought to give priority to our own children”, and it thus seems to imply that we ought to do (1) rather than (2), with the result that “both of our children will be harmed more”. So it looks as if CSM* creates but does not solve the above (kind of) PD-situation. And, as Parfit points out, there are many other situations where, due to its self-referential altruism, CSM* fails to solve PD-situations: “Similar remarks apply to all similar   Derek Parfit, “Is Common-Sense Morality Self-Defeating?”, in Samuel Scheffler (ed.), Consequentialism and Its Critics (Oxford: Oxford University Press, 1988): 173–86; originally published in Journal of Philosophy, 76 (1979): 533–45. Essentially the same text is found in Parfit’s Reasons and Persons, p. 95 ff. 39  Ibid., p. 174. 40  Ibid., p. 174. 41  Ibid., p. 176 f. Parfit gives three more schematic examples, but they add nothing essential to the discussion. (Curiously enough, Parfit says that many-person versions of Parent’s Dilemmas are “as common, and as varied as Many-Person Prisoner’s Dilemmas” (Reasons and Persons, p. 98). I would say that they are prisoners’ dilemmas. But Parfit perhaps restricts the use of this label to dilemmas due to the agents’ egoistic aims. 38

78

3  Deontological Ethics: Assessment

obligations—such as those to parents, pupils, members of our own trade union, or fellow countrymen.”42 Parfit admits that the kind of situation mentioned in the above example is unlikely to occur; it is a two-person PD-situation, and in such situations people usually can communicate and are therefore able to avoid the sub-optimal outcome. But in many-­ person situations this condition is seldom fulfilled; so in these situations, Parfit claims, CSM* is self-defeating. Is it an objection to CSM* that it is self-defeating? Parfit thinks so, at least if CSM* is (as he believes) a collective code, and he therefore suggests that we revise CSM*. Unlike CSM*, revised CSM* says that we should all co-operate in PD-situations, that is, act so as to realize the outcome which is best for the near and dear of each of us; thus in the above example each should do (2) rather than (1). Moreover, revised CSM* says that each should co-operate if enough others do so too, and that we should be disposed to act accordingly. Is Parfit right? Does CSM actually meet with PD-situations of the present kind, situations where the result of each agent’s providing for his near and dear is that each near and dear is worse provided for? Parfit, we have seen, claims that this is the case, but he does not furnish any real-life examples to support his claim, and I for one cannot think of any. If there are no such examples, CSM* is at most potentially self-defeating, and this, it could be argued, is not a fatal objection to it. But is CSM* even potentially self-defeating? I do not think so. I think that already CSM* says what Parfit’s revised CSM* says, viz. that we should all co-operate in PD-situations, and that each should co-operate if enough others do so too. This follows from the facts that, as I claimed in Sects. 2.2 and 2.3, CSM is both a contractualist code, devised to avoid sub-optimal outcomes in PD-situations, and a collective code. As Parfit himself says, in many-person PD-situations people are seldom able to co-operate and do what is best for their near and dear since they cannot communicate. So lack of communication explains the sub-optimal outcome of such interactions. In two-person PD-situations, however, people are usually able to communicate and will therefore, as Parfit himself says, together act so as to bring about what is best for their near and dear. So if, as Parfit claims, CSM* is self-defeating in such situations—telling people to act in a way that results in sub-optimal outcomes— people would here, according to Parfit, by avoiding suboptimal outcomes (consciously) disobey CSM*. I do not think they do. I think they obey CSM*, which tells them to co-operate in all PD-situations (if such exist). If they fail to do so in ­many-­person situations, this is due to lack of communication, not to CSM leading them astray.43 I am quite confident that similar remarks apply to other CSM* obligations stemming from CSM*’s self-referential altruism. But what about obligations stemming from CSM*’s agent-relativity, the claim that each person should see to it that he  Ibid., p. 179.  A similar objection, I have found, has been raised by Arthur Kuflik in “A Defense of CommonSense Morality”, Ethics, 96 (1986): 784–803. In his reply (“Comments”, ibid.: 832–72), Parfit says (p. 851): “I accept this objection.”

42 43

3.4  Is Common-Sense Morality Irrational?

79

himself carries out his obligations and commits no wrong actions? Will this feature of CSM* create PD-situations which CSM* cannot solve? Parfit thinks so. He gives the following general example: Suppose that each could either (1) carry out some of his duties or (2) enable others to carry out more of theirs. If all rather than none give priority to their own duties, each may be able to carry out fewer. Deontologists can face Each–We Dilemmas.44

I think that also this attempt to show that CSM* is self-defeating fails—and for similar reasons as the one discussed above. Firstly, I cannot think of any realistic examples of PD-situations of the present kind. Secondly, even if there exist such situations, I think that CSM* solves them. As I said above, CSM* says that we should all co-operate in PD-situations, and that each should co-operate if enough others do so too. (This is due to CSM*’s being a contractualist and a collective code.) This means that CSM* is not self-defeating in, what may be called, “deontological dilemma situations” either. To see this, consider agents A and B, each of whom has two actions at his disposal. A can either do a1 or a2, and B can either do b1 or b2. If both do their 1-actions, each thereby carries out three of his duties, and if both do their 2-actions, each thereby carries out two of his duties. If A does a1 and B does b2, A carries out one of his duties and B four of his duties; and if A does a2 and B does b1, A carries out four of his duties and B one of his duties.45 According to Parfit’s objection, CSM* would prescribe that A and B should do a2 and b2, each thereby carrying out only two of his duties. But CSM*, being a collective code, does not prescribe this. CSM* says that people should co-operate in PD-situations; it should therefore be the case that both do their 2-actions, each thereby carrying out three of his duties. (Each should do his 1-action if, and only if, the other does his.) If they do not act accordingly, their failure is due to lack of communications and/or lack of trust, not to CSM* leading them astray. Similar remarks apply to other putative deontological dilemmas, whether two-­ person or many-person. If such dilemmas really do obtain, CSM* is not directly collectively self-defeating in these situations.

3.4  Is Common-Sense Morality Irrational? For a certain kind of situations, CSM* prescribes norms that have the following features in common: (i) The outcome of an action prescribed by the norm seems to be better for no one, and worse for the agent, than the outcome of its

 Derek Parfit, Reasons and Persons, p. 98.  Another interpretation of the numbers is that they measure the (relative) moral strength of the agents’ duties.

44 45

80

3  Deontological Ethics: Assessment

non-­performance46; thus the total (positive) value produced by the action seems to be less than the total (positive) value produced by its non-performance. (ii) On the other hand, the outcome of the general performance of the prescribed kind of action has a total (positive) value that is greater than that of its general non-performance. But, one might wonder, how could this consideration matter in the case of an individual action? Also many of those who reject C find it difficult to accept a ban on Pareto-optimal actions, actions the outcomes of which are perceptibly worse for no one, but better for at least someone, than those of the prescribed alternative. The appeal of the question, “What if everyone did that?” is not sufficiently strong to destroy the conviction that an individual action forbidden by such norms cannot really be morally wrong. As a result, these norms, hence to a certain extent CSM* itself, seems to be quite irrational. I will call the kind of situations for which CSM* prescribes this kind of actions lawncrossers’ dilemma situations (LD-situations, for short), and the actions prescribed by CSM* solutions to this kind of situations. The schematic example of an LD-situation, responsible for the name, runs as follows. Consider a set of people S, each member of which faces the choice of whether or not to cross a certain lawn.47 By crossing the lawn each person produces some good and nothing (perceptibly) bad: he gains some time, and this is valuable for him, but he causes no (additional) damage to the lawn, at least no perceptible one, hence no (perceptible) negative aesthetic value. But if all members of S cross the lawn, the damage to the lawn will be considerable; in fact, the negative aesthetic value resulting from the damage will be so great that it outweighs the value of the combined gains of time. (To make the comparison of value easier we could spell it out in, for example, hedonistic terms: any gain of time gives the gainer some pleasure, and any deterioration of the lawn gives the spectator some pain, pleasure and pain being the only intrinsic values.) In this situation, it may safely be assumed that each member of S is strongly inclined to cross the lawn. And, it seems, there is no reason for anyone to have a bad conscience for doing that. For the outcome of his action is Pareto-better than its alternative: the consequences of his crossing the lawn are better for him and no worse for other people than those of his not crossing the lawn.48 So why shouldn’t he cross the lawn?  Some clarifications: (i) In this section, the term “action” also includes omissions. (ii) The nonperformance of an omission is, of course, an action. (iii) Instead of talking about the value “for the agent”, it might perhaps be more adequate to talk about the value “for the agent and/or his near and dear”, or even about the value “for someone”. I leave it an open question, however, whether there are any CSM* norms which would necessitate any such change. 47  It is perhaps more realistic to assume that each agent has a choice between regularly crossing and not crossing the lawn. Something similar may be true of certain real LD-situations too. To take this possibility into account we should understand the term “action” in the above definition as “action or sequence of actions”. 48  I said above that, even if there is a change in negative value between n people and n + 1 people crossing the lawn, such a change is imperceptible. This is what I am inclined to believe—also with respect to corresponding alleged changes in other LD-situations. But the belief is controversial, and I do not want to rest my case on it. If, therefore, the belief is challenged, I will retreat to a weaker position, claiming only that the change in negative value is very small, and much smaller than the corresponding change in positive value. This claim should not be controversial, and it is sufficient for my argument to go through. 46

3.4  Is Common-Sense Morality Irrational?

81

The problem is, of course, that, if each person reasons accordingly and is also strongly inclined to cross the lawn, they will all cross the lawn. And, as my above characterization of LD-situations implies, the outcome of this is worse than the outcome of no one crossing the lawn. It might be claimed that the optimum solution to the problem lies somewhere between these extremes: there is a number m of the n members of S (m  5). There is no reason why this would make the theory escape the argument from horrendous actions. Suppose, however, that it did, the theory thus far being acceptable to D. But then the theory would not be acceptable to C, since C requires less than the lives of n persons being saved in order to permit the killing of one person. Nor, it seems, does Frankena’s theory escape the argument from too heavy demands. It should be admitted that Frankena holds that the principle of beneficence does not state a moral duty, since, he says, in general people have no right to be treated beneficently. But he still claims that “in some wider sense of ‘ought’, I ought to be beneficent”.10 But this is exactly the question at issue—this is what is contended by consequentialists and denied by their critics.

6.2  Consequentialist Attempts at Reconciliation Consequentialists who want to reconcile C with D often have recourse to a two-­ level theory. Some version of C, they claim, gives the criterion of right action, but in order to apply the criterion we need a set of less abstract rules to guide our decisions, and these rules roughly coincide with D.11 (Usually it is utilitarianism (U) and  Ibid., p. 46.  A slightly different version of this reasoning is put forward by Joel Kupperman in The Foundations of Morality (London: George Allen & Unwin, 1983). Kupperman argues (i) that the goal of morality is “that events occur that have the best consequences” (p. 125), and (ii) that we can derive D (“common-sense morality”) from the goal of morality “plus the character of human life in aiming for that goal” (p. 128; “the character of human life […]” refers to our imperfect knowledge, our bias towards near and dear, etc.). But many adherents of D would not subscribe to (i), claiming instead that the goal of morality is, e.g., peaceful co-existence; and even if, due to its vagueness, they would subscribe to (i), they would then interpret the formulation quite differently from what adherents of C do (see Sect. 1.2). So Kupperman’s proposal does not reconcile C and D (CSM).

10 11

160

6  Possible Solutions

common-sense morality (CSM) that these attempted reconciliations try to bring in line with each other.) In the following I shall discuss three such two-level theories, viz. those of J. S. Mill, Henry Sidgwick, and R. M. Hare. In Utilitarianism, Mill considers the objection raised to U “that there is not time, previous to action, for calculating and weighing the effect of any line of conduct on the general happiness”.12 This is really a silly objection, Mill retorts, for there has been ample time, namely, the whole past duration of the human species. During all that time mankind have been learning by experience the tendencies of actions; on which experience all the prudence, as well as all the morality of life, is dependent. People talk as if the commencement of this course of experience had hitherto been put off, and as if, at the moment when some man feels tempted to meddle with the property or life of another, he had to begin considering for the first time whether murder and theft are injurious to human happiness. […] [M]ankind must by this time have acquired positive beliefs as to the effects of some actions on their happiness; and the beliefs which have thus come down are the rules of morality for the multitude, and for the philosopher until he has succeeded in finding better.13

According to Mill, then, CSM—or, as he calls it, “customary morality”—provides the decision method by means of which we approximate to the utilitarian criterion. That CSM is able to do that is, Mill suggests, no accident; it depends on the fact that CSM is at heart utilitarian. People have come to accept a body of moral rules—which, however, they take to have independent authority and mistakenly think to be opposed to U—because of the recognized tendency of the rules to promote general utility. Mill is thus the first in a series of utilitarians who hold that CSM is, as it has been called, “unconsciously utilitarian”.14 As the last sentence of the above quotation shows, Mill does not think that existing CSM necessarily contains the optimal rules for utilitarian decision-making; in Mill’s opinion, “mankind have still much to learn as to the effects of actions on the general happiness.”15 Moreover, Mill claims, the principle of utility too has an action-guiding role, but only in cases of conflict between secondary principles.16 What, then, does the primary principle, the principle of utility, amount to according to Mill? It is not easy to answer this question. The traditional picture of Mill takes him to be an act-utilitarian, and this interpretation still has adherents.17 The traditional interpretation was, however, famously challenged by J. O. Urmson, who

 J. S. Mill, Utilitarianism, Ch. 2, Par. 24; p. 69.  Ibid., p. 69 f. 14  As John Simmons points out, what Mill says in Ch. 2 of Utilitarianism might give the impression that Mill holds the (obviously false) belief that CSM is consciously utilitarian. But this false impression is, as Simmons observes, corrected by what Mill says in Ch. 5. See John Simmons, “Utilitarianism and Unconscious Utilitarianism”, in Harlan B.  Miller and William H.  Williams (eds.), The Limits of Utilitarianism (Minneapolis: University of Minnesota Press, 1982): 86–92. 15  Mill, op. cit., Ch. 2, Par. 24; p. 70. 16  Ibid., Ch. 2, Par. 25; p. X. 17  See, e.g., Roger Crisp, Mill on Utilitarianism (London: Routledge, 1997), esp. Ch. 5. 12 13

6.2  Consequentialist Attempts at Reconciliation

161

argued that Mill was a rule-utilitarian.18 The latter interpretation too has been ­challenged, and today there is no consensus among Mill scholars as to how Mill’s utilitarianism should be understood.19 A recent commentator on Mill’s moral philosophy suggests the following complex view. In some areas of morality, he wants rule-utilitarian criteria to be employed, thinking of morality as parallel to law but with less formal sanctions. In other areas, he wants act-­ utilitarian criteria to be employed, and there are degrees of flexibility between the two. […] I claim that all of this is consistent with the ultimate goal of promoting greatest happiness.20

This is not the place to enter into the debate concerning the correct interpretation of Mill’s utilitarianism. I will therefore restrict my criticism of Mill to an objection that is neutral as between the different interpretations. When, as we saw above, Mill says that the principle of utility should be applied only to cases of conflict between secondary principles, he undoubtedly means that in such cases we should act according to that secondary principle which (we think) has the best consequences. (This might be given either an act-utilitarian or a rule-utilitarian interpretation, depending on whether we take “best consequences” to refer to best consequences in the particular situation or to best consequences in that kind of situation.) Now consider once again the doctor’s choice. Here several secondary principles come into play, for example, the principle not to kill innocent people, as well as the principle to help (more) people survive. These principles come into conflict in this (kind of) case, in the sense of prescribing incompatible courses of action. But then, according to Mill, the principle of utility should decide the question of what to do. And, given the facts of the case, killing the person in Room 306 and distributing his organs to the five needy patients has the best consequences. So this is what the principle of utility and, hence, Mill’s moral theory, prescribes. But according to CSM this would be morally wrong. (And there are many other similar cases where Mill’s utilitarianism likewise comes into conflict with CSM.) So, obviously, Mill’s moral theory does not reconcile C and D. Let us turn to another attempt at reconciliation. In The Methods of Ethics, Henry Sidgwick attempts to reconcile U with CSM. Two different lines of argument may be discerned. The first starts with a very detailed and thorough examination of CSM. The result is negative: being vague, potentially conflicting, and lacking self-­ evidence, the rules of CSM cannot be taken as basic ethical principles. There are, however, Sidgwick finds, some seemingly self-evident principles embedded in the moral thinking of common sense, among them the following two: (i) The good of any one individual is of no more importance, […] from the point of view of the Universe, than the good of any other.  J.  O. Urmson, “The Interpretation of the Moral Philosophy of J.  S. Mill”, Philosophical Quarterly, 3 (1953): 33–40. 19  See, e.g., many of the articles in Wesley E. Cooper et al. (eds.), New Essays on John Stuart Mill and Utilitarianism (Canadian Journal of Philosophy, Suppl. Vol. V, 1979). 20  Henry R. West, An Introduction to Mill’s Utilitarian Ethics (Cambridge: Cambridge University Press, 2004), p. 79. 18

162

6  Possible Solutions

(ii) [A]s a rational being I am bound to aim at good generally,—so far as it is attainable by my efforts,—not merely at a particular part of it.21

By means of (i) and (ii) Sidgwick then derives (what he calls) “the fundamental principle of utilitarianism”: (iii) [E]ach 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.22

But Sidgwick is not content with only showing that CSM among its principles implicitly contains U. This, as Sidgwick certainly is aware of, does not suffice to reconcile U with CSM. For common sense also acknowledges some, in Sidgwick’s view, seemingly self-evident principles by means of which ethical egoism may be inferred, and Sidgwick does not claim to be able to show that these principles are not really self-evident. So in his second argument Sidgwick sets out to show, by means of numerous examples, that CSM is unconsciously utilitarian: its judgments approximate and roughly coincide with U, and when CSM is unclear, has unacceptable implications, or is in internal conflict, it appeals to U. And he further claims that only U is able to systematize and connect the principles of CSM into a harmonious system. In Sidgwick’s own words: Utilitarianism sustains the general validity of the current moral judgments, and thus supplements the defects which reflection finds in the intuitive recognition of their stringency; and at the same time affords a principle of synthesis, and a method for binding the unconnected and occasionally conflicting principles of common moral reasoning into a complete and harmonious system. If systematic reflection upon the morality of common sense thus exhibits the Utilitarian principle as that to which Common Sense naturally appeals for that further development of its system which this same reflection shows to be necessary, the proof of Utilitarianism seems as complete as it can be made.23

According to the second argument, then, there is a remarkable coincidence between U and CSM.  This coincidence, Sidgwick admits, is far from complete. Indeed, if it were, this would be “fatal […] as the adoption of the Utilitarian principle would then become a matter of complete indifference”.24 In order to achieve the attempted reconciliation of U and CSM, Sidgwick claims that

 Henry Sidgwick, The Methods of Ethics, p. 382.  Ibid., p. 382. In his “Utilitarianism and Unconscious Utilitarianism” (mentioned in note 415), John Simmons says that Sidgwick does not claim to find the self-evident principles from which he attempts to infer U in common-sense thinking but “elsewhere” (Simmons, p. 91). But immediately after having derived (iii) from (i) and (ii) Sidgwick says: “I think that a ‘a plain man,’ in a modern civilized society, if his conscience were fairly brought to consider the hypothetical question, 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,—without any counterbalancing gain to any one else—would answer unhesitatingly in the negative.” (Ibid., p. 382.) This, I think, strongly supports my interpretation of Sidgwick. 23  Ibid., p. 422. 24  Ibid., p. 425. 21 22

6.2  Consequentialist Attempts at Reconciliation

163

it is not necessary to prove that existing moral rules are more conducive to the general happiness than any others: but only to point out in each case some manifest felicific tendency which they possess.25

And this is only what Sidgwick actually establishes. But, needless to say, just showing that the rules of CSM possess some felicific tendency does not reconcile CSM with U, and does not lend any support to U as usually understood (and as understood by Sidgwick himself throughout most of the book), viz. as a set of comparative principles (see Sect. 4.1). At most, it may be thought, it supports a non-­ comparative version of U, consisting of principles of the following kind: An action is right if, and only if, it has some felicific tendency. But it hardly establishes even such a watered down version of U. For there are several alternatives to the above principle which, on the assumption that they are unconsciously embraced by common sense, would just as well explain the felicific tendencies of the rules of CSM.26 This is true of, for example, certain eudaimonistic versions of contractualism and a version of rule utilitarianism consisting of principles of the following kind. An action is right if, and only if, it is prescribed by a rule general observance of which has some felicific tendency.27 An interesting version of the Sidgwickean kind of attempted reconciliation between U and CSM has been propounded by Richard Hare in Moral Thinking.28 Hare distinguishes between two kinds of moral thinking. One takes place at (what Hare calls) “the critical level”, and the other at (what he calls) “the intuitive level”.29 The person who does the moral thinking at the critical level is, Hare stipulates,  Ibid., p. 425.  This is also argued by John Simmons, op. cit. For other discussions of Sidgwick’s thesis concerning the unconscious utilitarianism of common sense, see Gertrude Ezorsky, “Unconscious Utilitarianism”, The Monist, 58 (1974): 468–74; Allan Gibbard, “Inchoately Utilitarian Common Sense: The Bearing of a Thesis of Sidgwick’s on Moral Theory”, in Miller and Williams (eds.), op. cit.: 71–85; D. D. Raphael, “Sidgwick on Intuitionism”, The Monist, 58 (1974): 405–19; Peter Singer, “Sidgwick and Reflective Equilibrium”, ibid.: 490–516; J. B. Schneewind, “First Principles and Common Sense Morality in Sidgwick’s Ethics”, Archiv für Geschichte der Philosophie, Bd. 45 (1963): 137–56; J. B. Schneewind, Sidgwick’s Ethics and Victorian Moral Philosophy, (Oxford: Oxford University Press, 1977), esp. Ch. X. 27  According to G. E. Moore, however, the rules of CSM lend support to a comparative version of utilitarianism, since they are necessary for “the preservation of civilized society”, which in turn is necessary for “the existence, in any great degree, of anything which may be held to be good in itself” (G. E. Moore, Principia Ethica, p. 158). But it appears that Moore does not hold that the rules lend support to act utilitarianism. He only holds that [t]hey seem all to be such that, in any known state of society, a general observance of them would be good as a means” (ibid., p. 157). 28  R. M. Hare, Moral Thinking: Its Levels, Methods, and Point. A characteristic difference between Sidgwick and Hare is that, whereas Sidgwick finds support for U in “our intuitions”, Hare finds the support in “the logic of moral language”. 29  This two-level structure of moral thinking was first suggested by Hare in “Ethical Theory and Utilitarianism”, H. D. Lewis (ed.), Contemporary British Philosophy (London: Allen and Unwin, 1976): 113–31. 25 26

164

6  Possible Solutions

unbiased, clear-headed, in possession of all relevant factual knowledge, and only constrained by (what Hare claims to be) the two essential properties of moral language, prescriptivity and universalizability.30 The critical thinker is thereby, Hare thinks, optimally equipped to find out the correct answer, for any actual or possible situation, concerning what he really ought to do in that situation. Owing to his psychology, his unlimited factual information, and the two properties of moral language, the critical moral thinker, Hare claims, is bound to arrive at exactly those verdicts that are given by (act) utilitarianism. At the intuitive level, however, the moral thinker is constrained by all kinds of circumstances: he lacks relevant information, he is partial in many ways, and he often has not time to correct these deficiencies. As moral agents we are nearly always compelled to think at the intuitive level. But the moral agent may get help from critical thinking. Critical thinking is able, not only to decide what people really ought to do, but also to decide what dispositions to act have a best acceptance utility: that is, to pick out, for each person, which sets of (internalizable) dispositions are such that, if the person has any of these sets of dispositions, he will approximate the utilitarian ideal of maximizing good over bad consequences at least as much as if he had some other set of (internalizable) dispositions. Only an archangel, Hare admits, is able to think critically in a flawless manner. But, to varying degrees, human beings are able to approximate the archangel’s thinking and to train themselves and those in their custody to acquire the moral dispositions recommended by critical thinking. At the conscious level these dispositions are manifested as consciously held prima facie principles. To be able to guide a person’s actions even when he is tempted to act contrary to them, the principles in question have to be deeply integrated in his personality; the person will therefore feel compunction if he acts against them and will have a disposition to blame others who do. Although the principles will differ somewhat between different persons, they will in any case contain the core of CSM.31 Owing to the limitations of human nature, the prima facie principles (and their corresponding dispositions) that have a best acceptance utility are rather few and rather simple, paying attention to the most frequent recurrent kinds of situations in real life. So there will often be a clash between the prima facie principles best suited for moral thinking at the intuitive level and the utilitarian thinking at the critical level, although the intuitive principles are selected by critical thinking: what is right according to U will often be wrong according to the prima facie principles selected by utilitarian thinking, and vice versa. As a result, even those whose moral principles are the ones that would be selected by U are apt to denounce U as immoral if their thinking is wholly confined to the intuitive level.  Cf. the characterization of the ideal observer discussed in Sect. 6.5 below. That moral language is prescriptive means, roughly, that a statement of the form “You ought to do a” entails the imperative “Do a!”; that it is universalizable means, again roughly, that a statement of the form “You ought to do a” entails a statement of the form “For any agent x, if x is in a similar situation as you are in, then x ought to do a”. 31  Hare, Moral Thinking, p. 200 f. 30

6.2  Consequentialist Attempts at Reconciliation

165

This is in rough outline Hare’s view concerning the relation between C (U) and CSM. Does it really reconcile them? I do not think so. Suppose that, according to correct utilitarian critical thinking, an agent in some situation ought to do an action a, but that—according to correct intuitive thinking based on some set of prima facie principles that have a best acceptance utility for him—he ought to do action b instead. What, then, ought he to do according to Hare’s theory? Which is the really right action? As a utilitarian Hare has to say that in this situation a is the right action. But then Hare would have to say that prima facie principles having a best acceptance utility are at least sometimes mistaken: they state that actions which really are wrong are right. And, since CSM is included in this set of prima facie principles, it follows that CSM is, at least sometimes, mistaken. But defenders of CSM would not, of course, admit that CSM is mistaken just because it conflicts with U. So no reconciliation between U and CSM will have been brought about if these standpoints ever conflict. But they certainly do—or the present essay rests on a mistake. But Hare would not claim that in the above situation action b is simply wrong. In a comment to T. M. Scanlon’s discussion of his two-level theory, Hare admits that [when Scanlon] says that, if critical and intuitive thinking yield different answers, the judgment by [sc. correctly done] critical thinking ‘is always the unequivocally correct answer’, he comes even closer to my view.32

Hare then adds—what he apparently thinks is what is missing in Scanlon’s account—that also at the intuitive level there is a legitimate use of “right” (and “wrong”). The meaning of “right” at the intuitive level, Hare says, is not the central one, though even there we have to remember that ‘right’ is still prescriptive, because the standards and the dispositions are being subscribed to by the speaker. So ‘right’ may mean ‘prima facie or intuitively right’ […].33

It is not easy to grasp what this (non-Rossean) notion of prima facie right amounts to—except that it refers to what the agent operating at the intuitive level considers right (period). What is in this sense prima facie right may, of course, be right, period, according to CSM but, nevertheless, wrong according to U (C). So Hare’s proposal does not reconcile CSM and U (C). There are other reasons why Hare’s model is not able to provide a reconciliation between U and CSM. In order to do that, a necessary (though not a sufficient) condition is that the principles of (enlightened) CSM have a best acceptance utility. (As I mentioned above, this is something that Hare believes to be the case.) But this would be highly unlikely unless common sense is unconsciously utilitarian. This is, however, something that Hare claims to be the case. His reason is that ordinary moral  R. M. Hare, “Comments”, in Douglas Seanor and N. Fotion (eds.), Hare and Critics: Essays on Moral Thinking (Oxford: Clarendon Press, 1990): 199–293; p. 261. Scanlon’s discussion of Hare is in the same volume and entitled “Levels of Moral Thinking”. 33  Ibid., p.  261. Hare says something similar in his discussion of a paper by William Frankena, “Hare on Levels of Moral Thinking”, contained on p. 222 ff. in the same anthology. 32

166

6  Possible Solutions

concepts, in their central sense, are universalizable and prescriptive; so, given favourable circumstances, clear-headed moral thinking has to come to the same conclusions concerning what ought to be done that utilitarian thinking comes to. (As mentioned above, universalizability and prescriptivity are the two essential properties of moral language according to Hare.) “If I am right about the moral concepts”, Hare says, then it can be predicted that people will on the whole hold certain opinions. […] True, the opinions are the outcome, not only of the use of the concepts, but of factual beliefs (which may be mistaken) and thought processes (which may have gone astray through muddle and bias), as well of the inclinations of the parties affected (which may differ from time to another). However, all these disturbing factors can in principle be allowed for, so that it is normally possible to use actual moral opinions as a test of ethical theory.34

But this presupposes that universalizability and prescriptivity give us U. And this presupposition may be questioned.35 In previous chapters, especially in Chap. 2, I have argued at length that CSM endorses a contractual conception of morality. (And if this is so, CSM is not unconsciously utilitarian, even if its moral concepts are universalizable and prescriptive.) It seems, therefore, that the central moral concepts used by CSM do not constrain people to think in utilitarian ways.

6.3  Two Attempts at Partial Reconciliation Recall William Godwin’s story about Fénelon and your father being trapped in a burning house, mentioned in Sect. 1.1. Godwin said that if there was time to save only one person, you should save Fénelon, since this would have the best consequences. Many people have found this verdict shocking. It is evident, they have objected, that you have a much stronger obligation to save your father than to save Fénelon, a man whom you owe nothing and are not closely related to. Only a callous person could think otherwise. Since according to C you should save Fénelon, C is mistaken and should therefore be rejected. The objection to C, underlying the reaction to Godwin’s proposal, says that the obligation to benefit those whom you stand in a close relation to, such as members of your family, your friends, and your neighbours, obviously outweighs any obligation to benefit strangers and people not related to you in any of those ways. (A stronger version claims that, except in cases where the benefit is considerable and  R. M. Hare, “Comments”, p. 291.  In Hare’s argument for U, the following principle is essential: one ought to do that action which will maximize the satisfaction of people’s preferences if, and only if, it would be rational to do that in the imaginary case where one oneself has all these preferences. This principle, which does not follow from the (alleged) facts that moral terms are universalizable and prescriptive, is criticized by non-utilitarians as ignoring the separateness of persons (I discuss this issue at the end of Sect. 6.5 and at the beginning of Sect. 6.6.)

34 35

6.3  Two Attempts at Partial Reconciliation

167

the cost of providing it is very small, there is no obligation to benefit strangers (see Sect. 3.5).) The objection thus manifests that feature of CSM that, following Broad, I have called self-referential altruism (see Sect. 2.4). It may also be seen as an instance of the Argument from too many demands (see Sect. 5.2). It thus highlights the conflict between C and CSM (hence D). One way of (to some extent) reconciling C and D would therefore be to argue that, properly understood, C is not vulnerable to the objection: the world being what it is, C never, or very seldom, requires that we at all benefit people outside the circle of near and dear, or that we do that at the expense of benefiting those inside the circle. As I mentioned at the end of Sect. 5.2, this view was held by Bentham, Mill, and Sidgwick. But none of them gives any argument for it, and the view seems quite implausible. But Frank Jackson thinks otherwise. By means of an extended argument Jackson tries to reconcile C with CSM (and hence D) as far as the above objection concerns.36 Jackson takes the objection as an instance of the Argument from Too Heavy Demands.37 His thesis is that consequentialism—properly understood—is perfectly compatible with the right actions for a person being in many cases actions directed toward achieving good consequences for those persons and projects that the agent holds dear.38

In order to bring out this compatibility, Jackson claims, we must reject objective C in favour of subjective C. (The terms were introduced in Sect. 4.1 above.) We should not claim that an agent ought to do what has in fact the best consequences, but that he ought to do what has the best probable consequences. As Jackson once puts it, we should “maximize expected moral utility”.39 Jackson also gives some arguments why objective C should be rejected in favour of subjective consequentialism. I think these arguments mistaken, but I shall not attempt to show this. For I intend to show that Jackson’s defence of subjective C against the objection is mistaken. And if I succeed in showing this, it is easily seen that a similar defence of objective C would be equally mistaken. To rebut the objection, Jackson invites us to consider a police inspector who has been assigned the task of controlling a large crowd at a soccer match. The inspector has to choose between two plans, the scatter plan and the sector plan. According to the scatter plan, each member of the police squad moves through the crowd doing good (helping lost children, stifling violence, etc.) wherever he can. According to the sector plan, each member of the squad is assigned his own sector of the crowd as his special responsibility, with the proviso that he should give help in other sec Frank Jackson, “Decision-theoretic Consequentialism and the Nearest and Dearest Objection”, Ethics, 101 (1991): 461–82. 37  “It seems then that consequentialism is in conflict with what makes life worth living.” (Ibid., p. 461.) 38  Ibid., p. 461. 39  Ibid., p. 481. 36

168

6  Possible Solutions

tors when there is an urgent need. Jackson claims that, generally, the sector plan is, in consequentialist terms, the best plan to follow in situations having the following structure: (a) When getting to know certain individuals is important for achieving good results. (b) When achieving good results involves coordinating a series of actions. (c) When achieving good results depends on setting up mutual trust and respect and understanding between individuals. (d) When there is a significant chance of different squad member’s [sic] actions nullifying each other if directed toward the same people. (e) When there is an obvious way to assign police to separate sectors which coincides with their natural inclinations and enthusiasms, particularly when this fact is common knowledge.40 It is obvious, Jackson says, that, by and large, similar features characterize the situation of any individual with respect to other people. Since, given these features, the (generalized) sector plan is, in consequentialist terms, the best plan to follow, we, on consequentialist grounds, ought to follow it. But to follow this plan in our daily life involves directing our resources to our nearest and dearest. Hence, this practice is consequentially justified, and the present objection to C is thereby rebutted. My suggestion, then, is that the consequentialist can reply to the nearest and dearest objection by arguing that the kind of direction of attention toward those we hold dear which is so characteristic of a worthwhile life can be explained without attributing a biased value function. […] The suggestion is not of course that the kind of attention that we typically manifest in fact toward those we hold dear can be explained without attributing a biased value function. […] The suggestion is that a considerable degree of focus on our family and friends, enough to meet the demand that our lives have a meaningful focus, is plausibly consistent with living morally defensible lives according to consequentialism.41

But Jackson’s attempted (partial) reconciliation of CSM (D) with C succeeds, at most, if taken as a statement of what we as a group ought to do, not of what we individually ought to do. Granted that it would be consequentially best if everyone mainly cared for his family, friends, and those living in his close vicinity, it does not follow that, in the actual world, it would be consequentially best for anyone to channel his care and attention accordingly. And, as a matter of fact, it is not. For we live in a world where so many people are not adequately cared for by their nearest and dearest, often due to lack of resources. If all or most people were able to adequately care for their nearest and dearest and also generally did that, we would probably come rather near to acting optimifically. Since many people in the world are not given proper attention and care but are instead helpless victims of natural and social disasters, what C, even the subjective version endorsed by Jackson, prescribes for  Ibid., p. 474 f. Of course, in most real situations of this kind, “people” should be substituted for “police”. 41  Ibid., p. 475. 40

6.3  Two Attempts at Partial Reconciliation

169

many of us cannot be restricted to caring for his nearest and dearest. The moral life according to C is for many people a rather stern and self-sacrificing one. An interesting attempt to find a plausible middle ground between C and D has been made by Samuel Scheffler.42 C, as Scheffler notes, holds that it is always permitted to perform an optimal action, that is, do what would lead to a best outcome overall. Against this, D holds: ( 1) Some optimal actions are permitted but not prescribed. (2) Some (other) optimal actions are forbidden. Scheffler’s “hybrid” theory accepts (1) but not (2). Like C, it accepts the opposite principle: (3) No optimal actions are forbidden. To accept (1) is, in Scheffler’s terminology, to accept an “agent-centred prerogative”, and to accept (2) is to accept an “agent-centred restriction”.43 The argument for the hybrid theory consisting of (1) and (3) is that it is possible to find a rationale for the prerogative (that is, (1)), but not for the restriction (that is, (2)). The defence of the prerogative starts by reminding us of the fact that each person has an independent point of view, from which he sees the world, form projects and commitments, and acts upon them.44 The defence then continues: A moral conception that incorporates an agent-centred prerogative […] takes account of the natural independence of the personal point of view precisely by granting it moral independence: by permitting agents to devote energy and attention to their projects and commitments out of proportion to the value from an impersonal standpoint of their doing so. […] [G]iven this conception of the importance of the natural fact of personal independence, a moral view gives sufficient weight to that fact only if it reflects it, by freeing people from the demand that their actions and motives always be optimal from the impersonal perspective, and by allowing them to devote attention to their projects and concerns to a greater extent than impersonal optimality by itself would allow.45

Scheffler admits that a sophisticated version of C too can take account of the independence of the personal point of view. It can do this by admitting successful pursuance of personal projects and concerns as one of the values to be maximized. But, Scheffler claims, the prerogative account is at least as good as the consequentialist one. And this, he says, is sufficient for his purpose. For his project is, he reminds us,  Samuel Scheffler, The Rejection of Consequentialism (Oxford: Oxford University Press, 2000 (1982)). 43  Properly speaking, Scheffler says that an agent-centred prerogative or restriction “would have the function” of accepting, respectively, (1) and (2). 44  Scheffler argues that a failure to appreciate the independence of the personal point of view is also at the heart of classical utilitarianism’s unsatisfactory handling of distributive justice. One aim of his book is therefore to work out a theory of justice that is able to handle this point of view. I will not discuss this issue. 45  Ibid., p. 61 f. As Scheffler acknowledges, this argument is a development of Bernard Williams’s argument from integrity, set forth in “A Critique of Utilitarianism”, Ch. 5. 42

170

6  Possible Solutions

to explore the question whether agent-centred restrictions are as well-motivated as an agent-centred prerogative is. And for the purposes of this project, it appears that even if the conflict under consideration were resolved in the second way, a motivation of the requisite type for an agent-centred prerogative would still have been identified. For even given a resolution of this sort, it would have been demonstrated that such a prerogative does indeed embody a rational strategy for taking account of one significant feature of the person.46

The last step in the defence of the compromise theory is to show that there is no similar rationale for accepting agent-centred restrictions. To this end, Scheffler devotes a long chapter to examining prima facie plausible attempts to find such a rationale. The upshot is negative: none of the attempts succeeds. And, Scheffler concludes, “there are good reasons, in the absence of a plausible principled rationale for such restrictions, to worry about the adequacy of moral conceptions that include them”.47 I have two objections to Scheffler’s hybrid theory. The first is this: It may be thought that, since the agent-centred prerogative is meant to take care of the argument from too heavy demands, and the agent-centred restriction to take care of the argument from horrendous actions, the hybrid theory (which rejects the restriction) accepts horrendous actions. But it is not quite so. The prerogative, it is true, permits people not to perform optimal actions “when it would be unduly costly or burdensome for them to do so”.48 But notice that also horrendous actions are optimal actions that, sometimes and for some people, might be unduly costly. Suppose that horrendous actions of type H might be unduly costly for A, but not for B. In that case, not performing such an action is right for A, but not for B, according to the hybrid theory. But this is unacceptable: whether an action is right or wrong should not be determined (at least not only) by reference to the cost for the agent. It is of no avail to choose an “objective” criterion for being unduly costly, for example, by stating that a kind of horrendous actions are unduly costly if, and only if, the average person finds them so.49 For in this case too, an action’s normative status is determined by reference to the cost for the agent. My second objection concerns Scheffler’s rejection of the agent-centred restriction. Consider a case where an agent can minimize the total number of objectionable (horrendous) actions being performed if he himself performs such an action. (Consider, for example, Jim’s choice.) D, which accepts the restriction, forbids the agent to perform the action. Scheffler thinks that this prohibition is irrational, and he therefore rejects the restriction, hence D, in favour of his hybrid theory, according  Ibid., p. 65.  Ibid., p. 116. 48  Ibid., p. 20. 49  Scheffler evidently assumes a “subjective” criterion. He says that the prerogative would allow the agent “to promote the non-optimal outcome of his choosing, provided only that the degree of its inferiority to each of the superior outcomes he could instead promote in no case exceeded, by more than the [by the agent] specified proportion, the degree of sacrifice necessary for him to promote the superior outcome. If all of the non-optimal outcomes available to the agent were ruled out on these grounds, then and only then would he be required to promote the best overall outcome.” Ibid., p. 20.) 46 47

6.4  Rule Consequentialism

171

to which such actions are not forbidden. But, one may wonder, if the prohibition on such actions is irrational, and there is, as Scheffler evidently thinks, a conclusive reason for performing them, viz. the minimizing of harm, why are they not required? Scheffler’s answer is this: Hybrid conceptions do not systematically require the production of the best overall states of affairs, because by not doing so they are able to reflect the natural independence of the personal point of view.50

But, it may be thought, even if protecting the independence of the personal point of view is important, so is minimizing harm. That the former goal should always take precedence over the latter is an implausible absolutist view. I am fairly convinced that even many adherents of D think that one ought to have killed Hitler if killing him had meant that the Holocaust would never have happened. What, then, is the rationale of agent-centred restrictions. Scheffler could not find one. In the next chapter I shall try to show that a contractualist conception of morality can provide one.51

6.4  Rule Consequentialism It is obvious, I think, that rule utilitarianism and, more generally, rule consequentialism (RC) have been put forward in an attempt to reconcile C and D.  Thus Brad Hooker, himself a rule-consequentialist, says that there are, as several adherents of RC have recognized, two ways of arguing for this moral theory. The first way is to argue that it specifies an underlying principle that provides impartial justification for intuitively plausible moral rules [i.e., the rules of CSM], and that no rival theory does this as well.52

The second way starts from a commitment to consequentialist assessment, and then argues that assessing acts indirectly, e.g., by focussing on the consequences of communal acceptance of rules, will in fact produce better consequences than assessing acts directly in terms of their own consequences.53

Since then one argument for RC addresses itself to CSM and another to the consequentialist spirit, the prospects that this attempt to reconcile C and D will succeed seems promising. But, unfortunately, the attempt fails, since it is open to the same objection as Mill’s attempt (see Sect. 6.2 above). What seems to be its great virtue  Ibid., p. 107.  In “Prerogatives and Restrictions from the Cooperative Point of View”, Ethics, 105 (1994): 128– 52, R. H. Myers suggests a (broadly) contractualist defence of restrictions (and prerogatives). 52  Brad Hooker, “Rule-consequentialism”, The Stanford Encyclopedia of Philosophy (Spring 2004 Edition), Edward N. Zachta (ed.),URL = http://plato.stanford.edu/archives/spr2004/entries/consequentialism/rule, p. 12. 53  Ibid., p. 12. 50 51

172

6  Possible Solutions

is actually the source of its fundamental problem. To see this, consider the following version of RC, which I think is quite representative. (RC*) An act is right (wrong) if, and only if, it conforms (does not conform) to a set of rules general acceptance of which would maximize (expected) value.54 If RC* is to reconcile C and D, a necessary condition is that the set of rules in question coincide (at least roughly) with the rules of CSM. Otherwise RC* will not be acceptable to adherents of D. Let R be a set of rules general acceptance of which would maximize (expected) value; for the sake of simplicity, let us assume that it is the only set of rules which fulfils this condition. Consider the doctor’s choice. According to R, it is either permissible or forbidden to kill the man in Room 306 and distribute his organs to the five needy patients. If the former, RC* is not acceptable to adherents of D; if the latter, RC* is not acceptable to adherents of C. (A similar reasoning can be applied to many similar cases.) So, whatever verdict RC* gives in this and similar cases, neither the verdict, nor RC* itself is acceptable both to adherents of C and to adherents of D. Therefore RC* fails to reconcile C and D.55 The reason why RC cannot reconcile C and D is, then, that, in order to do that, it must appeal both to our CSM convictions and to consequentialist assessment. But, as I have argued at length in previous chapters, CSM is not a consequentialist morality; so general acceptance (following) of its rules does not maximize (expected) value. Hence, the rules referred to in the formulation of RC cannot be those of CSM.56 On the other hand, the rules general acceptance of which would really maximize (expected) value are not acceptable to adherents of D (Moreover, conformity to these rules, which are not generally accepted (adhered to), would, at least sometimes, be pointless, perhaps disastrous, and would not therefore be acceptable even to adherents of C.) A version of the above objection is the argument from rule worship. Probably J. J. C. Smart was the first to formulate the objection. (Smart directed the objection to rule utilitarianism, but it is obviously pertinent to other forms of RC.) Here is Smart’s formulation of the objection: [T]he rule-utilitarian presumably advocates his principle because he is ultimately concerned with human happiness: why then should he advocate abiding by a rule when he knows that it will not in the present case be most beneficial to abide by it? […] [T]o refuse to break a generally beneficial rule in those cases in which it is not most beneficial to obey it seems irrational and to be a case of rule worship.57  The formulation comes close, mutatis mutandis, to what David Lyons calls “Ideal RuleUtilitarianism”. See his Forms and Limits of Utilitarianism, p. 140. Let us stipulate that x maximizes (expected) value iff the outcome of x is at least as good as the outcome of any alternative to x. (A different version of RC is obtained by substituting “adherence to” for “acceptance of”.) 55  I am quite certain that what is true of RC* is also true of any other credible form of RC. 56  Hooker is therefore mistaken when he claims that “[t]he best argument for rule-consequentialism is that it does a better job than its rivals of matching and tying together our moral convictions”. (Brad Hooker, Ideal Code, Real World: A Rule-consequentialist Theory of Morality (Oxford: Oxford University Press, 2002 (2000)), p. 101.) 57  J. J. C. Smart, “An Outline of a System of Utilitarian Ethics”, p. 10. 54

6.4  Rule Consequentialism

173

There are many versions of RC, but the present objection holds against all of them, or so I will argue. In order to do this I will proceed as follows. I will first present a representative and elaborated version of RC, one set forth by Brad Hooker and show that it is vulnerable to the objection. I will then change the specific details of this version and argue that also the versions resulting from these changes are vulnerable to the objection. Here is then Hooker’s formulation of RC: An act is wrong if it is forbidden by the code of rules whose internalization by the overwhelming majority of everyone everywhere in each new generation has maximum expected value in terms of well-being (with some priority for the worst off). The calculation of a code’s expected value includes all costs of getting the code internalized. If in terms of expected value two or more codes are better than the rest but equal to one another, the one closest to conventional morality determines what acts are wrong.58

(Some comments on this formulation may be appropriate. As is evident from other places in Hooker’s book, the “if” on the first line should be read “if and only if”, and a right, or permitted, action is an action which is not wrong; the concept of ought does not feature in Hooker’s formulation of RC.) My objection to Hooker’s version of RC can be stated thus: Consider any situation where act consequentialism prescribes that the agent ought to do a certain action a, whereas, according to the present version of RC, a is wrong. Consider, for example, (once more) the doctor’s choice, introduced in Sect. 1.1. There is no doubt that the doctor’s action of cutting up the man in Room 306 and distributing his organs to the five patients is forbidden by the code of rules that satisfies Hooker’s version of RC and hence wrong. But the action would certainly maximize expected value. So why—on the strength of a bona fide consequentialist moral theory— should not the doctor perform it? To use Smart’s words, Hooker “presumably advocates his principle because he is ultimately concerned with human happiness: why then should he advocate abiding by a rule when he knows that it will not in the present case be most beneficial to abide by it?” The best argument for rule-consequentialism is not that it derives from an overarching commitment to maximize the good. The best argument for rule-consequentialism is that it does a better job than its rivals of matching and tying together our moral convictions, as well as offering us help with our moral uncertainties.59

But this answer is not plausible. If, as I have argued at length in Chap. 2, “our” moral convictions are mostly deontological, a moral theory that is committed to “matching and tying together” these convictions instead of maximizing the good is certainly not a consequentialist theory. Moreover, what Hooker says in his answer contradicts both his above formulation of RC and his characterization of C elsewhere in his book. Thus the formulation of RC quoted above takes closeness to conventional morality to be decisive only if “in terms of expected value two or more  Brad Hooker, Ideal Code, Real World: A Rule-consequentialist Theory of Morality, p. 32.  Ibid., p. 101. In Hooker’s encyclopedia article quoted above, this argument was said to be one of two possible (and presumably equally good) arguments for RC. Adding, as is now done, that it is the best argument amounts to abandoning C altogether.

58 59

174

6  Possible Solutions

codes are better than the rest but equal to one another”. And Hooker’s definition of “consequentialism” in general is in terms of the production of value: A theory is consequentialist if and only if it assesses acts and/or rules (or motives, social codes, virtues, or ways of life) in terms solely of the production of agent-neutral value.60

It seems, then, that Hooker’s version of RC is vulnerable to the argument from rule worship. And so are other versions of RC. In order to show this, let us generate new forms of RC by altering one or more of the distinguishing details in Hooker’s formulation of the theory. (As far as I can see, no credible version will thereby escape attention.) Thus (i) we may change “the overwhelming majority of everyone” either to the stronger “everyone” or to the weaker “most”; (ii) we may substitute “everywhere in each new generation” by “the society in question” or something similar; (iii) we may either weaken or strengthen the suggested requirement of equality, that of requiring “some priority for the worst off”; (iv) we may substitute “actual value” for “expected value”; (v) we may replace “internalization” with “compliance”; and (vi) we may disregard conventional morality altogether. It is easily seen that none of these changes, nor any combination of them, would save the resulting version of RC from the charge of rule worship.

6.5  J ustifying Ultimate Principles: Some Traditional Proposals A traditional way of arguing for the validity of a moral principle is to claim either that it is self-evident, or that it is entailed (in conjunction with some true empirical principle(s)) by some self-evident principle(s). Since in the present case we are dealing with ultimate principles, the latter alternative does not seem viable: by definition, an ultimate principle is not—not even partially—entailed by another moral principle, and it is generally agreed that a normative principle cannot be entailed by merely factual principles; hence, an ultimate principle is not entailed by any other principle(s).61 Neither does the former alternative give promise of success: since both C and D are rejected by competent judges—hence, not universally considered self-evident—it is not very convincing just to claim of either that, nevertheless, it is self-evident. Now it might seem strange—not to say utterly wrong—to say that D is an ultimate moral principle (or set of principles). Have we not recently witnessed how consequentialists like Sidgwick and Hare try to support D by means of C? And do not deontologists like Kant try to support D by means of other principles, such as  Ibid., p. 110. Cf. also the following statement concerning our chances of identifying the optimal code: “The most we can hope for is to find a code whose general internalization could reasonably be expected to result in at least as much good as could reasonably be expected to result from any other identifiable code.” (Ibid., p. 114.) 61  I disregard the fact that a moral theory, such as C and U, is entailed by the conjunction of its parts. 60

6.5  Justifying Ultimate Principles: Some Traditional Proposals

175

the categorical imperative? This is, of course, true, but, as far as I know, no one has claimed that D is entailed by any other principle (with or without the help of true empirical premises). The consequentialists in question have argued—not convincingly, as I have tried to show—that, given that C is true and the human condition is what it is, people should use D as their decision procedure. But this is something different from arguing that C entails D. And deontologists like Kant have given various arguments in support of D, but giving arguments for a moral principle does not mean that the principle is not (treated as) an ultimate principle. As we shall see, one might even argue for (and against) an ultimate moral principle by means of another ultimate moral principle. In recent years the foundationalist view of justification, where the point of departure is a set of premises (general or particular, as the case may be), claimed to be self-evident (or secure, certain, incontrovertible, incorrigible, etc.), has been largely abandoned in favour of coherentist views, according to which, as Simon Blackburn succinctly puts it, a body of propositions may be known without a foundation in certainty, but by their interlocking strength, rather as a crossword puzzle may be known to have been solved correctly even if each answer, taken individually, admits of uncertainty.62

The most well-known version of coherentism in moral philosophy is the method of (wide) reflective equilibrium, introduced by John Rawls and later elaborated by others, especially by Norman Daniels.63 Daniels describes the notion as follows: A wide reflective equilibrium is a coherent ordered triple of sets of beliefs held by a particular person, namely, a set of considered moral judgments, (a); a set of moral principles, (b); and a set of relevant background theories, (c). We collect the person’s initial moral judgments and filter them to include only those of which he is relatively confident and which have been made under conditions generally conducive to avoiding errors of judgment. We propose alternative sets of moral principles which have varying degrees of “fit” with the moral judgments. […] The agent may work back and forth, revising his initial considered judgments, moral principles, and background theories, to arrive at an equilibrium point which consists of the triple, (a), (b), and (c).64

Is it possible to solve the conflict between C and D by using the method of reflective equilibrium? Well, as we have just seen, the method is “person-relative”: it is taken to be used by particular persons to reach equilibrium with respect to their own beliefs. (It may, therefore, be the case that different persons will arrive at different equilibrium points, but, of course, it may also be the case that at least fully rational

62  Simon Blackburn, The Oxford Dictionary of Philosophy (Oxford: Oxford University Press, 1996 (1994)): “Foundationalism”. 63  John Rawls, A Theory of Justice; Norman Daniels, “Reflective Equilibrium and Archimedean Points”, Canadian Journal of Philosophy, 10 (1980): 83–103; and “Wide Reflective Equilibrium and Theory Acceptance in Ethics”, The Journal of Philosophy, 76 (1979): 256–82. A critical assessment of the method is to be found in Michael DePaul, Balance and Refinement: Beyond Coherence Methods of Moral Inquiry (London and New York: Routledge, 1993). 64  Norman Daniels, “Reflective Equilibrium and Archimedean Points”, p. 85 f.

176

6  Possible Solutions

and relevantly informed people will end up at the same point.65) The only way to use the method of reflective equilibrium in order to solve the conflict between C and D would therefore be to show either (i) that no rational adherent of C could go through the reflective process and still stick to his principles, or (ii) that no rational adherent of D could do that, or (iii) that neither could do that, but that there is some sort of compromise between C and D which a rational person would adopt. I shall discuss the first two possibilities in this section leaving the discussion of the third possibility to Sect. 7.5. A consequentialist, I take it, will admit that, like other people, he, at least to some extent, subscribes to some considered moral judgments which conflict with C in the sense that the former sometimes prescribe actions which are wrong according to the latter, or vice versa. Still, he could argue, when sitting down in a cool hour, it is evident to him that C is the valid moral principle, partly because there is a theory which explains why the conflicting considered moral judgments have such a strong hold on him. This is the theory mentioned at the end of Sect. 2.4, according to which CSM—into which he, like everyone else, has been indoctrinated since early childhood—is necessarily deontological. (This theory belongs to box (c) in Daniels’s schema, the set of relevant background theories, which includes “Theory of the Role of Morality in Society” and “Theory of Moral Development”.66) He could claim that the truth of this background theory casts grave doubts on the validity of his considered moral judgments, while admitting that, at the same time, he finds himself unable to wholeheartedly reject them; they have too strong a hold on him. So a rational adherent of C can go through the reflective process and still stick to C. Admittedly, he is unable to reach a perfect equilibrium, but, he could claim, this does not tell against C but rather against the naive idea that a person can always change his beliefs at will. The situation of our consequentialist is rather similar to that of the epistemological skepticist. The skepticist is convinced that he has no adequate ground for what he finds himself to believe about the external world. Still, like the rest of us, he is so constituted that he cannot, except momentarily, reject those beliefs, not even suspend his judgments. This was noted by Hume in a famous passage in Treatise, as well as by Thomas Nagel in a passage which, equally, deserves to become famous: Philosophical skepticism does not cause us to abandon our ordinary beliefs, but it lends them a peculiar flavor. After acknowledging that their truth is incompatible with possibilities that we have no grounds for believing do not obtain […] we return to our familiar convictions with a certain irony and resignation. Unable to abandon the natural responses on which they depend, we take them back, like a spouse who has run off with someone else and then decided to return67;

 Rawls confesses to be an agnostic concerning this issue. Such matters are, he says, “far beyond our reach” (John Rawls, op. cit., p. 50). 66  Norman Daniels, “Reflective Equilibrium and Archimedean Points”, p. 88. 67  Thomas Nagel, “The Absurd”, Ch. 2 in his Mortal Questions (Cambridge: Cambridge University Press, 1979), p. 19 f. Hume’s observation, quoted below, is to be found in A Treatise of Human Nature, Bk. 1, Pt. 4, Sec. 7, Par. 9; p. 175. 65

6.5  Justifying Ultimate Principles: Some Traditional Proposals

177

The deontologist, in contrast to the consequentialist, does not necessarily have conflicting moral beliefs and has therefore a greater chance of reaching reflective equilibrium. There is a risk, however, that he will fail to satisfy the requirements of reflective equilibrium in another way. With the exception of Kantianism, deontological moralities do not have much theoretical structure, and there is a risk, therefore, that the moral principles which the deontologist will arrive at, the principles in box (b) that are supposed to systematize and explain the considered moral judgments contained in box (a), are just generalizations of these judgments, with no independent support. But, as Daniels points out, such an equilibrium will not do. The theories in (c) must show the principles in (b) are more acceptable than alternatives on grounds to some degree independent of (b)’s match with relevant considered judgments in (a). Without such independent support, the principles have no support they would not already have had in a corresponding narrow equilibrium where no appeal to (c) is made.68

It could be argued, then, that the fit between deontological principles and considered moral judgments does not support the former, since deontologists do not have any background theories that would support their principles; their considered moral judgments and their moral principles are therefore not in a wide, only in a narrow, reflective equilibrium. Moreover, the consequentialist could point out, deontologists do not take due note of the background theory (mentioned above), according to which the morality instilled into us from early childhood, determining what we recognize as considered moral judgments, is necessarily deontological. This casts grave doubts on the validity of these judgements, since we then have an explanation of why we find them valid which has nothing to do with the question whether they are valid. (Something similar is not true of C.) On the other hand, it could be argued that it is possible that a deontologist could come up with background theories that would support his principles, and it is also possible that, despite their origin, our deontological common-sense morality is valid. But this, of course, has still to be shown. All in all, then, it seems that considerations connected with the method of reflective equilibrium do not decisively tell for or against either theory and therefore do not solve the conflict between C and D. A third traditional way of arguing for or against an ultimate moral principle is the (broadly) Kantian one, according to which a moral principle is valid only if it satisfies the requirements inherent in, or presupposed by, either the concept or the practice of morality. The most elaborated version of this view that I know of has been set forth by Philipps Griffiths.69 Formulating this Kantian view “in more contemporary language”, Griffiths says that [t]he only kind of solution which seems possible is one that shows that certain moral principles must be regarded as correct if moral discourse is to be possible at all, at least as an

 Norman Daniels, “Reflective Equilibrium and Archimedean Points”, p. 86.  Phillips Griffiths, “Ultimate Moral Principles: Their Justification”, in Paul Edwards (ed.), The Encyclopedia of Philosophy (New York: Macmillan, 1967), 8: 177–82. Griffiths sets forth his view in a more preliminary form in an earlier article, “Justifying Moral Principles”, Proceedings of the Aristotelian Society, N. S., 58 (1957/58): 103–24.

68 69

178

6  Possible Solutions

autonomous and objective form of practical discourse. An argument to this effect may be called a transcendental argument.70

Griffiths gives such transcendental arguments for the following principles: impartiality (“any action which it is right or wrong for one person to do is right or wrong for every person to do unless there are some special factors present in the other cases”); rational benevolence (“one ought in action to consider the interests of all beings in the universe”); and liberty (“one ought not to interfere, without special justification, in the chosen course of any rational being or impose on any rational being conditions which will prevent him from pursuing his chosen courses of action”).71 I will not give an account of Griffiths’s transcendental arguments for the principles. As he himself admits, these arguments “are very much oversimplified, and it could not be claimed that they have the force of demonstrations”.72 And even if they had the force of demonstrations, this would be of no avail for deciding our present issue since the principles in question are compatible with both C and D. (At first sight it may be thought that liberty is not compatible with C, but the qualification “without special justification” removes that suspicion. Likewise it may be thought that rational benevolence is not compatible with D, but an adherent of D might convincingly argue that he takes all interests into consideration, even though he does not take them into equal consideration.) And even if Griffith’s principles were not compatible with both C and D, this would not really matter; the concept of “morality” is vague, as well as essentially contested. (See above Sect. 1.5.) Suppose, however, that it could be strictly shown that the very possibility of morality entails the correctness of a certain principle, and that this principle is incompatible with some putative moral theory. Even then the adherent of the theory in question is not defeated. He could forgo the term “moral”, declaring that that his is a normative theory claiming superiority over all other normative theories whether termed “moral” or not.73 (Morality can decide that morality is superior to other normative theories only if morality is the superior normative theory; but that is just the point at issue.) A fourth traditional way of trying to justify a moral principle, it too having a Kantian pedigree, is to adduce the concept of “rationality”, rather than that of “morality”, in support of a principle. In this case, however, you cannot plausibly, when attacked, make the same move as was possible in the former case: you cannot give up the concept “rational” to your opponent, while still claiming that your ­theory

 Phillips Griffiths, “Ultimate Moral Principles: Their Justification”, p. 180.  Ibid., p. 180 f. Griffiths mentions that the second principle was stated by Sidgwick, but not that this is true of the first principle as well (but see Sidgwick, The Methods of Ethics, p. 209). However, Sidgwick’s reason for accepting these principles was another, viz. that they are self-evident. 72  Ibid., p. 181. 73  Similar considerations apply to arguments to the effect that certain moral principles are required by “the moral point of view”. See, e.g., Kurt Baier, The Moral Point of View (Ithaca, N. Y.: Cornell University Press, 1958). 70 71

6.5  Justifying Ultimate Principles: Some Traditional Proposals

179

is superior to his.74 Later in this section I shall discuss whether Kant is able to solve the conflict by means of the categorical imperative. My conclusion is that he is not, and, to my knowledge, no later attempts to solve the conflict by reference to the concept of rationality have elicited principles that would justify either C or D to the exclusion of the other. (This is also true of Alan Gewirth’s attempt to deduce “an egalitarian-universalist morality” from the concept of “rational agency”.75 As far as I can see, this morality is as compatible with C as with D.) And, pace Kant, it would indeed be astonishing if it could be shown that mere reason could decide the long-­ standing issue between C and D. At the beginning of this section I mentioned the possibility of arguing for and against ultimate moral principles by means of another ultimate moral principle. Such an attempt to solve the conflict between C and D would probably be a special case of (what I have called) elimination, an attempt to show that the actions which the new principle considers right and wrong, respectively, are, roughly, the actions that also one of C and D (but not the other) considers right and wrong. In order for such an attempt to be at all successful, the new principle must be more plausible than the principle with which it conflicts; otherwise adherents of the latter principle will have no reason to abandon their principle. (In the remainder of this section I will briefly review three attempts to solve the conflict between C and D with the help of such an “external” principle.) Let us first consider the categorical imperative (CI). Could CI adjudicate the conflict between C and D and support either? In Sect. 2.5, I claimed that CI (just as Kant himself claims) supports D. Suppose that adherents of C agree that CI actually supports D—something which I think that most of them actually do. Would they therefore admit that this is a reason for them to abandon C and switch their allegiance to D? In all probability they would claim that C is a more plausible principle than CI. (Not even Hare believed in C because he believed in CI.76) And it would not be easy to show that they are wrong. Suppose, on the other hand, that adherents of D became convinced by Cummiskey’s and Hare’s arguments that CI supports C, not D.77 Would they therefore switch to C? I do not think so. At least most of them would claim that D is much more plausible than CI, and, in this case too, I doubt that they could be proved wrong. Moreover, it is a matter of great controversy what exactly Kant’s moral theory amounts to, and even such a sympathetic Kant scholar as Onora Nell admits  You can, of course, give up some conception of “rationality” to your opponent, but certainly not the concept. For to claim that a moral principle (theory) is rational means, I take it, that it is supported by the best reasons. And you cannot claim that your principle (theory) is superior to that of your opponent while admitting that it is not supported by the best reasons. 75  See Alan Gewirth, “Moral Rationality”, in John Bricke (ed.), Freedom & Morality (Lawrence: University of Kansas, 1976): 113–50, and his later Reason and Morality (Chicago: Chicago University Press, 1978). 76  In “Could Kant Have Been a Utilitarian”, Utilitas, 5 (1993): 1–16, Hare suggests that Kant’s moral theory “is compatible with utilitarianism” (p. 8). 77  As for Cummiskey’s argument, see his “Kantian Consequentialism”, Ethics, 100 (1990): 586– 615, and Kantian Consequentialism (New York and Oxford: Oxford University Press, 1996). 74

180

6  Possible Solutions

that, although Kant has an impressive theory of moral worth, he has “vulnerable theory of right”.78 Another time-honoured ultimate principle is furnished by the ideal observer theory, going back to Hume, Hutcheson, and Adam Smith: an action is right (wrong) if, and only if, it would be approved (disapproved) by an ideal observer (or spectator).79 The ideal observer (spectator) is variously characterized, but most characterizations agree that he is perfectly rational, impartial, and benevolent, as well as in possession of all relevant empirical knowledge. Now it is often thought that the ideal observer theory supports (some version of) C. Thus Rawls says: The most natural way, then, of arriving at utilitarianism […] is to adopt for society as a whole the principle of rational choice for one man. […] [I]t is by the conception of the impartial and the use of sympathetic identification in guiding our imagination that the principle for one man is applied to society.80

The idea is that in any choice situation, sympathetically identifying with everybody concerned, the ideal spectator takes over their preferences and, on the basis of the combination of all these preferences, approves of the action that maximizes the satisfaction of the combined preferences.81 Thus the model of rational choice for a man who considers only his own interest is applied to rational choice involving the interests of many people, thereby getting a version of C.82 As far as I can see, however, there is nothing in the above characterization of the ideal observer that necessitates his opting for C rather than any of several other alternatives. He will opt for C only if he takes over the preferences of everybody concerned and approves of the action that maximizes the satisfaction of all these preferences. But why, as a matter of conceptual truth, must the ideal observer do that? Even if he sympathetically identifies with everybody concerned, he would still know that they actually are numerically different, so he does not have to treat their preferences as if they belonged to a single individual, viz. himself. As far as I can see, there is nothing that prevents the ideal spectator from, say, disapproving of maximizing preference satisfaction if that means that the worst-off people would  Onora Nell (now O’Neill), Acting on Principle: an Essay on Kantian Ethics, p. viii.  Most often the theory is advanced as a meta-ethical position. Thus Hutcheson says: “When we say one is obliged to an Action, we either mean, 1 […] Or 2. That every Spectator, or he himself upon Reflection, must approve his Action and disapprove his omitting it, if he considers fully all its circumstances. (Francis Hutcheson, An Essay on the Nature and Conduct of the Passions and Affections and Illustrations upon the Moral Sense, 3rd ed., (London: printed for A. Ward et al., 1742 (1728)); the quotation comes from D.  H. Monro (ed.), A Guide to the British Moralists, p. 268. Modern examples of taking it as a meta-ethical theory include Roderick Firth, “Ethical Absolutism and the Ideal Observer”, Philosophy and Phenomenological Research, 12 (1952): 317–45, and Richard Brandt, Ethical Theory, Ch. 7. It is, however, treated as a normative position in, e.g., John Harsanyi, “Morality and the Theory of Rational Behavior”, Social Research, 44 (1977): 623–56, and Shelly Kagan, Normative Ethics, Ch. 7. 80  John Rawls, A Theory of Justice, p. 26. See also ibid., p. 183 ff. 81  Sometimes, of course, more than one of the alternative actions do that. 82  The ideal observer is here described as endorsing preference utilitarianism. By varying what the ideal observer is supposed to take over, he will endorse other versions of C. 78 79

6.5  Justifying Ultimate Principles: Some Traditional Proposals

181

have it worse than the worst-off people given some alternative action. That might just be what benevolence, having a loving concern for everyone, requires. Might the ideal spectator endorse even D? I think he might. Being benevolent as well as relevantly informed concerning the limited benevolence and knowledge of human beings, he might approve of their conforming to D, and so approve of actions that do thus conform. This is not to say that the ideal spectator does not approve of people’s performing supererogatory actions, actions beyond the call of deontological duty; it is only to say that he does not disapprove of people’s not performing such actions. And it is to say that he disapproves of performing horrendous actions. Of course, I do not claim that the ideal spectator endorses D—any more than that he endorses C or some other morality. The ideal spectator is not sufficiently characterized to enable us to say which morality (within a certain range) he would endorse. And I think that any additional characterization of him that would decide the question in favour of one morality would be question-begging. (Why must the ideal observer have these additional characteristics?) A fairly obvious objection to the ideal spectator theory results from confronting it with an analogue to the Euthyphro dilemma: given that the ideal observer approves of the true morality, is this the true morality because the ideal observer approves of it, or does the ideal observer approves of it because it is the true morality? And the objection is that the obviously correct answer is given by the second alternative. Thus, the objection concludes, the ideal observer’s approval does not make the morality true. (Let us call this objection a version of the Euthyphro argument.83 A version of the Euthyphro argument can be—indeed, has been—levelled also against another time-honoured ultimate principle, the one furnished by the divine command theory: an action is right (wrong) if, and only if, it is commanded (forbidden) by God. In my opinion, both objections are well-taken, but there is no need to argue this here. If I am right in what I said about the ideal observer theory above, it cannot adjudicate between C and D. Nor can the divine command theory, since it is not obvious which morality God supports.84 Moreover, it is not an easy task to prove that (the required) God exists.

 In Sect. 1.4, I used the Euthyphro argument both against Hegel’s self-actualization theory and virtue ethics. 84  According to Sidgwick, God probably supports C: “If—as theologians agree—we are to conceive God as acting for some end, we must conceive that end to be Universal Good, and, if Utilitarians are right, Universal Happiness […].” (Henry Sidgwick, The Methods of Ethics, p. 505) And J. S. Mill expresses the same opinion: op. cit., Ch. 2, Par. 22; p. 68. According to Bishop Butler, however, God supports D: “Though the good of the creation be the only end of the Author of it, yet he may have laid us under particular obligations, which we may discern and feel ourselves under, quite distinct from a perception that the observance or violation of them is for the happiness or misery of our fellow-creatures. And this is in fact the case.” (Joseph Butler, Fifteen Sermons Preached at the Rolls Chapel (London: James and John Knapton, 1726), “Sermon XII”, Sec. 31, note; quoted from Five Sermons Preached at the Rolls Chapel and A Dissertation Upon the Nature of Virtue, ed. Stuart M. Brown, Jr. (Indianapolis: Bobbs-Merrill, 1950), p. 79.) 83

182

6  Possible Solutions

6.6  J ustifying Ultimate Principles: Some Non-Traditional Proposals One main hindrance to reconciling C with D is, I have argued, their different attitudes to horrendous actions: D, due to its contractual nature, condemns such actions as immoral, whereas C, due to its distinctive conception of practical rationality, accepts them. As an example, consider the following situation, where what is intrinsically good and bad can be distributed among three persons, X, Y, and Z in either of two ways.

D1 D2

X 0 −5

Y 0 3

Z 0 3

Assume that “–5” stands for something very bad and “3” for something quite good for the person in question. Assume further that D1 is the state that will obtain if you do not interfere. To interfere, thus bringing about state D2, is to cause something very bad to A. This is, however, what C says you should do, since the intrinsic value of the outcome of D2 is 1, while the intrinsic value of the outcome of D1 is 0. (We assume that there is no reasonable consequentialist conception of distributive justice which would alter this result, at least not alter the value ordering between D1 and D2.) One might view the situation as a reduced version of the doctor’s choice, X being the healthy person, Y and Z being two patients in need of new organs. The fault with C, critics have said—underlying, among other things, its prescribing the above kind of actions—is that it treats people as if they were different parts of a single person’s life. (As we saw in the previous section, this is often thought to result from accepting the ideal spectator theory.) Now suppose that “X”, “Y”, and “Z” instead stand for different moments of a single person’s life. Then, the critics would say, it might be morally right for the person in question to choose D2 instead of D1, and, if the person himself were relevantly informed, it would be rational for him to choose accordingly: although to allow (what may be called) “intrapersonal compensation” is not morally wrong, to allow (what may be called) “interpersonal compensation” is. To allow interpersonal compensation, Rawls says, is to ignore “the distinction between persons”, treating them as if they were temporal parts of a single person—but they are not.85 The objection, it might be argued, presupposes that human beings essentially are enduring mental substances; without this presupposition being true, the objection would lose much of its force, and C would appear more plausible. (Why this is (thought to be) so will be explained in the next paragraph.) Now this presupposition, it could be further argued, is in all likelihood a metaphysical illusion. The fact that  John Rawls, A Theory of Justice, p. 187. The quotation comes from a passage where Rawls criticizes U, but the criticism applies equally to any other version of C.

85

6.6  Justifying Ultimate Principles: Some Non-Traditional Proposals

183

we spontaneously and as a matter of course believe that we are enduring mental substances can probably be given a socio-biological explanation: the belief has great survival value both for individuals and for society. But, the argument continues, we have no reason to assume that there exist such substances: neither do we observe them, nor do we have to postulate them in order to explain what we do observe. What we have reason to assume is that there are human bodies as well as mental (and, of course, bodily) events. Any mental event “belongs to” some human body, and to each human body “belong” many mental events. The mental events belonging to a human body are temporally related and also stand in many other relations to each other, as well as to other things. But there are no underlying mental substances (“souls”), each standing in a one-many relation to some mental events, thus “binding them together”—at least we have no reason whatever to believe that there are.86 Let us call the view that human beings are enduring mental substances “the received view”, and the view that they are not “the radical view”.87 Now why does not the objection from interpersonal compensation tell against C on the radical view? The defender of C could argue as follows: “Consider once more the above matrix. According to the objection, what is wrong with harming X in order to benefit Y and Z is, I take it, that he who is harmed and those who are benefited are thought to be different mental substances; it is not considered wrong to act in a similar way in the case of intrapersonal compensation, where X, Y, and Z are not different mental substances. On the radical view, however, X, Y, and Z are not different mental substances; so, given this view, interpersonal compensation should be as acceptable as is intrapersonal compensation. In both cases the compensation concerns independent moments of mental life. A difference is, of course, that in one case the moments “belong to” the same body, whereas in the other case they “belong to” different bodies (and therefore are differently related to each other). But it is hard to see how this difference could be relevant.” Thus far the defender of C. Given that we are not enduring mental substances, C’s treatment of people as if they were parts of a single person seems then to be justified: if intrapersonal compensation is allowed, so is interpersonal compensation. The complaint that C ignores the distinction between persons has no more force, it seems, than the complaint that a theory of prudence ignores the distinction between temporal parts of a person’s life.  Sceptical views of this kind have been set forth in Indian religion and philosophy, e.g., in Buddhism, and in Western philosophy by, e.g., David Hume. A related idea which likewise denies that human beings are enduring mental substances is the Spinozist idea that there is only one enduring mental substance which, however, is instantiated in each human being. An interesting defence of (roughly) this idea is given by Arnold Zuboff in “Moment Universals and Personal Identity”, Proceedings of the Aristotelian Society, N. S., 77 (1977–78): 141–55 and in “One Self: The Logic of Experience”, Inquiry, 33 (1990): 39–68. 87  The received view and the radical view are closely akin to what Derek Parfit calls, respectively, the non-reductionist view and the reductionist view (Reasons and Persons, Part 3). The reason why I do not adopt Parfit’s terms is that his reductionist view is bound up with a certain view of what constitutes personal identity, and I want my counterpart term to be neutral on that issue. 86

184

6  Possible Solutions

Yes, so it seems, a critic might retort, but it is not really so. Even if the highly speculative metaphysics of the self put into the consequentialist’s mouth could be justified, it would not, in turn, justify what many consider the most plausible version of C, viz. preference consequentialism (PC), the view that the only thing that is intrinsically good (bad) is the satisfaction (frustration) of preferences. More than that, the radical view runs counter to PC. So a consistent adherent of PC has to subscribe to the received view. The critic could substantiate his claim as follows. “The received view, the belief that we are, each of us, a unique self throughout our life, is very deeply ingrained in us. Some of us, of course, may (sincerely) claim to be able to see through the veil of Maya and to reject the belief, but so deeply is the belief ingrained in us that outside our study we all take it for granted. The received view, having a crucial survival value, as well as being a sine qua non of social existence, would in all probability continue to have a hold on us, even if, on the theoretical level, we would come to disbelieve it.88 As even a critic of the view admits, ‘[i]t is more than just a belief, it is a manner of existence that reaches into every part of us.’89 Moreover, given this belief, we want to be treated as separate individuals: we are not, in general, willing to suffer that others may prosper. And, given PC, this preference ought, of course, to be satisfied. But, as the defender of C himself admits, in that case interpersonal compensation cannot be justified.” Our imaginary critic might add: “Perhaps a proponent of C would reply that what I have said is true if we consider actual preferences; but what, according to PC, we should strive to satisfy are ideal preferences, (roughly) those preferences people would have if they were relevantly informed.90 But [our critic continues] this does not help. For even if we were relevantly informed—even if we were to accept the radical view and thus believe that we are not enduring mental substances—our preference to be treated as separate individuals would probably not be eliminated. (It is, to repeat, based on “a manner of existence that reaches into every part of us”.) Hence this preference ought to be satisfied even given, what may be called, ideal PC. I think that what our critic says can be answered by means of introducing an improved version of ideal PC. In all likelihood, when formulating their theory, proponents of ideal PC have tacitly assumed that preferences based on false beliefs will disappear when the beliefs are rectified. (But our example shows that this is not always the case.) What they have wanted to claim is presumably that only those of a person’s preferences should be satisfied which it would be irrational (groundless)  In this respect, the radical view is like, for example, the conviction, so much insisted on by Heidegger, that we are mortal beings who might be dead in any moment. Although we vividly recognize its truth whenever we learn that someone we know has suddenly died, it has for the rest almost no practical impact on us. Cf. also my comparison of the consequentialist with the epistemological skepticist in the next section. 89  Arnold Zuboff, “Moment Universals and Personal Identity”, p. 154 f. 90  See, e.g., John Harsanyi, op. cit. Using “true” instead of “ideal”, Harsanyi defines a person’s “true preferences” as “the preferences he would have if he had all the relevant factual information, always reasoned with the greatest possible care, and were in a state of mind most conducive to rational choice” (p. 646). 88

6.6  Justifying Ultimate Principles: Some Non-Traditional Proposals

185

to have if he were fully relevantly informed. (Whether this is what proponents of ideal PC really want to claim is, of course, not relevant; what is relevant that this is a better version of ideal PC than the one taken for granted by our critic.) And, if a person is relevantly informed, it is irrational to prefer to be treated as a separate person and to resist interpersonal compensation. So people’s preferences to be treated as separate individuals ought not to be satisfied given (the improved version of) ideal PC. The conclusion of this extended discussion is, then, that the received view of the self tells against C, whereas the radical view tells against D. Which view is true? Well, I do not know, and I doubt that anybody else is able to give a convincing answer to the question. All in all, then, it seems that speculations concerning our true nature cannot decide the conflict between C and D. I now proceed to a second possible attempt to decide the conflict between C and D in favour of C. The conflict, it may be thought, is fundamentally a clash between two perspectives (points of view) which we are all able to take, the internal and the external. And the internal perspective, it might be argued, favours D, while the external perspective favours C; but, since it can be shown that the external perspective is more objective and, therefore, superior to the internal perspective whenever the two perspectives clash, C is thereby vindicated. I shall now spell out this argument in a little more detail, first giving an account of the two perspectives.91 Much of the time our attention is, more or less spontaneously, directed to the world outside us. We are not only aware of the outer world; we also form beliefs and attitudes concerning things in it. This is to occupy the internal perspective. Sometimes, however, we, as it were, take a step backwards and direct our attention also to ourselves and to our relations to the outer world, forming beliefs and attitudes concerning these objects of attention too. This is to take the more comprehensive external perspective. As Thomas Nagel reminds us, speaking of two perspectives that we can occupy is not wholly accurate. In reality there is a continuum of perspectives. At one end is the point of view of a particular individual, having a specific constitution, situation, and relation to the rest of the world. From here the direction of movement toward greater objectivity involves, first, abstraction from the individual’s specific spatial, temporal and personal position in the world, then from the features that distinguish him from other humans, then gradually from the forms of perception and action characteristic of humans, and away from the narrow range of a human scale in space, time, and quantity, toward a conception of the world which as far as possible is not the view from anywhere within it. There is probably no end-point to this process, but its aim is to regard the world as centerless, with the viewer as just one of its contents.92

 Cf. the spontaneous and the reflective points of view, mentioned in Sect. 3.2. They may be seen as less elaborated versions of the internal and the external perspectives. 92  Thomas Nagel, “Subjective and Objective”, Ch. 14  in his Mortal Questions (Cambridge: Cambridge University Press, 1979), p. 206. 91

186

6  Possible Solutions

(The two perspectives that I have in mind is, respectively, the first and the third one (that of abstracting from the features that distinguish one human being from other human beings).) Sometimes there is a clash between a belief or an attitude formed in the internal perspective and a belief or an attitude formed in the external perspective, either belief (attitude) appearing convincing from one perspective but unconvincing from the other. A classic account of a pervasive clash of this sort in metaphysics and epistemology is given by Hume. In the Treatise, Hume tells us how the skeptical conclusions of his philosophy drive him to despair, making him wonder, “Where am I, or what? From what causes do I derive my existence, and to what existence shall I return? […] What beings surround me? And on whom have I an influence, or who have any influence on me?”93 But when leaving his study Hume finds that “Nature” cures him of his “philosophical melancholy and delirium”. I dine, I play a game of backgammon, I converse, and am merry with my friends; and when, after three or four hours’ amusement, I would return to these speculations, they appear so cold, and strained, and ridiculous, that I cannot find in my heart to enter into them any further.94

In a like manner, it could be argued, the two perspectives give rise to the clash between C and D. One could argue as follows: Most of the time we occupy the internal perspective, to which the structure of D is as it were tailor-made, and from which the universalistic demands of C “appear so cold, and strained, and ridiculous”. When, however, we transcend this perspective and, from the external perspective, reflect on the specific features of D—its agent-reflexivity, self-referential altruism, etc. (see Sect. 2.4)—this morality looks arbitrary and irrational in comparison with C, with its dictum that every person is to count for one. As Thomas Nagel puts it, when he sets out “the difference between consequentialist and more agent-centered views of right and wrong”95: The real issue, therefore, is the relative priority, in regard to action, of two ways of looking at the world. On the one hand there is the position that one’s decisions should be tested ultimately from an external point of view, to which one appears as just one person among others. The question then becomes, ‘What would be best? Which of the actions within my power would do the most good, considering matters from out there, impersonally?’ [- - -] On the other hand there is the position that […] [t]he pursuit of what seems impersonally best may be an important aspect of individual life, but its place in that life must be determined from a personal standpoint, because life is always the life of a particular person, and cannot be lived sub specie aeternitatis.96

Nagel remarks that the opposition between the two positions “looks like a stale-­ mate because each of the points of view claims dominance over the other, by virtue of inclusion”.97 If, as seems to be the case, Nagel talks about an opposition between  David Hume, op. cit., Bk. 1, Pt. 4, Sec. 7, Par. 9; p. 175.  Ibid. 95  Thomas Nagel, op. cit., p. 202. 96  Ibid., p. 205. 97  Ibid. In his earlier The Possibility of Altruism Nagel’s opinion is that the former position wins over the latter. 93 94

6.6  Justifying Ultimate Principles: Some Non-Traditional Proposals

187

two moral positions—not between morality and some non-moral position, say, prudence—many would disagree. For, it might be said, although life cannot be lived sub specie aeternitatis, nevertheless the final court of moral appeal is “from an external point of view, to which one appears as just one person among others”. This is what has been claimed by many moral philosophers. Thus according to the ideal observer tradition (mentioned in Sect. 6.5), what is morally right is what would be approved of by a benevolent, impartial, and all-knowing observer, and the moral point of view (also discussed in Sect. 6.5) is often likewise characterized with reference to impartiality. Thus Hume in a well-known passage says that when a man uses the language of morals, he must depart from his private and particular situation, and must choose a point of view, common to him with others: He must move some universal principle of the human frame, and touch a string, to which all mankind have an accord and symphony.98

But even if the moral perspective is the external perspective, this does not, I think, favour C over D; for, pace Nagel, the external perspective does not automatically give pride of place to C. Even if you abstract from what distinguishes you from other human beings and view yourself “as just one person among others”, you may find the ban on horrendous actions, as well as other deontological requirements, compelling. Let us assume that, when taking the external perspective, the moral legislator is impartial (impersonal), viewing himself as just one person among others. But this is, of course, fully compatible with his legislating norms which themselves are not impartial but agent-centred, etc. In doing this the legislator is not partial to anyone, least of all to himself: every person is allowed to be equally partial to himself and to his nears and dears. It is an open question what norms (and actions) would be chosen from the external perspective. This was evident to Sidgwick. From (i) “the self-evident principle that the good of any one individual is of no more importance, from the point (if I may say so) of the Universe, than the good of any other”, and (ii) the, in his opinion, likewise self-­ evident principle “that as a rational being I am bound to aim at good generally”, he inferred (iii) “the maxim of Benevolence in an abstract form: viz. that each is morally bound to regard the good of any other individual as much as his own”.99 But, as he pointed out, “this reasoning only shows the Utilitarian first principle to be one moral axiom: it does not prove that it is sole or supreme”.100 The reason is, Sidgwick thinks, that as a rational being I am equally bound to aim at my good. (In Sect. 6.2, I gave a critical account of what reasoning Sidgwick thought does prove it.)

 David Hume, An Enquiry concerning the Principles of Morals, Ch. 9, Sec. 6; p. 148.  Henry Sidgwick, The Methods of Ethics, p. 382. 100  Ibid., p. 421. 98 99

188

6  Possible Solutions

6.7  Concluding Remarks In this chapter I have examined several attempts, actual as well as merely possible ones, to solve the conflict between C and D. The result is negative: none of the proposals discussed is able to solve the conflict in a satisfactory way. Some of the proposals have been suggested by other people; as I said at the beginning of the chapter, it is not possible for me to discuss absolutely every suggestion put forward, but I am fairly confident that no important proposal has been left out of account. Other proposals have been suggestions of my own; among these have been attempts to settle the issue by means of other ultimate moral principles, and I have examined (what I think are) the most plausible of these—save one. The remaining one is the contractualist principle. I think it will do the job.

Chapter 7

My Solution

The (perhaps too ambitious) aim of this work is to find out which is the (most) acceptable moral code, that is, the social morality we all should adopt (cf. Sect. 1.5 above). In earlier chapters I have scrutinized the proposals made by (what I claim are) the most plausible candidates, viz. consequentialism (C) and deontological ethics (D). I have argued that neither proposal is (fully) acceptable. Is it then possible to show that either C or D is (wholly) unacceptable? In the previous chapter I argued that the answer is No. I also argued that hitherto attempted reconciliations of C and D have not been successful. In this final chapter I will present my own proposal as to how C and D may be reconciled. The moral theory that lies behind and justifies the moral code I will propose is a contractualist one. I therefore begin (in Sect. 7.1) by sketchily relating my version of contractualism to other versions. In Sect. 7.2, I spell out and defend my version, called the Autonomy Condition. In Sect. 7.5, I state and defend a reconciliation of (my favoured versions of) C and D; in particular, I show that it is justified by the Autonomy Condition. In order to better accomplish this, I will first recapitulate my earlier assessments of C and D. This is done in Sect. 7.4. Section 7.3 is an interlude, in which I (i) compare my version of contractualism with the one defended by Derek Parfit and (ii) discuss Parfit’s claim that the latter entails rule consequentialism. (Since, as I argue; the two versions of contractualism are more or less equivalent, and I have rejected rule consequentialism (see Sect. 6.1), the question whether this claim is true is crucial to the consistency of my normative position.)

© Springer Nature Switzerland AG 2019 J. Österberg, Towards Reunion in Ethics, Philosophical Studies Series 138, https://doi.org/10.1007/978-3-030-12410-6_7

189

190

7  My Solution

7.1  Contract Theories The moral theory I will sketch in the following is, as I said, a kind of contract theory: what is right and wrong depends ultimately on what norms people would agree to.1 In the foregoing, I have repeatedly stressed that (the core of) CSM, and thus also D, are contractualist moralities. In Sect. 2.2, I termed this claim “descriptive (moral) contractualism”, and I characterized it as “the claim that (the most important part of) existing morality is, or is as if it were, the outcome of a contract or an agreement, which serves the interests of everyone who is party to it”. This claim, I also said, “should be distinguished from normative claims to the effect that a morality is justified if, and only if, it is, or is as if it were, the outcome of a contract agreed upon under such and such conditions”. Such a claim could be called “normative (or moral) contractualism”. It is a version of normative (moral) contractualism that I will defend in the following. It is sometimes remarked that there are two main kinds of contractualist theories, the Hobbesian one and the Kantian one. Lately the term “contractarianism” has come to be reserved for the Hobbesian kind, whereas the Kantian kind is called “contractualism”. The difference is often spelled out in something like the following way. Contractarianism has its roots in Hobbes, whose account is based on mutual self-interest. Morality consists in those forms of cooperative behaviour that is mutually advantageous for self-interested agents to engage in. [---] By contrast, any form of contractualism is grounded on the equal moral status of persons. […] According to contractualism, morality consists in what would result if we were to make binding agreements from a point of view that respects our equal importance as rational autonomous agents.2

But these characterizations are compatible with there being hybrid theories that (i) take (at least part of) morality to consist in mutually advantageous co-operation, but (ii), nonetheless, put an emphasis on equality and autonomy. Such a theory is the one that will be set forth in this chapter. (Whether to call this hybrid theory “contractualist” or “contractarian” is a matter of indifference, but a choice has to be made. I have chosen to use the former label, which I will also use to refer indiscriminately to (normative) contract theories of any kind.) To some extent it is a version of 1  It does not, it is true, involve a negotiated compromise between conflicting interests. According to this theory, people make no agreements—explicitly or tacitly—with each other; rather each person (hypothetically) agrees to (conform to) a certain moral code—on condition that others do so as well (see Sect. 7.2). But this is enough to make my theory a contractualist one. Requiring negotiated compromise for a theory to be contractualist would be to exclude, e.g., Hobbes, Kant, Rousseau, and Rawls from the contractualist camp. 2  Elisabeth Ashford and Tim Mulgan, “Contractualism”, Stanford Encyclopedia of Philosophy (Fall 2008 Edition, Edward N.  Zalta (ed.), URL = http://plato.stanford.edu/archives/fall2008/ entries/contractualism/), p. 3. Modern versions of Hobbesian contractarianism are developed in, e.g., James Buchanan, The Limits of Liberty: Between Anarchy and Leviathan (Chicago: Chicago University Press, 1975) and David Gauthier, Morals by Agreement; versions of Kantian contractualism are set forth in, e.g., John Rawls, A Theory of Justice and David Richards, A Theory of Reasons for Action (Oxford: Oxford University Press, 1970).

7.1  Contract Theories

191

Hobbesian contractarianism, although the stress on equality and autonomy gives it a Kantian stamp. It differs from Hobbes’s own theory in that, among other things, it does not assume that people are fundamentally selfish. It differs from David Gauthier’s Hobbesian theory by, for example, not assuming the maximizing conception of rationality, viz. that rational people maximize (expected) utility, that is, seek to realize the greatest possible satisfaction of their preferences.3 (This will become apparent in Sect. 7.2) As I said, it resembles the theory proposed by Derek Parfit. It also resembles the theory proposed by Thomas Scanlon, although the latter is more predominantly Kantian.4 (To further delineate the differences between other contract theories and my own theory, and to try to assess the merits and demerits of these theories in comparison with my own, is beyond the scope of this book; it would require a book of its own.) Against Hobbesian contractarianism Will Kymlicka has objected that, unlike Kantian contractualism, it does not embody one of our deepest held moral convictions, viz. that people have natural rights that should be respected. This shortcoming, Kymlicka says, makes it more of an (inferior) alternative to morality than an alternative morality. Hobbesians say that rights flow from the constraints necessary for mutual beneficial co-­ operation, even when the activity in which people co-operate is the exploitation of other individuals. Everyday morality, however, tells us that mutually beneficial activities must first respect the rights of others, including the rights of those too weak to defend their interests. [—] Mutual advantage cannot be the foundation of morality as we normally understand it, for there are moral claims prior to the pursuit of mutual advantage.5

The objection is perhaps pertinent to certain versions of Hobbesian contractarianism, for example, Gauthier’s, but not, I think, to mine. For, according to my theory, what is the “true” moral code depends on the choice made by rational and relevantly informed people. So, if there are valid moral claims prior to the pursuit of mutual advantage, then, being rational and fully relevantly informed, people will know that and will take it into consideration when choosing a moral code. (Whether there actually are such moral claims (rights) will be discussed in Sect. 7.5.) My theory bears some resemblance to the Golden Rule (GR) principle, discussed in Sect. 4.4 above. But, whereas GR enjoins the agent to do to others what he wants others to do to him, my theory implies that he should do to others what he requires them to do to him. The former principle is, no doubt, more demanding: what I want others to do to me by far outstrips what I require them to do. Suppose that I want others to do x to me. Then GR says that I ought to do x to others. Now suppose that 3  It is true that Gauthier (op. cit.) advocates, what he calls, “constrained maximization”: people should school themselves not to take advantage of those who will not take advantage of them. But the alleged reason for becoming a constrained maximizer is that it is straightforwardly maximizing to do so. 4  T. M. Scanlon, What We Owe to Each Other (Cambridge, MA and London, England: Harvard University Press, 1998). 5  Will Kymlicka, “The Social Contract Tradition”, in Peter Singer (ed.), A Companion to Ethics (Oxford: Blackwell, 1993), pp. 186–96; p. 190.

192

7  My Solution

I find this burdensome; since, however, I cannot stop wanting others to do x to me, I still ought to do x to others according to GR. But finding doing x to others too burdensome—I prefer neither side doing x to both sides doing x—I can stop requiring them to do that, thus not having to do x to others. In both cases, what concrete moral principles people should adopt depends on certain empirical facts, viz. what people want (prefer); if people were different from what they are, so would their moral principles. This might seem unsatisfactory. What actions are right and wrong, it may be thought, is not a contingent fact dependent on people’s wants and preferences. This is a view I strongly oppose. Morality should be in the service of people, not people in the service of morality. What people should do to each other depends on what people want. This is also presupposed by many moral theories, for example, by many versions of consequentialism. (This is, of course, not to deny that the ultimate principles of a moral theory are a priori.)

7.2  The Autonomy Condition My claim is that the fundamental condition of the acceptability of a moral code— what makes the code acceptable—is that it satisfies the following contractualist principle, the Autonomy Condition: (AC) A moral code is acceptable (ought to be accepted) if, and only if, it would be ideally agreed to (chosen) by any qualified person as being the code that he and other people should subscribe to.6 Let me first explain some key notions in the above rather compressed formulation and then state the rationale of the principle. (Later on I will defend the principle against possible objections.) That someone would ideally agree to a moral code means that he would accept the code if (i) he possessed relevant information and (ii) were sitting down in a cool hour. The relevant information consists of any piece of information relevant to making a choice between the codes, especially a rough-and ready understanding concerning, for each feasible moral code, how life would turn out if people subscribed to it. That a person is sitting down in a cool hour means that he is in a state maximally suitable for processing information: being alert, calm, rational, reasonable, etc. By “rational” I mean what The Concise Oxford Dictionary states as one of its meanings: “[being] sensible, sane, moderate, not foolish or absurd or extreme”. (Often being rational is equated with being a utility maximizer, but this, I think, captures just one (prevalent) conception of “being rational”, not the concept itself.) Being rational is then somewhat different from being reasonable, the meaning of which the same dictionary states as “[being] sound of judgment, sensible, moderate, 6  This means that I assume that there is exactly one code that satisfies the condition. Later on I will try to substantiate this assumption.

7.2  The Autonomy Condition

193

not expecting too much, ready to listen to reason”. A rational and reasonable person is thus sensitive to the legitimate claims of other people. That a person is qualified means that he is grown-up, possesses at least average intelligence, does not suffer from mental illness, and is not emotionally disturbed. And that he subscribes to a moral code means that he accepts it and, at least on most occasions, tries to follow (and act according to) it.7 According to AC, then, the criterion for an acceptable moral code is what qualified people would ideally agree to. This is, I think, quite in order. In my opinion, a moral code should be a mechanism for ensuring fair and efficient co-operation, and lack of such things as normal intelligence, mental health, and emotional sensibility makes a person ill-suited for identifying such a mechanism.8 But this is not, of course, to say that only qualified people are required to follow the code. This depends, to be sure, on what qualified people would ideally agree to, but it seems most likely that the code agreed to would require every person to follow it. Of course, the degree of blameworthiness for not following it would not be the same for everyone; thus, for example, a child or a mentally handicapped person is certainly to be treated more leniently than a qualified person. Nor are the beings that have moral standing necessarily narrowly restricted. In all probability an acceptable moral code gives not only human beings but also (many) animals moral standing. (Questions concerning the scope and the sphere of an acceptable moral code will be taken up in Sect. 7.5.) What, then, is the rationale of AC? The condition expresses the principles of (what may be called) moral autonomy and moral equality: what moral code people should regulate their lives by should be decided by them themselves, and in this decision everyone should have an equal say. Given that each (qualified) person has such a say, and that his decision is made under conditions which are, as far as possible, free from elements that are likely to impair his judgment, the moral code subscribed to will be an instrument for fair and efficient co-operation. And this, as I said, is what a moral code should be. But, it may be objected, if AC is justified by the principles of autonomy and equality, the moral theory here put forward is not a genuinely contractualist one; its basic principles are those of autonomy and equality. The objection is partly right, partly wrong. The basic principles of my theory are certainly those of autonomy and equality, but this does not mean, as the objection claims, that the theory is not contractualist. All (credible) versions of contractualism explicitly or implicitly invoke some principle or principles in its support. (The mere fact that people would agree to a moral code for whatever reasons does nothing to show that the code is an acceptable one.) In some cases the supporting principles are external to the contractualist principle, such as the principle pacta sunt servanda. (These versions of contractualism are perhaps vulnerable to the above objection.) In other cases, however,  The last qualification, I take it, excludes the psychopath as well as the amoralist (should he exist).  That an acceptable moral code ensures efficient co-operation follows from the claim, stated above, that morality should be in the service of people; the requirement that co-operation should be fair expresses the principle of moral equality, soon to be explained. 7 8

194

7  My Solution

they are internal to it. And this is the case with my theory. The principles of autonomy and equality are not conceptually independent of, and more basic than, AC but part and parcel of it; they do not give independent support to AC, they are expressed by AC. That the basic principles of AC are those of autonomy and equality means that there is an answer to the question how people can be bound by a merely hypothetical agreement, by an agreement that actually has never taken place. The answer is provided by Shelly Kagan. (Kagan speaks of “fairness” instead of “moral equality”, but that should not make any difference.) Kagan says: Suppose, then, that contractarianism is motivated by appeal to considerations of autonomy and fairness. According to a view of this sort, the very fact that suitably specified bargainers would agree to a set of rules means that those rules would properly respect the autonomy of all, and would constitute the terms of a fair bargain. From this point of view, it doesn’t matter at all whether or not such an agreement has ever actually taken place. The point is that the terms of such an agreement would be fair—fair by virtue of the very fact that everyone would agree to them under appropriate conditions. Thus, from the standpoint of this version of contractarianism, merely hypothetical agreement suffices: provided that the rules would be agreed to (by suitably characterized bargainers), we are bound to obey them.9

Let us consider other actual and possible objections to AC. One objection against contractual principles such as AC—principles which ground the acceptability of a moral code upon the attitudes of actual people—is that the resulting code just expresses people’s ideology or moral prejudices, not the true morality. The answer to this objection is that, of course, the content of the code mirrors people’s attitudes, including their moral ones. But this should not be held against the contractualist approach. A person without any pertinent attitudes would be quite incapable of having any opinion concerning which moral code that ought to be accepted. And the demand on ideal agreement by qualified people provides a sufficient guarantee that the requirement is not based on moral prejudices. (Such people under such circumstances are certainly capable of altering their earlier held moral beliefs.) A similar reply should be given to the objection that in the case of contractual theories, the present one being no exception, the contractual principle is so chosen that the desired moral code follows from it. (Even the very choice of contractualism as the normative starting-point, it could be claimed, settles the issue to some extent.) I do not think, however, that I have deliberately proposed AC just in order to support an independently favoured moral code. Of course, I find the code that (I think) AC supports appealing, and I would reject AC if I thought that it supported a code I would not accept. But that there is such coherence between different levels of one’s moral outlook is neither astonishing nor deplorable. (Moreover, I think that several other contractual principles support the moral code to be set forth in Sect. 7.5; there is thus no one-to-one correspondence between AC and this code. To show this is, however, beyond the scope of this book.)  Shelly Kagan, Normative Ethics, p. 255 f. (Notice that Kagan uses “contractarianism” to refer indiscriminately to both Hobbesian contractarianism and Kantian contractualism.) In my opinion (see Sect. 7.5), we are bound to obey the resulting rules only with respect to people who themselves obey them with respect to us. I do not think that Kagan disagrees.

9

7.2  The Autonomy Condition

195

Another possible objection takes its point of departure in the putative fact that AC conflicts with moral realism, the objection then being that a first-order moral theory should be neutral with respect to meta-ethical theories. Now moral realism is differently characterized by different authors. But, according to a rather common characterization, moral realism states, roughly, that moral facts are “mind-­ independent”. (And this, I think, is what may be thought to make AC conflict with moral realism.) Thus Jonathan Dancy says that realists hold that moral facts are independent of any beliefs or thoughts we might have about them. What is right is not determined by what I or anybody else thinks is right. It is not even determined by what we all think is right, even if we could be got to agree. We cannot make actions right by agreeing that they are, any more than we can make bombs safe by agreeing that they are.10

According to this account, it might seem that AC conflicts with moral realism. For does not AC, roughly, say that if people think that we should subscribe to a moral code, then we should subscribe to it, thus determining what is right (what we should do) by what people think it is right to do? And this, we were told, is just what moral realism denies. But appearances are here deceptive—there is no conflict between AC and moral realism. Moral facts may be held to be “independent of” or “determined by” moral beliefs in two different senses. We should distinguish between constructivism and subjectivism. Constructivism is a meta-ethical theory according to which moral facts are constituted by moral beliefs: there are no moral facts outside moral beliefs. Subjectivism (in the here relevant sense) is a first-order moral position according to which moral facts supervene on psychological facts, for example, on moral beliefs. Now moral realism obviously conflicts with constructivism, but not with subjectivism. AC is a subjectivist position, but it is not—any more than, say, ethical egoism or preference consequentialism—wedded to constructivism. Like other subjectivist positions it is meta-ethically neutral.11 It might be the case, however, that Dancy does not confound constructivism and subjectivism. Perhaps what he means is that since, according to moral realism, moral facts are mind-independent, they cannot supervene on psychological facts, such as that of being agreed to. For it is an essential feature of (our conception of) morality that it transcends our mental capacities: we might be mistaken about morality in a way that we cannot be mistaken about what it is that we agree to. This might be a pertinent objection to certain forms of contractualism, but not, I think, to AC. For, as I will shortly spell out, according to AC we can certainly be mistaken about what moral code we would ideally agree to. It might also be objected that Dancy thinks that contractualism is vulnerable to (a version of) the Euthyphro argument (mentioned in Sect. 6.3). The objection can   Jonathan Dancy, “Moral realism”, Routledge Encyclopedia of Philosophy (London and New York: Routledge, 1998), 6: 534–39; p. 534. 11  It might be claimed that no substantive moral principle can conflict with any meta-ethical principle: since you cannot derive an “ought” from a meta-ethical “is”, there are no logical or conceptual relations between principles of the first and of the second kind. 10

196

7  My Solution

be stated thus: A moral code is not acceptable because it would be ideally agreed to by qualified people; rather these people would agree to the code because (they think) it is acceptable—on independent grounds, not because (they think) they would agree to it. AC and other versions of contractualism are putting the cart before the horse. Now it is, of course, true that qualified people would agree to the code on grounds other than that they would agree to it: according to my theory, they would agree to it because they think that the code strikes a (fair) balance between what they are prepared to do for the sake of others and what they want (require) others to do for their sake. But this is quite compatible with the code’s being acceptable because they would agree to it. For the first “because”-sentence states a factual (explanatory) reason, whereas the second states a normative (justificatory) one. What tells against the claim objected to in the original theistic version of the Euthyphro argument, viz. the claim that something is right because (an amoral) God wills it, is that then anything, even the most horrendous actions, might conceivably be right. For God could conceivably have willed something other than what he actually wills. (Is there any guarantee that he, actually, does not will actions that we consider wrong?) But nothing similar could be true in the present case. People choosing a code regulating their interactions with each other cannot rationally will, and so consider right and acceptable, what is against their interests. Of course, they could rationally have willed something else than what they actually will—but only if they were differently shaped and had different interests. So the claim that, ultimately, what is right and wrong, hence what moral code is acceptable, depends on what people want (will) is not, as is the theistic version, an implausible claim. Thus AC is not vulnerable to the Euthyphro argument. A more substantive objection to AC says that no moral code could possibly satisfy AC: people are so different that no moral norm would, not even ideally, be agreed to by everyone. I disagree. Firstly, people are rather similarly shaped and have rather similar (kinds of) wants: (nearly) everyone wants, for example, not to be robbed, maltreated or killed. So (nearly) everyone would ideally accept norms prohibiting those kinds of behaviour. Secondly, the people whose agreement determines which is the acceptable code are qualified people—people who are rational, reasonable and fully relevantly informed (see above); they are therefore more likely to reach fair and mutually beneficial agreements than unqualified people. Moreover, notice how often in real life also relatively unqualified people are able to reach such agreements in the political and economic spheres. But, someone might rejoin, people are unequally altruistic. A person who is very little altruistic will opt for a moral code that requires less of him than does the moral code that a more altruistic person will agree to. So, again, no moral code will satisfy AC. But this reply overlooks the fact that opting for a less demanding code is two-­ edged: it requires less of you in your dealings with other people, but it likewise requires less of others in their dealings with you. So being limitedly altruistic is not in general a good reason for opting for a less demanding code. This is not to deny that limited altruism in combination with unequal circumstances may make people disagree about what moral code is acceptable in those circumstances. If, for e­ xample,

7.2  The Autonomy Condition

197

I am rich and you are poor, we might have different opinions about what norm of charity to adopt. (This issue will be discussed in Sect. 7.5 below.) But, it may be further objected, some people want norms tailored to their own advantage as being the norms we should all subscribe to, norms such as “Everyone ought to tell the truth—except I, who may lie whenever it suits my interest”. But other people will not, at least not ideally, accept such norms. So, once more, no norms will satisfy AC. Again I disagree. A person may, no doubt, want that the norms people should subscribe to be partial to him. But, as I stipulated above, to ideally agree to a set of norms is to agree to it when thinking rationally. And, thinking rationally, a person cannot possibly believe that there is a relevant difference between him and other people that will justify his being permitted to do what others are forbidden to do. So he cannot justify the claim that they should all subscribe to a code that is partial to him. And, being reasonable, a qualified person will not opt for such codes.12 What then about people whose factual beliefs differ widely from each other— could they really accept the same moral code? Is it really possible that, for example, an orthodox Christian and an orthodox Hindu could (ideally) agree to the same code? The Christian believes, let us assume, that all men are created equal and should therefore be treated equally, whereas the Hindu denies this, believing that the castes into which men are born differ from each other in the same way that the Christian believes that men differ from (other) animals. As Arthur Danto puts it: According to Hinduism, [t]he members of a different caste are, in effect, members of a different species: the distinction between castes is roughly the same as the distinction we recognize between humans and animals. […] We have a different ethic for dealing with animals than we have for dealing with one another. And apart from avoidance prohibitions, nothing in India ever was worked out that could serve as a general ethic overarching the differences between castes, treating men, as it were, as equals in the same way.13

In the face of such divergent outlooks due to deep-seated metaphysical beliefs, it might be thought that not even all qualified people would ideally assent to the same moral code: different metaphysical views will make them assent to different codes. And thus it appears that perhaps no moral code is acceptable. But once more appearances are deceptive. There cannot be more than one true metaphysical view of this kind. And, according to AC, the fact that a person’s acceptance of a moral code is based on false metaphysical beliefs might make the code unacceptable. So the fact that people with different metaphysical beliefs assent to different moral codes does

 Of course, in certain cases actions which are forbidden for one person are permitted for another. If, e.g., X is a policeman, an acceptable moral code may permit him to use violence to an extent that other citizens are not permitted to. 13  Arthur Danto, Mysticism and Morality: Oriental Thought and Moral Philosophy (Harmondsworth: Penguin, 1976), p. 43 f. According to Hinduism, Danto states, each one of us “must find his way, and though it is a plight that we have in common, essentially there is nothing we can do for one another” (ibid., p. 49). 12

198

7  My Solution

not show that there is not a uniquely acceptable moral code. (The same is, of course, true in the case of people with divergent empirical beliefs.)14 I cannot strictly prove that all qualified people would ideally assent to the same moral code. Nevertheless, I think that the belief that there is such uniformity is very plausible. (To some extent its plausibility depends on what norms it is suggested people will assent to. So the final verdict on the question must be postponed to Sect. 7.5, where such a suggestion is made.) Suppose, however, that the uniformity claim is false. Could AC be reformulated in a way that would preserve the underlying idea? I shall briefly sketch some proposals as to how this could be done. One possibility is to go for universal agreement by lowering the standard. Suppose that qualified people will ideally agree to different codes. In this case, the acceptable code, it might be suggested, is the least demanding of these codes. This code then has the following attractive property: everyone thinks that people should at least do what it prescribes, and there is no other code of which this is true. This proposal presupposes that there is exactly one least demanding code among the proposed codes, and this cannot, of course, be guaranteed. Alternatively it might be suggested that qualified people would negotiate with each other as to which code to select. Being both rational and reasonable, they seem to stand a good chance for reaching an agreement, but, again, this cannot be guaranteed.15 A third possible solution is to be content with majority agreement: the acceptable code is the one that most people will opt for, or, if there are ties, any of the codes most preferred. If majority voting is an acceptable way of reaching a compromise in many other areas, it should have some appeal in the present one. (But, as the majority becomes smaller and smaller, the legitimacy of the selected code, of course, diminishes.) A fourth possibility is to let the norms of the acceptable code be as indeterminate as is required in order to transform lack of universal agreement (should there be any) into universal agreement. As a matter of fact, the required degree of indeterminacy probably obtains in any case. For reasons stated in Sect. 2.1 with respect to CSM, the explicitly formulated norms of a moral code to be actually used by people for guiding their actions have to be rather imprecise. They might therefore, within a certain range, be differently interpreted by different people. So qualified people could, in a way, be said to assent to the same moral code, although they would interpret it somewhat differently. (This presupposes, of course, that the alleged d­ ivergence between qualified people concerning which code is the right one is not too great.)  Even if there is only one true metaphysical view of this kind, it is highly probable that people cannot agree as to which one it is. So someone may falsely believe that M is the true metaphysical view and, on account of this, falsely believe that C is the moral code he would ideally agree to. Thus, as I said some pages ago, according to AC we can be mistaken about what code we ourselves would ideally agree to. (In the present case one may, of course, question the underlying normative Hindu premise— that it is not the case that one ought to treat human beings of different castes as equals. Would qualified people really agree to this principle?) 15  It is, of course, important that the moral code is realistic, in the sense that people could (motivationally and otherwise) conform to it. 14

7.3  Parfit on Contractualism and Rule Consequentialism

199

Given AC, to ascertain what the acceptable moral code is requires, I have claimed, that one is a qualified person who has relevant information and is in a state maximally suitable for processing information. It is obviously very difficult to fully satisfy these requirements and next to impossible to know whether one really does. All claims concerning what norms the acceptable code contains are therefore somewhat uncertain and open to revision, especially when one goes into details. This, I think, tells in favour of AC. For, as we all know, it is not always easy to find out what (the true) morality requires: morality does not provide a foolproof algorithm for choosing the right action. But the insight that we are fallible moral investigators should not deter us from trying to answer the questions: What is the acceptable moral code according to AC? How much of C* and how much of D* does it contain?16 But before I answer these questions I will compare my theory to Derek Parfit’s version of contractualism and discuss his claim that the latter entails rule consequentialism.

7.3  Parfit on Contractualism and Rule Consequentialism My version of contractualism, AC, comes very close to the one defended by Derek Parfit.17 Indeed, given some non-controversial assumptions, either version entails the other. AC lays down a necessary and sufficient condition for a moral code’s being acceptable. (Let us assume that only one code fulfils the condition.) It is not, I think, controversial that everyone ought to follow the principles of the one and only acceptable moral code. Hence, everyone ought to follow these principles, which AC characterizes as the principles ideally agreed to by any qualified person as being those that he and other people should subscribe to. If someone agrees to some principles as being those that he and other people should subscribe to, he might be said to agree to the universal acceptance of the principles. That he would ideally agree to this means, I have stipulated, that he possesses relevant information and is rational. If, therefore, any person would ideally agree to the universal acceptance of some principles, it follows that everyone could rationally will the universal acceptance of the principles in question. Putting all this together, we arrive at the version of contractualism that Parfit wants to defend, viz. The Kantian Contractualist Formula: Everyone ought to follow the principles whose universal acceptance everyone could rationally will.18

 Remember that C* and D* are the, in my opinion, most plausible versions of, respectively, C (consequentialism) and D (deontological ethics). C* was introduced at the end of Sect. 4.3 and D* at the end of Sect. 3.4. (What I will say concerning C* and D* in Sect. 7.4, and many things I will say in Sect. 7.5, are, however, just as relevant to most other versions of C and D.) 17  Derek Parfit, On What Matters, Draft of 28 April 2008 (Derek Parfit’s Homepage). 18  Ibid., p. 268. Parfit mentions (on p. 314) a variant of this formula in which “universal acceptance” is substituted by “being universally followed”, but he does not attempt to decide which formula is to be preferred. Of course, also AC could be similarly changed. 16

200

7  My Solution

(By analogous reasoning it can be shown that the entailment also goes in the other direction.) Now Parfit claims that the Kantian Contractualist Formula (hereafter KCF) entails both Scanlon’s Formula: Everyone ought to follow the principles that no one could reasonably reject19;

and the following universal acceptance version of rule consequentialism, UARC: Everyone ought to follow the principles whose universal acceptance would make things go best20

I will not discuss the claim that KCF entails Scanlon’s Formula, but it is important for me to question the claim that KCF entails UARC.21 Parfit’s proof of this claim goes as follows. Kantians could argue: (A) Everyone ought to follow the principles whose universal acceptance everyone could rationally will, or choose. (B) Anyone could rationally choose whatever they would have sufficient reasons to choose. (C) There are some principles whose universal acceptance would make things go best. (D) These are the principles whose universal acceptance everyone would have the strongest impartial reasons to choose. (E) No one’s impartial reasons would be decisively outweighed by any set of relevant conflicting reasons. Therefore (F) Everyone would have sufficient reasons to choose that everyone accepts these UA-optimific principles.22 (G) There are no other significantly non-optimific principles whose universal acceptance everyone would have sufficient reasons to choose. Therefore (H) It is only these optimific principles whose universal acceptance everyone would have sufficient reasons to choose, and could rationally choose. Therefore These are the principles that everyone ought to follow.23

 T.  M. Scanlon, What We Owe to Each Other (Cambridge, Mass,: Harvard University Press, 1998), p. 4 f. (Parfit’s rendering and discussion of Scanlon’s Formula is on p. 318 f.) 20  Parfit, op. cit., p. 296. As with KCF, there is a variant of this formula, mentioned on p. 314, where “universally accepted” is substituted by “universally followed”. 21  In Sect. 6.1, I rejected rule consequentialism. But if (as I claim) AC entails KFC, and if (as Parfit claims) KFC entails UARC, then AC entails (a version of) rule consequentialism. In that case, I cannot consistently both accept AC and reject (all versions of) rule consequentialism. 22  The UA-optimific principles (“UA” being short for “universal acceptance”) are the principles mentioned by UARC, that is, the principles whose universal acceptance would make things go best. 23  Parfit, op. cit., p. 296 f. 19

7.3  Parfit on Contractualism and Rule Consequentialism

201

This argument, Parfit says, is valid. (A) is KCF, with “or choose” added. (B) states a conceptual link between rationally choosing a principle and having sufficient reasons for that choice. (It is assumed that rational choosing involves knowing all of the relevant facts and that you can have sufficient reasons both for choosing p and for choosing q, although p and q are incompatible.) (C) is a, not very controversial, existence assumption. (D) is, largely, true by definition, since Parfit defines the outcome that is “impersonally best in the impartial-reason-implying sense” as the outcome that “from an impartial point of view, everyone would have most reason to want”.24 (E), Parfit says, is the most controversial premise in this argument. In order to defend it, Parfit examines whether there are stronger, partial reasons to choose some alternative to the optimific principles. The result of his examination is negative: the impartial reasons are not outweighed by any conflicting reasons, so everyone could rationally choose that everyone accepts the optimific principles. The remaining premise, (G), strengthens the case: the impartial reasons outweigh any conflicting reasons; so no one could rationally choose that everyone accepts any non-optimific principles. I accept neither premise (E) nor premise (G), but will confine myself to questioning premise (G). In his defence of the premise, Parfit writes: If everyone accepted any such other principle, that would make things go in ways that would be impartially much worse than the ways in which things would have gone if everyone had accepted the optimific principles. In nearly all such cases, things would also go much worse for some unfortunate people. These people could not rationally choose that everyone accepts this non-optimific principle, since they would have both strong impartial reasons and strong personal reasons not to make this choice.25

I disagree. Consider the deontological features of CSM mentioned in Sect. 2.4 above. AC supports the principles corresponding to at least some of these features. And so does KCF if, as I argued above, it is entailed by AC. But, being non-­optimific, the principles in question are not supported by UARC. Hence, premise (G) is mistaken, and neither KCF nor AC entails UARC. The first deontological feature that I shall consider is constraints (see Sect. 2.4: (i) above). The corresponding principle runs: One shall not (hence, ought not) harm other people even if this is the only way to bring about the best results overall. As I said in Sect. 2.4, this principle is usually taken to have a pro tanto character, and whether it may be overridden in a given situation depends on the degree and kind of harm involved, as well as on the amount of good at stake. But in some cases, for example, the Doctor’s Choice, it is certainly not overridden. The “constraints principle” then characterizes such actions as wrong. Since the principle is, ex hypothesi, not optimific, it is not supported by UARC. Nevertheless, I am fairly confident that it is a principle which is supported by KCF: no one could rationally will to live in a world where at any moment he might be killed if this is considered the only way to bring about what is impartially best. 24 25

 Ibid., p. 291 f.  Ibid., p. 310.

202

7  My Solution

Another deontological feature of CSM is options (see Sect. 2.4: (iii) above): to a large extent you are permitted to fulfil your own projects and, in general, do what you prefer, even if this is not best from an impersonal point of view. I think that KCF here sides with CSM: we do not rationally will (choose) to be on moral duty all the time. (It might still be the case that, as M* says, you ought to do what is impersonally best.) But if everyone accepted the principle that you are sometimes permitted to fulfil your own projects even if this is not impersonally best, this, in Parfit’s formulation, “would make things go in ways that would be impartially much worse than the ways in which things would have gone if everyone had accepted the optimific principles”. So the principle is not supported by UARC. The last deontological feature that I will consider is special relationships, the claim that we have special obligations to our near and dear. As in the case of the two other principles, I think that this principle too is supported by KCF but not by UARC. But Parfit claims, somewhat surprisingly, that this is an optimific principle which, therefore, is supported also by UARC. He invites us to consider a situation where you could save either your own child or five other children who are strangers to you. UARC, Parfit claims, permits you to save your child. The optimific principles would not […] however, require you to save these five strangers rather than your child. Suppose that we all accepted and acted on some principle that required us to give no priority to saving our own children from death or lesser harms. In such a world, things would go in one way better, since more children would be harmed. But these good effects would be massively outweighed by the ways in which it would be worse if we all had the motives that such acts would need. For it to be true that we would give no such priority to saving our own children from harm, our love for our children would have to be much weaker. The weakening of such love would both be in itself bad, and have many bad effects. Given these and some similar facts, the optimific principles would in many cases permit us, and in many others require us, to give strong priority to our own children’s well-being.26

Contrary to Parfit, I think that UARC requires you to save the five strangers rather than your own child. For, accepting and acting on an optimific principle that required this would not mean that our love for our own children would be weaker than it actually is. As Parfit himself admits, if “we all accepted and acted on some principle that required us to give no priority to saving our own children from death or lesser harms”, then “more children’s lives would be saved and fewer children would be harmed”. If, therefore, we all accepted a principle always to save the larger group in situations of the above kind, the chances of your child’s surviving in such situations would increase. So, simply out of love for our children we ought to accept and follow such an optimific principle. (It is true, of course, that you should not—nor could you be motivated to—follow such a principle if (you know that) other people do not. But this is true of most of the other principles supported by UARC.27).  Ibid., p. 302. On p. 300, Parfit generalizes this claim so that it applies “both to our own wellbeing and to the well-being of those to whom we have close ties, such as our close relatives and those we love”. (But here our partiality is said to be permitted but not required.) 27  Cf. my discussion, in the first part of Sect. 3.3, of Parfit’s claim that CSM is directly collectively self-defeating. Even if each individual agent follows UARC or CSM, they sometimes fail to bring about the best outcome—due to lack of communication and/or lack of trust. 26

7.4  The Present Standing of C* and D*

203

7.4  The Present Standing of C* and D* In Chap. 3, I discussed objections to D (CSM). I found two of the objections valid and made some changes of D in order to meet these; the resulting theory I called D*. I claimed that D* can meet the other objections to D provided that (the true) morality is contractual. In Sect. 7.2 I argued that morality is contractual. So D* can meet the objections to it, and is therefore likely to be, at least part of, the moral code to be sketched in the next section. C*’s prospects of success are more bleak. In Chap. 5, I presented four objections to C*. One objection, raised in Sect. 5.2, was that C* makes excessively heavy demands, both quantitatively and qualitatively. But, as I admitted, a possible response to the objection is that there is no guarantee that the “true” morality is not very demanding. But this defence of C* does not work if, as I claim, the true morality is contractual. For it is quite unlikely that people would agree to such a demanding morality as C*. So this objection to C* still stands. Another objection, the Argument from Horrendous Actions, was discussed in Sect. 5.4. My discussion of the objection ended by questioning its force. I pointed out that the objection condemns certain actions, but not the “corresponding” omissions. (To kill one person in order to save the lives of five other people (the Doctor’s Choice) is considered horrendous, whereas not saving the life of one person in order to save the lives of five other people (Bystander 1) is not considered horrendous but right (permitted).) The objection therefore rests on the Acts and Omissions Doctrine (which, in turn, rests on a contractualist conception of morality). Adherents of C* could therefore confront adherents of D* with the following choice: (i) show that the distinction between actions and omissions is morally relevant, (ii) extend the objection also to omissions, or (iii) withdraw the objection. And, adherents of C* could argue, since adherents of D* cannot do (i) and are utterly unwilling to do (ii), they have to do (iii), to withdraw the objection. But even if this rebuttal of the argument is unobjectionable, the fact remains that people fear being the victims of horrendous actions in a way that they do not fear being the victim of “horrendous omissions”. Nearly everyone acquiesces to live in a society of horrendous omissions (such as ours), but desires not to live in a society of horrendous actions. Consider, for example, a society which has implemented “the survival lottery” proposed by John Harris.28 The aim of the lottery is to save as many lives as possible  John Harris, “The Survival Lottery”, Ch. 5 in his Violence and Responsibility (London: Routledge and Kegan Paul, 1980). It is, of course, somewhat misleading to say that Harris “advocates” the lottery. Harris thinks that the lottery is “sound in principle but [alas] self-defeating in practice” (p. 82) since it tends to lead to “a gradual deterioration of the health of any society which operates it” (p. 79). This, he thinks, is due to two circumstances. One is that “the computer [used to administrate the lottery] would select only healthy donors, thus […] gradually leading to a society in which those with healthy organs, and perhaps healthy living patterns, were weeded out” (p. 80). The second is that that it would remove “disincentives to imprudent action” (ibid.), since people could then carelessly count on having their diseased organs replaced.

28

204

7  My Solution

in a future society when transplant techniques are much more advanced than at present. According to the proposal, everyone is given a lottery number; whenever two or more people can be saved by a transplant and there are no organs in supply, a computer picks the number of a suitable donor at random; the donor will be killed and his organs given to the needy patients. Unlike the donor in the Doctor’s Choice, everyone in this case knows that he runs a risk of being sacrificed, and the resulting fear of one’s number being selected by the computer at any moment will, of course, detract from the overall positive value caused by the lottery practice. But, as Harris points out, “the chances of actually being called upon to make the ultimate sacrifice might be slimmer than is the present risk of being killed on the roads”.29 So it might very well be the case that the total value in a society is greater if it runs such a lottery than if it does not. But (almost) no one, I think, would opt for its implementation. In Sects. 5.4 and 5.5, I offered two arguments to the effect that C* (C) is collectively self-defeating. According to the first of these objections, C* is self-defeating in LD-situations (Lawn-Crossers’ Dilemma Situations). This is a very serious objection to C*. For, as I pointed out in Sect. 3.3, many of the fateful situations confronting us today—such as overpopulation, waste of non-renewable resources, the greenhouse effect, etc.—are LD-situations on a worldwide scale. That C* is self-defeating in these situations means then that, heedless of the imminent catastrophes, it gives the obviously wrong prescriptions in these situations. The second of the self-defeatingness objections argues that the immensely useful practices of truth-telling and promise-keeping are absent in the C* (and C) society, the society where there is common knowledge that (almost) everyone acts according to C* (C). This means that people in the C* society bring about less good than people in the D* (and D) society, the society where there is common knowledge that (almost) everyone acts according to D* (D).

7.5  How Should We Act? Now what collective moral code is then acceptable according to my autonomy condition AC? What code would be ideally agreed to by qualified people? I suggest the following preliminary answer (which I will elaborate in the following): (M*) One (i) shall not harm other people (beings); but (ii) shall benefit them if the benefit is great and the cost is not too high; and, moreover, (iii) one ought to benefit them as much as possible. The first two clauses of M* use the term “shall”, which in my terminology is stronger than the term “ought” used in the third clause: “shall” implies “ought”, but not I think that neither circumstance is in fact an obstacle to implementing the lottery. Firstly, it is seldom the case that all of a person’s vital organs are diseased and therefore unsuitable for being reused. And, secondly, people could be warned that those with unhealthy living patterns would be allowed only a small probability of themselves having a transplant should they need one. This, I am sure, would deter at least many people from debauchery. 29  Ibid., p. 70.

7.5  How Should We Act?

205

vice versa.30 As I use the terms, not doing what you shall do is morally wrong, but not doing what you merely ought to do is not. Not doing what you shall do makes you morally blameworthy, but doing what you shall do does not necessarily make you morally praiseworthy (cf. Sect. 4.2). By contrast, doing what you (merely) ought to do makes you, at least to some extent, morally praiseworthy, but not doing what you (merely) ought to do does not (normally) make you morally blameworthy. The first two clauses speak about obligatory actions and omissions, whereas the third clause speaks about supererogatory actions and omissions.31 The first clause forbids actions such as lying, cheating, breaking promises, injuring, and killing. Not harming people is morally more important than benefiting them. For, as I argued in Chap. 2, that people by and large abstain from cheating, injuring, and killing each other is a prerequisite for the very existence of civil society, whereas benefiting others is not. And, as I claimed in Sect. 5.5 and will argue at length in the Appendix, the absence of the practices of truth-telling and promise-­ keeping means the absence of efficient social co-operation. The obligation to benefit people stated in the second clause concerns cases such as giving first aid to injured people, saving a person from drowning by throwing a life-preserving buoy to him, and regularly sending some money to a charity organization. The third clause is meant to apply to all kinds of supererogatory actions, not only heroic and saintly ones. As J. O. Urmson points out: [T]here can be cases of disinterested kindness and generosity, for example, that are clearly more than basic duty requires and yet hardly ask for the high titles, ‘saintly’ and ‘heroic’. [—] It is possible to go just beyond one’s duty by being a little more generous, forbearing, helpful, or forgiving than fair dealing demands, or to go a very long way beyond the basic code of duties with the saint or the hero.32

Now for a defence of M*. As is easily verified, M* lies between C* and D*, though much closer to D* than to C*. I will first show that it is superior to C* and then that it is superior to D*. In order to show that M* is superior to C*, I will show, firstly, that it accords with AC (the Autonomy Condition introduced in Sect. 7.2), whereas C* does not, and, secondly, that two consequentialist arguments can be given in 30  It is sometimes alleged that, as used in moral contexts, saying that you ought to do something implies (or suggests, presupposes, etc.) that there are reasons for doing it, whereas saying that you shall do something does not imply this. If that is true, “shall” does not imply “ought” in these contexts. In my use of the terms, however, there is no such difference: both terms are meant to be reason-implying. 31  M* is thus rather similar to J. S. Mill’s theory in Ch. 5 of his Utilitarianism and also to the one proposed by J. O. Urmson in his “Saints and Heroes”. We should look for a theory, Urmson says, “that will allow for both absolute duties, which, in Mill’s phrase, can be exacted from a man like a debt, to omit which is to do wrong and to deserve censure, and which may be embodied in formal rules or principles, and also for a range of actions which are of moral value and which an agent may feel called upon to perform, but which cannot be demanded and whose omission cannot be called wrong-doing.” (J. O. Urmson, “Saints and Heroes”, in Joel Feinberg, Moral Concepts: 60–73; p, 67; originally published in A. I. Melden (ed.), Essays in Moral Philosophy (Washington: University of Washington Press, 1958): 198–216.) 32  Ibid., p. 65.

206

7  My Solution

support of it. The first task will be accomplished by, step by step, bringing C* closer to M* and, at each step, show that the move is sanctioned by AC. It will be based on the objections set out in Chap. 5 and summarized in the preceding section. (The first argument to the effect that M* is superior to C* thus rests on the plausibility of AC, but the second and third arguments do not.) Consider, then, a society where everyone subscribes to C*. As I argued briefly in Sect. 5.5 and will argue at length in the Appendix, Hodgson is right that such a society would lack the very useful practices of promising and truth-telling. It would therefore be at an enormous disadvantage as compared to a society where everyone subscribed to a morality—call it M1—in which the principles of promise-keeping and truth-telling took precedence over the principle of utility, but which was otherwise the same as C*. Since, obviously, qualified people would ideally prefer that they subscribed to M1 rather than to C*, M1 is therefore more acceptable than C* according to AC.33 But M1 is not a wholly acceptable moral code. Like C*, M1 prescribes horrendous actions (cf. Sect. 5.3), and it is beyond question that qualified people ideally (as well as actually) prefer living in a world where they do not have to be constantly afraid of becoming the victim of horrendous actions to living in a world where they have to. Let M2 stand for the moral code which forbids harming some people for the (greater) benefit of other people but is otherwise like M1. I think that qualified people ideally prefer M2 to M1: a world where people subscribe to M2 is a much safer place to live in than a world where people subscribe to M1. Now M2 has inherited from C* a very strong demand to benefit other people: one ought to do as much good as one can, giving up what is most valuable to oneself if necessary (see Sect. 5.2). For nearly everyone this certainly goes against the grain. On the other hand, everyone can rationally assume that there is some non-negligible probability that he will find himself in a situation where it is important for him to be helped. Striking a balance between the desire to get help when needed and the desire not to need to help when the cost is too high, people would ideally prefer M3, a code that is the same as M2 except that it distinguishes between obligatory and non-­ obligatory cases of benefiting: you ought to benefit people as much as possible, but you have an obligation to benefit them only if the benefit is great and the cost is not too high. So M3 brings us very close to M*. The gap is completely closed, I think, if M3 gives way to M4, the moral code that enjoins mutually beneficial co-operation in LD-situations (see Sects. 3.3 and 5.4) but is otherwise the same as M3. Also the last step is certainly something that qualified people would ideally agree to; so it is required by AC. But is it justified by M*? It is—provided that the term “one” in the formulation of the principle is taken to refer, not only to individual agents, but (perhaps somewhat strained) also to groups of individual agents. Thus M* furnishes solutions to LD-situations: it prescribes, firstly, that groups shall adopt mutually beneficial practices, and, secondly, that

 The meaning of “ideally” here is, mutatis mutandis, the same as in the above formulation of AC. If somebody ideally prefers a code X to any other code, then he ideally agrees to X.

33

7.5  How Should We Act?

207

every member of a group shall participate in such practices. (The justification of the latter prescription is the one given at the end of Sect. 3.3.) It might be objected that the preferences referred to in my above defence of M*are insufficiently altruistic; therefore, to claim that they are relevant and should decide the issue between M* and C* is to beg the question against C*, the altruistic morality par excellence. The answer to this objection is that, of course, people’s preferences are far from completely altruistic; they are mostly, to once more cite Broad’s apt characterization, self-referentially altruistic (see Sect. 2.4). But, firstly, things would not be different if people were complete altruists; if, for example, I am worried about horrendous actions that might suddenly befall some of my near and dear, I would not cease to worry, and the relevant preferences would not change, if all people were to become my near and dear. Only indifference, not altruism, could work such a change. And, secondly, there is, as far as I know, no version of C* according to which only altruistic preferences ought to be satisfied. So why should C* limit the set of relevant preferences to altruistic preferences in this case? This leads to my second argument to the effect that M* is superior to C*. It is an argument directed to adherents of preference satisfaction consequentialism (PSC): if you think that people’s (intrinsic) preferences ought to be maximally satisfied, and, as I have argued, (at least most) people prefer there being general compliance with M* rather than with C*, then you are committed to, and should opt for, M*. There are, of course, certain differences between PSC as ordinarily conceived and the version that figures in the present argument. One difference is that the former is concerned with any objects of preference, whereas the latter is only concerned with compliance with either of two moralities.34 But this means that the former version includes the latter one, and it cannot possibly have the consequence that the former is valid but the latter is not. Another difference is that, ordinarily, the objects of preference are consequences of individual actions, whereas in the present version they are consequences of lasting patterns of action. But, again, it is difficult to see how this could be a crucial difference.35 The third argument for M*’s superiority over C* starts by calling to mind that M* is the code that results from C* when it is modified in such a way that the troublesome objections to C*—the Hodgson objection, the Argument from Horrendous Actions, and the Arguments from Being Too Demanding—are disarmed. The argument then points out that it follows from the earlier discussion that the amount of preference satisfaction is greater, and the welfare (pleasure, happiness) level is higher, in a world where there is general compliance with M* than in a world where there is general compliance with C*. This is thus a strong consequen This is not, of course, quite true. According to some versions of PSC so-called anti-social (sadistic, malevolent, etc.) preferences do not count, whereas according to other versions only selfregarding preferences count. 35  An analogue to the present version of PSC, but concerned with individual actions, is the “selfinterest-vote” utilitarianism discussed by Ronald Glossop in “Is Hume a ‘Classical Utilitarian’?”, Hume Studies, 2 (1976): 1–16; p.  2  f. Like the present version, this version is not sensitive to intensities of preferences. 34

208

7  My Solution

tialist argument in favour of M* as against C*: on both hedonist and preference versions of consequentialism, people ought to conform to M* rather than to C*. (Some consequentialists allege other intrinsic values instead of, or in addition to, welfare and preference satisfaction. I think that, for any plausible objective list versions (as they are called) of C*, M* realizes more value than C*.)36 Now for a defence of M* as against D*. As I said above, M* is much closer to D* than to C*: in fact, the first two clauses of M* may be said to constitute the Core of D*. But the third clause of M* goes beyond D* and is much closer to C* than to D*. The third clause, remember, says that one ought to benefit other people as much as possible (within the restrictions resulting from the first clause), whereas D* is less demanding. Can this departure from D* be justified? The actions recommended by the third clause of M* are, as I said above, supererogatory actions. Those who perform supererogatory actions are, no doubt, prompted by many different motives, singly or in combination. A rather common motive is, I believe, the thought that something that I can do for another person will promote something good or prevent (remove) something bad for him. This thought is generated by that manoeuvre of putting oneself in other people’s shoes, variously called sympathy, empathy, compassion, etc.37 When identifying with you by taking your “perspective”, I take over your preferences as well and judge their satisfaction good, and their frustration bad. Not only is the thought that the satisfaction of anyone’s preferences is something intrinsically good and their frustration something intrinsically bad a powerful motive, it also provides a justification of supererogatory actions. Given that supererogatory actions bring about the satisfaction of preferences, and that this is something (intrinsically) good, it seems reasonable to claim that one ought to perform such actions, that is, that one ought to conform to (what may be called) the Core of C*.38 Thus the Core of C* is part of M*, the acceptable moral code according to AC. The acceptable moral code, according to (my version of) contractarianism, is thus a combination of the Core of D* and the Core of C*. I will now discuss a problem that pertains to all kinds of contractualism. Consider two groups of people (beings), A and B. Members of A are able to harm members of B, but members of B cannot repay in kind. (A might be, let us say adult people and  Note that the second and the third of the above arguments are quite different: the second is about preferences had by people in the actual world, the third is about preferences had by people in two merely possible worlds. 37  Of course, these terms are sometimes used in different, more narrowly circumscribed ways and are in these contexts not synonymous. 38  That one ought to bring about anything intrinsically good and ought not to bring about anything intrinsically bad means that (i) one ought to maximize the good and (ii) one ought to minimize the bad. Consequentialism as usually understood is the combination of (i) and (ii) with (iii), the sumranking principle, which says that the value of the consequences of an action is assessed by just adding up their positive and negative values. (See the “Introduction” to Amartya Sen and Bernard Williams (eds.), Utilitarianism and beyond (Cambridge: Cambridge University Press, 1982), p. 4.) What I call, maybe somewhat gratuitously, the Core of C* consists of (i) and (ii). (iii), which is what gives rise to the Argument from Horrendous Actions (see Sect. 5.3), is excluded by the first cause of M*. (I offer no substitute for the sum-ranking principle.) My theory, then, bears some resemblance to Frankena’s theory, discussed in Sect. 6.1. 36

7.5  How Should We Act?

209

B small children or non-human animals.) If, as contractualism maintains, morality is a device for obtaining efficient co-operation and securing mutual advantage, it seems that, according to contractualism, members of B have no moral standing; members of A lay under no moral restrictions with respect to members of B. But, most of us think, they certainly do. Hence, contractualism is not acceptable. Let us see whether M* can avoid this kind of objections. How does M* handle the objections against D (CSM) raised in Chap. 3? One of the objections was that D, being of a contractual nature, condones harming those who are not able to repay in kind, such as members of racial and other minorities and members of technologically less advanced societies. (As I said there, many wars, including colonial conquests and much exploitation may be due to, and supported by, this feature of D.) Like D*, however, M* escapes the objection since it states that people shall not harm other people (except in self-defence and in order to prevent them from harming other people—see below). One reason for this is that there would be no universal assent among qualified people to a moral code which allows harming other people (except in self-defence and in order to prevent them from harming other people). Probably no qualified person will ideally assent to a code which permits harming him (and his near and dear). A second reason is that harming a person is to frustrate his preferences, and, as stated above, this is something intrinsically bad and therefore to be avoided. A related objection is that contractual theories give no protection to children, the aged, and the mentally handicapped. M* escapes this objection. The answer I gave to the objection that M* gives no protection to members of minorities or of less advanced societies obviously applies to the case of the aged and the mentally handicapped too. An additional reason why M* protects the aged and the mentally handicapped is that any rational person recognizes that he runs (a high) risk of becoming old and defenceless and (some) risk of becoming mentally handicapped. He therefore wants to be “morally protected” should such things happen. (This reason does not, of course, apply in the case of children. But both of the reasons mentioned some paragraphs ago do.) Does M* give the same protection to animals? (In the above formulation of M*, I left it undecided whether the principle applies to all beings or only to human beings.) Whether and to what extent animals have moral standing is hotly debated in our days, and what answer M* gives is a question that belongs to the area of uncertainty mentioned in Sect. 7.2. (It is not easy to ascertain what norms qualified people would ideally agree to in the case of animals.) I preliminary suggest that all beings have moral standing according to M*, but that human beings have a better moral standing than non-human animals. (“All animals are equal, but some animals are more equal than others.”) Thus, I suggest, qualified people would agree to norms to the effect that animals may be (painlessly) killed and used for food, and that in the choice between, say, saving a human being and an animal, you should save the human being. But, as I said, the issue belongs to the area of uncertainty and my answer may certainly be contested. (To enter deeper into the issue would involve taking a stand on many controversial questions and cannot be undertaken here.)

210

7  My Solution

Another deontological feature of CSM is its Acts and Omissions Doctrine. This doctrine, as I noticed in Sect. 3.4, supports morally reprehensible verdicts in, for example, certain euthanasia cases: it is, according to CSM, morally permissible not to connect a person to a respirator (an omission), if he is judged unable to survive, but forbidden to disconnect him (an action), if, due to a mistaken diagnosis, he has already been connected to the respirator. Like D*, M* does not underwrite the application of the doctrine in such cases. What explains the doctrine is that in ordinary circumstances avoiding to actively harm anyone is much easier than helping everyone from being harmed. Thus CSM says that it is wrong to bring about a person’s death, but, in general, not (as) wrong not to save people’s lives. I think that the acts and omissions doctrine, when applied in ordinary circumstances, can be justified; almost no one can conform to a requirement to do everything he can to help people from being harmed. But extending the doctrine from ordinary circumstances to the medical field, as CSM unwittingly does, cannot be justified. For, even if the emotional stress involved in the action (switching off the machine) is certainly greater than that involved in the corresponding omission (not connecting the patient to the machine), it is in this restricted area no more difficult to help people from being harmed than to avoid causing them harm. So M* recognizes that, in this as in other areas, professional ethics might differ from general ethics. This is also the case with respect to another deontological feature of CSM, viz. its doctrine of agent-relativity: each person should only see to it that he himself does not commit a wrong action. Even if, as I agree, social morality must have this feature, there is no justification for extending it to, say, research ethics. But this is what adherents of CSM unwittingly do. In Sect. 3.4, I cited a social scientist who disavowed any responsibility for what use other people could make of his research results. Now what explains, as well as justifies, CSM’s being agent-relative is that, in general, holding people responsible for what other people do is to demand too much. But this does not mean that it is an unrealistic demand that, when acting in certain roles, people should assume such an extended responsibility. Thus it is possible to subscribe to CSM’s notion of restricted responsibility in general while insisting that certain professional codes should be less restrictive. And this, I think, is what qualified people would ideally agree to. All in all, then, it seems that CSM is not suitable to serve as a basis for professional ethics. The reason is that the situations to which professional ethics should apply lack certain of the features that characterize the situations which CSM is designed to deal with. There is, therefore, much that tells in favour of professional ethics being more consequentialist than CSM is. And this is what M* is. Now, given that we all should comply with M*, how ought one to act in a world— for example, the present one—where so many people do not comply with M*?39 In a morally imperfect world such as ours, people fail to conform to M* in two respects: many people harm other people, sometimes on a large scale, and most people ­benefit  The distinction presupposed by the question is the same as the distinction made by Rawls between strict compliance theory and partial compliance theory. (See Rawls, A Theory of Justice, p. 8 f.; cf. also Sect. 2.3.)

39

7.5  How Should We Act?

211

only their near and dear. Of course, most people abstain from the most flagrant kinds of harm, and some people benefit even people outside the narrow circle of their near and dear. But, on the whole, the lot of humanity is a depressing one: many people are wronged and many go unhelped. So, once more, how should the individual agent act in this world? M* states that one shall not harm other people. But, as mentioned above, there are, of course, exceptions to this: cases where the agent himself or other people are actual or prospective “innocent” victims of harm, and the only (or best) way to stop or prevent the harm is to harm the aggressor40 How much, and what kinds of, such “second-order” harm is prescribed or allowed by M* cannot be stated in a short formula; it depends on indefinitely many possible circumstances. But some general guidelines can be given: Note that harming X in order to stop or prevent him from harming Y is actually to benefit Y. So exceptions to the ban on harming other people fall either under the demand to benefit or under the right to self-defence.41 But whereas the demand to benefit is restricted (see below), the right to self-defence is not restricted to the same extent. I think that qualified people would ideally agree to a norm permitting people to use adequate means to protect themselves against being harmed. As Hume, on behalf of CSM professes, even a virtuous man who has fallen into the society of ruffians “must consult the dictates of self-preservation alone, without concern for those who no longer merit his care and attention”.42 In the ideal world where everyone conforms to M*, no one needs to benefit other people very much; if all help those who need help, the contribution demanded from each person will be very small. (There is, for example, in the ideal world no need for helping people from being (intentionally) harmed by other people.) But in the actual, non-ideal world, where so many people go unhelped, the need for help is enormous; there is almost no limit to what (almost) everyone could do in the way of benefiting other people. What norm would qualified people ideally agree to? I think that the M* norm for such cases is less demanding and also more indefinite than that of C*: very few people want to be on permanent moral duty, but it is difficult to give an exact measure of how much help people would ideally agree to give. The norm is certainly more demanding than the corresponding norm of CSM. Above all it has a wider scope: rational people are disposed to help strangers too. (This is not to deny that it is often both easier and more efficient to benefit one’s neighbours than people living far away.) Now, is M* itself best characterized as a deontological theory, as a consequentialist theory, or as something else? What speaks in favour of the first alternative is that the deontological part of M*, viz. its first two clauses, is a more genuinely moral one than is the consequentialist part, its third clause: its use of “shall” is a full-fledged normative one, whereas the “ought” of the consequentialist part, as I  That a person is innocent here means that, at the time in question, he does not harm or intends to harm other people. 41  Since benefiting other people is supererogatory, not obligatory (see the beginning of this section), there is a demand, not a duty, to benefit. 42  David Hume, An Enquiry concerning the Principles of Morals, p. 86. 40

212

7  My Solution

have argued in Sect. 4.2, is an evaluative one. Moreover, as I tried to show in Sects. 5.4 and 5.5, C* is collectively self-defeating and therefore by and large enjoins people to conform to D*. This makes C* parasitical on D*. And, finally, the C* part of M* is less like the standard versions of C than its D* part is like the standard versions of D, since it is subordinated to the first (deontological) part, and, moreover, does not have the property of sum-ranking. What speaks in favour of M*s being a consequentialist theory is, first, that, since its D* part is a condition for its C* part being effective, the C* part provides a rationale for the inclusion of the full-fledged normative part D* into M*: a moral code without a deontological part produces less value than a code with such a part. Moreover, as I tried to show, two consequentialist arguments can be given for M*’s being superior to C*. But if consequentialism justifies the inclusion of D* into M*, and if there are additional consequentialist arguments that support M*, M* is as much a consequentialist as a deontological theory. As a matter of fact, I think that it is not illuminating to characterize M* as either a consequentialist or as a deontological theory (moral code). In the language of chemistry, M* is a compound, not a mixture, of its two elements. My proposal for a moral theory, M*, is, we have seen, a compound of consequentialism (C) and deontological ethics (D). This makes it exposed to criticism from both camps, adherents of D not liking its C part, and vice versa. But, on the other hand, adherents of C (D) may find M* more acceptable than D (C) because it contains enough of their own theory. But be that as it may. I think that it speaks in M*’s favour that it contains elements from both C and D. As we have seen, both C and D have a long history, and so has the confrontation between them. It would be astonishing if either view rested on some hitherto undetected error, whether moral, factual, or conceptual. A much better explanation of the long-lived debate between adherents of C and D is that each side has grasped part of the truth, their error being one-sidedness, perhaps based on the belief that the truth cannot be complex. As a matter of fact, I think that a moral theory that does not do (at least some) justice to both C and D is almost certainly false.

Appendix: Hodgson’s Argument

Introduction In Consequences of Utilitarianism, D. H. Hodgson tries to show that act utilitarianism is self-defeating (in a sense to be explained later).1 By “act utilitarianism” Hodgson understands the following principle: (AU) An act is right if and only if it would have best consequences, that is, consequences at least as good as those of any alternative act open to the agent.2

By “consequences” are here understood actual consequences, but, says Hodgson, his arguments also apply to principles which refer to probable or reasonably foreseeable or foreseen consequences.3 All such principles, Hodgson claims, are self-­ defeating; that is, he claims that even correct application of act-utilitarianism, either by everyone in a community [where there is common knowledge that everyone is an act-utilitarian], or by individuals in a non-­ act-­utilitarian society like our own, would not necessarily have better consequences, and

1  D. H. Hodgson, Consequences of Utilitarianism: A Study in Normative Ethics and Legal Theory (Oxford: Oxford University Press, 1967). (Hodgson himself does not use the term “self-defeating”.) 2  Ibid., p.  1. Hodgson does not—like, e.g., Moore—distinguish between obligatory and merely right actions, and he apparently takes it for granted that an action is wrong if, and only if, it is not right. 3  Ibid., p.  13. According to Allan Gibbard, Hodgson’s arguments apply only to such principles: “Now theories of what is objectively right have no direct bearing on the problem Hodgson raises. Hodgson’s thesis concerns the behaviour of rational act-utilitarians who do not know what to expect from each other: they know they lack relevant information. In order to know what they will do, we need to know how they base their decisions on information they know is incomplete. (Allan Gibbard, “Act-Utilitarian Agreements”, in A. I. Goldman and J. Kim (eds.), Values and Morals (Dordrecht: Reidel, 1978): 91–119; p. 96.) As far as I can see, however, all this is consistent with taking AU as a theory of objective rightness.

© Springer Nature Switzerland AG 2019 J. Österberg, Towards Reunion in Ethics, Philosophical Studies Series 138, https://doi.org/10.1007/978-3-030-12410-6

213

214

Appendix: Hodgson’s Argument would probably have worse consequences, than would acceptance of specific conventional moral rules and personal rules.4

In both cases, Hodgson says, AU is in dire straits with respect to two very useful “practices” (“institutions”), viz. those of promising and truth-telling. According to AU, keeping a promise and telling the truth do not, per se, constitute reasons for acting, so these practices will be absent in a society where (there is common knowledge that) everyone accepts and conforms to AU. This, argues Hodgson, gives the utilitarian agent a severe handicap in his endeavour to make the world as good as possible, for the absence of these practices in a society means that co-operation is absent too. And the absence of co-operation makes the act-utilitarian society (henceforth the AU-society) totally different from all existing societies, depriving it of most of the fruits of civilization: “there could be no human relationships as we know them”.5 And this would, of course, deprive the AU-society of a great amount of value. But, as Hodgson also stresses, this is fully compatible with everyone’s fully conforming to AU. For this means only that the consequences would be the best possible in the circumstances; and since the circumstances (universal acceptance and rational application of act-­ utilitarianism, and common knowledge of this) preclude human relationships, the best possible consequences in these circumstances would be worse than consequences which are not the best possible in other more favourable circumstances.6

Likewise the consequences of being an act-utilitarian in a non-AU-society would be worse than those of being an adherent of common-sense morality (CSM). If Hodgson is right, AU is, to use Parfit’s notion, directly collectively self-­ defeating for people in the AU-society. (In what sense it is self-defeating for act-­ utilitarians in non-AU-societies will be considered in section “Truth-telling and promise-keeping in non-AU-societies” below.) A theory T is directly collectively self-defeating, Parfit says, when [i]t is certain that, if we [the members of some group] all successfully follow T, we will thereby cause our T-given aims to be worse achieved than they would have been if none of us had successfully followed T […].7 4  Ibid., p. 38. This is a very guarded formulation of the claim. To accord with the general tenor of Hodgson’s exposition, “probably” should be replaced by “almost certainly” or something of that sort. 5  Ibid., p. 45. It would not, of course, be a society in the proper sense of the word. 6  Ibid. 7  Derek Parfit, Reasons and Persons, p. 54. According to Parfit, AU, being a species of consequentialism (C), cannot be directly collectively self-defeating. For, Parfit argues, “[w]e successfully follow C when each does the act which, of the acts that are possible for him, makes the outcome best. If our acts do jointly produce the best outcome, we must all be successfully following C″ (ibid.). Now Hodgson does not deny that—given the circumstances, viz. universal acceptance and rational application of act utilitarianism, and common knowledge of this—people in the AU-society always jointly produce the best outcome and that, this being the case, each of them is successfully following AU. What he claims is that, if they had instead accepted and rationally followed a certain other morality, roughly CSM, their circumstances then being different, they would, on the whole, have jointly produced better outcomes than the ones that they did produce. And, if this is so, then

Truth-Telling in the AU-Society: Hodgson’s Argument

215

Hodgson’s arguments were hotly disputed by several prominent moral philosophers. Hodgson himself did not respond to the criticism, and I know of only three published defences of his theses.8 The general impression was, no doubt, that Hodgson was wrong and had lost his case. In my opinion, however, Hodgson was (mainly) right and his critics were (mainly) mistaken. I will therefore reopen the case and try to prove Hodgson right, not so much by adding new arguments for Hodgson’s claims as by criticizing the objections brought against it. I will begin with the case of AU in the AU-society and the fate of truth-telling in this society (Sections “Truth-telling in the AU-society: Hodgson’s argument” and “Truth-telling in the AU society: objections”). I will then discuss the fate of promise-keeping in the AU-society (Sections “Promise-keeping in the AU society: Hodgson’s argument” and “Promise-keeping in the AU society: objections”) and, rather briefly, the case of being an act-utilitarian in a non-AU-society (Section “Truth-telling and promise-­ keeping in non-AU-societies”). The Appendix ends with some concluding remarks (Section “Concluding remarks”).

Truth-Telling in the AU-Society: Hodgson’s Argument Hodgson asks us to consider a society in which everyone (i) accepts AU as his only personal rule; (ii) always tries to act in accordance with it; (iii) is highly rational and understands all relevant implications of the previous two conditions; and (iv) knows of the previous three conditions, knows that everyone else knows of them, and so on. In this society, Hodgson says, there would be no communication. In actual non-­ AU-­societies, such as ours, there are several good reasons for a person to think that what other people tell him is taken to be true by them; chief among these is the fact that truth-telling is required by CSM, which most people adhere to. In actual societies, therefore, telling the truth usually has better consequences than lying. For if truth-telling is required by their morality, members of a non-act-utilitarian society will ordinarily rely on what they are told; and because of that they will often make arrangements, based on what they are told, which will have bad consequences unless they are told the truth. Also, people will resent those who lie to them; they will blame them and count less on them in the future, all of which are harmful consequences of acts of lying. (pace Parfit) AU obviously can be directly collectively self-defeating. (Perhaps, however, the opposition between Parfit and me is due to our understanding the self-defeatingness condition differently.) 8  See Adrian Piper, “Utility, Publicity, and Manipulation”, Ethics, 88 (1978): 189–206 (discussed in Subsec. 3: (iv) below); Donald Regan, Utilitarianism and Cooperation (Oxford: Oxford University Press, 1984) (mentioned in Subsecs. 3: (ii) and (iii) below); and C. Provis, “Gauthier on Coordination”, Dialogue: Canadian Philosophical Review, 16 (1977): 507–9 (discussed in Subsec. 3: (i) below). In addition, Hodgson is shortly mentioned with approval in G.  J. Warnock, The Object of Morality (London: Methuen, 1971), p.  33  f., and in Dan Brock, Recent Work in Utilitarianism”, American Philosophical Quarterly, 10 (1973): 241–76; p. 258.

Appendix: Hodgson’s Argument

216

In the AU-society, however, A has a reason to tell B the truth (rather than what is false) only if he has reason to believe that B will take what he is told as true (rather than false). But, as A knows, B will take what he is told as true only if he has reason to believe that A has a reason to tell him the truth—which completes the circle. Obviously, the same circularity is involved in A’s attempt to communicate the truth by telling B what is false. Hodgson therefore concludes that in the AU-society no one would take what he is told as more likely to be true than false, or vice versa; attempts to communicate information would therefore be pointless. The central passage of Hodgson’s argument runs as follows: Being highly rational, the informant would know that the taking of the information as true rather than false was a condition precedent for telling the truth to have very best consequences; and so would not believe that it would have very best consequences unless he believed that the other would take the information as true rather than as false. Also being highly rational, the other would know this, and would not so take the information unless he believed that the informant believed he would so take the information. And this, of course, the informant would know. He could reason that if the other would take his information as true rather than false, it might have very best consequences to tell the truth, and that if he supposed that the other would so take his information and concluded that it would have very best consequences to tell the truth, then there would be good reason for the other so to take the information. But (as both would know), the informant could equally well reason that if he supposed that the other would take his information as false and concluded that it would have very best consequences not to tell the truth, then there would be good reason for the other to take the information as false.9

Before giving Hodgson’s critics a hearing, I will make a preliminary assessment of Hodgson’s argument. The argument may be represented by means of the following matrix (where “O1”–“O4” denote the four possible outcomes).

A tells the truth A does not tell the truth

B takes what he is told as true O1 O3

B takes what he is told as false O2 O4

How are the values of the outcomes related to each other? What Hodgson explicitly claims in his argument is that: (i) The value of O1 is better than that of O2. (A’s telling the truth has better consequences if B takes what he is told as true than if he takes it as false.) and that: (ii) The value of O4 is greater than that of O2. (B’s taking what he is told as false has better consequences if A does not tell the truth than if he tells the truth.)

9  Hodgson, op. cit., p. 44. Both Hodgson and his critics assume that the informant always knows the truth. They also assume that it is always better (in utilitarian terms) to believe what is true than what is false. I will later question both these assumptions.

Truth-Telling in the AU-Society: Hodgson’s Argument

217

How are the values of the other outcomes supposed to relate to each other? (Hodgson does not explicitly tell us, and (i) and (ii) imply nothing concerning that.) But, evidently, Hodgson must assume that: (iii) The value of O1 is greater than that of O3. (B’s taking what he is told as true has better consequences if A tells the truth than if he does not.) For otherwise—if the value of O3 were at least as great as that of O1—A’s not telling the truth would be (analogous to) what in game theory is called a weakly dominating strategy: given (ii), it would have better consequences than its alternative if B took what he was told as false, and, given the negation of (iii), it would have at least as good consequences if, instead, he took what he was told as true.10 Being highly rational, A would of course know this and would therefore—contrary to what Hodgson assumes—know that AU prescribed that he should always not tell the truth. Likewise Hodgson must assume that: (iv) The value of O4 is greater than that of O3. (If A does not tell the truth, the consequences of B’s taking what he is told as false are better than those of his taking it as true.) For otherwise—if the value of O3 were at least as great as that of O4—B’s taking what he is told as true would be a weakly dominating strategy: given (i), it would have better consequences than its alternative if A told the truth, and, given the negation of (iv), it would have at least as good consequences if, instead, he did not tell the truth. Being highly rational, B would of course know this and would therefore— contrary to what Hodgson assumes—know that AU prescribed that he should take what he is told as true. (Moreover, given the truth of both (i) and (ii), the negation of both (iii) and (iv) would mean that, according to AU, A should not tell B the truth, although B should take what he is told as true. It would be rather astonishing if this combination of strategies always had the best consequences.) Now, from (i)–(iv) it follows that: (v) The value of O1 is greater than both that of O2 and that of O3, and the value of O4 is greater than both that of O2 and that of O3. And this means that both O1 and O4 are (utilitarian analogues to) what in game theory is called equilibrium outcomes: Given that A tells the truth, B’s taking what he is told as true has better consequences than his taking it as false, and, given that B takes what he is told as true, A’s telling the truth has better consequences than his not telling the truth. Likewise, given that A does not tell the truth, B’s taking what he is told as false has better consequences than his taking it as true, and, given that  In game theory, an agent’s strategy S is weakly dominating if, and only if, its outcome is (i) better for him than the outcome of any alternative to S, given at least one combined choice of strategies by the other agents and (ii) as good for him as the outcome of any alternative to S given all other combined choices. Applying the notion to AU, “better for him” is replaced by “better”, period, and “as good for him” is replaced by “as good”, period.

10

218

Appendix: Hodgson’s Argument

B takes what he is told as false, A’s not telling the truth has better consequences than his telling the truth.11 This, in turn, means that if either O1 or O4 obtains, both A and B do what AU prescribes. What about the relation between O1 and O4 then? Hodgson’s thesis requires that (vi) The value of O1 is the same as that of O4. (The consequences of A’s telling the truth and B’s taking what he is told as true are just as good as the consequences of A’s not telling the truth and B’s taking what he is told as false.) This does not directly follow from (i)–(v), but (v) entails that either (a) exactly one of O1 or O4 is optimal, or (b) they are equally best. If, however, (a) were the case, A and B would not, contrary to Hodgson’s claim, be confronted with a co-­ordination problem. Being highly rational, they would know that (a) was the case and would act so as the secure the optimal outcome (whichever that was). The value of O1 is, therefore, the same as that of O4, and these outcomes are equally best (optimal). This, at any rate, is what Hodgson claims. Is the claim acceptable? This depends crucially on how A’s second alternative, viz. his not telling the truth, is to be understood. Hodgson assumes that AU prescribes that one should (at least often) successfully communicate the truth, and he claims that in the AU-society there are two equally good ways of doing that: telling the truth and taking what one is told as true, and not telling the truth and taking what one is told as false. Now suppose that B asks A what time it is, and that A, knowing that it is three o’clock, tells B that it is two o’clock. Suppose further that B takes what he is told as false. Here we have a case of not telling the truth—in the sense of telling something incompatible with the truth—and taking what one is told as false.12 But it is not successful; it results in B’s believing that it is not two o’clock, which is of course true, but not in his believing the more specific and interesting truth that it is three o’clock. This he would have believed if, instead, A had chosen the first way of communicating the truth, that is, 11  In game theory an outcome is in equilibrium (is an equilibrium outcome) if, and only if, for each agent, the consequences of his unilateral defection from the chosen strategy is not better for him. When applying the notion to AU, “not better for him” is replaced by “not better”, period. We might call this “equilibrium in the weak sense”, and define a notion of equilibrium in the strong sense by replacing “not better (for him)” by “worse (for him)” in the above formulation. It should be noted that O1 and O4 are equilibrium outcomes even in the strong sense. 12  It might seem that this example—as well as other similar examples given by Hodgson, myself, and other commentators, which will be introduced later on—begs the question. For according to AU, since truth-telling will not exist in the AU-society, no communication will take place there. It is, however, almost unavoidable not to use such examples when discussing Hodgson’s claims. One way out would be to follow Peter Singer’s advice and stipulate that just before the events assumed in the examples take place “everyone in an until-then-normal society is miraculously converted to act-utilitarianism” (Peter Singer, “Is Act-Utilitarianism Self-Defeating?”, Philosophical Review, 81 (1972): 94–104; p. 97. It is assumed that the habit of talking with one another would linger on for a while. But this proposal gives too much ground to Hodgson’s opponents: if habits from the pre-utilitarian society, such as talking with one another, still exercise their influence on people in the AU-society, so might the habits of taking what one is told as true or expecting that promises will be kept. A better expedient, it seems to me, is to take such examples as being preceded by a tacit counterfactual: “assuming that people in the AU-society were to communicate with each other, then …”

Truth-Telling in the AU-Society: Hodgson’s Argument

219

had told him the truth, and B had believed what he was told. Since in many cases it is important to learn the more specific truth, AU prescribes in these cases the first way of communicating the truth rather than the second way. If, therefore, not telling the truth is understood as just telling something incompatible with the truth, the value of O4 is often less than that of O1. If Hodgson’s claim that the value of O4 is the same as that of O1 be acceptable, not telling the truth must be understood as telling the denial (or negation) of the truth. If, for example, A instead tells B that it is not three o’clock, and B takes what A says as false, then B will believe the more specific truth that it is three o’clock. It seems plausible that the consequences of this way of communicating the truth that it is three o’clock are just as good as those of A’s telling that it is three o’clock and B’s believing what he is told. Given this understanding of not telling the truth, it seems prima facie reasonable that, quite generally, the value of O4 is exactly the same as that of O1. And this is how I will I will understand Hodgson’s notion of not telling the truth. (In the following, I will refer to the two interpretations as the contrariety and the contradictory interpretations, respectively, of “not telling the truth”.) Both Hodgson and his critics seem to assume that people in the AU-society would have no or very little reason to deceive each other. Hodgson himself does not even envisage this possibility, and in his critique of Hodgson Peter Singer says that “if everyone were an act-utilitarian most of the reasons, selfish or unselfish, which we would otherwise have for lying would not exist”.13 But this is not true, as I will try to show in some detail. (Most of my examples apply to act-utilitarians both in and outside the AU-society. Since, however, lying is not considered immoral in the AU-society, utilitarians in this society have more often (utilitarian) reasons to lie than have utilitarians outside it.) Act-utilitarians do not, of course, have selfish (justificatory) reasons for lying. But they have other, specifically utilitarian ones—given that the lies were (rightly estimated to be) believed.14 There are, for example, many cases where telling the truth is more harmful than lying, since believing the truth makes people unnecessarily upset (angry, sorry, depressed, etc.). The classic example of (what might be called) “beneficial deceit” is that of the doctor who dishonestly tells his patient that his condition is hopeful. In other cases telling the truth is indirectly harmful because, by making the addressee upset, it causes him to act suboptimally. Thus telling the person attacked by a tiger that he has only one cartridge left in his rifle may make him so nervous that he misses his target. A third kind of cases where AU justifies telling lies are cases where the person addressed holds certain erroneous beliefs which cannot be easily corrected. Suppose,  Peter Singer, op. cit., p. 100. A non-angelic act-utilitarian can, of course, have selfish motivating reasons for acting, although he cannot have selfish justificatory reasons for doing what he does. I take it that what Singer means is that the act-utilitarian lacks selfish justificatory reasons, not that he lacks selfish motivating reasons. Although the latter interpretation would strengthen Singer’s case, the former has the clear advantage of being true. 14  Remember that the following examples are taken to be preceded by the tacit counterfactual mentioned in the last note but one. 13

220

Appendix: Hodgson’s Argument

for example, that A wants his employee B to finish some important work by lunchtime the next day. As B is about to leave the office today, A asks him if the work is finished. It is not, but B knows that there will be plenty of time for him to finish the work tomorrow without detriment to his other tasks in the office; he also knows that A will not believe this, and will therefore order B to work overtime today if B tells him the truth. Since B does not want to miss tonight’s important football match between Djurgården and AIK, he lies to A, knowing that his lie will not be detected. (This is a case where the lie has a selfish motivation but, it may be assumed, a utilitarian justification.) A fourth kind of cases where AU holds that it is justified to lie are those where you can save several people by sacrificing one person with the help of a lie, or, more generally, cases where you can bring about good consequences for some people by bringing about (a lesser amount of) bad consequences for other people. Consider the following version of a stock example of this kind. (The incident is supposed to take place in our AU-society.) You are standing on a bridge watching a trolley hurtling down the tracks below you toward five innocent people. The brakes have failed, and the only way you can stop the train is to impede its progress by throwing some heavy object in its path. You yourself are not heavy enough, but luckily there is a fat man standing on the bridge next to you, and you could easily push him over the railing and onto the tracks below if only he takes another step forward. If he does and you then push him, he will die, but the five men will be saved. The fat man wants to take a step forward in order to have a better view, but, being aware that you are a fellow utilitarian, asks before he takes the further step, “You will not push me over the railing, will you?” “Of course not”, I say with a reassuring smile. The fat man—who, like Hodgson and his critics, thinks that people in the AU-society have no or very little reason to deceive each other—is completely reassured and takes the further step. I push him over the railing, thereby saving the five men on the tracks. It might be objected that I do not need to lie in the above situation; since the fat man is supposed to be a utilitarian, I only need to explain the situation to him, and he will gladly volunteer. But, first, there may be no time for me to explain. And, secondly, talking to him will probably not be effective: not even a convinced utilitarian might be sufficiently motivated to sacrifice his life at a moment’s notice, however much this is demanded by his morality.15 Because of the amount of justified lying in the AU society the third of the propositions discussed above is not true: sometimes O3 is better than O1, for, as we have seen, given that B takes what he is told as true, A’s not telling the truth sometimes has better consequences than his telling the truth. For the same reason proposition  Saying this is not to take a stand on the vexed issue whether akrasia really exists, that is, whether anyone ever voluntarily acts contrary to what he thinks he ought to do. If akrasia does not exist, then, on some meta-ethical views, not being able to force oneself to do what one thinks that utilitarianism says one should do shows that one does not “really” think that one ought to do it. But not being able to force oneself to do everything that one thinks utilitarianism says one should do does not show that one is not “really” a utilitarian.

15

Truth-Telling in the AU-Society: Hodgson’s Argument

221

(iv) is not true: given that A does not tell the truth, B’s believing what he is told sometimes has better consequences than his not believing it. Nor are propositions (i) and (ii) true: common knowledge of the fact that (iii) and (iv) are not generally true undermines their validity too. Suppose that A knows that p is the case, but that B’s believing that non-p is the case has better consequences than his believing the truth. Suppose also that A (correctly) believes that B believes that A, for utilitarian reasons, wants to deceive him by lying to him: B will therefore take what A says as false. A therefore tells B that p, whereupon B believes that p is false, that is, that non-p is the case. Hence (i) is not true. (By means of a similar example it can be shown that (ii) is not true either.) Further, if (i)–(iv) are not true, nor are (v) and (vi): sometimes O1 and O4 are not in equilibrium and are not both optimal. Does the fact that in certain cases propositions (i)–(iv) are false undermine Hodgson’s claim that people in the AU-society are confronted with an unsolvable co-ordination problem? No, it does not. For, firstly, the propositions are false only in cases where AU prescribes deceit. But, obviously, in most situations of communication AU does not prescribe deceit, so in most cases the propositions are true and hence give rise to a co-ordination problem. This means that communication of true information in the AU-society is blocked even when it is prescribed by AU. Secondly, the truth of the propositions is a sufficient, not a necessary, condition for there being a co-ordination problem: the problem exists also in many cases where (i)–vi) are false, that is, in many cases of deceit. Let us divide such cases into two main groups. The first group consists of those cases where B does not believe that A wants to deceive him. Those cases obviously pose the same problem as do cases where A does not aim at deceiving B: B does not know whether what is stated is meant to be taken as true or false, and A does not know whether it is best to tell the truth or to lie. The second group consists of those cases where B believes that A wants to deceive him. Those cases pose no co-ordination problem given that, in addition, (a) A believes that B believes that A wants to deceive him; (b) A believes that B does not believe that (a) is the case; and (c) B believes that A does not believe that B believes that A wants to deceive him. If (a)–(c) obtain and, say, p is true, the rational thing for A to do is to tell B that p. For, as A (correctly) believes, it is then rational for B to believe that p is false, that is, to believe that non-p is the case. In all other cases that belong to the second group there is a co-ordination problem. In nearly all cases of communication of information, whether true or false, in the AU-society, there is then an unsolvable co-ordination problem. In nearly all cases, therefore, the rational thing for A to do is to keep quiet, and, if A cannot refrain from talking, the rational thing for B to do is to suspend judgment.

222

Appendix: Hodgson’s Argument

Truth-Telling in the AU Society: Objections The critics of Hodgson’s theses have concentrated on his argument concerning truth-telling in the AU-society, in some cases treating the one concerning promise-­ keeping merely by implication. Since different critics raise different objections to Hodgson’s arguments I will discuss the objections separately, beginning with those attacking the argument concerning truth-telling. (i). Gauthier’s Objection David Gauthier discriminates between the two interpretations of “not telling the truth” mentioned in the preceding section.16 He notes that, according to the contrariety interpretation, O1 is better than O4 (the outcome of A’s telling the truth and B’s taking what he is told as true is better than that of A’s not telling the truth and B’s taking what he is told as false); given this interpretation, there is, therefore, no problem of co-ordination. But, Gauthier concedes, according to the contradictory interpretation, O1 and O4 are both optimal, and in this case there is a co-ordination problem. But, he thinks, this problem can be solved by means of the notion of salience. Stating that-p and believing what is stated is [a] more direct way of communicating that-p than stating that-non-p and believing the negation of what is stated. Hence salience attaches to the outcome of telling the truth and believing what is told. A second argument will reinforce this conclusion. There are circumstances in which it is possible to verify whether stating that-p is to be taken as a way of communicating that-p, or a way of communicating that-non-p. I say “The cat food is in the cupboard and the cat is not in the kitchen,” and you look and see whether the cat food is in the cupboard and the cat not in the kitchen, or whether the cat food is not in the cupboard and the cat is in the kitchen. If act-­consequentialists tell the truth in these situations, they thereby make telling the truth and believing what is told salient, not just for such situations but in general. In this way they develop the practice of communicating information by telling the truth.17

According to Gauthier, then, stating the truth and believing it is salient, since it is “a more direct way” of communicating the truth than stating the opposite of the truth and believing its negation. What then does it mean that the former is “more direct” than the latter? (Gauthier does not tell us.) Presumably that it is simpler and therefore requires less mental effort. But why is this thought to make it more salient than the latter? Presumably because it is held that people, ceteris paribus, prefer what is simpler and requires less mental effort to what is less simple and requires more mental effort. But this is certainly not a universal truth: people often have the opposite preferences. Would they have it in the present case? Hodgson thinks that

 David Gauthier, “Coordination”, Dialogue: Canadian Philosophical Review, 14 (1975): 195– 221; repr. in David Gauthier, Moral Dealing: Contract, Ethics, and Reason (Ithaca and London: Cornell University Press, 1990): 375–97; my references are to the reprint. 17  Ibid., p. 293. (By “act-consequentialists” Gauthier understands both act-utilitarians and adherents of ethical egoism.) 16

Truth-Telling in the AU Society: Objections

223

they would have preferences of both kinds and that these would cancel each other out. In a passage seemingly anticipating the present objection he says: The difficulty of telling a lie in a non-act-utilitarian society arises mainly because of the need to tell a ‘good’ lie, in order to avoid both detection and the bad consequences of someone’s being misled. In our postulated society, unless the act-utilitarian principle required the truth to be told, there would be no need to prevent detection, and no question of anyone’s being misled; and so the lie would not have to be a ‘good’ one. A minimal degree of inventiveness might perhaps still be required to tell a lie; but we may assume that our rational act-utilitarians would have this, and that if any disvalue were involved in the effort required to use this inventiveness, it would be balanced by the satisfaction of exercising the skill.18

Who is right, Hodgson or Gauthier? Before addressing this question I want to point out that if Gauthier is right, O1 and O4 (see above) are not, contrary to what he assumes, both optimal: the preferences for the former outcome tip the balance and make O1 the unique best outcome. Rational act-utilitarians would therefore opt for O1, not because it is salient, but because it is the unique best outcome. But would they? It should be admitted that it is not easy to decide whether Hodgson or Gauthier is right on this issue: our knowledge of the mental make-up of fully rational act-­ utilitarians is far from complete. There are, however, certain considerations that tell in favour of Hodgson’s position. (These considerations, it should be noted, are relevant to both of Gauthier’s arguments set forth in the above quotation from him.) In a society such as ours, where there is usually good reasons to believe what other people tell us, and people have formed habits to believe accordingly, it often requires more effort not to believe what someone says than to believe it. But there is no reason to think that people form these habits in our AU-society. (Simply to assume that people in this society habitually believe what other people tell them begs the question.) And what else could make it the case that it required more effort from people in this society not to believe what other people told them than to believe it? (It might be replied that some things that people tell us seem intrinsically more plausible than their contradictories and are therefore more easily believed than disbelieved. This is true, of course, but then other things that people tell us seem intrinsically less plausible than their contradictories.) But all this is actually beside the point. What Gauthier says suggests that he thinks that the alternatives are (i) believing what someone says and (ii) believing its contradictory. But obviously there is a third alternative, viz. that of suspending belief. And if there is (almost) just as much reason to believe what someone says as to believe its contradictory, this alternative seems to be the unique rational epistemic stance to take. Since our act-utilitarian agents ex hypothesi are highly rational, this

 Hodgson, op. cit., p. 43. Cf. what Rawls calls “the Aristotelian Principle”: “other thing being equal, human beings enjoy the exercise of their realized capacities (their innate and trained abilities), and this enjoyment increases the more the capacity is realized, or the greater its complexity.” (John Rawls, A Theory of Justice, p. 426.)

18

224

Appendix: Hodgson’s Argument

is then the stance they are expected to take—and this is so even if it would require somewhat more effort than just believing what they are told.19 The crucial question is whether people in the AU society would interpret the situations they confront according to the contrariety or the contradictory interpretation?20 If both A and B interpret a situation in the former way (and know of each other that they interpret it that way, and so on), then A will tell B the truth and B will take what he is told as true. (I temporarily disregard the kinds of situations, mentioned in the previous section, where A, for utilitarian reasons, lies to B.) Evidently there are situations which they would interpret according to the contrariety interpretation. To be a situation of that kind, what is required is that A believes that B believes that there are more than two alternatives any one of which might be true (and that B believes that A believes that, and so on). If in such a situation A tells B, concerning one of these alternatives, that it is true, then, as A knows, B has a reason to think he is told the truth. The alternative way of communicating the truth consists in A’s telling B what is not true and B’s taking what he is told as false. But, as both can easily verify, if the (according to B) possibly true alternatives are more than two, the former way of communicating the truth is more likely to be successful. And, as they both know, this gives A a reason to choose it. If, for example, as B believes, A believes that B believes that either C, D, or E has committed the crime, and A knows who did, the rational way of communicating the truth to B is telling the truth. If, however, B only suspected two people, the situation must be interpreted according to the contradictory interpretation, and both ways of communicating the truth would be equally good. The fact that people in the AU-society would sometimes view the situations they confront according to the contrariety interpretation certainly weakens Hodgson’s position, but not very much. For there are many situations that would be interpreted according to the contrary interpretation. Moreover, the kinds of situations where A has utilitarian reasons for lying to B are rather frequent. Hence, the situations where B really has good reasons to believe that A tells him the truth would probably not be many.21  This is also Donald Regan’s opinion. “But the easiest thing of all is to avoid the question of how to take A’s remark, by ignoring it entirely.” (Regan, op. cit., p. 35.) 20  As I pointed out in Sec. 2 above, Hodgson evidently interprets the situations according to the contradictory interpretation. 21  C. Provis, op. cit., objects to Gauthier’s solution to the following co-ordination problem: A and B want to meet each other either at x or at y, no matter at which place. For some reason, going to x is the salient option. Gauthier suggests that each agent should restrict his possible actions to (i) seeking the salient outcome and (ii) ignoring it. The agent seeks the salient outcome by going to x and ignores it by randomizing on an equal basis between going to x and going to y. It is easily seen that both agents seeking the salient outcome, that is, going to x, has a higher expected value than both agents ignoring salience. Being rational, they therefore both go to x, thus successfully coordinating their actions. Provis objects that Gauthier’s depiction of the situation is inadequate: the agents have a third alternative, viz. seeking the non-salient outcome, that is, going to y. And if both go to y, they likewise successfully co-ordinate their actions and will meet each other. 19

Truth-Telling in the AU Society: Objections

225

(ii). Singer’s Objections Peter Singer raises three objections to Hodgson’s argument.22 In the first, Singer asks us to consider the case of an office clerk B living in the AU-society, who on a particular day intends to work overtime. B’s only means of transportation home is by bus, and for some reason it is very important that he does not miss the last bus. B asks his colleague A when the last bus departs. A knows the answer, and he also knows that it would have best consequences to inform B of it. But what answer should he give? If Hodgson is right, Singer says, it is not more likely that A can inform B by telling the truth than by telling a lie. But, Singer objects, Hodgson is wrong: there is a utilitarian reason for A to tell B the truth. If A tells a lie, then, whether or not B believes it, he will (in all probability) not go to the bus stop in time. But there is a fifty-fifty chance that B will take the answer given by A to be true and, therefore, if told the truth, will go to the bus stop in time.23 This gives A a reason for telling B the truth. And, Singer says, this reason is actually a fairly strong one: Once there is some reason for A to tell the truth, there is more than enough reason for him to do so. For B, being highly rational, will have thought of the considerations just pointed to, and will be aware that there is a reason for A to tell him the truth, and A will know this, and so on. So we get the Hodgson spiral working in the other direction, and A will have the normal utilitarian reason for telling the truth—that is, that B will take the information to be true and make arrangements based on its truth.24

This objection presupposes for its validity the contrariety interpretation of Hodgson’s argument, viz. that A has to choose between telling the truth or telling any of a number of false answers. Given the contradictory interpretation—which, as I have argued above, is the one that Hodgson must subscribe to—the objection obviously does not work. If, as Singer says, “there is a fifty-fifty chance that B will take the answer given by A to be true”, there is a fifty-fifty chance that he will take it as false. If, therefore, the truth is that p, and A says that not p, there is a fifty-fifty chance that B will take the answer as false and hence believe that p, that is, believe the truth. As a matter of fact, Singer’s objection might be mistaken even given the contrariety interpretation. What is wrong with both Singer’s and Mackie’s criticism of Provis is, no doubt, right. But, as far as I can see, Gauthier could accept Provis’s objection and still have a good case—if, that is, he can defend his view concerning the importance of salience against my objections to it. 22  Peter Singer, op. cit. 23  It might be objected that the same beneficial consequences will also obtain, with the same probability, in certain cases of telling B a lie, viz. in those cases where A tells B that the bus will depart some time earlier than it actually will. For, if in those cases B believes that A tells him the truth, he will not miss the bus. In defence of Singer we might stipulate that coming to the bus “in time” means coming to the bus stop either just when the bus is due to depart or “shortly before” (admittedly a vague expression). If B arrives earlier than “in time”, then (we might further stipulate) the consequences are worse than if he arrives “in time”: he prefers go on working in his office to waiting for the bus longer than a “short while” before it is due to depart. 24  Singer, op. cit., p. 98.

226

Appendix: Hodgson’s Argument

Hodgson, Donald Regan says, is the assumption that the behaviour of an agent which is supposed to constitute a piece of communication will be taken by the other agent who perceives it to have some communicative effect. But there is no reason to assume that there will be any communicative effect at all.[- - -] Singer completely ignores the possibility that B will pay no attention whatever to what A says.25

(This accords with what I said above concerning Gauthier’s objection.) Singer’s first objection to Hodgson’s argument concerned the special case where information is given as an answer to a question. But in most cases information is given without its having been requested. In his second objection Singer turns to such cases, that is, to cases where information is volunteered. He takes as an example a situation where a stranger A comes up to “me” and says: “There is a very good film on the local cinema this week.” In such a case, Singer claims, I have a good reason to believe that A tells me (at least what he thinks is) the truth. Since by going through the business of inventing what A says to me—thinking to myself, “He says the film is good, but he may be telling a lie, so the film may be bad”—I am no more likely to arrive at the truth than if I take what A says at face value, why should I bother to invert it? Am I not just a fraction more likely to take it at face value? If I am, A, being highly rational, will know this, and will know that he is more likely to produce best consequences if he tells the truth, while I, being highly rational, will know this, and so expect A to tell the truth. […] and so we get the spiral unspiraling again, and we have all the reason we need for telling the truth.26

The crucial step in this argument is the claim that it requires more mental effort to disbelieve what someone says than to believe it (“why should I bother to invert it?”—my italics). The cogency of this claim was discussed in the preceding subsection and will not be repeated here. According to Singer’s third objection, if there were no social practices of truth-­ telling and promise-keeping in the AU-society, there would be act-utilitarian reasons for taking steps to establish such practices. Any steps toward the formation of these practices would have the good consequences of making desirable activities possible. Since telling the truth and keeping promises could help in the formation of these practices, while lying and breaking promises could not, this would give an additional reason for telling the truth and keeping promises. The spiraling effect would come into operation. This would ensure the rapid development of the practices. The informer or promisor would then have the dual reasons of preserving the useful practice and fulfilling expectations.27

As Singer notes, Hodgson is aware of the objection and tries to meet it; there are, Hodgson says, act-utilitarian reasons against taking the steps that would establish such practices. But, as Singer also notes, it is far from clear what these reasons are taken to be. The following is what Hodgson says by way of argument:

 Donald Regan, op. cit., p. 35. (As for Mackie’s criticism of Hodgson, see the next subsec.)  Singer, op. cit., p. 100. 27  Ibid., op. cit., p. 101. 25 26

Truth-Telling in the AU Society: Objections

227

Such steps could have good consequences, but, although perhaps justified by act-­ utilitarianism, they would amount to a partial rejection of act-utilitarianism and so would be inconsistent with our assumptions. These steps would amount to a partial rejection of act-­ utilitarianism, because the persons would be forming habits to do acts known not to be ­justified according to act-utilitarianism; and they would form these habits only if they resolved to refrain from applying act-utilitarianism in relation to these acts.28

Singer suggests that, according to the most plausible interpretation of this passage, what is claimed to be contrary to AU are the initial steps, those taken before there is any established practice and, therefore, any expectations. (I do not think that this is what Hodgson has in mind—I will come back to that later on—but that does not really matter; the important thing is whether, thus interpreted, the passage can be used to meet Singer’s third objection, not whether it is what Hodgson had in mind.) This interpretation, Singer thinks, is supported by the fact that in a later chapter of his book Hodgson raises a similar objection to the act-utilitarian justification of the practice of punishing: although an unbroken record of punishment may deter potential offenders, there is no act-utilitarian reason for starting the record in any given case rather than in the next. Singer objects that this reasoning is built on the faulty assumption that the only consequences of an action are those for which the action is a necessary or a sufficient condition. But, Singer says, an action may have consequences—“may contribute to a result”—for which the action is neither a necessary nor a sufficient condition, and such consequences are, of course, relevant to AU.29 The contribution that my vote makes toward the result I judge to be best in an election is a relevant consideration in deciding whether to vote, although it is, almost certainly, neither a necessary nor a sufficient condition of that result; for if this were not so, the act-utilitarian view would leave us with a result which was unconnected with the actions of any of the voters, since what is true of my vote is equally true of any individual vote. [- - -] In the cases we were considering originally, an act of telling the truth or keeping a promise will normally have greater utility than would its opposite, because it has a reasonable chance of contributing to the beneficial consequences of setting up a desirable practice.30

According to Singer, then, if in an election my vote for a candidate is considered as not making a contribution to the candidate’s being elected, this outcome would be “unconnected with the actions of any of the voters, since what is true of my vote is equally true of every individual vote”.31 Evidently Singer finds the conclusion of this reasoning paradoxical and takes it to show that, therefore, my vote must be considered as making a contribution to the outcome. But I cannot find anything paradoxical in the claim that in cases of overdetermination, like the one we are now  Hodgson, op. cit., p. 48.  Singer, op. cit., p. 103. 30  Ibid. 31  Singer says that my voting for the successful candidate is, “almost certainly, neither a necessary nor a sufficient condition of that result” (my italics). Since the version of AU that Hodgson discusses is primarily concerned with actual consequences (see above p. 1), “almost” should be omitted: to be relevant, the case we are dealing with must be a case of actual over-determination, not one of possible over-determination. 28 29

228

Appendix: Hodgson’s Argument

considering, the outcome is causally unconnected with any individual action— although, of course not, with all of them. That is why, if in the above voting situation I had to choose between voting and performing another action, AU would prescribe that I ought to perform the other action even if my candidate’s being elected was very valuable, whereas the consequences of the other action was only of little (positive) value. In the case of participating in starting a practice of promise-keeping or truth-­ telling there is not, as in the case of casting one’s vote, a certain definite outcome that one either is, or is not, instrumental in achieving; rather it is a matter of participating in the gradual realization of something—let us call it “trust”—among members of one’s society. (So the parallel drawn by Singer is quite misleading and of little help for his argument.) But the contribution that a single individual can give to establishing the practices of promise-keeping or truth-telling in his society is probably negligible: the practice would most certainly be established (and preserved) whether or not, say, individual A contributed. Notice that in our present society some people are untrustworthy, and known to be so, without this tending to abolish the practices of promise-keeping and truth-telling. (It might be objected that A’s not keeping his promises or telling lies is usually directly harmful to those who trust him. But this objection overlooks the fact that, if Hodgson is right, nobody trusts A in the AU-society. If, on the other hand, and contrary to what Hodgson argues, the AU-society would be sufficiently transformed into a non-AU society for other people to trust him, then there are utilitarian reasons for A to match up to the trust. See below section “Truth-telling and promise-keeping in non-AU-societies”.) But, someone might protest, even if there might be a practice of promise-keeping or truth-telling in a society although A does not participate in it, it would be minimally better, in utilitarian terms, if he did. Similarly, even if A’s contribution to starting such a practice might not be necessary, it would still be better if he participated; for then the useful practice would be established (if only minimally) faster. So there is a utilitarian reason for each person to contribute to establishing the beneficial practices of promise-keeping and truth-telling. Hodgson’s answer, I think, would be the following (and this is how I think the above quotation from him should be interpreted): In our AU-society, establishing the practices in question means taking steps to form habits of keeping promises and telling the truth, habits having roughly the same scope and strength as those prevailing among members of actual societies. The taking of such steps by at least the great majority of people in the society would certainly have good consequences—it would establish the practices in question, and these are very useful—and it would “perhaps”, that is, if AU concerned itself with collective actions, be justified by AU. But the habits in question are habits to do individual actions, which are wrong according to AU.32 So if any person in the AU-society acts so as to form the habits,  In Sec. 5.5 above, I pointed out and discussed some problems connected with taking C as a (wholly or partly) collective moral theory.

32

Truth-Telling in the AU Society: Objections

229

he acts contrary to what AU prescribes. As Hodgson says: “they would form these habits only if they resolved to refrain from applying act-utilitarianism to these acts.” (iii). Mackie’s Objection J.  L. Mackie suggests that members of the AU-society could manage without using sentences having truth-values; instead they could use what he calls “belief-­ imperatives”, imperatives of the form “Believe that p”. (They could even simply use “p” instead of the longer form.)33 Mackie makes two claims on behalf of this device. The first is that there is no great psychological difficulty involved in believing what one is thus ordered to believe: “For most things that it is in accordance with utility for people to believe are truths”.34 The second claim is that this device is better than the ordinary practice of truth-telling. This use and acceptance of belief-sentences may not be exactly what we ordinarily call the communicating of information, but it is practically equivalent to this, and in some ways superior to what we have in all actual societies. For though we have conventional rules of truth-telling, we well know that they are often violated, occasionally for benevolent reasons but more often in support of divergent interests. Would it not be better to be sure that your neighbour was always telling you to believe what it would be best for you to believe than to be uncertain whether he was telling you the truth or deceiving you for his private ends and against your interests?35

I disagree with both of these claims. As I argued at length in section “Truthtelling in the AU-society: Hodgson’s argument” above, I think that there are many kinds of situation where AU says that you should lie to other people. Recognizing this, people in the AU-society would often hesitate to believe what they are ordered to believe—even if they believed that believing it and acting upon the belief would have better consequences than not believing it. In the AU-society there would, for example, be many situations where AU prescribed that the interests of one person should be sacrificed for the sake of better consequences on the whole. In some such situations one would have to deceive somebody in order to accomplish this. Of course, no miscalculation being suspected, the victim would think that the sacrifice was justified. But, unless human nature were drastically transformed, people in the AU-society would live in constant fear of suddenly being sacrificed ad maiorem gloriam utilitatis. In not a few cases, then, there would be “great psychological difficulty involved in believing what one is thus ordered to believe”. So Mackie’s first claim is false. As for the second claim, the claim that the device proposed is better than the ordinary practice of truth-telling, it is evident that the proposal does not even meet Hodgson’s objection. Hodgson claims that when A states that p, B does not know  J.  L. Mackie, “The Disutility of Act-Utilitarianism”, Philosophical Quarterly, 23 (1973): 289–300. 34  Ibid., p. 297. 35  Ibid., p. 298. 33

230

Appendix: Hodgson’s Argument

whether A wants to communicate that p is the case or that non-p is the case. It seems that B has exactly the same problem if A, instead, would say, “Believe that p”. B knows that A either asks him to believe that p is the case or that non-p is the case, but one is not more probable than the other. Switching from the indicative to the imperative mood does not solve the problem posed by Hodgson.36 (iv). Lewis’s Objection David Lewis illustrates Hodgson’s thesis by means of the following example. Two highly rational act-utilitarians, “you” and “I”, are put in separate rooms, each having a red and a green button at his disposal. If, and only if, we both push either the red or the green button, we bring about the Good; otherwise we bring about the Bad. We know all this, we know that we know, and so on. You manage to send me a message, “I pushed red”. Do I then have a reason to push red?37 Not if Hodgson is right. According to Hodgson, Lewis says, I must reason as follows. I have not the slightest reason to believe you unless I have reason to believe that you think that I have reason to believe you. But I know that you—knowledgeable and rational creature that you are—will not think that I have reason to believe you unless I really do. Do I? I cannot show that I have reason to believe you without first assuming what is to be shown: that I have reason to believe you. So I cannot, without committing the fallacy of petitio principii, show that I have reason to believe you. Therefore I do not. Your message gives me not the slightest reason to believe that you pushed red, and not the slightest reason to push red myself.38

But this is absurd, Lewis says, so there must be a flaw in the argument. Lewis thinks that the flaw comes with the step taken in the italicized sentence, the step where, he says, I tacitly assumed that my reason to believe you must be found only in facts about the situation and us, “our utilitarianism and rationality, our knowledge of these, our knowledge of one another’s knowledge of these, and so on”.39 But, Lewis asks, why must my reason to believe you be limited to these facts? To show that I have such a reason, I could start with any premise that gives me a reason to believe you, provided it is available to me and consistent with, as well as indepen-

 Donald Regan’s criticism of Mackie’s proposal is more radical. “If verbal communication were not established, people would not make vocal noises with the intent to communicate, and the only vocal noises I would hear would in fact flow from other motives.” (Regan, op. cit., p. 36.) This accords with what I said concerning Gauthier’s objection in Subsec. 3: (i) above. 37  David Lewis, “Utilitarianism and Truthfulness”, Australasian Journal of Philosophy, 50 (1972): 17–19; p. 17. The example seems to presuppose the contrariety interpretation, since each of us has four alternatives: push red, push green, push both buttons, and push neither button. But, since we are utilitarians and know that pushing either button has greater expected utility than pushing either both buttons or neither button, we take only the former alternatives into consideration. So the example actually presupposes the contrary interpretation. 38  Ibid., p. 17. A similar argument, Lewis points out, could be applied to promising: “for an example of this, just change the message in my example to ‘I will push red’” (ibid.). 39  Ibid., p. 18. 36

Truth-Telling in the AU Society: Objections

231

dent of, the facts about the situation and us. And, Lewis says, there is actually such a premise. The premise that you will be truthful (whenever it is best to instill in me true beliefs about matters you have knowledge of, as in this case) is just such a premise. It is available to me. At least, common sense suggests that it would be; and our only reason to suppose that it would not is the Hodgsonian argument we are disputing. […] On the one hand it is consistent with our rationality and utilitarianism, our knowledge thereof, and so on. […] On the other hand, it is not implied by our rationality and utilitarianism, our knowledge thereof, and so on.40

Admittedly, the premise seems available to me—at least if “you think it is best” is substituted for “it is best”. As Lewis says, common sense suggests that it is. But, first, common sense has little experience of thinking as a highly rational act-­ utilitarian: even if the premise is available to common sense, it might be unavailable to the utilitarian. And, secondly, common sense might easily be misled by what is omitted in Lewis’s presentation of the example: Lewis never mentions that you and I are members of the AU-society; it is therefore easily imagined that we are two act-­ utilitarians living in a predominantly non-utilitarian society, who usually deal with, and are accustomed to, people who adhere to the norms concerning truth-telling prescribed by CSM. As far as I can see, however, Lewis’s objection begs the question at issue. Of course, if I accept the premise, then I must, on pain of contradiction, believe what you said. (For I know that (you know that) it would be best to instill true beliefs in me.) But what reason do I have to accept the premise?41 If Hodgson is right, the premise is false, and I have no reason to accept it. Lewis does not prove that Hodgson is wrong, he just takes it for granted.42 (v). Hoerster’s Objection Yet another critic of Hodgson is Norbert Hoerster.43 Hoerster argues that it is possible, as well as justified on act-utilitarian grounds, to introduce the practice of truth-telling in the AU-society allegedly lacking it. At first one gets the impression that Hoerster wants to defend the collective version of act utilitarianism (see section  Ibid., p. 18.  Lewis says that I know that the premise is true. (The reason to believe your message and, therefore, to push red myself is, he says, “premised on further knowledge that I do in fact possess (ibid., p. 19).) True, if I know that the premise is true, then, as a matter of conceptual truth, I must have good reasons for believing it and should be able to state them, but Lewis does not tell us what these reasons are. 42  Adrian Piper (op. cit.) criticizes Lewis’s attempt (as well as a similar attempt by Allan Gibbard in his (unpublished) Ph.D. dissertation) to refute Hodgson. If I have understood Piper correctly, his main criticism of Lewis (and Gibbard) is that they beg the question by assuming what Hodgson implicitly denies, viz. that communication of any sort would be possible in the AU society. I think that Lewis could meet the objection. It is true, he might retort, that my example seemingly presupposes that people in the AU society are able to communicate with each other. But so do the examples given by Hodgson himself. In neither case, however, is such an assumption really made. The examples should be read: “assuming that people in the AU society were able to communicate with each other, then ...” 43  Norbert Hoerster, “Is Act-Utilitarian Truth-Telling Self-Defeating?”, Mind, 82 (1973): 413–16. 40 41

232

Appendix: Hodgson’s Argument

“Truth-telling in the AU society: objections”: (ii) above). For he says that a member A of the AU society is pro tanto obliged to tell the truth if, by telling the truth, A contributes to creating the expectations necessary for establishing the practice of truth-telling in his society. And this condition, Hoerster says, is ordinarily satisfied. For if A and all other individuals in similar situations will actually tell the truth, whenever, but for the consideration of turning truth-telling into a practice, its utility is indifferent, then truth-telling will become more frequent than lying and as a result people will generally expect to get true rather than false statements from their fellow beings.44

But later Hoerster seems to recognize that this answer does not address the objection raised by Hodgson, which is directed against individualistic act utilitarianism. For he imagines that Hodgson might object to the above answer by pointing out that “[e]ach separate extra true statement will only ‘contribute’ to creating a new expectation, if it does not remain alone”.45 So Hoerster turns his attention to this objection, that is, to an objection against AU actually entailed by what Hodgson says. It is not easy, however, to say what exactly the answer that Hoerster gives to Hodgson is. Hoerster admits that a single act of truth-telling cannot be justified on act-utilitarian grounds by its alleged contribution to establishing the practice of truth-telling in the society. For whether this practice will ever be established is, as he admits, just the point at issue. The solution lies, however, in describing the specific act to be tested by the act-utilitarian formula not as telling the truth on some occasion, but rather as creating some individual’s expectation to be told the truth in the future. Thus each individual can be shown to have an obligation usually to tell the truth to those of his fellows whom he frequently contacts. And, in this way, the practice of truth-telling is established, at least in its most important aspect, namely between people not being strangers to each other.46

I take it that what Hoerster wants to claim is the following: Suppose that A and B are two members of the AU-society who regularly meet each other and also have some need to communicate with each other. At the beginning, they do not trust each other to speak the truth. But, by starting to be truthful to B, A gradually causes B to trust him to speak the truth. Each single act of A’s telling B the truth contributes to this happy result. There is therefore an act-utilitarian reason for A to be truthful to B on almost every occasion. And what is true of A is true of all or most members of the AU-society in relation to their nears and dears. I disagree. There are several reasons why B will not come to trust A.47 One is that in many cases B does not have the opportunity to find out, without much ado, whether A speaks the truth or not. In order to do that, B must make some extra effort. And why should he do that if he initially does not trust A and therefore does not care  Ibid., p. 414; my italics.  Ibid., p. 415. 46  Ibid., p. 416. 47  In the rest of this section I will, for the sake of brevity, speak of “trusting someone”, instead of using the longer expression “trusting someone to speak the truth”. 44 45

Truth-Telling in the AU Society: Objections

233

about what A says? Secondly, even if A speaks (what he thinks is) the truth, and B tries to check out whether what A has said is true, B may still believe that A has told what is false. For either A or B, or both, may be mistaken concerning relevant facts: A may falsely think he speaks the truth, and/or B may falsely think that what he is told is false. And, thirdly, even if B is quite sure that A has spoken the truth, this, as he is aware of, does not give him a reason to trust A in the future. For, as Hodgson points out, there is no reason to think that members of the AU-society do not sometimes tell the truth: they might sometimes tell what is false, sometimes what is true. A fourth reason why B will not trust A is that A sometimes, on strict utilitarian grounds, will lie to B. As we saw in section “Truth-telling in the AU-society: Hodgson’s argument” above, there are several kinds of situations where AU prescribes that people tell lies—on condition that those whom they are addressing trust them. So if A thinks that B trusts him, he will lie to B if he thinks that they find themselves in such a situation. If B, whether truly or falsely, does not believe that they find themselves in such a situation, he will believe what he is told. If later he finds out that A lied to him, his trust in A will be diminished. Moreover, it may be questioned whether, as Hoerster thinks, A really has “an obligation usually to tell the truth to those of his fellows whom he frequently contacts”. There are several reasons why A does not have such an obligation. First, if Hodgson is right, A has at most an obligation (a utilitarian reason) to communicate the truth to other people. Now, as been repeatedly mentioned, there are two ways of doing this: telling the truth and not telling the truth. And if Hodgson is right, one way is as good as the other. Suppose, however, that Hodgson is wrong: it is better to communicate the truth by means of telling it than by means of not telling it. But this does not mean that A usually has an obligation to tell B the truth. For, secondly, there are, as I have repeatedly said, many cases where AU prescribes that A should lie to B. Moreover, as I argued above, B will not trust A. So even in cases where it would be better, in utilitarian terms, that B believed what is true than what is false, it might be impossible for A to accomplish this. Given that “ought” implies “can”, A does not then have an obligation (a reason) to communicate (or to tell) B the truth. At most, therefore, what can be said is that A has an obligation sometimes to try to communicate the truth to B. But, unfortunately, we know that he will most probably fail. And even if people in the AU-society would succeed in establishing the practice of truth-telling between people who regularly meet each other and need to communicate, all is not well. The AU-society would still be severely handicapped in comparison with our society. For even if it is true, as Hoerster claims, that it is most important that there is a practice of truth-telling among non-strangers, it is also very important that the practice extends to strangers too. In all but the most primitive kind of society people need to trust even those whom they are not acquainted with. And Hoerster’s proposal, as he himself admits, cannot establish that this is the case in the AU-society.

234

Appendix: Hodgson’s Argument

Promise-Keeping in the AU Society: Hodgson’s Argument What is true of truth-telling is, Hodgson claims, also true of promise-keeping. In the AU society promising would be pointless, so there would exist no such practice.48 For the only reason for keeping a promise in this society is that it would have best consequences to do what one has promised. But if it would not have best consequences to do the action apart from the promise, it would not have best consequences given the promise. The fact that one has promised to do an action would add to the utility of performing the action only if this fact gave rise to additional expectations that the action would be done. But if the promisee knows that the promisor is an act-utilitarian, knows that the promisor knows that the promisee knows that, and so on, then the fact that the promisor has promised to do the action does not increase the promisee’s expectations that the promisor will do it. And, of course, the promisor knows that, knows that the promisee knows that the promisor knows, and so on. In Hodgson’s own words: So, a promised act could have greater (comparative) utility (than it would have had if it had not been promised) only if the promisee has a greater expectation that it would be done (than he would have had if it had not been promised); but there would be a good reason for such greater expectation only if (in the promisor’s belief) the act would have such greater utility. Being highly rational, the promisor would know that the greater expectation was a condition precedent for the greater utility; and so would not believe that the act would have greater utility unless he believed that the promisee had greater expectation. Also being highly rational, the promisee would know this, and so would not have greater expectation unless he believed that the promisor believed that he had greater expectation. And this, of course, the promisor would know.49

The argument that Hodgson gives concerning promise-keeping in the AU-society has then a different structure than that concerning truth-telling. The argument is roughly that in the AU-society the consequences of keeping a promise are not better than that of not keeping it. So there is no point in promising in the AU-society, and there will therefore be no such practice. But is this really the whole story? Does not promising involve saying something that is either true or false, or does it not at least give rise to a (true or false) belief? If so, it seems that the Hodgsonian argument concerning truth-telling in the AU-society applies to the practice of promising as well. I will try to show that this is really the case. Suppose that A and B are members of a non-AU-society, and A tells B, “I will visit you tomorrow”. Let us assume that both take this as a promise. Does it have a truth-value? No, says the received view: a promise is a performative (a performative utterance), and performatives are neither true nor false.50 I think that the received view is mistaken in the case of promising, but I will not argue this here. It suffices  Let us in the following by the practice of promising understand the practice of making, as well as (normally) keeping, promises. 49  Hodgson, op. cit., p. 41. 50  See, e.g., Justus Hartnack, “Performative Utterances”, in Paul Edwards (ed.), The Encyclopedia of Philosophy. 48

Promise-Keeping in the AU Society: Objections

235

for my purpose if it be admitted that, if B trusts A, what A says to B gives rise to a belief in B to the effect that A will visit him tomorrow, and that A knows that it does. B’s belief is either true or false: if A visits B tomorrow, the belief is true, otherwise it is false. Whether or not A visits B, B will for some time have entertained exactly the same belief that he would have entertained if instead another person whom he trusts, had said, “B will visit you tomorrow”, which is clearly not a performative.51 If, then, promising normally gives rise to beliefs, the ways of communicating true beliefs by promising in the AU-society is essentially similar to that of ordinary communication: either (i) the promisor makes a promise that he will keep, and the promisee believes that he will keep it, or (ii) the promisor makes a promise that he will not keep, and the promisor believes that he will not keep it. For reasons similar to those adduced in section “Truth-telling in the AU-society: Hodgson’s argument” above, the promisee would have no way of knowing whether the promisor will or will not keep the promise. So there would be no point in promising in the AU-society, and the practice of promising would not exist. By two different routes, then, we arrive at the same conclusion. As far as I know, there have been four replies to this conclusion, three of which attempt to refute it, and one which concedes it but thinks that it is no objection to AU. I shall discuss these replies in turn, starting with the latter. But before that I shall introduce a helpful division of different kinds of promises, a division borrowed from Russell Hardin.52 Hardin divides promises into three categories: (i) exchange promises, promises that “facilitate exchanges made over time”; (I give you now my x and you promise to give me your y later.) (ii) co-ordination promises, promises that “facilitate our getting together or otherwise accomplishing some joint venture”; (iii) “gratuitous” promises, promises that “are the promissory equivalent of acts of beneficence: there is no quid pro quo, no evident benefit to the promisor”.53

Promise-Keeping in the AU Society: Objections (i). Mackie’s Objection Mackie concedes that promising would have a very restricted role in the AU-society sketched by Hodgson; its only role there, Mackie claims, is to help people make combined choices between (actual or presumed) utility maxima—that is, to make choices between several outcomes of combinations of actions, each  In the case of promising, the performative justifies a belief that “the performer” will perform another action—e.g., paying a visit to the promisee—such that its non-performance shows that something was wrong with the performative. In the case of most other performatives—such as apologizing, naming, and inviting—something similar is not true. This, I think, explains why, e.g., promising has a communicative effect different from, e.g., that of apologizing. 52  Russell Hardin, Morality Within the Limits of Reason (Chicago and London: The University of Chicago Press, 1988). 53  Ibid. p. 60. 51

236

Appendix: Hodgson’s Argument

combination consisting of one action by each of several agents, outcomes which are, or are considered to be, equally best.54 (I will come back to this role in the next subsection.) But, Mackie says, Hodgson is wrong in inferring that a society, which (almost) completely lacked the practice of promising, would therefore necessarily be in dire straits. The other purposes served by promising in our society would be served by other means in the AU-society: it would be served by the utilitarian morality of its inhabitants. The main point of promising in our present society is that it enables people with divergent aims to co-operate to some extent. The point of a promise is that it helps to construct a compromise. Alf, say, would like best to get his work done and not pay any wages for it; Bill would like best to be paid wages and do no work; but Alf would rather pay wages and get the work done than not get the work done and pay no wages, and Bill would rather work and be paid than neither work nor be paid. They can reach a compromise between their divergent interests if Alf promises to pay Bill if he first does the work, and Bill trusts this promise, and does the work, and Alf then keeps his promise and pays Bill. But if there had been no divergence between their aims, if they had each been concerned only for their common welfare, no promising or promise-keeping would have been necessary.55

I have two comments to make on this. First, Mackie’s example does not show that act-utilitarians can do equally well without promising. True, if “they had each been concerned only for their own common welfare”, Bill could perhaps trust Alf to pay him when he has finished his work. But in the AU-society, although there is “no difference between their aims”, Alf and Bill are certainly not “concerned only for their own common welfare”; being utilitarians, they are concerned for the general welfare. So, if Alf, before he pays Bill, would realize that he could spend his money in some way that had better consequences than handing it over to Bill, he would do that. Being an act-utilitarian. Bill would not, of course, have any reason to complain. But knowing that Alf is an act-utilitarian, he would not trust Alf to pay him for his work. And, if it were important for him to get money for work done, he would probably never start working for Alf unless being paid in advance—but it may be doubted that Alf would trust Bill sufficiently to pay him in advance. (Bill’s decision not to work for Alf is consistent with AU. It might, of course, be the case that it would have better consequences if Bill worked for Alf and Alf gave the money to someone else than if he gave them to Bill. But it might also be the case that it would have even better consequences if Bill worked for someone else, whether or not Alf hired someone else to do the job or did not get the job done.) My second point concerns Mackie’s claim that “[t]he main point of promising in our present society is that it enables people with divergent aims to co-operate to some extent”. Of course, some promises are of this semi-contractual nature, for example, the one in Mackie’s example. But there are other kinds of promises, for example, gratuitous promises (see the end of section “Promise-keeping in the AU society: Hodgson’s argument” above. Whereas co-ordination promises are often, at least partially, made (and kept) out of (mutual) self-interest (the promisor wants to 54 55

 J. L. Mackie, op. cit.  Ibid., p. 296.

Promise-Keeping in the AU Society: Objections

237

secure some benefit for himself), the latter are made (and kept) out of “other-­ interest” (the promisor wants to secure some benefit for someone else). Examples of such promises are promises to assist one’s friend with money in case of emergency, to visit a sick relative before long, to discuss this appendix with me when it is finished, and so on. Due to the discrepancy between the promisee’s welfare and the general welfare, such promises could seldom be trusted in the AU-society. They would therefore be pointless and, hence, not made—to the detriment of the members of the society. Now if the promising practice is useful but, as Mackie concedes, has a very restricted role in the AU society, it may be wondered whether there could not be some replacement for it. (Act-utilitarians, it should be noted, would not object to transforming a kind of situations in such a way that the outcomes of the actions required by AU in the transformed situations were better than the outcomes required in the untransformed situations, the cost of transformation included.) Such a replacement would guarantee, for example, that Alf pays Bill for his work even if in the untransformed situation it is better that Bill works and Alf does not pay him than that Bill works and Alf pays him. Now whereas in some cases co-ordination promising may be replaced by other (more clumsy) devices—Alf, say, paying Bill continuously during the time the latter works—the chances of finding out such replacements for gratuitous promising seem bleak. This is due to, what may be called, the “non-­ reciprocity” of such promises: in the case of these promises, as distinguished from co-ordination promises, it is not the case that the promisor should keep his promise on condition that the promisee does something else. The promisee has therefore no hold on the promisor like the one Bill has on Alf, no device, that is, which could be used as a means for creating a replacement for such promising. In the case of gratuitous promising there is no essential feature of the promising situation that can be used to restructure the value-ordering in such a way that it becomes better, in utilitarian terms, to keep the promise than not to keep it. (ii). Narveson’s Objection As we saw, Mackie thinks that the only role that promising would play in the AU society is to help people to make combined choices between several (actual or presumed) utility maxima. Jan Narveson thinks that, in addition, promising would help people to make individual choices between such maxima.56 And these two roles, he claims, are far from being restrictive; together they range over the whole field of promising as ordinarily practiced. Moreover, he says, there would be just as much promising in the imagined AU society as in actual societies.57 Let us consider the two alleged roles for promising, starting with the second.

 Jan Narveson, “Promising, Expecting, and Utility”, Canadian Journal of Philosophy, 1 (1971): 207–23. 57  One gets the impression that Narveson thinks that promises of these two kinds are rather frequent. As far as my experience goes, they are rather uncommon, especially promises of the latter kind. 56

238

Appendix: Hodgson’s Argument

Sometimes, Narveson says, an agent finds himself in a situation such that he believes that two or more alternatives have equally best outcomes. Suppose that B believes that a utilitarian A is in a situation where actions a1–an appear to have equally best outcomes. Then B cannot possibly know which of these actions A will do; barring further information, he knows only that, for each action, there is a possibility of 1/n that A will do it. Suppose further that it is important for B to know what A will do in order to be able to know what is best for himself to do. There is, then, a utilitarian reason for A to pick one of these actions, promise to do it and do it. We get the same result in the case of combined choices, Narveson says. Suppose that two people want to have lunch together. There are n possible restaurants to meet at. Each person wants very much to have lunch with the other, but neither cares which restaurant they go to. Suppose further that the best outcome (in utilitarian terms) is that they meet at some—no matter which—of the n restaurants. There is, then, a utilitarian reason for them to pick one of the restaurants, promise each other to go there and keep the promise. Picking one of the restaurants and agreeing to go there is, of course, using an arbitrational device, but, Narveson asks rhetorically, are we to infer from the fact that some arbitrational device is necessary here that such a device requires an extra-utilitarian justification, or that it has the status of a moral rule, independent of its utility? I should think not—any more than adherence to the rules of cricket or baseball needs a justification independent of their utility.58

There is, Narveson says, no other role for promising, whether in the imagined AU-society or in actual societies. If there is no uncertainty concerning what an agent will do apart from the promise—if, for example, we (correctly) believe that a utilitarian agent has one best action with a best outcome at his disposal (and he knows that we believe this, and we know that he knows, and so on)—there is no use for promising, and it will not take place. If everybody in a group knows exactly what everyone else, and himself, wants, at all times, then the supposition that an institution of promise-keeping, with its attendant language of obligation, would be of any use at all would be quite baseless.59

I think that Narveson is mistaken in several respects. First, he arbitrarily restricts the range of promising to situations where there are several utility maxima. Thus he overlooks both the kind of semi-contractual promises that Mackie thought are the most important and gratuitous promises.60 For an example of the former kind of promises, see Mackie’s example cited above. For an example of the latter, suppose that you have to pay the rent for your flat within a week or be evicted, but you do not have the money. Having learned of your financial trouble I promise to give you the needed money. This example, as well as the one given by Mackie, is two of countless promises that do not satisfy the several-utility-maxima condition.  Ibid., p. 225.  Ibid., p. 227. 60  The promises considered by Narveson may be seen as solutions to co-ordination games situations, whereas those considered by Mackie may be seen as solutions to co-operative games situations, especially PD-situations. (Gratuitous promises fall outside the scope of game theory.) 58 59

Promise-Keeping in the AU Society: Objections

239

Would any of the kinds of promises not recognized by Narveson take place in the AU-society? Semi-contractual promises would not, there being no difference between people’s aims in the AU-society. Moreover, in the case of many such promises, the outcome agreed upon is not necessarily optimal (in utilitarian terms), taking into account only the welfare of the agents in question instead of the general welfare. (The forming of economic cartels does not usually benefit the public at large.) As for the existence of gratuitous promising in the AU-society, I can only repeat what I said in the preceding subsection: due to the discrepancy between the promisee’s welfare and the general welfare, such promises could seldom be trusted in the AU-society. They would be pointless and, hence, not made. There are, then, many cases of promising in actual societies that would not take place in the AU-society. There are also many cases of promises that are normally kept in actual societies, but would not be kept in the AU-society. This is something that Narveson cannot admit, since in his opinion what has given rise to, and justifies, the norms concerning promise-keeping is our (unconscious?) utilitarian convictions. The only merit of Hodgson’s argument, he says, is that, by making us reflect on the matter, “it does help us to see more clearly the role of promising in human affairs, and likewise to see that its obligational force is entirely a function of its utility”.61 This is, I think, a second issue about which Narveson is mistaken. To see this, consider a case where something unforeseen happens between the giving of the promise and the time when it is to be kept. If the unforeseen thing had not happened, then, all things considered, the outcome of keeping the promise would have been better than that of not keeping it. Since it happened, however, the two outcomes are equally good, and both keeping and not keeping the promise are therefore equally right according to AU. I do not think, however, that this is what most people think, especially not if what raised the value of the outcome of not keeping the promise were benefits to the promisor. According to CSM, a promise should not, of course, be kept come what may; but it should be kept unless the consequences of keeping it are considerably worse than those of not keeping it, especially if keeping it is important to the promisee.62 My third objection to Narveson is that he begs the question at issue. Consider the two kinds of promises which are most favourable to Narveson’s position, and the only kinds of promises whose existence he recognizes, viz. those made in situations where several of the agent’s (agents’) possible actions appear to have equally best outcomes. In such situations, whether they occur in actual societies or in the AU-society, there is, Narveson states, a utilitarian reason to pick one of the actions, promise to do it, and do it. But according to Hodgson there is no such reason in the AU-society. Hodgson’s argument for his claim, let us recall, is (roughly) the following: As both A and B know, there would be a reason for A (the promisor) to do a (the  Ibid., p. 227.  This is also the standard view of deontologists. Primarily, of course, they claim that it is the true view, but in their attempts to justify this verdict, they explicitly or implicitly claim that the true view is also the view of common sense. See, e.g., David Ross, The Right and the Good, passim, and Foundations of Ethics, pp. 87–113.

61 62

240

Appendix: Hodgson’s Argument

promised action) only if he believed that B (the promisee) would expect that. But, as A knows, B would expect that only if he believed that there was a reason for A to do a. Thus there is no utilitarian reason to do what one has promised to do just because one has promised to do it. Hence, members of the AU-society would not make any promises. This argument, if valid with respect to any promises, is certainly valid with respect to (putative) promises to do one of several actions with equally best outcomes. In order to show that Hodgson is wrong, therefore, it is not enough just to asseverate that he is; one should show what is wrong with his argument. (iii). Gauthier’s Objection David Gauthier does what Narveson fails to do: he tries to show what is wrong with Hodgson’s argument.63 According to Gauthier, “the primary function of the practice of promising is to serve as a device for coordination”.64 (Thus he takes the practice of promising to be equally restricted as Narveson does.) Suppose, Gauthier says, that you and I want to meet tomorrow. I promise you that I will be at my office at 2 p.m. Prior to the promising there were many best alternatives as to when and where to meet, but, acting independently of each other, we were unable to secure any of them. My promise helps us to single out one of the outcomes—to our mutual benefit. If I promise you that I will perform some action, then I make the outcome of my doing that action salient. In promising I change, not the utilities in the situation in which we find ourselves, but our conception of that situation. My reason for performing the act promised is that in the situation conceived in terms of the promise, the promised act leads to the unique best equilibrium outcome. And this reason is provided by the making of the promise, for in the situation apart from the promise, no act leads to a unique best equilibrium outcome.65

I think Gauthier is right to some extent: if A and B, two inhabitants of the AU-society, believe that they confront a co-ordination situation with more than two alternatives (and also believe of each other that they believe that, and so on), A’s promising to do his part for bringing about a certain one of the alternatives provides him with some reason to do it and B with some reason to expect it. (This is analogous to what was said with respect to truth-telling at the end of section “Truthtelling in the AU society: objections”: (i) above.) This weakens Hodgson’s position, but not very much. For, first, A knows that B knows (and so on) that A is a utilitarian and therefore holds that the only reason to keep a promise is that it maximizes utility. So even if, at the time of promising, A thinks that keeping the promise has better consequences than not keeping it, much that might happen until it is time to keep it could tip the balance. So B’s reason to trust A cannot be very strong. Knowing this, A’s reason for keeping the promise is not very strong either.  David Gauthier, op. cit. Gauthier wants to show that “at least a rudimentary form of promising is not only possible but also rational and desirable for all act-consequentialists” (p. 294). 64  Ibid., p. 295. 65  Ibid. My promise, Gauthier says, does not change the utilities. This is, no doubt, true. But it changes the expected utilities, and that is what is important. 63

Promise-Keeping in the AU Society: Objections

241

Moreover, if, as I argued in section “Promise-keeping in the AU society: Hodgson’s argument” above, promising is a way of (knowingly) raising true or false beliefs in the promisee, promising inherits all the utilitarian reasons for deceiving people by lying to them discussed in section “Truth-telling in the AU-society: Hodgson’s argument” above. (In the case of promising, deceit will, of course, be effected, not by (strictly speaking) lying, but by not keeping the promise.) Finally, and quite decisively, as I said in the previous subsection, promising to bring about a certain one of several best outcomes far from exhausts the field of promising; it is in fact only a small part of it. (iv). Gibbard’s Objection Allan Gibbard bases his objection to Hodgson on David Lewis’s theory of convention.66 A crucial element in this theory is the claim that people will rationally keep an agreement if, and only if, it is common knowledge in their society, or just among the parties to the agreement, that people have generally kept their agreements in the past; in other words, people will rationally keep their agreements if, and only if, past history has established a convention. With one exception (to be introduced later in this section) this is so, Gibbard argues, even in the AU-society. I do not have space to critically discuss at length Gibbard’s complex and very subtle defence of his thesis. I shall content myself with pointing out some weak points in Gibbard’s argument, points which, I think, cast grave doubts on the tenability of his position. But before that I want to point out that, even if tenable, Gibbard’s objection does not cut very deep. So even if his objection were well-­ founded, it would not destroy, only to some extent weaken, Hodgson’s thesis concerning the fate of promises in the AU-society. Firstly, Gibbard too discusses only one kind of promises, viz. co-ordination promises (see section “Promise-keeping in the AU society: Hodgson’s argument” above). And, as I said above, even if such promises were given and held in the AU-society, it does not follow that other kinds of promises too are given and held there. Secondly, as Gibbard himself admits, a condition for there being in the AU-society an agreement justified by AU is that “the parties share their relevant experience, […] [that] they agree on all probabilities relevant to the effects of making the agreement binding and carrying it out.”67 If this condition is violated, then it is as if they worked from different utility scales altogether. Even if they agree on their ultimate ends, they may disagree on what more immediate ends would foster those ultimate ends. […] In those circumstances, different people may find different agreements optimal.68

Since the parties do not always have the opportunity to make sure that they agree on all relevant probabilities, the condition is probably far from always fulfilled.  Allan Gibbard, op. cit.; David Lewis, Convention: A Philosophical Study (Cambridge, Mass.: Harvard University Press, 1969). 67  Gibbard, op. cit., p. 107. 68  Ibid., p. 109. 66

242

Appendix: Hodgson’s Argument

So, as I said above, Gibbard’s criticism at most weakens Hodgson’s thesis. But does it even accomplish that? I do not think so. For there are, in my opinion, (at least) two unconvincing points in Gibbard’s argument. One of these may be introduced as follows: As I said above, Lewis claims that it is rational to keep one’s agreements if, and only if, past history has established a convention. But then, Gibbard points out, it may be objected that [i]f it is rational to follow a proto-convention only when it has a supporting history, then it must have been irrational for anyone to follow it in the first place, and so in a society of rational agents, a supporting history could never arise.69

But the objection is dismissed by Gibbard: I agree that if it is rational to follow the proto-convention only when it has a supporting history, then a supporting history could never arise in the first place. It may, however, have been rational to follow the proto-convention before it had a supporting history precisely because a history of its being followed could later make it rational to follow it and a history of its not being followed—an ‘undermining history’, I shall say—would later make it irrational to follow it.70

So it is rational to start a proto-convention because this is to start giving it a supporting history, which, in turn, makes it later rational to follow the proto-convention. And, Gibbard adds, to start following the proto-convention is rational because this has good expected consequences, whereas not to start following it has bad expected consequences. In the first situation to which the proto-convention applies, where it has neither a supporting nor an undermining history, parties’ knowledge [of the situation] gives them reason for following the proto-convention, because establishing the beginnings of a supporting history has good expected consequences and establishing the beginnings of an undermining history has bad expected consequences.71

Now suppose that John and Harriet for the first time make an agreement to meet in the park at noon. (This is the example that Gibbard uses throughout his article.) To do what they have agreed to do has good expected consequences, Gibbard claims, since it contributes to establishing a supporting history. But, it might be objected, the contribution of a single case to this goal is, if at all noticeable, quite negligible. A supporting history will, or will not, become established whether or not, say, John keeps this agreement. (The same is true of every single agreement, whether or not a supporting history has been established at the time when the agreement is made. Consider how conventions survive occasional deviations from them in our actual non-AU-society.) This means that many circumstances that may attract John’s attention after the agreement has been made will, for utilitarian reasons, tell against going to the park. Thus, for example, John might discover that one of his favourite  Ibid., p. 101. A proto-convention is a convention without its supporting history. A supporting history for a convention is the common knowledge that the convention has been followed in the past. 70  Ibid., p.101. 71  Ibid., p. 102. 69

Promise-Keeping in the AU Society: Objections

243

programs will be on TV at noon, or that it unexpectedly starts to rain just when he is about to go to the park. Harriet knows that such things might happen, and John knows that Harriet knows this, and so on. So if a supporting history will ever be established—which I doubt—the proto-convention it supports will be a very weak one and hardly to be much trusted. It will not be as useful as a corresponding convention in our non-AU-society. But, it might be objected, I have left out of account a circumstance that definitely speaks in favour of John’s going to the park, viz. that Harriet will go to the park as agreed and will be disappointed if John does not come. This circumstance, it might be hold, gives John a strong utilitarian reason (and a corresponding motivation) for keeping the agreement. But does it? I do not think so. Harriet is in exactly the same situation as John. So she too might become aware, after the agreement has been made, of circumstances that tell against her going to the park. And John knows this. Harriet knows that John knows, and so on. This further reduces both Harriet’s and John’s utilitarian reasons for going to the park: it is far from certain that the other person will come to the park. (This too is common knowledge between John and Harriet, and it further reduces their reasons for going to the park.) If the probability that the other will stay home is estimated to be higher than 0.5, the utilitarian rational thing for each of them to do is to stay home. And, as far as I can see, this might very well be the case. I have hitherto presupposed that the best outcome is that of both John and Harriet going to the park, the next best that of both staying home, and the worst outcomes are those of only one going to the park. This is also how Gibbard describes the example: The best outcome they can achieve is to meet as agreed, and the next best outcome is for both to stay home and read. Because each would find it distressing to come and not find the other, the worst outcome they can achieve is for one to come to the park and the other to stay at home.72

Does this mean that John and Harriet going to the park would be better than both staying home even without an agreement to go to the park, or that it is better only given an agreement? I think that the first alternative, although it is in some respects the more plausible interpretation, is not the intended one: since it is common knowledge among John and Harriet that they are rational act-utilitarians who know about their situation, they do not have to make an agreement to do what both know is the best thing to do. The second alternative is also problematic. Note that it presupposes that, without an agreement, both going to the park and both staying home are equally good. For, as we just saw, the first outcome cannot be better than the second. Nor can the second outcome be better than the first: if staying home is better than going to the park, it would be irrational to agree to go to the park. But if, without an agreement, going to the park and staying home are equally good, then it seems that John and Harriet should stay home. For meeting each other in the park at noon requires making an 72

 Ibid., p. 91.

244

Appendix: Hodgson’s Argument

agreement, but staying home is the default position and requires nothing of that sort. Since making the agreement and carrying it out involves unnecessary costs with respect to the time and effort spent, the rational thing to do is to abstain from making any agreement—which means staying home. I have now dealt with what, in my opinion, is one of two unconvincing points in Gibbard’s argument. The other point concerns Gibbard’s claim that the conventions which, according to him, exists in the AU-society are not “conventional moral rules” in Hodgson’s sense. If they were, Gibbard admits, his argument would not refute Hodgson: Hodgson stipulates that in the society we are to consider, there are no “conventional moral rules.” If this rules out AU-conventions, then in showing that members of an openly actutilitarian society would keep promises if they had an appropriate AU-convention, I shall not be refuting Hodgson.73

But, Gibbard argues, Hodgson’s stipulation does not rule out AU-conventions: Something is a conventional moral rule in Hodgson’s sense only if deviations from it “are generally regarded as lapses or faults open to criticism.” AU-conventions do not need to be sustained by criticism. They are sustained by common knowledge that each person chooses the most favourable prospect when he acts, and that he reasons inductively. Hence there can be an openly act-utilitarian society which satisfies Hodgson’s stipulation that there be no conventional moral rules, and which still has AU-conventions.74

AU-conventions, Gibbard then claims, differ from conventional moral rules in that the former do not have to be sustained by criticism of deviations from them, while the latter have to be thus sustained. This, I think, is wrong. Conventional moral rules are not primarily sustained by criticism of deviations from them—nor does Hodgson say so. What sustains them are successful moral indoctrination and/ or moral insight. And the same is certainly true of AU-conventions. Moreover, deviations from conventional moral rules are, no doubt, criticized, but so are in all probability deviations from AU-conventions too. Thus, unlike Gibbard, I cannot see that there are any relevant differences between the AU-conventions and conventional moral rules.75

Truth-Telling and Promise-Keeping in Non-AU-Societies As mentioned at the beginning of the appendix, Hodgson claims that also the correct application of AU by members of a predominantly non-AU society like our own would (probably) have worse consequences than would acceptance of CSM. Hodgson argues as follows: Let A be a highly rational act-utilitarian living in a  Ibid., p. 98. An AU-convention is, roughly, a proto-convention such that it is common knowledge among those who are parties to it in the AU-society that it has been followed by them in the past. 74  Ibid., p. 98. 75  Hodgson’s treatment of exchange promises (see Sec. 4 above) is criticized by Russell Hardin, op. cit., p. 61 f. But Hardin’s discussion contributes nothing of interest. 73

Truth-Telling and Promise-Keeping in Non-AU-Societies

245

non-AU-society. If those with whom he interacts know that A always tries, possibly successfully, to act in accordance with AU, many of the problems with respect to promise-keeping and truth-telling that affect members of the AU-society appear; even if other people tolerate A’s moral conviction, they do not trust him, and, as we have seen, lack of trust causes great disutility.76 If, on the other hand, A does not avow his acceptance of AU, other problems appear on account of A’s then having to deceive other people. He will feign to accept CSM and will conform to it in all cases where there is a risk that non-conformance will be detected. But then A is insincere: he only seems to recognize the notions of personal obligation, apologizing, and blaming, notions which are essential to CSM.  All this deceit will destroy his candour and openness of character. And Hodgson concludes: There are alternatives which are open to those who accept personal rules approximating to the conventional rules of their society, but which are not open to those whose only personal rule is that of the act-utilitarian principle. Even though the former persons might not always choose, from the acts open to them, those with the best consequences, nevertheless the consequences of the acts which they do choose might be better than the best consequences which the latter persons could bring about through the alternatives open to them. Not only is this possible, but the cases we have considered suggest that it is probable.77

I disagree. Certainly, if A makes it known that he is an act-utilitarian, he is in trouble. But what if he does not? Of course, his candour will be destroyed. But this should not worry an adherent of AU. I fully agree with what Mackie says on this issue:  In “Consequences of Utilitarianism”, Dialogue: Canadian Philosophical Review, 7 (1969): 639– 42, L. W. Sumner criticizes Hodgson’s account of the fate of an act-utilitarian in a non-utilitarian society with respect to promise-keeping. It is possible for the promisee, Sumner contends, to shape the situation in such a way that a promise given by the act-utilitarian agent can be trusted. (Sumner assumes that it is common knowledge that the agent in question is an act-utilitarian.) All that the promisee has to do, Sumner says, is to put himself to some trouble having some disutility unless the promise is kept. “By doing so he may manufacture the conditions necessary for the act utilitarian to keep the promise.” (p. 641) (Sumner evidently assumes that what the promisee does will be known to the promisor.) This is, no doubt, a workable strategy. But, obviously, it does not make the act-utilitarian agent trustworthy, in the sense of “trustworthy” here under discussion, viz. being such as to keep his promises even when it does not maximize expected utility. Moreover, the suggested proposal does not square with Sumner’s view that the disutility of the trouble that the promisee puts himself to because of his expectation that the promise will be kept “will ordinarily precede rather than follow the breaking (or keeping) of the promise and thus cannot be a consequence of it” (ibid.). (But the fact that Sumner’s proposal does not square with this view does not really matter, since Sumner’s view is not true: the trouble may precede the breaking of the promise, but the disutility is certainly a consequence of it.) 77  Hodgson, op. cit., p. 58 f. Howard Sobel, who seems to agree with Hodgson, has pointed out an interesting analogy between Hodgson’s argument and David Gauthier’s argument for the rationality of constrained (egoistic) maximization. (Jordan Howard Sobel, “Kent Bach on Good Arguments”, Canadian Journal of Philosophy, 19 (1989): 447–54.) As Sobel notes (p.  451), Gauthier states that “[t]he essential point in our argument is that one’s disposition to choose affects the situations in which one may expect to find oneself”. (David Gauthier, Morals by Agreement (Oxford: Clarendon Press, 1986), p. 183. 76

246

Appendix: Hodgson’s Argument

A thorough-going act-utilitarian would be impervious to the social pressures that condemn even benevolent deceit; knowing that he is deceiving his fellows only for the sake of common good, he will feel a glow of conscious virtue each time he takes them in.78

So AU prescribes that A should not avow his adherence to AU but should conform to CSM whenever non-conformance might be detected.79 And if he does, he will produce at least as much utility as he would have done had he accepted CSM instead. Therefore AU is not self-defeating for its adherents in actual, non-AU-societies. Suppose, however, for the sake of argument, that Hodgson is right: it would have been better, on act-utilitarian grounds, if A had accepted CSM instead. This means that AU is, in Derek Parfit’s terms, indirectly individually self-defeating in A’s case. A moral theory T is thus self-defeating, Parfit says, when it is true that, if someone tries to achieve his T-given aims, these aims will be, on the whole, worse achieved.80

It is evident that if, according to AU, A ought to have accepted CSM instead of AU, then AU is often self-defeating in this sense for its adherents in non-AU-societies. However, as Parfit argues, being indirectly individually self-defeating does not tell against a moral theory. An agent’s less than optimal achievement of his T-given aims is not the result of his doing what T tells him to do. It is the result of his being disposed to act in a certain way, often due to his belief in T. And if this is the case, these dispositions and this belief are not sanctioned by T. If T tells him anything in this respect, it tells him to change his dispositions and belief, and thereby his aims, if he can, and adopt other ones. Thus AU would tell A to adopt the motivations and beliefs which, of those open to him, would have the best consequences.81 It may be the case, however, that the agent cannot change his dispositions and moral beliefs. Perhaps, for example, A cannot stop believing in AU. This is, however, just a sad fact about reality, comparable to A’s being, say, poor. For if A had been rich instead of poor, he would, let us assume, have produced more value than he does now. In neither case is AU to be blamed. So whether or not an act-utilitarian living in a predominantly CSM-society abides by his moral theory or abandons it in favour of CSM, it creates no problem for AU. But, someone may ask, why is AU directly (collectively) self-defeating, and not only indirectly (individually) self-defeating—and thus innocuously self-defeating—for people in the AU-society as well. In that case too, it might be claimed, the less than optimal outcomes of people’s actions are the result of their having beliefs and dispositions not sanctioned by AU. The answer is that in this case the less than optimal outcome of an action is not, as in the case of act-utilitarians living in nonAU-societies, the result of the agent’s beliefs and dispositions; it is the result of  Mackie, op. cit., p. 299.  For a classic defence of this position, see Henry Sidgwick, The Methods of Ethics, pp. 485–92. 80  Parfit, op. cit., p. 5. Directly individually self-defeatingness is introduced at the end of Sec. 3.4. 81  Cf. Parfit, op. cit., Sec. 18. 78 79

Concluding Remarks

247

other people’s beliefs and dispositions, notably their beliefs concerning the agent’s beliefs and dispositions. To verify this, suppose that A and B are two inhabitants of the AU-society. One day A rejects AU and becomes an adherent of CSM. If, however, B believes that A is still an adherent of AU (and A believes that B believes this, and so on), what AU prescribes that A should do with respect to telling the truth and keeping his promises to B is obviously no different from what it would have been if A had still been an adherent of AU. And there is no reason to think that the converted A produces more utility than the unconverted A would have done. So A’s conversion was not prescribed by AU.

Concluding Remarks In this appendix I have discussed D. H. Hodgson’s claims that AU is self-defeating, both for people in the AU-society and for act-utilitarians in non-AU-societies. I have defended the first claim but argued against the second. When defending the first claim, I have also briefly analysed and tried to assess the arguments adduced by Hodgson, viz. that truth-telling and promise-keeping would be absent in the AU-society. In addition, I have pointed out the great amount of deceit, prescribed by AU, which would take place in the AU-society, something that strengthens Hodgson’s position. For the most part, I have been busy defending Hodgson’s first claim. The defence has almost entirely consisted in criticizing the objections that have been raised against it. I am unpersuaded by these objections, and have tried to show, for each of them, what is wrong with it. Of course, even if, in each case, my criticism is wellfounded, this does not show that Hodgson is right; there might be other objections, waiting to be raised, that would show that Hodgson’s claim is false. But, although my defence of the claim does not amount to a proof of it, I hope that my defence has strengthened its plausibility. But, it may be asked, assuming that Hodgson’s claim is true, what does this show? Not very much, it seems. Why could not an adherent of AU with equanimity accept that his morality is self-defeating for people in the AU-society, a society that will never be realized? The important thing, he could say, is that Hodgson’s second claim is not true: it is not true that AU is self-defeating for people in actually existing societies. I disagree. If Hodgson’s first claim is true, this shows that AU is not, as its adherents claim, the fundamental moral theory. The fundamental moral theory is not, I take it, self-defeating in any (logically) possible world. But this claim is controversial and has to be argued for at some length. I shall therefore content myself with a less controversial claim that serves my purpose equally well: the fundamental moral theory is not self-defeating in that possible world where everyone accepts the theory and nearly always does what the theory tells them to do. In this case it cannot plausibly be held that the theory’s being self-defeating shows that something is wrong with the world rather than with the theory—unless knowledge that other people

248

Appendix: Hodgson’s Argument

hold the same moral convictions as oneself is taken as showing that something is wrong with the world. The truth of Hodgson’s first claim, together with the falsity of the second claim, also indicates that AU is parasitic on CSM, in the sense that the former is “successful” with respect to truth-telling and promise-keeping only if most people conform to the latter.82 For, as we have seen, an act-utilitarian A living in a non-AU-society ought, according to AU, to conform to the norms of CSM with respect to truth-telling and promise-keeping. And the consequences of A’s conforming to CSM in these respects are, on the whole, beneficial. If, on the other hand, A had lived in the AU-society, where no one adheres or conforms to CSM, he would have been unable to bring about such beneficial consequences. It might be objected that this inability does not depend on people’s general adherence to AU, but on their common knowledge of this adherence; if everyone in the AU-society thought that everyone else adhered to CSM, then everyone, though still adhering to AU, would nearly always conform to the norms of CSM, thus bringing about the beneficial consequences of truth-telling and promise-keeping. The short answer to this objection is that even in this case is AU parasitic on CSM: the beneficial consequences of anyone’s conforming to AU with respect to truth-telling and promise-keeping depend on everyone’s conforming to CSM. The discussion in this Appendix has been concerned with two moral practices, truth-telling and promise-keeping. In both cases, I have argued, AU is self-defeating unless most people conform to CSM.  The question is whether this result can be generalised. In other words, is act-utilitarianism generally parasitic on commonsense morality? I discussed this question in Chap. 7.

 It is not, of course, required that they conform to CSM in toto, only that they conform to its norms concerning truth-telling and promise-keeping.

82

Bibliography

Acton, H.B., ed. 1969. The Philosophy of Punishment. London: Macmillan. Albee, Ernest. 1902. A History of English Utilitarianism. London: Swan Sonnenschein. Annas, Julia. 2004. Being Virtuous and Doing the Right Thing. Proceedings and Addresses of the American Philosophical Association 78: 61–75. Anscombe, G.E.M. 1958. Modern Moral Philosophy. Philosophy 33: 1–19. ———. 1990. Modern Moral Philosophy. Ethics 101: 42–63. Aristotle. Nicomachean Ethics (Find references). Ashford, Elisabeth, and Tim Mulgan. 2008, Fall. Contractualism. In Stanford Encyclopedia of Philosophy, ed. Edward N.  Zalta. http://plato.stanford.edu/archives/fall2008/entries/ contractualism/. Augustine. 1948. The Lord’s Sermon on the Mount. In Ancient Christian Writers: The Words of the Fathers in Translation, ed. Johannes Quasten et al. Westminster: Newman Press. Baier, Kurt. 1958. The Moral Point of View. Ithaca/New York: Cornell University Press. Baron, Marcia. 1997. Kantian Ethics. In Three Methods of Ethics: A Debate, ed. Marcia Baron, Philip Pettit, and Michael Slote, 3–91. Oxford: Blackwell. Baron, Marcia, Philip Pettit, and Michael Slote. 1997. Three Methods of Ethics: A Debate. Oxford: Blackwell. Beauchamp, Tom. 1998. Editor’s Introduction. In An Enquiry concerning the Principles of Morals, ed. David Hume’s. Oxford: Oxford University Press. Beck, Lewis White, (ed. and tr.). 1976. Immanuel Kant: Critique of Practical Reason and Other Writings in Moral Philosophy. Reprint. New York: Garland. Bennett, Jonathan. 1995. The Act Itself. Oxford: Oxford University Press. Bentham, Jeremy. 1823. Not Paul but Jesus. Issued under the pseudonym “Gamaliel Smith, Esq”. London: John Hunt. Bentham, Jeremy. An Introduction to The Principles of Morals and Legislation (Find references). Bergson, Henri. 1977. The Two Sources of Morality and Religion. Trans R.A.  Audra, and C. Brereton. Notre Dame: University of Notre Dame Press. Bergström, Lars. 1996. Reflections on Consequentialism. Theoria 62: 74–94. Blackburn, Simon. 1996 (1994). Foundationalism. In The Oxford Dictionary of Philosophy. Oxford: Oxford University Press. Brandt, Richard. 1958. Blameworthiness and Obligation. In Essays in Moral Philosophy, ed. A.I. Melden. Seattle: University of Washington Press. ———. 1959. Ethical Theory: The Problems of Normative and Critical Ethics. Englewood Cliffs: Prentice-Hall. Bricke, John, ed. 1976. Freedom & Morality. Lawrence: University of Kansas.

© Springer Nature Switzerland AG 2019 J. Österberg, Towards Reunion in Ethics, Philosophical Studies Series 138, https://doi.org/10.1007/978-3-030-12410-6

249

250

Bibliography

Brink, David. 1989. Moral Realism and the Foundations of Ethics. Cambridge: Cambridge University Press. Brinton, Crane. 1959. A History of Western Morals. New York: Harcourt/Brace and Company. Broad, C.D. 1930. Five Types of Ethical Theory. London: Routledge & Kegan Paul. ———. 1942. Certain Features in G. E. Moore’s Ethical Doctrines. In The Philosophy of G. E. Moore, ed. Paul Schilpp, 43–67. Chicago/Evanston: Northwestern University Press. Brock, Dan. 1973. Recent Work in Utilitarianism. American Philosophical Quarterly 10: 241–276. Brown, Thomas. 1820. Lectures on the Philosophy of the Human Mind, 4 vols. Edinburgh: Tait. Brown, Stuart M., Jr., ed. 1950. Five Sermons Preached at the Rolls Chapel and a Dissertation Upon the Nature of Virtue. Indianapolis: Bobbs-Merrill. Buchanan, James. 1975. The Limits of Liberty: Between Anarchy and Leviathan. Chicago: Chicago University Press. Butler, Joseph. 1726. Fifteen Sermons Preached at the Rolls Chapel. London: James and John Knapton. Bykvist, Krister. 2003. Normative Supervenience and Consequentialism. Utilitas 15: 27–49. Carlson, Erik. 1995. Consequentialism Reconsidered. Dordrecht: Kluwer. Chan, Wing-tsit, ed. 1963. A Source Book in Chinese Philosophy. Princeton: Princeton University Press. Clarke, Samuel. 1987. A Discourse Concerning the Unchangeable Obligations of Natural Religion, and the Truth and Certainty of the Christian Revelation. 2nd series of the Boyle lectures, delivered at St. Paul’s in 1705; quoted from Mackie, J.L. 1980. Hume’s Moral Theory. London: Routledge and Kegan Paul. Collins Dictionary of the English Language. Glasgow: Collins, 1985. Commins, Saxe, and Robert Linscott, (eds.). 1947. The World’s Great Thinkers, Vol. 3: Man and the State. Trans. Samuel Moore. New York: Random House. Cooper, Wesley E. et al. (eds.). 1979. New Essays on John Stuart Mill and Utilitarianism. Canadian Journal of Philosophy, Suppl. Vol. V. Copp, David, ed. 2006. The Oxford Handbook of Ethical Theory. New York: Oxford University Press. Crisp, Roger. 1997. Mill on Utilitarianism. London: Routledge. Crisp, Roger, and Michael Slote, eds. 1997. Virtue Ethics. Oxford: Oxford University Press. Cummiskey, David. 1990. Kantian Consequentialism. Ethics 100: 586–615. ———. 1996. Kantian Consequentialism. New York/Oxford: Oxford University Press. Dancy, Jonathan. 1993. Moral Reasons. Oxford: Blackwell. ———. 1998. Moral realism. In Routledge Encyclopedia of Philosophy, vol. 6, 534–539. London/ New York: Routledge. Daniels, Norman. 1979. Wide Reflective Equilibrium and Theory Acceptance in Ethics. The Journal of Philosophy 76: 256–282. ———. 1980. Reflective Equilibrium and Archimedean Points. Canadian Journal of Philosophy 10: 83–103. Danielsson, Sven. 1988. Konsekvensetikens gränser (The Limits of Consequentialism). In Filosofiska utredninga. N.p.: Thales. ———. 1998. Filosofiska utredningar. N.p.: Thales. Danto, Arthur. 1965. Basic Actions. American Philosophical Quarterly 2: 141–148. ———. 1976. Mysticism and Morality: Oriental Thought and Moral Philosophy. Harmondsworth: Penguin. de Waal, Fran. 2006. Primates and Philosophers: How Morality Evolved. Eds. and Intr. Stephen Macedo and Josiah Ober. Princeton/Oxford: Oxford University Press/Princeton University Press. DePaul, Michael. 1993. Balance and refinement: Beyond coherence methods of moral inquiry. London/New York: Routledge. Dickie, George. 1984. The New Institutional Theory of Art. In Proceedings of the Eighth International Wittgenstein Symposium, Part I, ed. Rudolf Haller, 57–64. Vienna: Holder-Pichler-Temsky.

Bibliography

251

Dickie, George, et al., eds. 1977. Aesthetics: A Critical Anthology. 2nd ed. New York: St. Martin’s Press. Doberstein, John. 1959. Luther’s Works. Trans. and ed. Philadelphia: Fortress. Donagan, Alan. 1977. The Theory of Morality. Chicago: University of Chicago Press. Edwards, Paul, ed. 1967. The Encyclopedia of Philosophy. New York: Macmillan. Eriksson, Björn. 1994. Heavy Duty: On the Demands of Consequentialism. Stockholm: Almqvist & Wiksell International. Ezorsky, Gertrude. 1974. Unconscious Utilitarianism. The Monist 58: 468–474. Falk, W.D. 1986. Morality, Self, and Others. In Ought, Reasons, and Morality: The Collected Papers of W. D. Falk, 198–231. Ithaca: Cornell University Press. Firth, Roderick. 1952. Ethical Absolutism and the Ideal Observer. Philosophy and Phenomenological Research 12: 317–345. Fischer, John Martin, and Mark Ravizza, eds. 1992. Ethics: Problems and Principles. Fort Worth: Harcourt Brace Jovanovich. Fletcher, Joseph. 1966. Situation Ethics. Philadelphia: The Westminster Press. Flusser, David. 1990. The Ten Commandments and the New Testament. In The Ten Commandments in History and Tradition, ed. Gershon Levi. Jerusalem: The Magnes Press/The Hebrew University. (1985). Foot, Philippa. 1988. Utilitarianism and the Virtues. In Consequentialism and Its Critics, ed. Samuel Scheffler, 224–242. Oxford: Oxford University Press. Originally published in Mind 94 (1985): 196–209. Frankena, William. 1973. Ethics. 2nd ed. Englewwod Cliffs: Prentice-Hall. ———. 1990. Hare on Levels of Moral Thinking. In Hare and Critics: Essays on Moral Thinking, ed. Douglas Seanor and N. Fotion. Oxford: Clarendon Press. Freud, Sigmund. 1953–1964. Civilization and Its Discontents. First published in 1930, it is included in Vol. XXI of The Standard Edition of the Complete Psychological Works of Sigmund Freud. London: The Hogarth Press. Fried, Charles. 1978. Right and Wrong. Cambridge: Harvard University Press. Fromm, Erich. 1947. Man for Himself: An Inquiry into the Psychology of Ethics. New  York: Rinehart. Galling, Kurt, (ed.). 1958. Goldene Regel. In Die Religion in Geschichte und Gegenwart: Handwörterbuch für Theologie und Religionswissenschaft. Tübingen: J.  C. B.  Mohr (Paul Siebeck). Garnett, A.C. 1960. Ethics: A Critical Introduction. New York: Ronald Press. Gauthier, David, ed. 1970. Morality and Rational Self-Interest. Englewood Cliffs: Prentice-Hall. ———. 1975. Coordination. Dialogue: Canadian Philosophical Review 14: 195–221. ———. 1979. David Hume: Contractarian. Philosophical Review 88: 3–38. ———. 1986. Morals by Agreement. Oxford: Oxford University Press. ———. 1990. Moral Dealing: Contract, Ethics, and Reason. Ithaca/London: Cornell University Press. Geach, Peter. 1956. Good and Evil. Analysis 17: 33–42. Gensler, Harry. 1996. Formal Ethics. London/New York: Routledge. Gert, Bernard. 2005 (1988). Morality: Its Nature and Justification. New York: Oxford University Press. Gewirth, Alan. 1976. Moral Rationality. In Freedom & Morality, ed. John Bricke, 113–150. Lawrence: University of Kansas. ———. 1978. Reason and morality. Chicago: Chicago University Press. Gibbard, Allan. 1978. Act-Utilitarian Agreements. In Values and Morals, ed. A.I. Goldman and J. Kim, 91–119. Dordrecht: Reidel. ———. 1982. Inchoately Utilitarian Common Sense: The Bearing of a Thesis of Sidgwick’s on Moral Theory. In The Limits of Utilitarianis, ed. B.Miller Harlan and William H. Williams. Minneapolis: University of Minnesota Press.

252

Bibliography

———. 1990. Wise Choices, Apt Feelings: A Theory of Normative Judgment. Oxford: Oxford University Press. Glossop, Ronald. 1976. Is Hume a ‘Classical Utilitarian’. In Hume Studies, vol. 2, 1–16. Glover, Jonathan. 1975. It Makes No Difference Whether or Not I do It. Proceedings of the Aristotelian Society 49 (Supp): 171–190. Godwin, William. 1793. Enquiry Concerning Political Justice. Dublin: Luke White. Goldman, A.I., and J. Kim, eds. 1978. Values and Morals. Dordrecht: Reidel. Green, Thomas. 1798. An Examination of the Leading Principle in the New System of Morals Greene, Joshua D., et  al. 2001. An fMRI Investigation of Emotional Engagement in Moral Judgment. Science 293: 2105–2108. Griffin, James. 1986. Well-Being: Its Meaning, Measurement, and Moral Importance. Oxford: Oxford University Press. Griffiths, Phillips. 1957/58. Justifying Moral Principles. Proceedings of the Aristotelian Society, N. S. 58: 103–124. ———. 1967. Ultimate Moral Principles: Their Justification. In The Encyclopedia of Philosophy, ed. Paul Edwards, vol. 8, 177–182. New York: Macmillan. Grote, John. 1870. An Examination of the Utilitarian Philosophy. Cambridge: Deighton, Bell & Co. Halévy, Elie. 1901. La formation du radicalisme philosophique. Paris: Alcan. Trans. Mary Morris as The Growth of Philosophical Radicalism. London: Faber and Gwyer, 1928. Haller, Rudolf, ed. 1984. Proceedings of the Eighth International Wittgenstein Symposium, Part I. Vienna: Holder-Pichler-Temsky. Hardin, Russell. 1988. Morality Within the Limits of Reason. Chicago/London: The University of Chicago Press. Hare, R.M. 1963. Freedom and Reason. Oxford: Oxford University Press. ———. 1975a. Abortion and the Golden Rule. Philosophy and Public Affairs 3: 201–222. ———. 1975b. Euthanasia: A Christian View. Proceedings of the Center for Philosophic Exchange 6: 43–52. ———. 1976. Ethical Theory and Utilitarianism. In Contemporary British Philosophy, ed. H.D. Lewis, 113–131. London: Allen and Unwin. ———. 1981. Moral Thinking: Its Levels, Methods, and Point. Oxford: Oxford University Press. ———. 1990. Comments. In Hare and Critics: Essays on Moral Thinking, ed. Douglas Seanor and N. Fotion. Oxford: Clarendon Press. ———. 1993. Could Kant Have Been a Utilitarian. Utilitas 5: 1–16. Harman, Gilbert. 1977. The Nature of Morality: An Introduction to Ethics. New  York: Oxford University Press. Harman, Gilbert, and Judith Jarvis Thomson. 1996. Moral Relativism and Moral Objectivity. Oxford: Blackwell. Harris, N.G.E. 1972. Nondeliberative Utilitarianism. Ethics 82: 344–348. Harris, John. 1980. Violence and Responsibility. London: Routledge and Kegan Paul. Harsanyi, John. 1977a. Morality and the Theory of Rational Behavior. Social Research 44: 623–656. ———. 1977b. Rational Behavior and Bargaining Equilibrium in Games and Social Situations. Cambridge: Cambridge University Press. Hart, H.L.A. 1961. The Concept of Law. Oxford: Oxford University Press. Hartnack, Justus. Performative Utterances. In The Encyclopedia of Philosophy, ed Paul Edwards. Hastings, James, ed. 1908–1926. Encyclopedia of Religion and Ethics, 13 vols. Edinburgh/New York: T. and T. Clark/Charles Scribner’s sons. Epstein, I., ed. 1987. Hebrew-English Edition of the Babylonian Talmud. London: Soncino, Cop. Hegel G.W.F. 1991. Elements of the Philosophy of Right. Trans. H.B. Nisbet, ed. Allen W. Wood. Cambridge: Cambridge University Press. Henry, Sidgwick. 1907. The Methods Of Ethics. 7th ed. London: Macmillan.

Bibliography

253

Herodotus. 1963. Histories. 3:38; The Loeb classical library, ed. and tr. A.D.  Godley, Vol. 2. Cambridge, MA: Heineman (1921). Hodgson, D.H. 1967. Consequences of Utilitarianism: A Study in Normative Ethics and Legal Theory. Oxford: Oxford University Press. Hoerster, Norbert. 1973. Is Act-Utilitarian Truth-Telling Self-Defeating? Mind 82: 413–416. Holmes, Jr, and Wendell Oliver. 1896. The Path of the Law. Harvard Law Review 10: 457–478. Hooker, Brad. 2002 (2000). Ideal Code, Real World: A Rule-Consequentialist Theory of Morality. Oxford: Oxford University Press, . Hooker, Brad. Rule-Consequentialism. In The Stanford Encyclopedia of Philosophy (Spring 2004 Edition), ed. Edward N. Zachta. http://plato.stanford.edu/archives/spr2004/entries/consequentialism/rule, p. 12. Hume, David. 1998. An Enquiry concerning the Principles of Morals. Oxford: Oxford University Press. ———. 2004 (2000). A Treatise of Human Nature, ed. David Fate Norton and Mary Fate Norton. Oxford: Oxford University Press. Hurka, Thomas. 1990. Two Kinds of Satisficing. Philosophical Studies 59: 107–111. Hursthouse, Rosalind. 1999. On Virtue Ethics. Oxford: Oxford University Press. Hutcheson, Francis. 1742. An Essay on the Nature and Conduct of the Passions and Affections and Illustrations upon the Moral Sense. 3rd ed. London: printed for A. Ward et al. (1728). Isocrates. 1928. Nicocles. In Loeb Classical Library. Cambridge: Harvard University Press. Jackson, Samuel M., ed. 1949. The New Schaff-Herzog Encyclopedia of Religius Knowledge. Grand Rapids: Baher Book House. Jackson, Frank. 1991. Decision-Theoretic Consequentialism and the Nearest and Dearest Objection. Ethics 101: 461–482. Jarvis, Thomson Judith, and Gerald Dworkin, eds. 1968. Ethics. New York: Harper & Row. Jeffrey, Wattles. 1996. The Golden Rule. Oxford: Oxford University Press. Joel, Feinberg, ed. 1969. Moral Concepts. London: Oxford University Press. Kagan, Shelly. 1989. The Limits of Morality. Oxford: Oxford University Press. ———. 1991. Précis of The Limits of Morality. Philosophy and Phenomenological Research 51: 897–901. ———. 1998. Normative Ethics. Boulder: Westview Press. Kalin, Jesse. 1970. In Defense of Egoism. In Morality and Rational Self-Interest, ed. David Gauthier, 64–87. Englewood Cliffs: Prentice-Hall. Kant, Immanuel. 1873. Kritik der praktischen Vernunft; quoted from Kant’s Critique of Practical Reason and Other Works on the Theory of Ethics. Trans. T.K.  Abbott. London: Longmans, Green. ———. 1948. Grundlegung zur Metaphysik der Sitten, 404; the quotation is from H. J. Paton’s translation, The Moral Law. London: Hutchinson. ———. 1976. On a Supposed Right to Lie from Altruistic Motives. In Immanuel Kant: Critique of Practical Reason and Other Writings in Moral Philosophy, edited and translated by Lewis White Beck, 346–350. Reprint. New York: Garland. Kapur, Neera. 1991. Why Is It Wrong to Be Always Guided by the Best: Consequentialism and Friendship. Ethics 101: 483–504. Keynes, John Maynard. 1949. My Early Beliefs. In Two Memoirs. New York: Augustus M. Kelley. Knud, Haakonssen. 2002. Article on Mackintosh. In Dictionary of Nineteenth-Century British Philosophers, ed. W.J. Mander and Alan P.F. Sell, vol. 2, 715–719. Bristol: Thoemmes Press. Korsgaard, Christine. 1998. Teleological Ethics. In Routledge Encyclopedia of Philosophy. London/New York: Routledge. Kuflik, Arthur. 1986. A Defense of Common-Sense Morality. Ethics 96: 784–803. Kupperman, Joel. 1983. The Foundations of Morality. London: George Allen & Unwin. Kymlicka, Will. 1988. Rawls on Teleology and Deontology. Philosophy & Public Affairs 17: 173–190.

254

Bibliography

———. 1993. The Social Contract Tradition. In A Companion to Ethics, ed. Peter Singer, 186– 196. Oxford: Blackwell. Leibniz, G.W. 1981. New Essays on Human Understanding. Trans. Peter Remnant and Jonathan Bennett. Cambridge: Cambridge University Press (1765). Levi, Gershon (ed.). 1990 (1985). The Ten Commandments in History and Tradition. Jerusalem: The Magnes Press, The Hebrew University. Lewis, David. 1969. Convention: A Philosophical Study. Cambridge: Harvard University Press. ———. 1972. Utilitarianism and Truthfulness. Australasian Journal of Philosophy 50: 17–19. Luce, Duncan, and Howard Raiffa. 1957. Games and Decisions. New York: Wiley. Lukes, Steven. 1973. Individualism. Oxford: Blackwell. Lyons, David. 1965. Forms and Limits of Utilitarianism, 1970. Oxford: Oxford University Press. Mackie, J.L. 1973. The Disutility of Act-Utilitarianism. The Philosophical Quarterly 23: 289–300. ———. 1977. Ethics: Inventing Right and Wrong. Harmondsworth: Penguin. ———. 1987. Hume’s Moral Theory. London: Routledge and Kegan Paul. (1980). Mackintosh, James. 1836. A Dissertation on the Progress of Ethical Philosophy, Chiefly During the Seventeenth and Eighteenth Centuries. Edinburgh: Adam and Charles Black. Mander, W.J. and Alan P.F.  Sell (eds.). 2002. Dictionary of Nineteenth-Century British Philosophers, 2 vols. Bristol: Thoemmes Press. Marx, Karl. 1947. The Communist Manifesto. In The World’s Great Thinkers, ed Saxe Commins and Robert Linscott, Vol. 3: Man and the State; translated by Samuel Moore. New  York: Random House. ———. 1977. The German Ideology. In Karl Marx: Selected writings, ed. David McLellan. Oxford: Oxford University Press. Mautner, Thomas, ed. 1997. Dictionary of Philosophy. 2nd ed. Harmondsworth: Penguin. McCloskey, H.J. 1965. A Non-Utilitarian Approach to Punishment. Inquiry 8: 249–263. McLellan, David, ed. 1977. Karl Marx: Selected Writings. Oxford: Oxford University Press. McNaughton, David, and Piers Rawling. 2006. Deontology. In The Oxford Handbook of Ethical Theory, ed. David Copp, 424–458. New York: Oxford University Press. Melden, A.I., ed. 1958. Essays in Moral Philosophy. Seattle: University of Washington Press. Midgley, Mary. 1993. The Origin of Ethics. In A Companion to Ethics, ed. Peter Singer, 3–13. Oxford: Blackwell. Mill, J.S. 1998. In Utilitarianism, ed. Roger Crisp. Oxford: Oxford University Press. Miller, Harlan B., and William H. Williams, eds. 1982. The Limits of Utilitarianism. Minneapolis: University of Minnesota Press. Monro, D.H., ed. 1972. A Guide to the British Moralists. London: Collins. Moore, G.E. 1903. Principia Ethica. Cambridge: Cambridge University Press. ———. 1912. Ethics. London: Oxford University Press. Moore, G.  E. 1942. Autobiography. In The Philosophy of G.  E. Moore, The Library of Living Philosophers, Vol. IV, ed. Paul A. Schilpp. Evanston: Northwestern University Press. Muirhead, J.H. 1932. Rule and End in Morals. London: Oxford University Press. Myers, R.H. 1994. Prerogatives and Restrictions from the Cooperative Point of View. Ethics 105: 128–152. Nagel, Thomas. 1970. The Possibility of Altruism. Oxford: Oxford University Press. ———. 1979a. Moral Luck. P. A. S., Supp. Vol. L (1976): 137–155; Reprinted in Mortal Questions. Cambridge: Cambridge University Press. ———. 1979b. Mortal Questions. Cambridge: Cambridge University Press. ———. 1986. The View from Nowhere. Oxford: Oxford University Press. Nansen, Fridtjof. 1890. The First Crossing of Greenland, 2 vols. London: Longman/Green & Co. Narveson, Jan. 1971. Promising, Expecting, and Utility. Canadian Journal of Philosophy 1: 207–223. Nell, Onora. 1975. Acting on Principle: An Essay on Kantian Ethics. New York/London: Columbia University Press. Niebuhr, Reinhold. 1949. The Nature and Destiny of Man. New York: Charles Scribner’s Sons.

Bibliography

255

Norman, Richard. 1983. The Moral Philosophers: An Introduction to Ethics. Oxford: Oxford University Press. Olson, Mancur. 1965. The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge: Harvard University Press. Olson, Robert. 1967a. Deontological Ethics. In The Encyclopedia of Philosophy, ed. Paul Edwards. New York: Macmillan. ———. 1967b. Teleological Ethics. In The Encyclopedia of Philosophy, ed. Paul Edwards. New York: Macmillan. Olson, Jonas, and Frans Svenson. 2003. A Particular Consequentialism. Utilitas 15: 194–205. Österberg, Jan. 1988. Self and Others: A Study of Ethical Egoism. Dordrecht: Kluwer. ———. 1999. The Virtues of Virtue Ethics. In Philosophical crumbs, ed. Rysiek Sliwinski. Uppsala: Uppsala Philosophical Studies. Parfit, Derek. 1984. Reasons and Persons. Oxford: Oxford University Press. ———. 1988. Is Common-Sense Morality Self-Defeating? In Consequentialism and Its Critics, ed. Samuel Scheffler, 173–186. Oxford: Oxford University Press. Originally published in Journal of Philosophy, 76 (1979): 533–45. ———. Draft of 28 April 2008. On what matters. Derek Parfit’s Homepage. Philippidis, L.J. 1929. Die “goldene Regel” religionsgeschichtlich untersucht. Leipzig: Adolf Klein Verlag. Piper, Adrian. 1978. Utility, Publicity, and Manipulation. Ethics 88: 189–206. Plato. 1987. The Republic. Trans Desmond Lee. Harmondsworth: Penguin. Postow, B.C. 1977. Generalized Act Utilitarianism. Analysis 37: 49–52. Provis, C. 1977. Gauthier on Coordination. Dialogue: Canadian Philosophical Review 16: 507–509. Rabinowicz, Wlodek, and Jan Österberg. 1996. Value Based on Preferences: On Two Interpretations of Preference Utilitarianism. Economics and Philosophy 12: 1–27. Rachels, James. 1974. Two Arguments against Ethical Egoism. Philosophia 4: 297–314. ———. 1999. The Elements of Moral Philosophy. 3rd ed. Boston: McGraw-Hill. Railton, Peter. 1988. Alienation, Consequentialism, and the Demands of Morality. In Consequentialism and Its Critics, ed. Samuel Scheffler, 93–133. Oxford: Oxford University Press. Originally published in Philosophy and Public Affairs 13 (1984): 134–71. Raphael, D.D. 1969. British Moralists 1650–1800, 2 vols. Oxford: Oxford University Press. ———. 1974. Sidgwick on Intuitionism. The Monist 58: 405–419. Rawls, John. 1971. A Theory of Justice. Cambridge: Harvard University Press. Regan, Donald. 1984 (1980). Utilitarianism and Cooperation. Oxford: Oxford University Press. Regan, Tom. 1986. Bloomsbury’s Prophet: G.  E. Moore and the Development of His Moral Philosophy. Philadelphia: Temple University Press. Reich, Wilhelm. 1969. The Sexual Revolution: Toward a Self-Governing Character Structure. New York: Farrar, Straus and Giroux. Reiner, Hans. 1983. Duty and Inclination: The Fundamentals of Morality Discussed and Redefined with Special Regard to Kant and Schiller. The Hague: Martinus Nijhoff. Rescher, Nicholas. 1966. Distributive Justice: A Constructive Critique of the Utilitarian Theory of Distribution. Indianapolis: Bobbs-Merrill. Richards, David. 1970. A Theory of Reasons for Action. Oxford: Oxford University Press. Ross, W.D. 1930. The Right and the Good. London: Oxford University Press. ———. 1939. Foundations of Ethics. London: Oxford University Press. Rousseau, Jean-Jacques. 1973. The Social Contract and Discourses. Trans G.D.H. Cole. London: Everyman. Routledge Encyclopedia of Philosophy. London/New York: Routledge, 1998. Runciman, W.G., and Amartya Sen. 1965. Games, Justice and the General Will. Mind 74: 554–562. Rydén, Lars (ed.). 1990. Etik för forskare: en antologi med utgångspunkt i arbetet med Uppsalakoden (“Ethics for Scientists: An Anthology with a Point of Departure in the Working with the Uppsala Code”). Stockholm: UHÄ/foU skriftserie, 1990: 1.

256

Bibliography

Scanlon, T.M. 1988. Levels of Moral Thinking. In Hare and Critics: Essays on Moral Thinkin, ed. Douglas Seanor and N. Fotion. Oxford: Oxford University Press. (Oxford: Clarendon Press, 1990). ———. 1998. What we owe to each other. Cambridge/London: Harvard University Press. Scarre, Geoffrey. 1996. Utilitarianism. London/New York: Routledge. Scheffler, Samuel, ed. 1988a. Consequentialism and Its Critics. Oxford: Oxford University Press. ———. 1988b. Introduction. In Consequentialism and Its Critics, ed. Samuel Scheffler. Oxford: Oxford University Press. ———. 2000 (1982). The Rejection of Consequentialism. Oxford: Oxford University Press. Schilpp, Paul, ed. 1942. The Philosophy of G.  E. Moore. Chicago/Evanston: Northwestern University Press. Schneewind, J.B. 1963. First Principles and Common Sense Morality in Sidgwick’s Ethics. Archiv für Geschichte der Philosophie, Bd. 45: 137–156. ———. 1977. Sidgwick’s Ethics and Victorian Moral Philosophy. Oxford: Oxford University Press. ———. 1990. The Misfortunes of Virtue Ethics. Ethics 101: 42–63. Seanor, Douglas, and N. Fotion, eds. 1988. Hare and Critics: Essays on Moral Thinking. Oxford: Oxford University Press. Sedgwick, Adam. 1832. Discourse on the Studies of the University of Cambridge. Cambridge: Cambridge University Press. Selby-Bigge, L.A., ed. 1897. British Moralists. Oxford: Oxford University Press. Sen, Amartya. 1970. Collective Choice and Social Welfare. San Francisco/Edinburgh: Holden-­ Day/Oliver & Boyd. Sen, Amartya, and Bernard Williams, eds. 1982. Utilitarianism and Beyond. Cambridge: Cambridge University Press. Shakespeare Hamlet. Act 5, Scene 2, lines 129–132. Shaw, George Bernard. 1919 (1903). Maxims for a Revolutionist. In Man and Superman. London: Constable. Simmons, John. 1982. Utilitarianism and Unconscious Utilitarianism. In The Limits of Utilitarianism, ed. Harlan B. Miller and William H. Williams, 86–92. Minneapolis: University of Minnesota Press. Simon, Herbert. 1959. Theories of Decision Making in Economics and Behavioral Science. American Economic Review 49: 253–283. Singer, Marcus G. 1967. Golden Rule. In The Encyclopedia of Philosophy, ed. Paul Edwards, vol. 3, 365–367. New York: Macmillan. Singer, Peter. 1972. Is Act-Utilitarianism Self-Defeating? Philosophical Review 81: 94–104. ———. 1974. Sidgwick and Reflective Equilibrium. The Monist 58: 490–516. ———. 1981. The Expanding Circle: Ethics and Sociobiology. Oxford: Oxford University Press. ———, ed. 1993. A Companion to Ethics. Oxford: Blackwell. ———. 2005. Ethics and Intuitions. The Journal of Ethics 9: 331–352. Slote, Michael. 1985. Common-sense Morality and Consequentialism. London: Routledge & Kegan Paul. ———. 1997. Virtue Ethics. In Three Methods of Ethics: A Debate, ed. Marcia Baron, Philip Pettit, and Michael Slote. Oxford: Blackwell. ———. 2001. Moral from Motives. Oxford: Oxford University Press. Smart, J.J.C. 1956. Extreme and Restricted Utilitarianism. The Philosophical Quarterly 6: 344–354. ———. 1973. An Outline of a System of Utilitarian Ethics. In Utilitarianism: For and Against by J.J. C. Smart and Bernard Williams, Ch. 1. London: Cambridge University Press. Smart, J.J.C., and Bernard Williams. 1973. Utilitarianism: For and Against. London: Cambridge University Press. Sobel, Jordan Howard. 1989. Kent Bach on Good Arguments. Canadian Journal of Philosophy 19: 447–454.

Bibliography

257

———. 2008. Walls and Vaults. New York: Wiley. Spooner, W.A. 1913. The Golden Rule. In Encyclopedia of Religion and Ethics, ed. James Hastings. Edinburgh: T. & T. Clark. Stephen, Leslie. 1900. The English Utilitarians. London: Duckworth. Strang, Colin. 1960. What If Everyone Did That? Durham University Journal 53: 5–10. Strawson, P.F. 1961. Social Morality and Individual Ideal. Philosophy 36: 1–17. ———. 1970. Social Morality and Individual Ideal. In The Definition of Morality, ed. G. Wallace and A.D.M. Walker, 98–118. London: Methuen. Sumner, L.W. 1969. Consequences of Utilitarianism. Dialogue: Canadian Philosophical Review 7: 639–642. Svensson, Frans. 2006. Some Basic Issues in Neo-Aristotelian Virtue Ethics. Uppsala: n. p. Thieme, Karl. 1949. Consilia Evangelica. In The New Schaff-Herzog Encyclopedia of Religius Knowledge, ed. Samuel M. Jackson. Grand Rapids: Baher Book House. Hobbes Thomas. 1968. Leviathan, ed. C.B. MacPherson. Harmondsworth: Penguin. Thomson, Judith Jarvis. 1985. The Trolley Problem. The Yale Law Journal 94: 1395–1415. Reprinted in Ethics: Problems and Principles edited by Fischer, John Martin and Mark Ravizza, 279–292. Fort Worth: Harcourt Brace Jovanovich, 1992. ———. 1997. The Right and the Good. The Journal of Philosophy 94: 273–298. Tiberius, Valerie. 2006. How to Think About Virtue and Right. Philosophical Papers 35: 247–265. Turnbull, Colin. 1972. The Mountain People. New York: Simon and Schuster. Unger, Peter. 1996. Living High and Letting Die: Our Illusion of Innocence. Oxford: Oxford University Press. Urmson, J.O. 1953. The Interpretation of the Moral Philosophy of J. S. Mill. The Philosophical Quarterly 3: 33–40. ———. 1969. Saints and Heroes. In Moral Concepts, ed. Joel Feinberg, 60–73. London: Oxford University Press. Originally published in Essays in Moral Philosophy edited byA. I. Melden, 198–216. Washington: Washington University Press, 1958. von Wright, G.H. 1963. The Varieties of Goodness. London: Routledge & Kegan Paul. Waldron, Jeremy. 1994. Kagan on Requirements: Mill on Sanctions. Ethics 104: 310–324. Wallace, G., and A.D.M. Walker, eds. 1970. The Definition of Morality. London: Methuen. Warnock, G.J. 1976 (1971). The Object of Morality. London: Methuen. West, Henry R. 2004. An Introduction to Mill’s Utilitarian Ethics. Cambridge: Cambridge University Press. Westermarck, Edward. 1908. The Origin and Development of the Moral Ideas, 2 vols. London: Macmillan. Whewell, William. 1836. Preface to A Dissertation on the Progress of Ethical Philosophy, Chiefly During the Seventeenth and Eighteenth Centuries by James Mackintosh. Edinburgh: Adam and Charles Black. William, Bishop of St. Davids. 1679. The Comprehensive Rule of Righteousness: Do as You Would Be Done By. Cornhill: William Leach. Williams, Bernard. “A Critique of Utilitarianism”. In Utilitarianism: For and Against, by J.  J. C. Smart and Bernard Williams. London: Cambridge University Press, 1973. ———. 1981a. Moral Luck. P.  A. S. Supp. Vol. L (1976): 115–135. Reprinted in Moral Luck: Philosophical Papers 1973–1980. Cambridge: Cambridge University Press. ———. 1981b. Utilitarianism and Moral Self-Indulgence. In Moral Luck. Cambridge: Cambridge University Press. ———. 1981c. Moral Luck. Cambridge: Cambridge University Press. Wood, Allen W. 1990. Hegel’s Ethical Thought. Cambridge: Cambridge University Press. Zuboff, Arnold. 1977–1978. Moment Universals and Personal Identity. Proceedings of the Aristotelian Society N. S 77: 141–155. ———. 1990. One Self: The Logic of Experience. Inquiry 33: 39–68.