From reasons to norms
Library of Ethics and Applied philosophy Volume 22
Editor in Chief Marcus Düwell, Utrecht University, Utrecht, NL
Editorial Board Deryck Beyleveld, Durham University, Durham, U.K. David Copp, University of Florida, USA Nancy Fraser, New School for Social Research, New York, USA Martin van Hees, Groningen University, Netherlands Thomas Hill, University of North Carolina, Chapel Hill, USA Samuel Kerstein, University of Maryland, College Park Will Kymlicka, Queens University, Ontario, Canada Philip van Parijs, Louvaine-la-Neuve (Belgium) en Harvard, USA Qui Renzong Peter Schaber, Ethikzentrum, University of Zürich, Switzerland Thomas Schmidt, Humboldt University, Berlin, Germany
For other titles published in this series, go to www.springer.com/series/6230
From Reasons to Norms On the Basic Question in Ethics
by
Torbjörn Tännsjö University of Stockholm, Sweden
Torbjörn Tännsjö Department of Philosophy University of Stockholm Sweden
[email protected]
ISBN 978-90-481-3284-3 e-ISBN 978-90-481-3285-0 DOI 10.1007/978-90-481-3285-0 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2009942134 © Springer Science+Business Media B.V. 2010 No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, microfilming, recording or otherwise, without written permission from the Publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the work. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)
Acknowledgements
This book has grown out of a brief correspondence, some 10 years ago, with Derek Parfit and Wlodek Rabinowicz, on points where I disagree with these philosophers. It is the result, however, of more than 30 years of reflection on metaethical questions, and it can be considered a credo. It is a short book, but in it many bold theses are put forward. Together they constitute a consistent, systematic, and well considered, I hope, moral philosophy. It is a mixture of reductionist (reasons are for example reduced to explanations of particular obligations) and nihilist theses (I claim that there are no sui generis norms to do with epistemology, prudence, or aesthetics, for example) and one strong defence of a realist position (I argue that there is indeed a unique answer to the question what to do, simpliciter). Why such a mixture of reductionist, nihilist, and realist claims? The only and equally natural answer I can give to this very natural question is that in the book I describe the world, as I tend to believe that it happens to be. I should also add that, while I make both semantic and ontological claims in my book, I have been more eager to defend my ontological claims than the semantic ones. There is no way in a short book to account in all detail for the semantic aspects of my reductivist and nihilist claims. The number of ways of using words such as ‘reason’, ‘prudence’, ‘justified belief’ and so forth is legion. I cannot truly say that I have expertise in all the fields into which I have had to enter in my defence of all the claims I make, and in particular, there are shortcomings in my knowledge of the relevant literature in some areas. However, many people have helped me to update my reading, first and foremost the members of the seminar of practical philosophy at the department of philosophy at Stockholm University, the members of the seminar of practical philosophy at Uppsala University, and the members of a research project on relativism conducted by the philosophy departments in Stockholm and Gothenburg together with the department of the study of ideas in Gothenburg. One of the members of the Stockholm seminar in particular, Jens Johansson, has helped me to a better version of Chapter 8 on naturalised epistemology, the field where I have least knowledge. As a matter of fact, he raised 24 objections to an early draft of the chapter, all of them worthy of consideration. Lars Bergström, from whom I have learnt more in philosophy than from any other source, has read and commented upon the chapter as well and corrected several mistakes. Interestingly v
vi
Acknowledgements
enough, my ambition to naturalise epistemology started as a reaction to Bergström’s belief that it was inherently normative; while I was working on the theme, I noted that Bergström simultaneously and somewhat unexpectedly changed his view. He helped me to relevant literature in a field. Professor Isaac Levi also made helpful comments. My views as here put forward are obviously very much inspired by things that Professor Levi has had to say about the subject. This does not mean that he agrees with every single point I make, quite to the contrary. So even if I claim no originality in this chapter, I thank Professor Levi for having pressed for a clearer presentation of some of the rather subtle points where we disagree; it goes without saying that I take full responsibility for whatever mistakes that may remain in the chapter. Both Jens Johansson and Jonas Olson have carefully read a penultimate draft of the ms and given me more help than I can acknowledge without being boringly repetitive. I have profited from further correspondence and intense discussions with Wlodek Rabinowicz and John Broome. With Broome I share an idea of what it means to have a normative (moral) reason to do something; I disagree, however, with his idea that there is such a thing as valid practical reasoning (the subject of our correspondence). John has also read and commented on several drafts of Chapters 2 through 5. The book has been written, as was noted, while I have been working within a research project on relativism. This project has been kindly subsidised by The Bank of Sweden Tercentenary Foundation. I thank the foundation for making this very stimulating research project possible. In the spring of 2008 I gave a course on the book ms together with Attila Tanyi, at Stockholm University. I thank the students, and Attila in particular, for many valuable comments. Attila has commented on the entire penultimate version of the ms. The manuscript was finalised when I spent a semester (the fall of 2008) as a research fellow at Swedish Collegium for Advanced Study (SCAS), together with, among others, Gustaf Arrhenius, Folke Tersman, Wlodek Rabinowicz, John Broome, Krister Bykvist, Anandi Hittiagadi, and Goeffrey Brennan, all of whom gave helpful comments at this final stage. John Bishop, who visited SCAS for a workshop on reasons and the explanation of action, also gave helpful comments on Chapter 6. The manuscript has been written with the intention of producing a book, but a discussion about moral explanation in Chapter 5 has appeared as ‘Understanding Through Explanation in Ethics’ in Theoria (2006) and a section about moral relativism in Chapter 6 has appeared as ‘Moral Relativism’ in Philosophical Studies (2007), and a small point on the causal theory of action made in Chapter 2 builds on my recent ‘On deviant causal chains – no need for a general criterion’ (Analysis, 2009). I am grateful to the publishers for permitting me to use the material also in this context.
Preface
Metaethics is the inquiry into the nature of morality (or ethics, I use the words ‘morality’, ‘morals’, and ‘ethics’ as synonyms). When we pass moral judgements, what kind of claims are we then making? I speak of this as the semantic metaethical question. Are there moral facts, to be discovered by us and existing independently of our thoughts and conceptualisation? I speak of this as the ontological or metaphysical metaethical question. And, if there are, can we know about them; and, if we can, how do we get this kind of knowledge? I speak of this as the epistemic metaethical question. All these metaethical questions, the semantic, the ontological, and the epistemic ones, are raised and discussed in this book, but they are not the core questions raised. I have been more concerned with another kind of questions, which deserve to be called metaethical as well: what are the problems of morality? Are there many different moral questions, or, do they all, in the final analysis, reduce to only a few, or perhaps just one? This question is of special importance to a non-naturalist objectivist and realist like the present author, who believes that we do make truth-claims when we pass moral judgements and who believes that there is a truth in these matters so that we must face the possibility that even our most cherished moral judgements may be false. If this general belief is correct it becomes important to know whether there is just one kind of moral problem to ponder, or whether there are many. Is there more than one kind of moral truth? G.E. Moore wrestled with this question during his entire philosophical career. Apart from his contribution, it has been amazingly absent from the discussion. Here it is raised once again and answered. The answer given to the question is that there is just one kind of moral problem to worry about, to wit, the normative one. What ought to be done is the central (only) moral question. When we know what in each situation we ought to do simpliciter – and why, one should add – then we know all there is to know in morality. I identify morality with an authoritative answer to this question. This is the basic question of ethics. When I speak of it as ‘moral’, then this is a mere matter of definition on my part, and it is important to keep this in mind when reading this book. My moral realism is realism with respect to this problem, and this problem only. However, it might still be asked whether this in itself is a unique question. Are there not different kinds of ought-judgements? Could we come to face conflicting obligations? vii
viii
Preface
This is not so. Hence a further claim made in this book is that there is just one source of normativity, just one kind of ‘ought’-question, which lends itself to an objectively correct and authoritative answer. This is the question what I ought to do, simpliciter. We never face genuine conflicts between the calls from different obligations. For example, we never face a genuine conflict between the call from morality, prudence, or epistemology. There are no prudential or epistemic sui generis norms that can come into conflict with the moral ones. And even if there were, one may add, why bother with them, if we already have an answer to the question what we ought to do simpliciter? This is the main theme of the book, then, the insistence on the thesis that there is exactly one source of obligation. I do approach this theme in an indirect fashion, however. I start my discussion with a query about practical reasons. This has to do with a deeply felt frustration on my part. The source of this frustration is the observation that many moral philosophers now seem to be preoccupied with practical reasons rather than with obligation. In the book I try to see through and expose this talk about reasons and to exorcise this entire intellectual fashion. This is indeed a book on practical reason, but the main thrust of my argument is the thesis that, in the final analysis, we have no real need for the notion of a practical reason. What is needed is a Copernican counter-revolution rendering it possible for us to have our thoughts revolving around the unique, deep, and real question: What ought I to do and why? We ought to make a move in our thinking, then, from reasons back to norms. Only if we understand why we should do so can we get a clear understanding of our metaethical predicament. The turn to reasons is a recent phenomenon; many practical philosophers tend now to focus on practical ‘reason’ rather than on ‘obligation’. The trend is perhaps inaugurated already in Thomas Nagel’s The Possibility of Altruism, and it is consummated in books such as Joseph Raz, Practical Reasons and Norms, Derek Parfit’s Reasons and Persons, Thomas Scanlon’s What We Owe to Each Other, and Jonathan Dancy’s Moral Reasons (followed up by his Practical Reality). This recent turn to practical reasons leaves me at unease. The notion of an action being right or wrong or obligatory are much more central to moral theory than the notion of a reason. Moreover, as we will see, a fruitful notion of an objective moral reason can be defined in terms of obligations and the explanations of obligations (or permissions). There is no way that we can move in the other direction, that is to define obligations in terms of objective reasons. There exist some obligations, for which there are no reasons, or so I will argue. The turn to reasons also tends to allow us to forget about, or not realise, some radical implications of some moral positions. One example of this the present idea in vogue that we can solve the putative problem that moral theories seem to place too heavy burdens on our shoulders with reference to a theory of practical reasons. On this view, it might well be true that a moral theory such as utilitarianism or Kantianism makes heavy demands on us. We ought to give all we own to charity according to utilitarianism, and we ought to tell the truth even if it means the destruction of ourselves, our family and everything we care about, according to Kantianism, but this is not the end of the matter. We must also query how strong
Preface
ix
reasons we have to abide by these theories in situations such as these. Perhaps our reasons are not as stringent, then, as they at first seemed to be.1 If I am right this is a non-starter. There are objective moral reasons explaining our obligations simpliciter. If utilitarianism is true, say, then we ought to give all our wealth to charity, since by so doing we make the world a better place (let us make this empirical assumption). But then we have a conclusive reason to give all our wealth to charity, as I will show, since our (moral) reason is the explanation of this obligation (utilitarianism together with the empirical assumption that by so doing we make the world a better place explains why we ought to do it). Moreover, to a moral realist, who believes in the existence of objective moral reasons, the very idea that morality, if it makes heavy demands on us, may violate our integrity, invade our right to privacy, and so forth, is a non-starter. If morality is conceived of as socially constructed it makes sense to say that, if (the other members of) society require a lot from us, our integrity is at jeopardy. But if moral realism is true, the claim that morality makes heavy demands on us is of the same order as the claim that the law of gravity renders life difficult for us, in so many ways. Another example of the same phenomenon is given by particularism. According to particularism, there are no true moral principles. If this is so, then we may still face strict particular obligations, even heavy ones, but we have no reasons whatever to abide by them, or so I will argue. If there are no true moral principles, then there are no moral reasons. This does not show that particularism is false, but it does indeed show that it has very radical implications. Moreover, the turn to reasons has a tendency to engender confusion. In fact, this talk about reasons seems to prepare the ground for much too many types of reasons. We ought to try to see through and expose this talk. Many practical philosophers, misled by their interest in reasons, seem to believe that there are many sources of normativity, corresponding to different types of practical reasons. They speak of such things as norms of, or reasons stemming from, prudence, rationality, justice, epistemology, and the idea of a suitable and fitting emotion – as well as from morality. This position is deeply problematic. If these philosophers are right, we may have to acknowledge that, in a situation, we may come to face conflicting obligations. Not only is it the case that, from one point of view, there is one thing we ought to do (or believe, or feel) while, from another point of view, we ought not to do it. This is trivial. But it seems also, according to this view, to be the case that, each point of view is its own source of obligation. So, on the face of it, it seems as though there are several ‘truths’ about what it is we ought (categorically) to do in the situation. This phenomenon, which could be called a kind of practical ‘relativism’, leaves me, as it once left Sidgwick, disturbed. In the first edition to his Methods of Ethics he claimed that it meant that ‘the Cosmos of Duty is reduced to Chaos’2 However,
Sarah Stroud has toyed with this idea, in ‘Overridingness and Moral Theory’, and Paul Hurley has defended it in ‘Does Consequentialism Make Too Many Demands?’. 2 Quoted in Jerome Schneewind, Sidgick’s Ethics and Victorian Moral Philosophy, p. 352. I have the reference from Owen McLeod, ‘Just Plain “Ought”’, p. 271. 1
x
Preface
while Sidgwick accepted this relativism, and spoke of it as a ‘dualism’ of the practical reason,3 it transpires that there is a way to transcend it. I suspect, then, that the turn to reasons represents no intellectual advance; it is nothing but a new way of formulating the same old problems – good old wine, only in new bottles. And it does engender the mistaken beliefs that there are easy ways out of well-known difficulties with normative theories, and that there are many competing sources of normativity. However, the situation is even worse than that. The turn to reasons tends to bring with it conflation of various different considerations that should be kept apart. The fact that the term ‘practical reason’ is ambiguous, the fact that there exist two separate notions of practical reason, is often lost track of in the discussion with lack of clarity as an obvious result. In fact, the turn to reasons seems to bring with it a tendency to conflate two kinds of problems, that do both allow that we speak of them in terms of ‘reasons’, but which are entirely different in kind: the problem of how to explain and rationalise (by the lights of an agent) a certain action, on the one hand, and, on the other hand, the problem about what to do and why, that is the problem of explaining obligations (or permissions). When we tend to say that the reason that a person performed a certain action was his beliefs and desires that did rationalise what he did, then we may speak of ‘Humean reasons’. But we may also be interested to know why in the first place he had a certain obligation. Our explanation of why he had the obligation in question may also be seen as giving (him) reasons for action. These two notions, the Humean one and the normative one, are quite distinct. And yet, there is a tendency not to keep them apart. It would have been much better if we had never heard of practical reasons in the first place. But it is quixotic to try to get rid of the notion, since it has a firm grip on our thought. However, if we are prepared to allow ourselves the use of the notion, then we had better make a clear definition of it. In the present book I will do so. There are indeed two separate notions of practical reason, one moral one, according to which reasons explain obligations (or permissions), and a Humean one, according to which reasons can explain (or cause) actions. To go for two different notions may seem to be to accept a complication. However, once these two notions are in place, it is possible to provide a very simple and general account of practical reasons, covering not only reasons to act, but also reasons to believe (epistemic justification) and reasons to feel and desire. This gain in generality makes the complication well worth its price. Moreover, once these notions are in place we become able to see through the talk of practical reasons: it is really a talk, either about our obligations and their moral explanation, or about our actions and their possible explanation. So in the end we gain also simplification. Some philosophers (moral nihilists) deny that there exist any moral reasons. They can accept my notions of moral and Humean reasons, but they must deny that there are any moral reasons. They hold an even simpler view than I do, then. However, as I will try to show, the ontological complication we countenance, when
In his preface to the second edition of his Methods of Ethics p. xii.
3
Preface
xi
we accept moral realism, is worth its price. We do have some moral obligations and there are moral explanations of these obligations. Or, so I tend to believe, at any rate, and I will indicate, in the book, what has brought me to this belief. Other philosophers (moral relativists) believe that there are several sources of normativity. They defend a view according to which there are, besides moral obligations, also obligations stemming from rationality, from justice, from aesthetics, from axiology, and what have you. I will reject this kind of moral relativism and claim that there exists exactly one source of normativity. Torbjörn Tännsjö Stockholm, Sweden
Contents
1 Introduction................................................................................................ 1.1 A War Tribunal................................................................................... 1.2 Reasons for Action.............................................................................. 1.3 Two Notions of Practical Reasons...................................................... 1.4 The State of the Art............................................................................. 1.5 Plan of the Book..................................................................................
1 1 3 3 5 7
2 The Humean Notion of Practical Reasons............................................... 2.1 Introduction......................................................................................... 2.2 Humean Reasons Are Concrete Entities............................................. 2.3 A Requirement of Rationality............................................................. 2.4 Are Humean Reasons Something We Argue from?............................ 2.5 Non-Humean Explanation of Action.................................................. 2.6 Mistaken Beliefs as Reasons............................................................... 2.7 Are Humean Reasons Always Instrumental?..................................... 2.8 The Temptation to Idealise.................................................................. 2.9 Conclusion..........................................................................................
11 11 12 14 16 17 18 22 23 25
3 The Moral Notion of Practical Reason..................................................... 3.1 Introduction......................................................................................... 3.2 Prima Facie Reasons........................................................................... 3.3 Are There Moral Reasons That We Cannot Know About?................ 3.4 Are Moral Reasons Always a Part of Humean Reasons?................... 3.5 On the Existence of Objective Moral Reasons................................... 3.6 The Principle of Charity...................................................................... 3.7 The Normative Constraint................................................................... 3.8 The Meaning of Moral Realism..........................................................
27 27 29 32 34 35 36 37 38
4 In Defence of Moral Realism.................................................................... 4.1 Introduction......................................................................................... 4.2 The Argument from Queerness........................................................... 4.3 The Argument from Relativity............................................................
41 41 43 50 xiii
xiv
Contents
4.4 Harman’s Empiricist Argument.......................................................... 4.5 Moral Knowledge............................................................................... 4.6 Conclusion..........................................................................................
52 55 58
5 Some Consequences of Moral Realism.................................................... 5.1 Introduction......................................................................................... 5.2 Moral Explanation.............................................................................. 5.3 Nihilism and Moral Explanation......................................................... 5.4 Quasi-realism...................................................................................... 5.5 Cornell Realism and Moral Explanation............................................. 5.6 The Normative Question..................................................................... 5.7 Broome’s Answers.............................................................................. 5.8 Korsgaard’s Answer............................................................................ 5.9 Conclusion..........................................................................................
59 59 59 60 62 64 67 68 70 74
6 Reasons from Prudence and Rationality................................................. 6.1 Introduction......................................................................................... 6.2 Rationality and Prudence.................................................................... 6.3 Person-Relative Practical Reasons...................................................... 6.4 Indeterminate Cases............................................................................ 6.5 Different Sense of ‘Ought’.................................................................. 6.6 Taking Relativism Seriously............................................................... 6.7 A Further Stretch of Our Moral Ontological Imagination.................. 6.8 Phronesis............................................................................................. 6.9 Conclusion..........................................................................................
75 75 76 80 82 84 85 88 90 94
7 Reasons from Justice and Aesthetics........................................................ 95 7.1 Introduction......................................................................................... 95 7.2 Theories of Justice.............................................................................. 95 7.3 Aesthetics............................................................................................ 100 7.4 Suppose I am Wrong........................................................................... 102 7.5 Conclusion.......................................................................................... 103 8 Reasons to Believe...................................................................................... 8.1 Introduction......................................................................................... 8.2 The Epistemic Goal............................................................................. 8.3 Coherence and Justification................................................................ 8.4 Is Epistemic Justification a Feasible Goal?........................................ 8.5 The Sui Generis Approach.................................................................. 8.6 Is Justification a Blessing?.................................................................. 8.7 Conclusion..........................................................................................
105 105 109 114 116 118 121 122
9 Reasons to Desire....................................................................................... 125 9.1 Introduction......................................................................................... 125 9.2 Fittingness........................................................................................... 125
Contents
xv
9.3 Morality Without Values................................................................. 9.4 Values Without Normative Import.................................................. 9.5 The Meaning of ‘Good’.................................................................. 9.6 Recent Attempts to Defend the Idea That There Are Reasons to Prefer or Desire...................................................... 9.7 The Simple View............................................................................. 9.8 Must We Will the Means to Our Ends............................................ 9.9 Broome on Practical Reasoning...................................................... 9.10 The Value of Prospects.................................................................... 9.11 Ought We to Avoid Cyclical Preferences?...................................... 9.12 Is There a Common Emotional Goal?............................................ 9.13 Conclusion......................................................................................
128 129 131
10 Conclusion................................................................................................ 10.1 Introduction..................................................................................... 10.2 Prudence, Rationality...................................................................... 10.3 Justice and Aesthetics..................................................................... 10.4 Epistemology.................................................................................. 10.5 What Are We to Believe?................................................................ 10.6 Reasons to Feel and Desire............................................................. 10.7 The War Tribunal – The True Story................................................
151 151 152 153 154 154 155 156
133 135 137 138 143 145 146 149
Bibliography..................................................................................................... 159 Index.................................................................................................................. 163
Chapter 1
Introduction
1.1 A War Tribunal The Judge: You admit that you drove the villagers into the camp? The Accused: I do. The Judge: Did you realise that they would most probably be killed in the camp? The Accused: I don’t think I considered that possibility. I didn’t think I had reasons to bother with that kind of thought. After all, it was none of my business. The Judge: But considering what you must have known about other camps, you must have had reasons to believe that the same thing might happen this time? The Accused: I suppose there were reasons for me to believe that they might get killed. But I also had reasons rather to think about other things. The Judge: If you realised what might happen to them, why didn’t you release them instead of driving them into the camp? The Accused: I’m not certain I realised that. Not at the time. Anyway, I had my reasons to do what I did. The Judge: What reasons? The Accused: Well, even though I wasn’t happy with the situation in my country, I thought there was little I could do about it. As you know, the Americans supported the government. And I knew that if I did not obey my orders, I might myself have had to face some sort of consequences. The Judge: Are these reasons really good ones? You could have saved several lives, and the cost to you would probably not have been devastating. I take it from what happened to others in your situation that, at worst, you would have lost your job.
T. Tännsjö, From Reasons to Norms, Library of Ethics and Applied Philosophy 22, DOI 10.1007/978-90-481-3285-0_1, © Springer Science + Business Media B.V. 2010
1
2
1 Introduction
The Accused: Why should I not have acted as I did? I think it was quite rational for me to do as I did. My job was dear to me. And more serious consequences, if I did not comply, were not totally out of the place. The Judge: Perhaps it was rational, in some sense, for you to do as you did. But certainly, you had moral reasons to abstain. The Accused: What reasons? The Judge: You could rather easily have saved the lives of the villagers. The Accused: You are the judge, you are on the winning side, and your morality prevails now, so I suppose you are right. I concede that I must have had moral reasons not to do what I did. The Judge: Then how come you did it. The Accused: As I said, I had my reasons; I did not want to sacrifice either me or my family. I thought it was only prudent of me to do what I did. Even if I had not lost my life if I had released those people, I would certainly have lost my job and my family would have suffered. The Judge: But you admit that from a moral point of view, you should not have driven the people into the camp. Morally speaking, their lives must certainly be more important than your job. The Accused: I suppose so. At least according to the morality that now prevails. The Judge: For certainly, if you can save innocent lives without making terribly severe sacrifices, you ought to do so! The Accused: From a moral point of view, I suppose that is what I ought to do. But from a legal point of view, I saw no reasons not to comply with the orders I was given. These orders accorded with the law at the time. And, as I have said, I thought it was only rational to comply. The Judge: But doesn’t moral reasons weigh heavier than prudential ones, do you think? Even heavier than legal ones? The Accused: Now and here they do, obviously. I didn’t think like that, then. Even if I had been aware that, from a moral point of view, what I did was wrong – but I’m not certain I saw things like that, then – I don’t think it fair to say that my action was irrational. I have explained to you why I did as I did. I had my reasons to do it. I’m not crazy. I’m no monster. I’m not a criminal. I didn’t kill anyone. There were reasons for me to do as I did. I had reasons to do what I did. The Judge: Would you be prepared to do it again? The Accused: I don’t know. If you had been in my shoes, do you think you would have acted any differently?
1.3 Two Notions of Practical Reasons
3
1.2 Reasons for Action The subject matter of this book is practical reasons, reasons for action. As is clear from the anecdote above, the word ‘reason’ can be used in bewildering many ways. My ambition is to clarify these uses, allowing us to understand more clearly what it is that is going on in a discussion such as the one between the Judge and the Accused in the anecdote above. Here ‘action’ is conceived of broadly, to include also belief and desire. So the subject matter of the book is reasons to do, believe and desire certain things.
1.3 Two Notions of Practical Reasons A central claim made in the book is that there exist two radically different notions of practical reasons. The word ‘reason’ is ambiguous between them. This ambiguity is one source of confusion in discussions where the term ‘reason’ appears, as in the anecdote above. These two notions of practical reasons are, on the one hand, something I will speak of as the ‘Humean’ notion of practical reasons, and, on the other hand, the moral or normative notion of practical reasons. To simplify, I will sometimes allow myself to speak of Humean and moral (normative) reasons, even though this is slightly misleading. We are not here dealing with two types of one and the same thing, to wit, practical reasons, but with two radically different notions of practical reasons. In the anecdote the Judge leans heavily towards the moral notion while the Accused tries to stick to the Humean one. On the Humean notion of practical reasons, practical reasons are desires and beliefs. Such desires and beliefs can cause actions. And our Humean reasons, or rather, the fact that we have them, can explain our actions. This kind of (action) explanation is special in that the content of our desires and beliefs, cited in the explanation, rationalises (by our own lights) our actions. A rational person acts on his or her Humean reasons. There are some ambiguities here to which I will return below in that we speak of desires and beliefs both as psychological events as well as psychological facts (our having them, or the events occurring). Our beliefs, moreover, have propositional content. Moral reasons are different. Or, to put it more correctly, on the moral notion of practical reasons, practical reasons are different. They are facts, explaining our obligations. I have moral reasons to perform a certain action only if I ought to perform it. The moral reasons are the facts, if such facts exist, that explain why I ought to perform it. A methodological claim made in the book is that no other notions of practical reasons than these two are needed if we want to account for all there is to account for in moral psychology and moral philosophy. I do not deny that there are many ways of drawing distinctions in this field and many different notions around, but
4
1 Introduction
each one of these other notions of practical reasons easily translates into one or the other of the two notions here characterised – or it is superfluous in the sense that there is noting in reality corresponding to it. This does not mean that there are no other uses of the word ‘reason,’ however. We may say that the reason a person wanted lunch at 4.00 am was that here circulation clock was disturbed.1 However, we do not count this fact as a practical reason. It is a mere explanation. The kind of explanations that rely on practical reasons could also be said to be part of ‘folk psychology’. The reference to a circulation clock goes beyond folk psychology. But, of course, when this person goes to lunch at 4.00 am there is probably also a folk psychological explanation available. He was hungry and he thought that an immediate lunch would satisfy his appetite. Sometimes the English word ‘reason’ means no more, no less, than ‘explanation.’ And this use of the term is beyond the scope of this book. Moreover, when someone asks me for my reasons for a certain belief, what I am supposed to present are the arguments rendering the content of my belief plausible or true. This is a notion of a reason, that is tantamount to an argument, and it too is beyond the scope of this book. My thesis, then, is partly methodological. My (positive) methodological claim is that the two notions of practical reason I will explicate, to wit, Humean and moral ones, are sufficient when we want to account for everything there is to say about practical reasons within normative ethics, moral psychology and even epistemology (with regard to justified belief). I don’t claim that, by ‘reasons’, people usually mean exactly what I mean by this term. However, when they use the word, if they use it in a way that makes sense, we can explain their use of it with the help of my terminology. Furthermore, when we acknowledge that there are two different notions available in the field, we have a better chance not to conflate things that should not be conflated. I will return to this point. Strictly speaking, however, no notion of a reason is necessary to our understanding of practical discourse. As a matter of fact, we can say anything there is to say within normative ethics, moral psychology, and epistemology, without having recourse to any notion of a reason at all. Had we not introduced the notion of practical reasons, I submit, we would probably have been better off in the sense that we would have been able to think more clearly about these problems! We can say everything we want to say in terms of desires and beliefs causing actions (and we can explain actions with reference to the fact that people hold the desires and beliefs they do) and we can say whatever we want to say about our obligations and their explanations without ever using the word ‘reason.’ However, since the use of the term ‘practical reason’ is well established, it is not a good idea to try to get rid of it altogether. A better idea would be to make an ‘explication’ of it (in Carnap’s sense of this word), and this is what I will attempt to do. It then transpires that what we need to have recourse to, what we need to explicate, are two different notions, the ones indicated above. As we will see, it is of utmost importance to keep them apart. The example was suggested to me by John Broome.
1
1.4 The State of the Art
5
However, when we accept to use the term ‘practical reason’ we should be careful not to allow the use of it to lure us into multiplying entities beyond what is, strictly speaking, necessary. And this leads me to my second claim in the book. This claim is a negative, ontological one. My ontological claim is that there exists only one source of normativity. We should reject the notion, entertained by the Accused in the anecdote, that there exist, besides moral reasons, reasons from prudence, rationality, justice, aesthetics and, what have you, capable of conflicting with morality. The ontological claim, that there exists only one source of normativity, has an important epistemic implication. If the claim is correct, then we cannot independently pursue the truth in different areas, asking what, from the point of view of prudence (all things considered) we ought to do, what, from the point of justice (all things considered) we ought to do, and so forth, in order eventually to face the problem what, tout court, it is we ought to do, where we have somehow to adopt a stance to all the putative and conflicting ‘oughts’ we may have arrived at. There is just one normative question for us to address. When we search for the sole (moral) normative truth, we have to take a stand to things like: How does the action affect me? In what sort of distributive pattern does it result? We have to assess whether the answers to these questions are relevant or not. The relevance, however, must be assessed with respect to one unique question: What ought to be done? I speak of this as the fundamental moral (normative) question. We cannot proceed by first asking what rationality (prudence) requires of us, then what justice requires of us, and so forth. There is only one kind of normative requirement there for us to consider. The claim that there exists one unique source of normativity is based on the idea that there exist obligations in the first place, and that they exist independently of our conceptualisation and beliefs (moral realism).
1.4 The State of the Art The literature in the field treated in this book is so extensive that it is beyond survey. Everything written, for and against, the Humean belief-desire model of the explanation of actions is relevant to the understanding of the Humean notion of practical reasons, of course. Perhaps Donald Davidson2 is the most important proponent of this view of practical reasons in modern times. I will here rely mainly on his early views on what has been called the causal theory of action. Many speak of this simple model of explanation, as I do, as ‘Humean’, but it is debatable whether Hume held it and, if he did, if he was original in this respect. According to John Cooper, it was Hobbes who originated the view.3 I will neither try to settle such scholarly matters nor go into all the subtle points in the recent discussion about the
Most famously in Essays on Actions and Events. Cf. his Reason and Emotion, p. 12.
2 3
6
1 Introduction
model as such. I will refer in my discussion to important contributions to this debate only when it serves my argumentative purposes to do so: I will certainly not try to do justice to the entire discussion. Some moral or normative notion of practical reasons is often vaguely alluded to in the discussion of practical reasons, but an exact form of this notion has been given only rather recently; when I worked on the notion I found that John Broome had given a similar definition that I do in this book. To my knowledge, this (his) definition is original. The definition makes the existence of moral (normative) reasons dependent on whether moral realism is a correct view or not. I defend moral realism. In particular, I scrutinise here what I find to be the strongest arguments against moral realism put forward in the recent discussion (the last 25 years, or so). Here John Mackie’s4 arguments from queerness and relativity are of importance, and so is Gilbert Harman’s empiricist argument for the same conclusion.5 I take issue with these arguments in Chapter 4. In Chapter 5 I discuss some important consequences of moral realism. My discussion about the various putative sources of normativity, which goes on through Chapters 6–9, leads me to the conclusion that there is just one source of normativity, and hence just one sort of objective (moral) reason. In this claim I differ from Broome, of course. We have a common understanding of the notion of an objective reason but differ in our views about what kind of reasons there are. My argument where I try to show that there is just one source of normativity is hampered by the fact that the views I attack are never worked out in detail by their advocates; they are merely suggested. It is said by many authors that there are many sources of normativity; there are reasons to do or not to do certain things that are stemming from rationality, or prudence, or justice, or aesthetics, or axiology, or what have you, but such claims have never been worked out clearly in any unequivocal manner. So I have to rely here on whatever I have been able to find in the literature, such as suggestions from Henry Sidgwick, Wayne Sumner, Derek Parfit, Wlodek Rabinowicz, and others. My discussion on the putative existence of norms stemming from justice draws on what philosophers such as Rawls and Temkin have had to say recently about these things. And my discussion about aesthetics as a source of normativity takes as a point of departure what G.E. Moore had to say about the value of beauty (with a comparison with the view of recent ecological ideas to the same effect). But since the discussion of these various different putative sources of normativity is not much developed, I feel that I have to break new grounds here through all these chapters. In my discussion on epistemology I take advantage of an ongoing discussion about how epistemology could get ‘naturalised’. Now, epistemology naturalised, in the hands of a philosophers such as Quine, is an entire philosophical program with many tenets. My interest is more restricted. I restrict myself to the claim that the notion of epistemic justification is not a normative (not an evaluative) one. Some philosophers claim is that, to be ‘justified’ in holding a certain belief, merely means having arrived at it through reliable methods, having it cohere with In his Ethics. Inventing Right and Wrong. Put forward in The Nature of Morality. An Introduction to Ethics.
4 5
1.5 Plan of the Book
7
the rest of ones beliefs, and so forth. On such an understanding of epistemic justification, there is no temptation to think of epistemology as a source of normativity at all. I am not happy with this idea, however. It is more natural, I think, to see reliability and coherence as criteria of justification, not as something that gives the meaning of the term. If this is correct, does it not then mean that justification is normative, after all? No, it does not. There exists an interesting pragmatist tradition, dating back at least to William James, and developed clearly in present times by Isaac Levi, where epistemic justification is conceived of in Humean terms.6 The question what it means to be justified in holding a belief is answered with reference to a notion of an epistemic goal, turning epistemic justification into a kind of instrumental (Humean) justification. I avail myself of the resources of this tradition in my discussion in Chapter 8. In my discussion about epistemic justification it transpires that, even if there is no epistemic source of normativity in its own right, since epistemic justification is instrumental, we all happen to share, for evolutionary reasons, a common epistemic goal. This means that it is possible to rationalise our beliefs in relation to this goal. With respect to our emotions, the situation is very different. There exists no common emotional goal. Some philosophers believe that if something is valuable, then we ought not only to produce it, but also to take up a special stance towards it, we ought to desire it. This idea has been common within aesthetics; it was held, for example, by Immanuel Kant in his Critique of Judgement and it is defended also by G.E. Moore (in addition to his idea that beauty has value in itself). This idea has also been taken up in ethics where the notion of intrinsic or final value has been analysed in terms of fitting emotions, and it dates back at least to Brentano and his The Origin of Our Knowledge of Right and Wrong. It was elaborated upon by Ewing in The Definition of Good and, in recent times, it has been defended by Derek Parfit and Wlodek Rabinowicz. Their defence of this idea triggered my work on this book in the first place. In Chapter 9 I take up, examine and reject this idea. Values do not constitute any special source of normativity. In fact, there are no problems in their own right to do with intrinsic or final value. There is no such subject as axiology proper.
1.5 Plan of the Book I begin my study of practical reasons by, in Chapter 2, explaining what I mean by a ‘Humean’ notion of practical reason. I go on in Chapter 3 to define the moral (normative) notion of practical reasons and to a discussion about the existence of such reasons. Even if there are some people who deny that moral (normative) reasons exist, it is hard to do moral philosophy without having recourse to them – or, rather,
See his Gambling with Truth and The Enterprise of Knowledge.
6
8
1 Introduction
to objective categorical norms and their explanations. In Chapter 4 I examine the best arguments I know of in defence of the claim that independently of our thought no moral (normative) reasons exist and, tentatively, I conclude that such reasons do exist. In Chapter 5 I go on to discuss some implications of moral realism. The most important consequence is, of course, that we are allowed to assume that objective moral reasons exist. We can assume that particular actions are right and wrong and that there exist moral explanations of their rightness and wrongness. It is shown that a nihilist doesn’t have access to the notion of a moral explanation. Not even a moral particularist, who is a nihilist with respect to moral principles, has access to a coherent notion of an objective moral reason, or so I argue, at any rate. Another claim made in the same chapter is that the same goes for moral naturalism of various different kinds; the moral naturalist cannot make sense of a moral explanation. In this chapter it is also contended that the approach here defended allows us to answer what Christine Korsgaard has called the normative question. As was indicated above, many practical philosophers seem to believe that there are many sources of normativity, and corresponding types of practical reasons, such as norms of, or reasons stemming from, prudence, rationality, justice, aesthetics, epistemology, and the idea of a suitable and fitting emotion – as well as from morality. If these philosophers are right, we may have to acknowledge that, in a situation, we face conflicting obligations or, as Sidgwick once put it, a ‘dualism’ of practical reason. This phenomenon, which could properly be called a kind of practical ‘relativism’, is the main target of my argument in Chapters 6–9. In Chapter 6, I argue against the idea that prudence or rationality provides us with a source of normativity, in Chapter 7, I argue against the view that justice or aesthetics does so, in Chapter 8, I argue against the view that epistemology does so, and in Chapter 9, I argue against the claim that there are special obligations per se to feel and desire certain (valuable) things. Here I put forward and defend a strong version of what has been known as the ‘buck passing theory of value’. In fact, I argue that it would be a good thing to get rid of all sort of talk about ‘intrinsic,’ or ‘final,’ value. To the extent that such talk makes sense, and in many cases it does, of course, it can be reduced to talk about our obligation to do so and so. However, there are no obligations, as such, to feel one thing or the other. We ought to adopt certain feelings, in certain situations, but the answer to when and why we ought do so, belongs to normative ethics proper. If utilitarianism is on the right track, for example, something I believe to be a fact but for which I do not argue in this book, then we ought to do so when by so doing we see to it that the sum-total of wellbeing in the universe is maximised. The main thrust of the argument in all these chapters, then, is that there is only one source of normativity. And theories such as utilitarianism, ethical egoism, Kantian deontology, and Lockean moral rights theory are in one and the same competition for a systematic answer to it. Now, since my ontological thesis is partly a negative one (there is at most one source of obligation), in order to render it plausible, I have to consider various different putative kinds of practical normative reasons and try to show that, upon
1.5 Plan of the Book
9
closer inspection, they turn out to stem from one and the same source. This is exactly what I will do in these chapters The discussion in these chapters proceeds on the positive ontological assumption that there are objective normative reasons, existing independently of our conceptualisation. And the claim is that there exists only one source of normativity. There is, in each situation, one, and only one, truth about what we ought to do, and our moral (normative) reasons explain this obligation. To the extent that we seem to be faced with other practical reasons to do or feel or believe these other practical reasons are Humean in nature. This means that they lack normative force altogether. They can explain why we do, or feel, or believe as we do, but they cannot explain why we ought to do so. The book ends with a short Chapter 10 where my argument is concluded and the anecdote about the War Tribunal (above) is retold in clear terms, without any mentioning of any reasons whatever.
Chapter 2
The Humean Notion of Practical Reasons
2.1 Introduction On the ‘Humean’ notion of practical reasons, practical reasons are beliefs and desires a person has when acting, which could cause the action; the fact that the person has them, if they cause the action, can both rationalise his or her action (by her own lights), and explain it. I have practical reasons of the Humean kind to write this book. I have a feeling that I have a point to make, I want to make it, and I believe the most efficient way of making it would be to write this book. This is why I write it. When I call such reasons ‘Humean’ this has to do with the fact that my notion accords with a standard interpretation of how Hume conceived of explanations of actions. It is noteworthy, however, that Hume himself would not call desires (or ‘passions’) reasons. While he would agree, I suppose, that both desires and beliefs cause actions (bearing in mind his special view of causality) he would not call the desire a ‘reason’. There is a point in including desires (passions) among reasons for action, however. When we ask for the reasons for an action – or for the explanation of it – we tend to be selective. There is a pragmatic aspect of the question. We tend to stress some among all the factors that contributed to the action, leaving others in the background as mere ‘enabling conditions’ (to borrow a phrase from Jonathan Dancy). Sometimes these aspects that we stress are certain beliefs. ‘Why did you drive the villages into the camp?’, we ask. And the answer: ‘I thought it would help me to keep my job’ gives a plausible answer, describing a reason for the action in question. Note that the (content of the) same reason could have been expressed by the person who answered the question: He or she would then have answered: ‘In order to keep my job.’ In other situations we mention a preference or a passion or motive, when we are asked for reasons. ‘Why did you abuse this child sexually?’ It makes perfectly good psychological sense to answer: ‘I have a sexual appetite for young children.’ It is a sad fact about our language, however, that it hardly allows us any possibility to express our feelings or desires in the way we express the content of our beliefs. To the extent that we express feelings or desires, we do so by having recourse to seemingly autobiographical language, intended in the context, not to report, but to express our state of mind: “I’m pleased to hear that.”
T. Tännsjö, From Reasons to Norms, Library of Ethics and Applied Philosophy 22, DOI 10.1007/978-90-481-3285-0_2, © Springer Science + Business Media B.V. 2010
11
12
2 The Humean Notion of Practical Reasons
2.2 Humean Reasons Are Concrete Entities I prefer to speak of Humean reasons as concrete mental entities, that is beliefs and desires are things that obtain at a certain time, ‘in’ a certain person. These beliefs and desires are capable of causing action. But we may also speak, rather than in terms of causation, in terms of the explanation of actions. If we do, then we rather say that it is the fact that a person holds certain beliefs and desires that explains this person’s action. The explanation is still causal, I submit, but it would be odd to say that the explanans is a cause; the cause is the events described in the explanans. Donald Davidson has brought to our attention deviant causal chains as a problem for causal theories of action. It seems as though not all events that can be explained with reference to beliefs and desire, which cause and rationalise them, are actions. And yet he seems to have been unperturbed by the seeming impossibility of solving the problem of sorting out which causal chains are in order and which are deviant. Why was he so unperturbed? It is too late to ask him, but this is how he could have argued: It is not meaningful to seek a general criterion, which can tell the difference between actions and mere events in the abstract. Each event type requires its own answer to the question what kind of causal relations are the right ones, and what kind are ‘deviant’. Consider Davidson’s own example: A climber might want to rid himself of the weight and danger of holding another man on a rope, and he might know that by loosening his hold on the rope he could rid himself of the weight and danger. This belief and want might so unnerve him as to cause him to loosen his hold, and yet it might be the case that he never chose to loosen his hold, nor did he do it intentionally.1
The loosening of the hold on the rope is an action if we can describe it (subsume it under an action type) in a way that renders it intentional under that description. Is there a way of describing the climber’s loosening of his hold of the rope as intentional? Did he intentionally loosen his hold? It seems not. For all we are told, he may have wanted his hands not let go of the rope. And even if this was not so, even if he held a secret desire that his hands should let go, this desire was not, in the circumstances, causally efficacious. His hands would have loosened their grip, even if he had wanted to hold on to the rope. In relation to this kind of act we do require control of the agent’s hands. With respect to other acts, such as crossing the street, it is not necessary that the agent wills every step he or she takes, or that the agent’s will with respect to the exact movements of the legs is at all efficacious. The sense morale, then, is that we need no general solution to the problem what distinguishes actions from mere events; moreover, in relation to specified action types, in a context, we can tell which causal relations are deviant and which are in order. If this were a treatise in action theory there would be more to be said on this subject, but I leave it with this observation.2 Action and Events, p. 79. The literature on this problem is so extensive that the best reference to give here must be to the relevant entry in The Stanford Encyclopedia of Philosophy: George Wilson, ‘Action’. See also my ‘On deviant causal chains – no need for a general criterion’.
1 2
2.2 Humean Reasons Are Concrete Entities
13
Need we insist on speaking of human reasons as beliefs and desires. Could we not instead speak of psychological facts, constituting the explanans, as Humean reasons? I have no objection to this manner of speaking. It works just as well as the one I have adopted. I may even slip into it now and then. But note that it makes no difference from the point of view of ontology, which way we decide to talk. Once this remark has been made, there is no room for misunderstanding. My standard way of putting things is as follows: Humean reasons can cause actions, and the fact that a person has Humean reasons to perform a certain action may explain (causally) that he or she performs it. It is not controversial that such (subjective) reasons for action exist, although many details in the picture here adumbrated are disputed. Many will concede that the fact that we have Humean reasons explains some of our actions, but some of them will say that even if the fact that we have these reasons explains our action, these reasons do not cause our action; these people will deny that it is possible to give causal explanations of actions. Other will say, and I agree, that explanations of action with reference to the fact that a person has certain reasons is a causal one, or no explanation at all. And while some may want to argue that all our actions can be explained with reference to Humean reasons (including both a description of desires and beliefs held by the agent) others will claim that some actions can be explained with reference only to the beliefs of the agent, leaving the desires or the necessary motivation in the background (as an ‘enabling condition’). This may be a reasonable way of speaking in some situations, where it is obvious which desire was operative, but it is only reasonable if we grant that a full (complete) explanation of the action must include a reference to the desire as well as the belief (and to the assumption that the agent was rational). It is finally a moot question if the beliefs and desires of the agent must be manifest at the time when the action to be explained is performed, or whether it is sufficient that they are there as mere dispositions, and so forth. Hume is sometimes taken to have stressed a certain asymmetry between desires and beliefs. While the former are active in generating an action, the latter are inert. While the desire constitutes a force, the role of the belief is to ‘channel’ this force, and so forth. I don’t know whether this metaphor really represents Hume’s intentions correctly.3 I doubt that it does. Be that as it may, I think this metaphor distorts the true picture. Both desires and beliefs are important to the explanation of actions. They act together, to produce actions. Together they are active, in isolation they are each inert. This fact is consistent, of course, with the standard observation that there is a difference with regard to the ‘direction of fit’ between on the one hand beliefs and the world and, on the other hand, desires and the world. Or, a caveat may be in place here. Even if it is true of many actions that they are caused by a belief and a desire operating together, this need not be true of all actions. There may exist exceptions from this in so-called basic actions, where we For a discussion, see Elizabeth Radcliffe, ‘Hume on the Generation of Motives: Why Beliefs Alone Never Motivate’.
3
14
2 The Humean Notion of Practical Reasons
just desire something and hence do it (such as raising our arm). We can perhaps rationalise and explain such an action with reference only to a desire in the situation on the part of the agent that his or her arm be raised.4 A further controversy pertains to the idea that, by the lights of the agent, a Humean explanation of an action ‘rationalises’ the action. This must not be understood to mean that the presence of Humean reasons somehow shows that the agent did the right thing – not even by the agent’s own lights. However, the idea that a Humean explanation rationalises an action is certainly true in the sense that, if the action can be thus explained, then the agent has been rational in performing it – in a purely descriptive sense of the word ‘rational’ to be explained below. I will not go any further into any of these complicated matters right now.5 Here it must suffice to note that most people, who have thought hard about matters such as these, seem to agree that at least some actions can be explained with reference to the desires and beliefs of the agent. To explain the action, and to rationalise the action (by the agent’s own lights), come to one and the same thing. On my own favoured description, we can sometimes explain why a certain action came about with reference to the reasons held by the agent. On this description, the agent’s (relevant) beliefs and desires (reasons) are what caused the action.
2.3 A Requirement of Rationality Do we also have to include in the explanation of the action, given with reference to a belief and a desire, the premise that the person who performed the action was, in the relevant sense, ‘rational’? This addition is crucial. If we wish, we may of course follow Davidson in his idea that by imputing beliefs and desires to a person, we also implicitly impute this kind of rationality to this person.6 As a matter of fact, it is hard to make sense of a person who holds beliefs and desires, but who doesn’t, in most (normal) cases, exhibit this kind of rationality. To have beliefs and desires is to be, at least somewhat, rational. However, this means precisely that the rationality assumption is crucial to the explanation. So why not make it explicitly? This means that a Humean explanation takes roughly the following form:
For a defence of the view that we can sometimes explain action without recourse to beliefs, see Alfred Mele, ‘Effective Reasons and Intrinsically Motivated Actions’. 5 A recent critical discussion and defence of the Humean view of the explanation of actions can be found in Björn Petersson, Belief & Desire. 6 See his Essays on Actions and Events, pp. 273–4. A special trait with Davidson’s view is that he thinks that the relevant rationality assumption is true merely of a single individual, not of people in general. But, certainly, he must think that, anyone who is similar to a certain person in relevant respects, would act in relevantly similar ways. 4
2.3 A Requirement of Rationality
15
1. P’s strongest desire (with respect to the alternatives P believes are open to him or her in the situation) is to F 2. P believes that, in order for P to F, it is necessary for P to G 3. P is ‘rational’ in the sense that P acts on what he or she recognises as his or her strongest desire 4. Therefore, P does G This is merely a sketch. We may want to add further premises or qualifications to the ones mentioned. What are we for example to think of hidden desires? Should we include a clause that P is aware of (1) to get rid of such desires? Personally, I think we should do so. I may act on some repressed and sub-conscious desire, and feel the action I perform is not genuinely mine. Then I would hesitate to say that my action is rational, at any rate. And I would hesitate even more to say that, if I happen not to act on this repressed and sub-conscious desire, I am irrational. Yet, others may think otherwise and, if by this kind of explanation we want to rationalise actions by the lights of their agents, we should, in each explanation, rely on the notion of rationality favoured by the agent him- or herself. I take it that, with respect to rationality, different agents may have different ideas of what kind of person they want to be. Is it at all possible to act on a desire which is not one’s strongest one (in the situation)? Some may find this contention strange. They may want to argue that, if we act on a certain preference, then we show that it is the strongest one. But this move is awkward. It seems to rule out the possibility of weakness of will (akrasia). However, weakness of will is a well-known phenomenon that should not be ‘explained away’. What typically happens, when I act on a desire that is not, in the situation (with respect to the alternatives I believe are open to me) the strongest one, is that I get distracted. I have, for example, some very strong desire which in the situation (with respect to the alternatives open to me right now) is irrelevant, but which I cannot put to one side. And the result is that I make a bad choice in the present situation. I fail to concentrate on the present choice that is before me. Instead I contemplate another choice to be made later on. In the situation I fail to be rational. Sometimes conflicts between what we do, because we want to do it, and what we believe we ought to do, are also seen as examples of weakness of will (akrasia). Here ‘weakness of will’ is a misnomer, however. In these cases the agent does indeed act on his or her strongest desire, which happens to be at variance with what the agent conceives of as his or her moral duty. This person is rational and the action this person performs can indeed be explained in an ordinary action explanation. I return to a discussion about the relation between beliefs about obligation and the forming of intentions in Chapter 6, where I argue in defence of the position here just stated. Note that the notion of rationality employed here is ‘natural.’ No implicit evaluation is made. We only assume that a certain pattern is exhibited in P’s actions or, better put, we assume that there exists a kind of ‘fit’ between P’s beliefs, desires and actions.7 7 In my treatment here of the notion I stipulate that the notion is non-normative. For a recent and elaborate defence of the view that rationality should be naturalised, see Niko Kolodny, ‘Why Be Rational?’.
16
2 The Humean Notion of Practical Reasons
In a simple case it is fairly straightforward to say what it means for a person to be ‘rational.’ He wants something, there is just one set of means available to him, ascertaining that he gets what he wants, he adopts these means and obtains what he wants. He is rational. But the situation may be more complicated. The agent may have to act under risk or uncertainty. Now the literature abounds with suggestions as to what a ‘rational’ agent would do. However, these are just different stipulations. Given one notion in particular (to be rational when acting under risk, say, is to maximise expected utility) we can say whether a certain agent is or is not rational in his or her actions. There is no need here to go into all these complications. An agent can opt for one notion in particular (he or she may decide that this is the kind of rationality he or she wants to exhibit) and, given this notion, we can say of his or her actions whether they are rational or not. If we want to rationalise actions by the lights of their agents, this is how we should proceed. Moreover, given this notion, we can say which action this person has Humean reasons in a situation to perform. Does the fact that this kind of explanation can be made also imply that the agent (P) has conducted (at least implicitly) a piece of valid practical inference (from the expression of relevant intentions and beliefs, to the intention to perform the relevant action)? No, it does not. A person who intends to F, believes that, unless she will do G, she will not do F, but doesn’t intend to do G, is ‘irrational’, in a purely descriptive sense of the word, and this person will probably perform actions that cannot be given any simple Humean explanation. Yet, for all that, this person need not have made any logical mistake. Or, at least, this is something I will try to show in Chapter 7. A thesis put forward there is that there is no such thing as valid practical reasoning. There is no such thing as a valid practical syllogism.
2.4 Are Humean Reasons Something We Argue from? Some may want to complain that reasons are something we reason from. But this is not true of Humean reasons. Judith Jarvis Thomson in particular makes this complaint. This [a Humean account of reasons] won’t do.… Following Hume, what explains a person’s doing a thing is his wanting this and believing that. But the fact that he wants this and believes that is not likely to be his reasons for doing the thing.8
I agree that, typically, we do not argue from our Humean reasons. First of all, this is true on trivial grounds. Humean reasons are psychological entities, not propositions (facts). However, even if we identify Humean reasons with the fact that we have certain desires and beliefs, rather than with these desires and beliefs themselves, it is still true that we rarely argue from them. Is this a problem for my account? I think not. I admit that there are reasons also in the sense Thomson mentions. As a matter of fact, we see here that two notions Goodness and Advice, p. 31.
8
2.5 Non-Humean Explanation of Action
17
of practical reasons are really needed. One notion, the moral one to be defined in the next chapter, does satisfy Thomson’s requirement. The other (Humean) one doesn’t. But this is how it should be. There is no notion that can satisfy all the desiderata that exist in this field. Moral reasons, we shall see, are propositions (true propositions, as Thomson insists that practical reasons should be). Moreover, such (moral) reasons do function as premises of an argument.9 We can argue from them, once we have grasped them. But Humean reasons are different. The presence of Humean reasons can explain actions and rationalise them by the lights of the agent. This is an entirely different notion of practical reasons.
2.5 Non-Humean Explanation of Action Not all actions can be explained in this way – or, at least, not all explanations of actions take this form. Because of weakness of will, as was noted above, a person may come to act, not on his or her strongest desire, but on some other desire. If this is the case, there is no way to give a Humean explanation of her action. This also means that, when a person has reasons to do something, but does not do it, this person is, in a sense we are familiar with from standard causal theory of action, ‘irrational.’ There is a lack of fit between what he or she desires, believes, and does. But note once again that this use of the term ‘rational’ is purely descriptive. Note also that such lack of fit between beliefs, desires and actions may be due to actual difficulties; we often go wrong when we want to put together our various different beliefs and desires in the appropriate way; we fail to keep track of our beliefs, for example, when things get complicated. There is an abundance of empirical evidence to this effect. The very fact that, from our own personal perspective, a certain alternative dominates all other alternatives, may go unnoticed.10 If an action is not rational, does this also mean that it is altogether inexplicable? This is the claim Judith Jarvis Thomson makes: Suppose we accept the.… Humean account of explanation of action … Thus suppose we accept that what a person will in fact do is what he believes will most efficiently satisfy his wants. Suppose, last, that Alfred’s wants are appropriately restricted, and that he believes that refraining from paying his grocer’s bill will most efficiently satisfy them. The friend of the familiar theory of rationality says that for Alfred to pay his grocer’s bill would be irrational. But that can’t be right if Hume is right. For if Hume is right, then Alfred’s paying his grocer’s bill would not be irrational. It would be unintelligible. It would be inexplicable.11 9 It is not quite clear, on Thomson’s account, however, what we argue to, when we give reasons for actions. For, certainly, what we argue to cannot be an action. On my account of moral reasons, what we argue to (what we explain), once we have grasped the reasons, is a normative fact. Perhaps the best way of making sense of Thomson is to take her to claim that what we argue to, once we have grasped the reasons, is an evaluative fact, the fact that a certain action would be good in some way in particular. For, on Thomson’s view, we have a reason to perform an action if our performing it would be good in some way. 10 Cf., for example, Daniel Kahneman and Amos Tversky, ‘Choices, Values, and Frames’ about this. 11 Judith Jarvis Thomson, Goodness and Advice, ibid. pp. 79–80.
18
2 The Humean Notion of Practical Reasons
But this is certainly not correct. Alfred’s action would indeed be irrational. And this means that it would not be open to Humean explanation. And this is what it means for it to be irrational. This doesn’t mean that it would be inexplicable. There are other ways of explaining actions. If an action is irrational, and if we want to explain it, then we must explain it in a non-Humean way. If it is not rational, we cannot rationalise it (not even by the agent’s own lights). We say then that a certain person did as he did because he was drunk, or we say that he did as he did because he had forgotten about a certain promise he had given, and so forth.12 The number of such possible explanations of action is legion. We are perfectly familiar with these sorts of explanation. But this means, of course, that the claim that ‘what a person will in fact do is what he believes will most efficiently satisfy his wants’ is not true without exceptions. In particular, it is not true of all counterfactual – merely possible – actions. When the action is not rational, we may want to know not only why the agent did not act on his or her strongest desire, or neglected some piece of information actually available to the person, but also on what other desires and beliefs it was that the agent acted. This means that we what to know what desires and beliefs were causally efficacious in the case under scrutiny. However, in many cases, people do act on their strongest desires. They possess the relevant factual knowledge, that renders this possible for them, and they exhibit a usual amount of strength of will, allowing them to do so. Then their actions are open to a Humean explanation. And the resources used in such an explanation are what are needed for the explication of the notion of a Humean reason. For, of course, even a person who doesn’t act on his or her strongest reasons (desires) has reasons to do so. Unless he or she does act on them, he or she is simply irrational. But the reasons here referred to are Humean ones. Such reasons generate no categorical norms in their own right. The notion of rationality, as here used, is, it should once again be remembered, purely descriptive. It signifies no more, no less, than a certain ‘fit’ or ‘pattern’ between desires, beliefs, and actions. It is no coincidence, however, that this fit is often there. Evolution has sought for it to be there. People, who are not by and large ‘rational’ in this sense of the word, are not very good at propagating their genes. But this is a biological, not a moral, fact.
2.6 Mistaken Beliefs as Reasons What are we to say of ignorance? What are we to say of a person who wants to get safely home, believes that a good idea is to go by car, but will, if he or she does, get killed in an accident? Does not the fact that this person will be killed constitute a Humean reason not to go by car?
The example is given in S. Darwall, Impartial Reasons, p. 29.
12
2.6 Mistaken Beliefs as Reasons
19
We may say so if we like. But it is far more appropriate, I think, to say that, if this person came to know about this fact, his or her knowledge (belief) would constitute a reason for this person not to travel by car. In the circumstances, where the person does not know this, the fact is no (Humean) reason whatever for action for this person. This is how I prefer to put things, at any rate. And this is consistent with this person having a moral reason not to take the car. I return to this kind of reasons in the next chapter. But those who want to say that a person has a reason not to travel by car in the circumstances are free to say so. And it is easy to make sense of their use of the word ‘reason’ – in terms of the notion here characterised. According to them, a person has a ‘reason’ to F if, had this person possessed all the relevant knowledge of his situation, he would have had a Humean reason to F. Their talk of reasons, then, easily translates into mine. What about seriously false but deeply entrenched beliefs, then? What about beliefs generated in an unreliable manner, such as, for examples, beliefs resulting from psychosis? Do they really constitute Humean reasons for action? What are we to say of a person who (falsely) believes that if he enters the street he will be killed by snipers? Does he really have a Humean reason not to enter the street? This is what I am prepared to say, with some authority, having a child who, for a while, because he was psychotic, was in this situation, but some may not want to concur in this use of the term ‘reason.’ Some may want to say that he has no reason to stay at home. I would rather say that, indeed he has. Even the actions of a psychotic person can to some extent be rationalised. And according to the notion here defined, he does have a Humean reason to stay at home. I am happy to acknowledge this conclusion. But it may well be the case that this person has also got a good Humean reason to give up this belief. He may know that the belief has been generated in a not reliable manner. Then he has reasons to give it up (to the extent that he wants to have correct beliefs about the world). However, the belief may be difficult to give up. He may be in the grips of his psychosis, and it may even take medication to get rid of it. If it does, and if he realises this, which is possible, even in these circumstances, he has Humean reasons to take his medicine. Yet, for all that, while the belief is still there, he has Humean reasons not to enter the street. Or so I prefer to put matters. Those who deny that this person has a ‘reason’ to stay home may go on using the word as they see fit. However, it is not difficult to see that their way of using the word ‘reason’ can be easily explained with reference to the notion of a Humean reason, as here defined. When they say that this person has a ‘reason’ to enter the street, they mean that, had this person had correct beliefs about the situation, he would have had what I have called a Humean reason to enter the street. However, even if this is no big issue, I think my notion is in many ways preferable to theirs. In particular, given the notion of a Humean reason, we can explain (rationalise) this man’s reluctance to enter the street with reference to this belief. And this is the belief he would himself cite, if we ask why he does not enter the street. There may even be a therapeutic point in using the word ‘reason’ in the way I do.
20
2 The Humean Notion of Practical Reasons
For, with reference to his belief, we can, with Donald Davidson’s words, understand this man in his ‘role of Rational Animal’: In the light of a primary reason … the agent is shown in his role of Rational Animal.… Thus there is a certain irreducible – though somewhat anaemic – sense in which every rationalization justifies: from the agent’s point of view there was, when he acted, something to be said for the action.13
It should be obvious by now that, in the present context, a Humean reason is taken to be a desire or a belief or a combination of a desire and a belief. This means that reasons are psychological entities. It might be objected that this is not true of all reasons. It makes perfectly good sense to say that the reason I was late was that I held the belief that the train should leave at 10 a.m. when, in fact, it did leave at 9.30 a.m. But it does also make sense to say that the reason I was late was that my car broke down, so I had to catch a train. Now, these kinds of reason seem to be propositions or facts. Does this mean that there are two kinds of reasons, some being concrete and psychological in nature, and others being abstract? We may say so if we like. However, the choice is merely a matter of terminology. When we say that the fact that my car broke down was a reason for me to catch the train, we can understand this in my terms. This fact was a reason for me since my acknowledging it made me take the train. Had I not held the belief that the car was broken I would not have tried to catch the train. I call this belief my reason. Others want to say that the content of the belief (the proposition) is the reason – irrespective of whether it is true or false. Others yet want to say that it is the fact (i.e. only the true proposition held by me), which constitutes my reason. Furthermore, as was noticed above, some want to say that the fact that I believed the care was broken, was my reason. But we can easily translate between these terminologies. The crucial thing is that, even those who want to say that reasons are abstract propositions or facts have to suppose that the agent holds the relevant proposition, or acknowledges the fact in question, in order to be able to provide a Humean explanation of any action of the agent. Typically, they say that, even if the reason as such is a proposition or a fact, the agent’s having it consists in the agent’s believing in the relevant proposition. If we want, we may introduce a special category to account for the propositional content of what I have here called Humean reasons. We may speak of them as ‘deliberative’ reasons. Since this means no addition to the ontological claims made in the present context I have no objection.14 However, the term ‘deliberative’ may be misleading. It would be better to speak of the content of those of our beliefs that constitute practical reasons, as ‘rationalising’ reasons. Certainly, when I perform an action, this has sometimes been preceded by deliberation. But to deliberate is to perform an action as well. So this action too should be amenable to a Humean explanation. Essays on Actions and Events, p. 9. For an elaboration of this idea, see Olson and Svensson, ‘Regimenting Reasons’.
13 14
2.6 Mistaken Beliefs as Reasons
21
However, deliberation is special in that it means that I entertain propositions and make inferences. Are these propositions practical reasons? I think not. I have already put to one side propositions intended in general to prove (the truth) of other propositions as something that is beyond the scope of this book. We sometimes speak of arguments, and pieces of evidence, as ‘reasons’ but this has noting to do with practical reasons – unless they take the form of a moral reason to be explained more thoroughly in the next chapter. The arguments to which we here refer are typically ordinary naturalistic arguments. Think of Hitler contemplating whether he should attack France as a way of opening World Ware II. He makes the following inference. If I attack France the UK will declare war. If the UK declares war, I may loose the war. Now, since he wants to win the war, he comes to reject the idea of a sudden attack on France. In the false belief that he will get away with it he opts instead for Poland. There is some deliberation going on here, obviously, but the inferences imputed to Hitler are ordinary factual ones. Note that his new desire, to attack Poland, is a causal result of his deliberation. However, neither the premises of his reasoning, nor the inference as such, is part of his Humean reasons when he attacks Poland. He does so because he wants to win the war and because he believes this is the best way of achieving his goal, period. Besides this we have moral arguments as well, I submit. They do belong to the topic of this book, but they will be discussed in the next chapter. It might be thought that, by saying that the reason that a person did something was that he thought so and so, we do not always do justice to the first-person perspective. If a person claims that she broke up with her boy friend because he cheated on her, this person would certainly be offended by the comment: ‘I see, you broke up with him because you thought he cheated on you.’ But the reason that this is offensive is that the person would not herself put the explanation like this. If she did, it would sound as though she wasn’t sure. Perhaps he didn’t cheat on her after all. Yet, when she says, ‘I broke up with him because he cheated on me’ she expresses (the content of) her belief that he cheated on her. So she is really using (the content of) her belief in giving her own explanation of her own action, I would say. While, if she had said: ‘I broke with him because I thought he was cheating on me’, she would have expressed (the content of) another belief, as it were, a belief she did not hold, to wit, the belief that, perhaps he wasn’t cheating on her after all. The agent, then, has normally access to two ways of explaining his or her actions, either by describing his or her relevant mental states, or by expressing (the content of) them. From a third person perspective, however, there exists only one way of representing the relevant beliefs and desires, to wit, by describing them. This does not mean that those who want to say that propositions or facts are reasons cannot say so, but it shows that there is no problem in using the terminology here defined. So, even if the ontological status of reasons has been discussed ad nauseam, the question of whether (Humean) reasons are psychological or abstract entities is not a genuine one. The question is terminological and it can be settled either way through an act of fiat.
22
2 The Humean Notion of Practical Reasons
Furthermore, to the extent that we are prepared to doubt whether abstract entities like propositions and facts really form a part of the fabric of the universe, while we are prepared to accept that there are concrete events such as beliefs, desires, and actions, abiding by this act of fiat gives us the right to the claim that there are Humean reasons.
2.7 Are Humean Reasons Always Instrumental? Here is a final terminological remark on Humean reasons. Humean reasons are sometimes referred to as ‘instrumental.’ I’m not entirely happy with this way of putting things even if, sometimes, I will allow myself to do so. The reason I am not entirely happy with this way of putting things is that, to speak of Humean reasons as ‘instrumental’ is to invite misunderstanding. If we speak of Humean reasons as ‘instrumental’ we give the impression that, whenever there is intentional action, there is a kind of planning going on, a hidden agenda and a further purpose, of which we may, in certain circumstances, be suspicious. We may feel that, in the circumstances, people should allow themselves to act in a spontaneous manner and not to plan. Those who make plans while making love, say, may not be loveable people. But the Humean model doesn’t say that there must be any further point of doing what, as a matter of fact, we are doing. If I am whistling, say, I may be whistling just for the fun of it; and we may give a Humean explanation of my action in terms of my desire for whistling, period. I form the lips in the way I do, I blow wind between them in the way I do, just for the fun of it, just because I like to whistle. So even if, in a sense, Humean reasons are instrumental reasons, they are ‘instrumental’ only in a very weak sense. There need not be any further purpose, for which I am whistling. The same is true of making love, of course. And the same may be true of some truly nasty actions. Suppose a young man, who works as a guard in a camp of prisoners of war, finds an old woman, just brought back from a session with her torturers, lying bleeding on the floor of her cell. He looks at her for a moment and then kicks her head. In a few hours, he knows, he will be part of an execution patrol squad, shooting at the very same woman. It may be the case that this man has a further goal: he may want to convince himself that this woman is inferior to himself, rendering easier his job as an executioner. But it may just as well be that he is doing what he does just for ‘the hell of it.’ With reference to his desire to kick her for no further reason (for fun) we can make sense of (rationalise by his own lights) what he does. We can give a Humean explanation of what he does. And note that, in this case, we give a Humean explanation without mentioning of any belief. I have admitted that this is a possibility in some limiting cases, when it comes to basic actions. If, on the other hand, a person is simply whistling or kicking, without any desire to whistle or kick, because he suffers, say, from Thourette’s syndrome, then his behaviour cannot be explained in Humean terms. But this is as it should be since
2.8 The Temptation to Idealise
23
this kind of whistling or kicking is not intentional; it is not intentional under any description of it. For a person with Thourette’s syndrome to whistle or kick, in the manner indicated, is not for this person to perform any action at all. And the same is true of a person who cannot help stealing (a kleptomaniac) or drinking (an addict) or telling lies (a mythomaniac) – if such persons exist. To act out of habit may come close to the Thourette (the compulsive) case. Even if I have deliberately chosen to develop a certain habit it may well be that, in the circumstances, when I act in conformity with it, my action is not intentional – so it is doubtful whether we should really speak of it as an action in the first place. When I see a car in front of my car, and turn to the right to avoid running into it, it is doubtful whether my ‘action’ can be explained in Humean terms. I’m simply conditioned to turn to the right in situations such as these; this is a behaviour I have adopted, for good normative reasons, certainly, but a behaviour, which means that, in the circumstances, I react rather than act.
2.8 The Temptation to Idealise Some people, sometimes called ‘reason internalists’, believe that Humean reasons, or reasons that can easily be defined in terms of Humean reasons, are the only reasons there are for action. Davidson seems to be one among them. And Bernard Williams, in his influential paper, ‘Internal and external reasons’, states that the truth of the sentence ‘A has a reason to F’ implies very roughly ‘that A has some motive which will be served or furthered by his F-ing, and if this turns out not to be so the sentence is false; there is a condition relating to the agent’s aims, and if this is not satisfied it is not true to say … that he has a reason to F’. Williams then goes on to qualify this statement, however, since, after all, he wants to ascertain that our reasons are based, not on our actual aims, but on our corrected aims (where aims based on false beliefs, a less than complete understanding, or poor reasoning have been rectified); and yet, even on the more refined version of his theory of practical reasons, there exists nothing like objective moral reasons as will be defined in the next chapter. This is his general idea about what could constitute a reason for action: If something can be a reason for action, then it could be someone’s reason for acting on a particular occasion, and it would then figure in an explanation of that action.15
Why does Williams qualify his notion of reasons in this manner? Why does he go for corrected aims (where aims based on false beliefs, a less than complete understanding, or poor reasoning have been rectified) rather than actual aims? I suggest the following explanation. Williams believes that there are no normative reasons (of a kind to be defined in the next chapter). His notion of a reason is cast in purely naturalistic terms. And yet, he feels that there is some kind of positive In Bernard Williams, Moral Luck, p. 106.
15
24
2 The Humean Notion of Practical Reasons
emotive meaning attached to the word ‘reason.’ He hesitates to say of a person who rapes young children, say, that this person has reasons to do what he does (even if this person does have a sexual appetite for children), since he cannot sympathise with this appetite. He hopes that the appetite would go away, if the person were more reasonable. Yet, this won’t do. For some people may well stick to their strange appetite even if they become better informed. If there are no normative notions in the definiens of ‘practical reason’ we had better, therefore, to get rid of the positive emotive meaning attaching to the definiendum. If we don’t we may end in an unstable compromise. We arrive at a notion that is useless for explanatory purposes (since it has been idealised) but which is also incapable of fulfilling any normative purposes (since even a rational person with idealised wishes may behave in a less than decent manner). Many more philosophers who deny that there are any objective moral reasons tend to seek ways to show that, even so, a rational person should be expected to behave rather decently. Here R.M. Hare comes to mind. He never advocates utilitarianism; he never suggests that utilitarianism gives the true answer to the question what we ought to do and why we ought to do it. He attempts rather to show that we, at least to the extent that we tend to be rational (and to the extent that we tend to have a correct grasp of moral language), tend to pass moral judgements in accordance with what utilitarianism dictates, that is, he attempts to show that, to the extent that we live as we preach, we have Humean reasons to behave decently towards one another. For those who want to take moral language seriously, there is no way to avoid universalising one’s particular moral reasons along the lines described by Hare, he claims. In a similar vein David Gauthier attempts to show that rational egoists, for egoistic (Humean) reasons, tend to behave rather decently towards one another.16 In the end, however, I believe all such attempts must fail. There will always be cases where some perfectly rational person, who has all the correct relevant beliefs, who masters moral language, and who has only desires that have stood up to cognitive psychotherapy, to borrow an idea from Richard B. Brandt,17 will perform some nasty action. In the circumstances, we can explain and rationalise this action by the agent’s own lights, but we will (should) feel no inclination to approve of his action. This is as it should be. This agent has reasons for his action. These reasons rationalise the action by the agent’s own lights. Yet, we can say correctly that there were good reasons why he should not have acted as he did. These reasons, however, are of a different kind. Here a different notion of practical reasons is operative. This other notion will be defined in the next chapter. We had better be clear about what notion we are using, when we speak of practical reasons, then. Once we realise that there are two separate notions of practical reasons, there should be no problem for us to acknowledge that there are evil but perfectly rational people. When I say of a person, who commits atrocities, that he or she is rational, when I rationalise his or
See his Morals By Agreement for a statement of this view. Cf. his The Theory of the Good and the Right.
16 17
2.9 Conclusion
25
her actions (by his or her own lights), I make no normative claim whatever. My notion of a Humean reason, and my notion of rationality, are entirely naturalistic.
2.9 Conclusion When I claim that all these thinkers mentioned above deny that there exist objective moral reasons I do not want to deny that several among them are prepared to say that ‘moral’ reasons exist and are, in a sense, special. However, what these thinkers have in common is the insistence that moral reasons too are Humean, that moral reasons are somehow based on individual goals or desires. What qualifies a practical reason as a ‘moral’ one, according to these thinkers, then, is only the mode of desire upon which it is built (a second-order desire, a ‘rational’ desire, or a desire that is overriding other desires, and so forth). But this manner of speaking doesn’t make things clear. On the contrary, to speak of ‘moral’ reasons in this, naturalistic, manner is to conflate things. It is to borrow prestige from the notion of genuine normative reasons, without paying the ontological price one has to pay, if one is to endorse them. In the next chapter I turn to the definition of the notion of an objective moral reason, and I will conjecture that, besides Humean reasons for action, there exist also objective moral reasons for action.
Chapter 3
The Moral Notion of Practical Reason
3.1 Introduction Moral reasons for action exist, if they do exist, independently of whether we know about them or not. If I have a (conclusive) moral reason to perform a certain action, then, as a matter of fact, I ought to perform it. The ‘ought’ in question is a categorical one. The point is not that, if I want to be moral, I ought to perform this action (this would be a Humean reason for me to perform it). The point is that, irrespective of what I want, I ought to perform this action. The explanation of why I ought to perform this action is my moral reason to perform it. Some notion of a moral reason along these lines is common place. However, the definition here used was first put forward only recently by John Broome.1 The idea that the notion of a conclusive moral reason is basic is my own, however. Broome does not share it. This idea fits nicely with a traditional covering law notion of an explanation in general, however. I will get back to this. In contradistinction to Humean reasons, which are psychological in nature, moral reasons are abstract true propositions (facts) capable of explaining a normative fact. To use Judith Jarvis Thomson’s phrase, moral reasons are something we reason ‘from’. What do we reason to, then? This was not clear on her notion of reasons. However, here we receive a clear answer to that question: Moral reasons form the premises (the explanans) of an inference to a normative fact (the explanandum). Not all inferences to a normative fact provide us with moral reasons, of course. For example, the following inference provides us with no moral reason: 1. Either you ought to shoot the Prime Minister, or you ought to go to bed. 2. It is not the case that you ought to go to bed. 3. Hence, you ought to shoot the Prime Minister.2
John Broome, ‘Reasons’. Wlodek Rabinowicz drew my attention to this paper when he read an early version of this ms. 2 I owe this example to Jonas Olson. 1
T. Tännsjö, From Reasons to Norms, Library of Ethics and Applied Philosophy 22, DOI 10.1007/978-90-481-3285-0_3, © Springer Science + Business Media B.V. 2010
27
28
3 The Moral Notion of Practical Reason
Why doesn’t this (valid) inference provide us with moral reasons? The explanation is that it is not an explanation. And why is it not an explanation? There is something missing in it. It would be an explanation if it had included essentially among the premises a moral principle. Moral principles play, then, the same role in moral explanations that laws of nature play in scientific explanations of a standard ‘Hempelian’ model. Which moral reasons exist, then? This is what normative ethics is all about. In normative ethics we confront with each other various different putative categorical normative principles (or criteria of right action) such as the ones given by utilitarianism, egoism, contractualism, moral rights theories, deontology and virtue ethics, and we try to find out which one is most plausible. The assumption behind the search is that there exists a true answer to this question. These theories all compete in suggesting to us conflicting answers to the question what actions we ought to perform, or not perform, and why we ought to perform, or not perform, them. Suppose utilitarianism is on the right track. Then this means that, if I ought to perform a certain action (if I have moral reasons to perform it), then this action maximises the sum-total of well-being in the universe. The fact that it maximises the sum-total of well-being in the universe, together with utilitarianism, stating that we ought to maximise the sum-total of well-being in the universe, explains why I ought to perform it. My moral reason to perform this action is the fact that it maximises the sum-total of well-being in the universe. Reasons explain normative facts. When I speak of ‘explanation’ this should be understood in analogy with how the term is used in science, where we explain an event by subsuming it under a covering law. So this means that utilitarianism (and competing moral theories) must be taken, not only to state sufficient and necessary conditions of our obligations, but a kind of ‘law-like’ connections. The fact that an act maximises the sumtotal of well-being, according to utilitarianism, makes it obligatory. On my account, the fact that I ought to perform a certain action is no reason for me to perform it. This is as it should be. When I ought to perform a certain action, I have a moral reason to perform it. However, the reason is not the fact that I ought to perform it, but the fact that explains why I ought to perform it, such as the fact that it maximises the sum-total of well-being, etc. One might wonder whether this means problems for Immanuel Kant’s idea that we should ‘act from duty’. Does not Kant want to say that our reason for doing our duty should be that it is our duty? I don’t think that this is the most charitable interpretation of Kant. It is more plausible to assume that what he is after is that the person who acts in accordance with his or her duty should have as a reason for his or her action, not that it is obligatory, but that it means that a promise is kept, a truth is told, and so forth. If this is how Kant is interpreted, then facts such as these can be conceived of as (putative) ‘moral reason’, as here conceived of, for this person to perform this action. The fact that the action is the fulfilment of a promise or telling the truth explains why it is obligatory.3
For an interpretation of Kant along these lines, see Marcia Baron, ‘Acting from Duty’.
3
3.2 Prima Facie Reasons
29
To sum up. If I have a moral reason to perform an action, then it is a fact that I ought to perform it (and the reason explains this fact). Is it true as well that, if a normative fact exists, there exists also a corresponding reason? Do all normative facts have explanations? I tend to hold a ‘deterministic’ view of normative facts. Unless this view is warranted, there will be no moral explanations and hence no moral reasons. I alluded to this above where I indicated that we couldn’t have a moral explanation unless a moral principle figures in an essential way among the premises of the inference to the moral fact to be explained. However, in this book, I will leave it an open question if this belief in moral determinism is true. So for all I claim in this book there may exist particular normative facts that cannot be explained with reference to general principles. Can there still be reasons why we ought to perform them? According to ‘particularists’, there may exist aspects of the situation, where these particular actions are performed, explaining why they ought or ought not to be performed. But what do these particularists mean by an explanation? Can there be explanations if there are no principles? I have not been able to spot a satisfactory answer to this question in their presentation of their favoured view. Hence I conclude that, if they are right, then, contrary to what they believe, there are no moral reasons. Or, to put the point more cautiously, if there are no true moral principles helping us to explain moral obligations in the manner here indicated, then there exist no moral reasons in the sense here discussed. I discuss particularism and moral explanation elsewhere, where I claim that particularism carries a high epistemic price since it lacks a notion of moral explanation. This means that we cannot improve our moral understanding through inferences to the best explanation of our particular moral intuitions.4 Finally, even if all particular obligations can be explained with reference to universal moral principles given a deterministic view of ethics, there are some normative facts that seem to lack explanation, to wit, the most general ones (the most basic moral principles themselves). These normative facts are unexplained explainers. In this they are – once again – similar to the most basic laws of nature.
3.2 Prima Facie Reasons We should also acknowledge that, beside conclusive moral reasons, there may exist ‘prima facie’ moral reasons for and against actions as well, each of which should, at least in a concrete situation, be given a certain consideration, when we are looking for an answer to the question not only what we ought to do, but why we ought to do it. Given the notion of a conclusive moral reason it is not difficult to define what
See my Hedonistic Utilitarianism, Chapter 2, about this.
4
30
3 The Moral Notion of Practical Reason
it means for a reason to be of a prima facie character. These reasons say that, caeteris paribus, if a certain fact obtains, then the agent ought to perform a certain action. An example may be helpful here. It is prima facie wrong to inflict pain in your fellow human beings. This means that, in ordinary circumstances, where there are no reasons to inflict pain in others, we should not to so. But in some circumstances you can improve the situation of people by inflicting pain in them, or help many persons a lot by inflicting pain in a few; then it may be right to do so. In these circumstances you have reasons to do what is prima facie wrong. If this is so, then, this means that we have sometimes (prima facie) reasons to perform actions that we ought not to perform. We may conceive of these in counterfactual terms. If some aspects of the situation had been different, the action in question had been obligatory, and the ‘prima facie’ reason would then have explained, or at least been part of an explanation of, why it was obligatory. Note that, on this conception, a ‘prima facie’ reason is an epistemic notion. Does that mean that, in a situation, where we have a prima facie reason to do X, but where we ought not to do X, our prima facie reason to do X has been, to use McDowell’s5 term, ‘silenced’? Or, should we say that the prima facie reason to do X is still there, ‘speaking to us’, and having a certain weight, in spite of the fact that, upon closer inspection, it should not be acted upon? I think that, unless we want to exercise in spurious profoundness, we should say that the reason has been silenced. In fact, we should reject the metaphor that reasons have a certain ‘weight’. We may well weigh interests against one another, but reasons are not the kind of entity possessing weight. Rather, in a situation, a reason is decisive or not, period. For, certainly, what has happened when we have a prima facie reason to do something we ought not to do is not that the prima facie reason did not turn out to be heavy enough to decide the matter. What has happened is rather that, upon inspection, it has turned out that the prima facie reason in question depended on a hypothesis that was not, in the situation, fulfilled. So the prima facie reason is not there any more, speaking to us in a low voice; on the contrary, it has gone away, it has been silenced. Or, perhaps an even better way of putting it is that, it has never actually entered the picture. This is not to say, of course, that, if I had a prima facie reason to help a friend but, as it were, had no opportunity to do so, since I happened upon a situation where instead I had to help a stranger, then I have no remaining duties to my friend. It would be strange to claim that I should apologise to my friend, since, after all, what I did was the right thing. I would be offended if my friend told that he had forgiven me. There doesn’t seem to be any room for what Barnard Williams has called ‘agent regret’ either, since, once again, what I did was absolutely right and merely prima facie wrong.6 Still it may be true that I owe my friend an explanation of my behaviour and an expression of the sympathy I feel for him.
Cf his ‘Are Moral Requirements Hypothetical Imperatives?’, p. 26. Bernard Williams, ‘Moral Luck’, in Moral Luck, pp. 20–39.
5 6
3.2 Prima Facie Reasons
31
Moreover, even if, in the situation, I did the right thing, I may come to suspect that I did the right thing only too easily. There is more to morality than rightness and wrongness of particular actions. We are also responsible to some extent for our set of motivations, our traits of character. It may be a fact that a true friend should not act in the way I did. A true friend should act wrongly in this situation. And it may be a fact that it is better to be a true friend (who acts wrongly in rare situations such as this one) than a not so genuine friend, who acts rightly in situations such as this one, but wrongly in many other situations. I may be to blame for having done the right action in the situation, then. Or, at least for being the kind of person who did that. If this is a fact, than I may have reasons to think through what kind of person I am and perhaps to seek improvement (even therapy). All this has to do with the fact that there exist both blameless wrong-doing and blameful right-doing.7 This fact can explain, and hence debunk, the belief many people have had that there is a remaining ‘pull’ from prima facie norms that did not apply in a particular situation. This hypothetical (counterfactual) idea of prima facie reasons, relying on a caeteris paribus-clause, is not entirely clear, but this only reflects a corresponding unhealthy vagueness in standard talk of reasons for actions that ought not to be performed.8 However, I need not go any further into these difficulties in the present context – these problems should be dealt with within normative ethics. If eventually we are successful (and lucky!) in that field, then we will end up with fundamental principles answering (in principle) in each case what we ought to do and why we ought to do it. If we are less successful (or less lucky), then we may have to stay satisfied with mere ‘prima facie’ principles, leaving room for a kind of ‘judgement’, when, in concrete situations, moral reasons for action are given. But it should be noted that, if the kind of intuitionism defended by W.D. Ross is correct, that is if there are true moral prima facie principles, but these principles do not determine the normative status of particular actions, then there are no moral reasons, strictly speaking, as they have here been defined, since then there are no explanations of particular obligations. This is how Ross himself describes our predicament: 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.9
Or, to be more cautious, explanations exist only in situations where we face no conflicting prima facie norms. Given the plethora of prima facie norms in Ross’ moral universe I doubt, however, as does Ross himself in the quoted passage, that we ever, in real life, face such situations.
I write about this in ‘Blameless Wrongdoing’. For a more detailed explanation of prima facie reasons, or ‘pro tanto’ reasons, as he calls them, see John Broome’s highly instructive article referred to above, ‘Reasons’. 9 The Right and the Good, p. 18. 7 8
32
3 The Moral Notion of Practical Reason
It is also noteworthy that, if after all I would find myself in a situation where all prima facie duties pull in the same direction, then my obligation would be overdetermined. I could explain it with reference to any one of the prima facie obligations applicable in the situation, with the addition that, in the situation, there are no prima facie reasons against following it. When Ross speaks of ‘prima facie’ obligations, it should also be noted, he has a different notion in mind than the one I have used in my definition of prima facie reasons for action. I use the term merely to express an epistemic notion, and hence, the term ‘prima facie’ obligation is really in place here, while, in Ross’ treatment of prima facie obligations, the term carries an ontological (indeterministic) import. The situation is even worse, from the point of view of the existence of moral reasons, as was pointed out above, if particularism gives the correct account of our moral predicament. Then we are not even near a situation where we can give moral explanations.
3.3 Are There Moral Reasons That We Cannot Know About? On the notion of an objective moral reason here put forward, we can have moral reasons to perform an action even if we do not know this. This is perhaps not surprising. However, should we also make the stronger claim, that we can have moral reasons to perform an action even if there is no way we can find out which they are? I think we should make this move. Some have found this counter-intuitive, however. This is how Jonathan Dancy puts the point: Consider the question on which one has reason to bet on in a race. If the answer to this question depends only on matters of fact, and not on probabilities, it must be that one has most reason to bet on the horse that will in fact win and no reason to bet on any other horse at all. Nobody thinks this is the case.10
Now, it may be wondered why one should bet on the winning horse in the first place. But suppose that, if I do, and if I win, then I will send the money to Oxfam. Then it is reasonable to suppose, I think, that I ought to bet on the winning horse. Dancy, however, thinks otherwise. He thinks that ‘nobody’ believes what I believe. What are we to make of this? Certainly, Dancy’s claim that nobody thinks that one ought to bet on the winning horse is far too strong. It is falsified by my belief that this is what one ought to do. And many (objective) utilitarians share my view. And yet, for all that, the view may be wrong, of course. Is Dancy right in his insistence that one doesn’t have such a duty?
Practical Reality, p. 66. The idea comes originally from Lars Bergström, ‘Reflections on Consequentialism’, to which Dancy makes a reference. 10
3.3 Are There Moral Reasons That We Cannot Know About?
33
Obviously, whether Dancy or I am right on this point depends on what objective duties we have. Dancy thinks we cannot have a duty to perform an action of which we have no chance of knowing that it is right. But on his objective notion of moral reasons, how should he avoid this conclusion? This is how he answers this question: … how can the objectivist avoid [this conclusion]? We avoid it by appeal to the epistemic filter that all reasons must pass.11
This is strange. Can epistemic considerations decide what obligations there are? Must it not be mistaken to take ontological matters to be settled by epistemic considerations, at least in this straightforward way? To me it seems to be perfectly in accordance with moral phenomenology as it actually is to acknowledge that, in many situations, even if we have done our best to find out what to do, we may have failed. I would be surprised to find that our moral obligations were ever quite accessible to us. Moreover, some facts about the same case may be accessible to one person but not to another. I suppose the idea of a filter then implies that the epistemic situation of the agent is decisive to what reasons there are. But suppose I know that horse A with high odds will win since I have arranged for this to happen. Would it then be false if I were to claim that the agent had most reasons to bet on horse A? This does not seem to be a correct characterisation of the situation. Dancy defends the existence of the filter with the help of two arguments: ... the filter can be given an independent rationale. Part of the sense of the principle that ‘ought’ implies ‘can’ is that practical requirements that bind agents must be ones that it is not impossible for those agents to live up to – otherwise the requirements are, as we might put it, shouting in the dark. A similar thought applies to the suggestion that some of the grounds for our reasons might be ones that we are utterly incapable of discerning. The grounds for our reasons, like the reasons themselves, must lie within our capacities for recognition, if they are to be capable of being practically relevant for us.12
To answer Dancy’s first argument, we can bet on the winning horse. We can bet on any one of them, so we can also bet on the winning one, even if we do not do so under this description. In some situations we can be pretty sure as to which horse this is. We are lucky, then. In other situations we are in a worse epistemic position. What a pity – but not at all inconceivable! And it doesn’t mean that our obligation goes away. The idea that ‘ought’ implies ‘can’ is only plausible if it means what it says! To Dancy’s second argument, to the effect that the grounds for our reasons must, in order to be practically relevant for us, lie within our capacities for recognition, a moral realist and objectivist is free to answer that we have no guarantee that moral truth must (always) be applicable in practical life. The assumption that it must is based on an unwarrantedly simplistic picture of our moral predicament. Furthermore, on a week notion of practical relevance suggested in the literature, a moral principle
Ibid. Ibid., p. 59.
11 12
34
3 The Moral Notion of Practical Reason
need only guide choices in the sense that an omniscient deliberator should be able to apply it.13 This week requirement is satisfied under the envisaged circumstances with the race. Finally, if we learn that there are objective reasons of which we cannot know, then this may indirectly have practical relevance for us. It may lead us to the idea that we should distinguish between right-making characteristics of an action and a method of moral decision-making. I will get back to this idea in Chapter 6. It should be noted that even some utilitarians tend to argue that, rather than having a duty to bet on the winning horse, we have a duty to bet on the horse with the best odds. I find that equally odd, and for the same reasons. It doesn’t seem to fit with moral phenomenology. It is strange to think that, once we have done our very best to ascertain what action we should opt for, our moral problem has been solved. But this is not the place to settle disputes such as this one. Once we have pursued normative ethics in a successful manner, we will know the answer to the question. Nothing in the present context hinges on whether Dancy or I am right on this point. However, Dancy’s (and Bergström’s) claim that the objective account, according to which there are reasons we cannot know about, can be set to one side on simple a priori (epistemic) grounds, is mistaken.
3.4 Are Moral Reasons Always a Part of Humean Reasons? Clearly, there are cases where I believe that I ought to F where I want to do what I ought to do, and thus decide to F. Some philosophers seem to believe that this is always the case where someone acts for a reason. Unless a person believes she has some kind of obligation to fulfil, her action cannot be intentional. So it makes sense to ask a person, who has given her reasons for a certain action, whether these reasons are really good reasons, whether what the person took to be an obligation really was an obligation. All desires are based on normative beliefs (about their appropriateness), it has been claimed; and a natural corollary to this thought has been the claim that our normative reasons to have the desires we have translates into reasons for the action we perform, satisfying the desire in question.14 This is how the point has been made by Frederick Stoutland: The proper way to understand the notion that reasons explanations must be given from the agent’s own point of view is, in my view, ... to understand it in terms of the agent’s taking a consideration to be a reason for her to act. This is the crucial concept linking reasons as normative with reasons as explanatory: if the reason for which S A’s is R, then S must both take R to be a (normative) reason for her to A and take R to be the (explanatory) reason
Erik Carlson, Consequentialism Reconsidered, p. 101. There exists an extensive discussion about this. Examples of authors making similar are Johnatan Dancy, in Practical Reality, Warren Quinn, in ‘Putting Rationality in Its Place’, and Joseph Raz, Engaging Reasons, and Thomas Scanlon in What We Owe to Each Other. 13 14
3.5 On the Existence of Objective Moral Reasons
35
for which she A-ed. If the former taking is in error, then she has made a normative mistake (which is common); if the latter taking is in error, she has made an explanatory mistake (which is not common).15
But the point, if taken literally, is implausible. Even if there are examples of our performing actions because we believe we ought to do so, and want to do what we ought to do, it is more common that we act as we do without any concern for our obligations. We perform actions to satisfy desires and wants, period. And even where we have reasons to have these desires, these reasons need not pertain to the object of them, and hence they need not carry over to the act itself when we satisfy them. I may have developed a taste for Austrian expressionism in order to impress my girlfriend. I now like to watch expressionist painting. My reason for having the taste I have gives me no further reason to watch the paintings. The reason I watch them is that I like to do so. And even if I like to watch them, I may hold no value opinions about them whatever. I may very well be an aesthetic nihilist. Those who are of a moralistic bent may regret the fact that we often act on desires, which we have no reason to hold, or, if we have reasons to hold them, we hold them for reasons irrelevant to the act we perform, but, even so, this is a fact.16 And if we take up a less moralistic stance to the world we may even welcome this fact. What a bore it would be always seeing one’s life through moral (normative) spectacles.
3.5 On the Existence of Objective Moral Reasons Some philosophers, as we saw, deny that there are objective moral reasons, existing independently of whether we know about them, and also existing independently of our ways of conducting our lives. According to these philosophers there is no way to transcend conventional common sense morality. Other philosophers, such as, for example, Henry Sidgwick, G.E. Moore, A.C. Ewing, and Thomas Nagel, have thought that such ‘transcendental’ reasons exist. Others yet, claim that only objective moral reasons exist. Humean reasons are no reasons whatever. This was, as we saw, a view put forward by Judith Jarvis Thomson. A philosopher who argues in defence of it is, once again, Jonathan Dancy. This is how he puts the point: … indeed, I will eventually argue that all our practical reasons are objective.17
Who are right about this? Do such moral reasons exist or not? I think it obvious that Dancy’s point is exaggerated. What he insists on, I suggest, is not that Humean reasons (as here defined) do not exist, but merely that they
Frederick Stoutland, The Belief-Desire Model of Reasons for Action, p. 17. I can’t help finding this point obviously true. For an elaborate and tortuous defence of the obvious, see Ruth Chang, ‘Can Desires Provide Reasons for Action?’. 17 Jonathan Dancy, Practical Reality, p. 49. 15 16
36
3 The Moral Notion of Practical Reason
are no real reasons, they have no genuine normative force. He adheres to an intuition, which seems to be shared by many. It is not enough that a reason rationalises an action by the agent’s own lights, they feel. Unless they are prepared to sympathise with it, unless they are prepared to acknowledge it as correct, they will not call it a ‘reason’. This may seem a mere terminological difference, and I suppose it is. But it is indeed convenient to have recourse to Humean reasons as well as normative reasons. Indeed, unless we have recourse to both notions, the temptation to conflate the two ideas with each other may become irresistible. Those who deny that objective normative reasons exist will tend to ‘idealise’ the Humean notion, in the hope that what rationalises actions by the lights of the agent will also rationalise the action by the lights of a decent spectator. And those who deny that there are Humean reasons will tend to believe that people will somehow be drawn to the correct moral beliefs; acting on objectively existing normative reasons as if there did not exist deeply immoral persons.
3.6 The Principle of Charity Does it not boggle the mind to admit that a person has a reason to do a certain thing, when, in no way, I can sympathise with this reason? One might think that Donald Davidson’s principle of charity, if we want to abide by it, should make us reluctant to acknowledge that a person can have a reason with which we cannot sympathise. But this is not a proper use of the principle of charity. Considering how the notion of a Humean reason has here been defined, there is nothing strange about finding that a person acts for a reason with which we cannot sympathise. As we have seen, this is a view of reasons adhered to by Davidson himself (who seems to reject that there are any objective normative reasons.) His view is perfectly consistent. It is certainly true that, when pursuing radical interpretation, we may have to assume, not only that an alien person of whom we want to make sense, shares some of our most mundane factual beliefs, but also that he, to use Davidson’s phrase, is … a lover of the good (all by our own lights, it goes without saying).18
But once we have mastered an alien language, we may come to realise that, a person speaking this language, with whom we have been confronted, such as Adolf Hitler, holds both false factual beliefs and perverse aspirations, explaining his actions. We now say that these factual beliefs and these aspirations explain his action. We may be prone to believe that he holds these aspirations because he believes he ought to do so. Perhaps he wants to exterminate communists, Gypsies, homosexuals and Jews because he thinks these people ought to be exterminated. And, to put the matters from his own perspective, perhaps he just wants to do what
Donald Davidson, ibid., p. 222.
18
3.7 The Normative Constraint
37
is right for him to do. We may safely assume this – and yet feel no inclination to share his values, beliefs, or aspirations in these respects. It has been claimed that radical interpretation is normative as such, since in radical interpretation we assume a degree of rationality of the person speaking the target language. The number of philosophers who have made the claim that this makes radical interpretation partly a normative business is legion. It started already with Davidson’s own talk about ‘the constitutive ideal of rationality’19 and another early example is Saul Kripke, who claimed that ‘... the relation of meaning and intention to future action is normative, not descriptive’.20 As should become obvious to the reader of this book, I do not share this view. The notion of rationality we rely on when performing radical interpretation, and in semantics in general, is naturalistic. However, when we attribute moral beliefs to a person, it is true that we must make some moral assumptions. We must assume that the person speaking the target language get some moral matters right; otherwise we cannot attribute moral beliefs to this person in the first place. This is true, as we have seen, even of Hitler. But as far as I can see, it is quite plausible that Hitler got some moral matters right. He seems to have been a caring person in his relation to Eva Braun, for example. In his dealings with animals he didn’t show any signs of cruelty. So even by our lights, in some basic respects, he may have been a lover of the good. If this diagnosis is correct, then we can make sense of Hitler’s moral opinions and we are free to attribute many false moral beliefs to him, as it seems natural to do. If Hitler did not after all get even these basic moral facts right, however, then we have to conclude that Hitler held no moral beliefs. What is special with Hitler, then, is perhaps not that he made moral mistakes – we all do – but that he made moral mistakes with such terrible consequences. There may be other examples of this, however. Just think of the fundamentalist Christian belief that the use of contraceptives, as well as the performance of abortion, is immoral. These beliefs are false (or so I believe). There is no denying that they have had devastating consequences on the lives of many generations of women, particularly on women living in abject poverty in the Third World.
3.7 The Normative Constraint On my account normative reasons are not capable as such of explaining action. Jonathan Dancy would see this as a deficiency. And it is certainly at variance with his ‘normative constraint’, that is the claim that if S has a (normative) reason to f, then it must on any plausible theory be at least possible that the reason for which
‘Mental Events’, p. 223. Wittgenstein on Rules and Private Language, p. 37. Italics in the original.
19 20
38
3 The Moral Notion of Practical Reason
S f’s is that very (normative) reason she had to f. Otherwise it would not be possible for S or anyone else to act for good reasons – and that, he maintains, is surely absurd.21 But this constraint is not plausible. Normative reasons, on his account and on my account, are facts. But facts cannot be something for which we act. We may perhaps identify facts with true propositions. And we can act for these propositions, that is act on our belief in them. And we can act, then, for (a belief in) propositions that happen to be true. But this de re way of describing things doesn’t seem to capture what we want to say when we say that we act for something. However, my belief that the fact obtains can explain my action. And I can express (the content of) my belief by saying: I did as I did because, say, it was helpful to that old gentleman. I then express the content of my belief. We may even say, if we like, that, that my action was helpful (the proposition), was a ‘deliberative’ reason for my action.22 Now, the explanation why my reason, with this content, led to action must also, to be complete, include a reference to some desire of mine (a desire to be helpful, or to do what I believe I ought to do, and so on and so forth). My account is much simpler than Dancy’s, then, and it does not take me into any dubious ontological problems. And it does allow that sometimes the Humean reasons for a person’s action may, in a figurative manner of speaking, have normative reasons as their content. But do such objective normative reasons exist, in addition to subjective Humean reasons? To this question I will turn in the next chapter, where I defend moral (normative) realism. First a few words about the meaning of the kind of moral realism I will set out to defend.
3.8 The Meaning of Moral Realism The thesis I take to be true, and in defence of which I will give some arguments in the next chapter, is to the effect that there are objective normative facts, existing independently of our conceptualisation and thinking. If this thesis is true then there exists a ‘source’ of normativity, out there, to be discovered, and not of our making. As will be elaborated on in this essay, this is also a unique source of normativity. Moreover, it is also unique in that it has a special authority. Furthermore, normative facts are not only one kind of moral facts among many; as I will attempt to show, there exists only one kind of problem in morality, to wit, problems to do with what we ought to do, with right and wrong action, simpliciter. All categorical normative facts of this kind, existing independently of our conceptualisation and thought, are
Practical Reality, pp. 103–108. See Jonas Olson and Frans Svensson, ‘Regimenting Reasons’ for an instructive criticism of Dancy on this point. 21 22
3.8 The Meaning of Moral Realism
39
moral facts, and there are no moral facts in their own right other than normative facts of this objective kind. Problems to do for example with intrinsic value are not moral problems additional to the normative one. They are not genuine problems in their own right. What we search for in morals is a criterion of rightness and wrongness, simpliciter, of actions, cast in natural terms. We may speak of value, and articulate for example utilitarianism as the view that we ought to maximise the sumtotal of intrinsic value in the universe, or the Priority View as the view that we ought to maximise the sum of intrinsic value weighted somehow in relation to absolute levels of intrinsic value, but here ‘intrinsic value’ is just a place-holder for what it is that we ought to maximise (happiness, desire satisfaction, or what have you). Unless we can replace this placeholder by a clear statement cast in naturalistic terms, we have not answered any moral question whatever.23 We often hear that moral normative claims ‘override’ other claims. Is this claim true? In a sense this is true. They do have a special authority. It is false, however, if it is taken in a psychological sense. Often we realise what we ought to do but we do not do it. Moral reasons do not provide us, automatically, with Humean ones. Moral reasons do not override other reasons psychologically speaking. It is true furthermore that moral norms are ‘categorical’, and it is tempting to identify being overriding with being categorical. However, being categorical does not capture the sense in which moral norms override other norms. As will be explained in chapters to come, even norms of etiquette or grammar may be cast in an absolute and categorical form. But such norms do not exist independently of our conceptualisation and thinking. This means that they lack the authority typical of objective moral norms. When we realise that, from the point of view of etiquette or grammar, we ought to do A while, from the point of view of morality, we ought not to do A, the latter ‘ought’ overrides the former. We may also speak of it, as an authoritative plain ought simpliciter. My classification of it as ‘moral’ is just a manner of speaking. This kind of absolute authority, which is built into morality, understood in the way I do in this book, is part of what it means for a normative fact to be an objective moral one, then. Of course, it is possible to doubt that there exist such normative facts and it is possible to doubt that normative facts of any kind exist, but if moral realism, in the version here discussed is true, then such normative facts do exist. If morality is just one system among others which, together with etiquette and normative grammar, we construct, then it is possible to hold that sometimes a moral ‘ought’ overrides an ‘ought’ issued from the point of view of etiquette while, at some other occasion, and ‘ought’ issued from the point of view of etiquette
In my short introduction to normative ethics, Understanding Ethics, I try to show how virtue ethics, and the ethics of care, are only ethical theories in their own right to the extent that they contain criteria of rightness in competition with Kantianism, Utilitarianism, Egoism, and so forth. In the present context I will not comment any further on these theories. 23
40
3 The Moral Notion of Practical Reason
overrides an moral ‘ought’. Given the kind of moral realism here defended, this claim makes no sense. Authority is built into morality. When we believe in moral realism we believe there are authoritative answers to normative questions to be discovered out there. So let us now see what kind of considerations could lead to the belief that moral realism is true.
Chapter 4
In Defence of Moral Realism
4.1 Introduction Moral realism as here conceived has two tenets, one semantic thesis to the effect that when we pass moral judgements, we make truth-claims of an irreducible kind, and one ontological thesis to the effect that there are objective normative facts (truths) out there, sui generis, to be discovered. Some moral propositions are true. I have defended moral realism elsewhere (in my book Moral Realism) and recently Russ Shafer-Landau has also published a book with the same title. The main difference between my argument here, and in my earlier book, is that I now defend normative realism – I earlier focused rather on intrinsic value. The difference is perhaps more verbal than real, since I did hold then, and still hold, that there is only one problem to be pondered in ethics. I now find it more natural to speak of it as the normative problem, however. My main disagreement with Schafer-Landau concerns how we should answer Mackie’s argument from queerness, in its part concerning supervenience, but we also disagree on matters to do with moral epistemology. I have not tried to relate my new argument in detail either to my own previous view or to Shafer-Landau’s more recent argument; my argument here stands on its own feet. And the question I ponder is why we should believe in moral realism. Now, the semantic aspect of this question seems to me simple. It is obvious, that when we pass moral judgements to the effect that certain actions are right, other actions wrong, and so forth, then we intend to make objective judgements capable of being true or false. It is also clear, I think, that we make judgements that are not merely descriptive of empirical realities. For example, moral judgements are not elliptical, saying things like, according to the norms existing in my society, this action is right or wrong or, this action is liked, or disliked, by me. It is noteworthy that even a philosopher like John Mackie, who thought that there are no objective moral facts, thought that, in issuing moral judgements, we imply (wrongly) that such facts exist. Why should we draw the conclusion that our moral judgements are objective (in intent)? Well, it is sufficient to observe carefully how we make these judgements, and how we react to moral phenomena such as moral disagreement, and (putative) moral mistakes.
T. Tännsjö, From Reasons to Norms, Library of Ethics and Applied Philosophy 22, DOI 10.1007/978-90-481-3285-0_4, © Springer Science + Business Media B.V. 2010
41
42
4 In Defence of Moral Realism
When we run upon conflicting moral judgements we believe that both parties cannot be right. Why don’t we believe that they can? The best explanation is that we take them to make objective and contradictory judgements. But, if we run upon inconsistent judgements, at least one judgement must be false. And we often feel that we come to the conclusion that what we once believed was a proper moral judgement was, upon closer inspection, not a proper one. We then tend to believe, not only that we judged the case differently before, but that, before, we judged it wrongly. This means that we assume that there is a fact of the matter to be right or wrong about. Or, upon reflection, this claim may be too strong. It is not certain that everyone, all the time, is using a moral vocabulary in the objective sense here defined. Perhaps this is even a rare, and fairly recent phenomenon. It may well be true that some people, who have been indoctrinated into a particular moral creed, perhaps with the argument that it has been dictated to us human beings by God, are merely intending to report what is required by this conventional moral creed, with which they sympathise, when they pass moral judgements. We may suspect that this is a case if there is little or no moral disagreement in a society. In that case, a moral semantic (indexical) relativism of the kind defended by Wong1 and Harman2 may be true of the moral utterances made by people in this society. Other people, living in other societies, may have revolted against the morality into which they have been socialised, but not because they have started to make claims about a moral reality they believe exists independently of their conceptualisation; instead they may be just evincing their strong emotions (even their gut feelings) when they pass moral judgements. Hence emotivism or expressivism of the kind defended by Stevenson3 and Gibbard4 may be true of their moral utterances. However, even if there are such cases, and even if they are what we often meet with, when we observe how people pass moral judgements, there are also cases where people do intend their moral judgements to capture a moral reality that is seen by them as sui generis. This is certainly how Sidgwick and Moore conceived of their own moral judgements, and this is how Moore conceived of his disagreement with Sidgwick, and this is how contemporary moral realists such as Parfit and myself conceive of our own moral judgments and disagreements. And even if it is possible that we delude ourselves, even if it is possible that we are mistaken in our understanding of our own words, this possibility is far-fetched, considered the way at least we moral realists conceive of the subject-matter of morality. Once again, we pursue the truth, we want to avoid inconsistency (since we believe that two inconsistent beliefs cannot both be true), we sometimes feel that we were previously wrong when we held opinions different from the ones we now hold. Hence, is reasonable to believe that Mackie is right when he claims that an objectivist moral semantics is true at least of our moral utterances.
See, for example, his Moral Relativity. See, for example, his contribution to Gilbert Harman and Judith Jarvis Thomson Moral Relativism and Moral Objectivity. 3 See his Ethics and Language and Facts and Values. 4 See his Wise Choices, Apt Feelings. 1 2
4.2 The Argument from Queerness
43
However, John Mackie is not only famous because he thought that our moral judgements are objective, he also, as was alluded to above, thought that they were all false. He argued that ‘although most people in making moral judgements implicitly claim, among other things, to be pointing to something objectively prescriptive, these claims are all false’.5 I will speak of this ontological view as ‘moral nihilism’, even if this was not the term Mackie himself used. Mackie’s semantic view, to the effect that we refer to objective moral properties when we make moral statements, together with his ontological nihilism, to the effect that nothing possesses these properties – Mackie’s error theory – has by some been considered problematic. If the claim that torture is wrong is false, does this not mean that torture is not wrong? But does not this entail that torture is all right? And is this not a strange moral implication, which appears to follow from Mackie’s error theory? Moreover, if it is also false that torture is all right, does not this imply that torture is wrong? But then torture is both wrong and right (not wrong), which seems to be not only morally speaking strange but incoherent. Appearances are deceptive. If Mackie’s error theory is correct, then both the claim that torture is right and the claim that torture is wrong are false. We realise this when we realise that only together with the premise that all actions have normative status (they are right or wrong) are we allowed to conclude that, if an action is not wrong, then it is right. A moral realist accepts the first premise, I submit. All actions have normative status, the moral realist is likely to claim. And our common sense moral metaphysics is realist as well. But Mackie denies the first premise, of course. According to his theory, all actions lack moral status. According to his theory, all positive moral judgements to the effect that an act is right, wrong, or obligatory, are hence false.6 I suppose he is also making the weaker claim, now made popular by so called particularists, that all moral principles are false. All this means that his theory is consistent and it has no ethical implications. Yet, it is a strange theory, or a strange combination of views. But what is strange about the combination is perhaps not the semantic objectivist part, but the corresponding nihilist ontological assumption. What lead Mackie to his moral nihilism? His most important arguments to this effect are his arguments from ‘queerness’ and from ‘relativity’. I will examine them in order.
4.2 The Argument from Queerness Mackie assumed that, when we make moral judgements, not only do we ascribe an objective property to the actions we are passing moral judgements upon, but we also attribute to it a property which is inherently prescriptive. He thought that Ethics, p. 35. This is a very natural and plausible interpretation of Mackie’s error theory; for an elaboration on it, see Sinnott-Armstrong, Moral Skepticisms pp. 34–36, even though Sinnott-Armstrong does not generalise it to permissions.
5 6
44
4 In Defence of Moral Realism
we think, when we say of a certain action that it is right, not only that it has a certain objective property (rightness) but also that this property has the property in its turn to motivate us to act, at least when we think it is present. Moral properties, then, have ‘intrinsic to-be-pursuedness’ built into them. But such a property would be very queer indeed. This is how he states his argument: If there were objective values, then they would be entities or qualities or relations of a very strange sort, utterly different from anything else in the universe. Correspondingly, if we were aware of them, it would have to be by some special faculty of moral perception or intuition, utterly different from our ordinary ways of knowing everything else.7
I think Mackie is right when he insists that this would be a very strange (queer) property. However, I think he is wrong when he believes that we attribute such a property to an action when we claim that it is right. We do attribute an objective property, I concede. However, we do not attribute any inherently motivating property. And yet, for all that, Mackie’s observation is close to the mark. As a matter of fact, his mistake, when he believes that, when we pass moral judgements we want to attribute an inherently motivating property is highly instructive, so let us try to see how he goes wrong. It is certainly true that most people when they believe that an action ought to be done have some incentive to perform this action. But this does not mean that the property of obligatoriness as such has any inherent magnetic force attached to it. It is because most people care about doing their duty that they are somewhat inclined to do a certain action, once they acknowledge that it is obligatory. This means that Mackie was not completely wrong when he claimed that normative properties are in a manner of speaking motivating. However, this motivating force is not always present, and it is certainly not inherent in the moral properties as such. This property should rather be explained along the following lines: It is a remarkable fact that some normative judgements, such as most typically the expression of obligations, have, not only truth-conditions, but satisfaction conditions as well. The proposition that I ought to do X can be satisfied or not satisfied. It is satisfied, if I do X, and it is not satisfied, if I don’t. And we take an interest in these satisfaction-conditions. These satisfaction-conditions, and our interest in them, explain why we feel that moral properties are, in a sense, ‘prescriptive’. In fact, the notion that norms have satisfaction values takes care of one aspect of Mackie’s idea of prescriptivitity. He explains what he means by prescriptivity, with a reference to Plato’s idea of the Forms, and stresses two rather different aspects. The Form of the Good is such that knowledge of it provides the knower with both a direction and an overriding motive; something’s being good both tells the person who knows this to pursue it and makes him pursue it.8
The fact that norms have satisfaction values explains why they give us direction. This means that the proposition that I ought to do F can be, not only true or false, but
Ibid. p. 38. Ibid. p. 40.
7 8
4.2 The Argument from Queerness
45
satisfied or not satisfied. In this sense moral properties are prescriptive. In this a normative claim is different from other truth-claims, such as the claim that the sky is blue. If there are moral properties, then this means that there are properties that are different from ordinary natural ones, of course. But there is nothing queer about that. I have also taken moral realism to imply that answers to the question what I ought to do simpliciter has a special authority. Is that queer? I think not, we are perfectly familiar with this phenomenon. This follows from the fact that moral properties are prescriptive and real. However, if these properties also provide us with an overriding motive, then this is different. If this is the case, then it is queer. It is very queer indeed to believe that the fact that an action ought to be done should make me prone to perform it. This would be queer if for no other reason that I may be ignorant of its normative status. But even the idea that my belief that an action is obligatory should, as such motivate me, would be queer, I submit. But why believe in this kind of internalism? Let us return to the satisfaction values and truth-values of a norm and consider them more in detail. Consider a simple example: the circumcision of a young girl. Let us assume that one person claims that the circumcision of this girl ought to take place. This claim is true, on my objective understanding of it, if, and only if, the girl in question ought to be circumcised. This is a way of explaining the truth-condition of this normative judgement. But the same judgement has a satisfaction-condition as well. The judgement that the young girl ought to be circumcised is satisfied if and only if the young girl is circumcised. Most people have an interest in having their normative judgements (about obligations) satisfied. We have been socialised to have this interest. This means that many moral discussions have a practical dimension. When we disagree about the truth conditions of normative judgements we also have a practical conflict about the satisfaction conditions of the very same judgements. If you are interested in having a certain girl circumcised (you believe that she ought to be circumcised and you want to have this norm satisfied) while I believe that she ought not to be circumcised (and want this norm to be satisfied), then we face a practical conflict. This captures the grain of truth in Mackie’s thought that normative properties are motivating. However, not all normative judgements have necessary satisfaction-conditions in this straightforward manner. In order to see which judgements do, and which judgements don’t have both sufficient and necessary satisfaction-conditions, we may account for the normative categories in the following manner. An action is right if it is not wrong. An action is wrong if it is not right. An action ought to be performed if and only if it is wrong not to perform it. As we saw above, the first two claims about the relations between actions being right and wrong are not logical truths, but for a moral realist, it is natural to assume that they are true at our world and at worlds that are similar to ours. Obviously, judgements about obligation (you ought to do so and so) and about wrongness (it is wrong to do so and so) have satisfaction-conditions (the latter is satisfied if, and only if, you do not perform the action in question), but judgements about rightness (it is right to do so and so)
46
4 In Defence of Moral Realism
do not have any necessary satisfaction conditions. They only have sufficient ones, and they can be viewed as permissions. Suppose you ought to F, then your obligation is satisfied if you F otherwise it is not satisfied. This prepares the room for the following kind of practical conflict. If you believe that you ought to F and want to do what you believe you ought to do, then you want your obligation to be satisfied. Suppose you come across a person who believes that you ought not to F. This person too wants to see to it that what she believes to be obligations be satisfied. She wants you not to F. So she has a practical conflict with you. There is no way to satisfy both your conflicting (and inconsistent) norms. Here is another example. One person believes that it is right to F. Another person believes that it is wrong to F. According to the latter person, it is obligatory not to F. If you F, then you are in conflict with this person. He does not get his favoured obligation satisfied. However, if this person, who believes that it is wrong to F, does not F, you believing that it is right to F, need not have any problem with this. And if, for some reason, you don’t F, the person who believes that it is wrong to F, is perfectly happy with this. So not all intellectual moral conflicts translate easily into practical ones. However, many do, if not directly, so at least indirectly, and I believe that this is this phenomenon that Mackie has conflated with the queer property of an inherently motivating force of moral judgements. Furthermore, not all people want to have their favoured norms (the norms they believe are true) satisfied. We all know a kind of person (we know him or her from our philosophy seminars) who is exclusively interested in the truth-conditions of various different norms. He or she doesn’t give a damn about the satisfaction of these norms. We may look askance at such a person and yet, for all that, this kind of person does exit. But the mere possibility that there are such persons — and we may safely infer that they are possible, once we have met with them – shows that there is no inherently motivating force in moral properties, in Mackie’s sense – that is no ‘intrinsic to-be-pursuedness’ – built into them. They give us direction, all right, and this is what it means for them to be ‘prescriptive’, that is they have satisfaction-conditions – but they do not move us to action. But even if our moral (normative) judgements are objective, having both truth-conditions and (some of them) both necessary and sufficient satisfactionconditions, and even if we believe that most people, most of the time, are interested in having their normative judgements satisfied, why should we believe that, by issuing such judgements, we hit upon anything existing objectively, to be captured by our judgements? Why not believe, as Mackie did, that all moral judgements are false? Well, if there is nothing strange with the prescriptivity of moral (normative) properties (we are perfectly familiar with the phenomenon and we can it explain it with reference to satisfaction-conditions), why think that they do not exist? Why assume that our moral judgements do not correspond to anything in reality?
4.2 The Argument from Queerness
47
In his argument from queerness Mackie adds another respect in which he finds that objective moral properties are queer, the fact that they are ‘consequential’ or ‘supervenient’: Another way of brining out this queerness is to ask, about anything that is supposed to have some objective moral quality, how this is linked with its natural features.… The wrongness must somehow be ‘consequential’ or ‘supervenient’; it is wrong because it is a piece of deliberate cruelty. But just what in the world is signified by this ‘because’? … It is not even sufficient to postulate a faculty which ‘sees’ the wrongness: something must be postulated which can see at once the natural features that constitute the cruelty, the wrongness, and the mysterious consequential link between the two.9
This is most plausibly taken as two different objections, one to do with the word ‘because’, and one directly concerned with supervenience. I will start with the query about the word ‘because’, which seems to be Mackie’s main worry. Is it difficult to understand what is referred to by this word? I think not. The answer to Mackie’s query what in the world is signified by this ‘because’, is given with reference to the fact that in the world there are true moral principles, capable of explaining (together with certain facts) particular obligations. And these principles are not different with respect to the necessity implicit in them, than laws of nature. They state right- and wrong-making conditions. Our notion of objective moral reasons hence helps us to answer Mackie’s argument. The answer to his more general question about supervenience must be given with reference to some metaphysical principle. I once toyed with the idea that we should give up supervenience altogether. Two worlds could be similar in all nonmoral respects, and yet be different in moral respects. I happened to discuss this idea, many years ago, with John Mackie. He was not convinced. As a matter of fact, he was angry with me! He could not understand how on earth I could hold such a bizarre view.10 And I realise that if moral properties do not in any way supervene upon non-moral properties, it is hard to see how we could have any knowledge about them. If we jettison supervenience, we end up as moral sceptics. Now I tend to believe that the most natural response from the moral realist to Mackie’s argument would be, not a denial of supervenience, but a denial of the claim that supervenience as such is any strange phenomenon. We know it from many other parts of our human understanding. The mental seems to be supervenient upon our neurophysiology, causal relations seem to presuppose supervenience, and so forth. The details of this thesis, however, are far from clear. There is a strong version of supervenience, according to which moral principles are necessarily true. So what makes something right or wrong in the actual world makes it right or wrong in all possible worlds. This claim may appear to be queer. Can the claim that moral properties are both sui generis and supervenient upon natural properties in this sense be upheld?
Ibid. p. 41. It must have been around 1980, when he visited our philosophy department in Stockholm.
9
10
48
4 In Defence of Moral Realism
Frank Jackson has put forward an argument with reference to Simon Blackburn to the effect that this is not possible. I here answer Jackson’s concern. I am at loss with respect to Blackburn’s own version of the argument, since I cannot understand why he finds it difficult to explain why there cannot be a ‘mixed’ world, where sometimes a naturalistic characteristic is right-making, and at another time not right-making. That this is impossible follows from an ordinary understanding of what a moral principle (connecting the two properties) is. Unless it holds without exceptions in a world, and with counter-factual implications, it is no principle (the same is true of putative laws of nature).11 The view that moral properties are both sui generis and supervenient upon natural properties can be upheld, if we adopt a fairly weak view of supervenience, according to which a certain natural property brings with it a corresponding moral property only in all possible worlds that are sufficiently like ours. This seems to me a fairly plausible metaphysical view. I do not want to rule out the possibility that there are possible worlds where different laws of nature and morality obtain. Even if utilitarianism is actually true, or close to the truth, there may be worlds where Kantianism is true. There may even be worlds where moral nihilism is true, and there may be worlds where moral particularism is true (worlds where particular actions are right or wrong but where there are no true moral principles). If I am right about all this, then this means that the moral realist is under no pressure to explain how it is possible that certain moral properties and certain natural properties co-exist in all possible worlds and are yet, for all that, distinct. We have no reason to believe in the claim of co-existence in all possible worlds in the first place. We may even concur in the assessment that this claim is queer. On a weak notion of supervenience, we can accept the claim that if a world is exactly like ours in all natural respects, it must also be like ours in moral respects. And this weak notion of supervenience is consistent with the claim that, in each world, if there is a moral difference between two cases, there must also be a natural (non-moral) difference. Why, then, should we not believe in strong supervenience? Why should we not believe that true moral principles are necessarily true (true in all possible worlds)? The rationale behind the belief that moral principles have (merely) the same kind of necessity built into them that natural laws have, helps us not to have moral realism collapsing into naturalism. And the reason to believe in it is the observation that moral principles play the same role in moral explanations that natural laws play in scientific explanations. Furthermore, as will be noted below, we seem to arrive at our belief in moral principles in a manner that is similar to how we arrive at our belief in laws of nature, that is through a hypothetical deductive method of investigation. Hence, if laws of nature can differ between possible worlds so could moral laws (principles). This view of the kind of necessity exhibited by moral principles allows us to acknowledge that, even if moral realism is true, it could have been false.
See Blackburn, ‘Supervenience Revisited’.
11
4.2 The Argument from Queerness
49
Finally, there are views in normative ethics inconsistent with strong supervenience. I think for example of actualism in population ethics, the view that only the fate of actual people matter to the moral rightness and wrongness of actions. On this view, an action performed in the actual world may be wrong while, at the same time, there is a possible world, from the point of view of which the very same action is right. This view may sound strange, I believe it is false, but it does not strike me as incoherent. This standard and rather weak notion of supervenience, combined with the idea of (merely) a weak necessity built into moral principles, implies that there is a possible world where it is all right to torture an innocent child for no reason at all. I suppose that this world is very different from ours, but my imagination cannot help me to a concrete understanding of what it would be like. Moreover, when this is said, it is crucial that we note that this talk of possible worlds is a mere metaphor. No existential claim with respect to merely possible worlds is here being made. If the metaphor is taken literally, if we claim that all possible worlds exist, then we are in deep moral trouble. I have discussed this elsewhere,12 so I will not elaborate on this point in the present context. So much for the standard notion of supervenience. To go for strong supervenience, according to which moral and natural properties are coextensive in all possible worlds, threatens, of course, to have one’s view collapsing into the naturalistic claim of the Cornell variety that moral properties are identical to some natural properties. I will say more in the next chapter about why we should avoid moral naturalism in all its varieties. To sum up, moral (normative) properties are not queer. They are not queer in the sense that they can directly make us want or desire certain things (internalism). The answer to internalism is externalism: they don’t have this kind of motivational force built into them. They are not queer because they have prescriptivity built into them either. They do have prescriptivity built into them, but we know this phenomenon quite well. It has to do with the fact that normative judgements have both satisfaction-values and truth-values. We are familiar with, and can specify, criteria of both truth and satisfaction for any particular norm. The notion of supervenience associated with moral properties, may well involve us in deep metaphysical problems, but we are familiar with the very notion from many other areas of science. Finally, and most importantly, we know what it means to say that an action is right ‘because’ so and so. By saying this, we imply that there are true moral principles helping us to explain its rightness. Of course, the particularists may be right when they insist that there are no moral principles. We still know what it means to say that it is ‘because’ of the existence of a true moral principle that something is right or wrong. However, if the particularists are right, all such claims are false. Moreover, if, as I think, the particularists are wrong, and there are indeed true moral principles, then there are such claims that are
In my ‘The Moral Import of Modal Realism’.
12
50
4 In Defence of Moral Realism
not only meaningful, but also true. On the other hand, when the particularists speak of actions being right or wrong ‘because’ so and so, Mackie’s argument from queerness does seem to stick. They (the particularists) cannot explain to us how they use the term ‘because’. But that is their problem, not mine. So why believe that moral properties are projections of ours? It doesn’t seem as though Mackie’s argument from queerness has forced us to accept that conclusion. But Mackie’s has also put forward another argument, to the same nihilist effect, his argument from relativity.
4.3 The Argument from Relativity The argument from relativity is supposed to show that our positive moral judgements, pretending to be objective, are all false. This is how he states it: The argument from relativity has as its premiss the well-known variation in moral codes from one society to another and from one period to another, and also the differences in moral beliefs between different groups and classes within a complex community … radical differences between first order moral judgements make it difficult to treat those judgements as apprehensions of objective truths.13
This argument is problematic in many ways, however. Mackie himself is well aware of one of them. Even if this disagreement exists, this is not different from what we see in sciences such as history, or biology, or cosmology, he concedes. We do not, because of this disagreement, conclude that there is no objective truth in history, biology, or cosmology. But there is a difference, Mackie claims: … such scientific disagreement results from speculative inferences or explanatory hypotheses based on inadequate evidence, and it is hardly plausible to interpret moral disagreement in the same way. Disagreement about moral codes seems to reflect people’s adherence to and participation in different ways of life.14
Could this be correct? I do not deny that many moral judgements are, in this way, conventional. However, this is consistent as well with the possibility that some people have been able to transcend their actual conventional moral universe and pose the more fundamental questions about right and wrong action. This is the possibility I want to defend. Here it is not far-fetched to assume that, while there is much disagreement about speculative moral principles, intended to ‘explain’ particular moral judgements, just as there are disagreements in history, or biology, or cosmology, there is also much agreement. There are indeed moral facts, and, if we succeed in taking up an impartial stance, we are sometimes capable of recognising them.
Ibid. p. 36. Ibid.
13 14
4.3 The Argument from Relativity
51
All decent and impartial people, who consider the question, would agree that it is wrong to torture an innocent child, at least unless there are any pressing reasons of any sort to do so. The latter clause is elliptical and though of as taking care of objections such as: ‘But by torturing this child you make the world a better place in the long run’, ‘But by torturing this child you do what you have undertaken to do’, and so forth. If we come across a person, who doesn’t share our judgement about this case, we would say that he doesn’t understand what it means for an action to be wrong – he has not been able to take up a critical stance with respect to the norms that have been inculcated in him, or he suffers from some serious personality disorder. Perhaps there are only a few examples of this kind that can be produced, and perhaps they are a poor foundation for speculative inferences or explanatory hypotheses. In that case, we may here see the explanation of why people who have thought about them, and tried to find general explanations of them, have moved in different directions. And yet, for all that, these cases exist. And we may add that the kind of transcendental and distanced approach to morality that I am here speaking of is perhaps a rather new phenomenon. As Derek Parfit has insisted, ‘both human history and the history of ethics may be just beginning’.15 It should be noted that when I claim that it is obvious that it is wrong to torture an innocent child ‘for no reason’, this additional ‘no-reason’ clause is crucial to my argument. For according to some ethical theories, held by clever people who have thought hard about these things, it is indeed all right to torture an innocent child – if the consequences are good enough to compensate for the suffering felt by the innocent child. I have myself argued that it might be right to torture an innocent child, if this means that a great many people, who are all of them already very happy, will undergo a sub-noticeable increase in their respective hedonic situation.16 This is a controversial view, of course. All decent and clever people who have thought hard about it do not agree. Derek Parfit, for one, has nicknamed my idea ‘the Ultra Repugnant Conclusion’. The fact that he and many others disagree means that I have to concede that I may be wrong about this conclusion. However, this is not enough for me to give it up. After all, this conclusion follows from, and is explained by, a theory (hedonistic utilitarianism) for which I think there exists adequate evidence. Yet, for all that, this conclusion is far from self-evident, in the way it is self-evident that it is wrong to torture an innocent child for no reason whatever. I find it hard to see that there is any way that I could come to give up this belief – unless I would become morally perverted or go insane. But is it not because we have been socialised to believe that this is wrong, that we find it wrong to torture an innocent child for no reason? This is the main point in Mackie’s argument from relativity, even if it has little to do with relativity as such, and everything to do with the origin of our moral beliefs.
Reasons and Persons, p. 453. See my Hedonistic Utilitarianism, p. 77.
15 16
52
4 In Defence of Moral Realism
4.4 Harman’s Empiricist Argument It is a remarkable fact that the argument I have here attributed to Mackie, to do with how we should explain our moral beliefs, was also put forward by Gilbert Harman, in the opening chapter of his book, The Nature of Morality. Mackie’s and Harman’s books were published the same year, in 1977. I now turn to this argument. It should first be noted, however that, this kind of argument was alluded to already by R.B. Brandt when, in a foot-note to his Ethical Theory, he compares the role of intuition or feeling in morality to observation in science and makes the following qualification: There is one difference. Physical theory, taken with a description of the experimental setup, may logically imply ‘The ammeter will point to 30’ and we can observe whether or not this is the case. Whereas, although in ethics we may reject ‘There is no obligation to do X’ by appeal to the fact that we feel a strong obligation to do X … we cannot say that ethical principles entail anything about how we shall feel – at least not in any direct way.17
According to Harman’s more elaborated development of the argument, it relies on an explicitly stated empiricist principle to the effect that we ought not to countenance something as real unless we need to have recourse to it in the best explanation of some observation. And, according to Harman, when we explain why people make moral observations we never need to have recourse to moral facts. It is sufficient (we get a better explanation) if we refer rather to psychological facts to do with the agent, such as his moral sensibility, and so forth. This is the somewhat more elaborate way in which Harman makes his point: … observation plays a role in science that it does not seem to play in ethics. The difference is that you need to make assumptions about certain physical facts to explain the occurrence of the observations that support a scientific theory, but you do not seem to need to make assumptions about any moral facts to explain the occurrence of the so-called moral observations … In the moral case, it would seem that you need only make assumptions about the psychology or moral sensibility of the person making the moral observation. In the scientific case, theory is tested against the world.18
It might seem that the assumption upon which Harman’s argument relies is problematic. It seems itself to be normative. Harman assumes that we should not acknowledge something as real, unless we have to refer to it in our best explanation of some judgement we make (such as the report of an observation). But as will be shown in Chapter 8, we may think of rules such as this one as stating mere Humean, instrumental norms. They are to the effect that, if you want to further an epistemic goal you hold, if you want to have a realistic picture of the world, then you ought to abide by the rule. This is how I will understand it. Under this interpretation I accept it. The moral realist should bite the bullet and answer Harman’s (and Mackie’s) challenge, then. So it is pertinent to find out if Mackie is correct when he claims that the moral codes we adhere to reflect our various different ways of life. R.B. Brandt, Ethical Theory, p. 249n. The Nature of Morality, p. 6.
17 18
4.4 Harman’s Empiricist Argument
53
And we need to know whether Harman is right when he claims that our moral observations depend on our moral psychology, not on moral facts. These assumptions seems to me extremely far-fetched, when applied to the example we now discuss. It is more plausible to assume that we take it to be wrong to torture innocent children for no reason whatever because it is wrong. And it is very plausible to assume that we have been socialised into believing that this is wrong because it is wrong. Can we not find a better, evolutionary explanation, why we hold this belief, than concluding that we have been able to grasp its truth? In general, if we can find an evolutionary explanation of most our moral beliefs, this means that we should become agnostic with respect to their truth. This means less of a trouble than if Mackie is right in his insistence that normative notions are queer. If he is right, then all our moral beliefs are false. It is also less serious, than if expressivists of Stevenson’s and Gibbard’s variety would succeed in showing that all our moral expressions are used, not to express propositions with truth-values, but our emotions or something of the kind. If they were right, then our moral beliefs would lack cognitive content altogether. And yet, an evolutionary explanation would be a ‘debunking’ one indeed. It would show that we should be less confident in our moral beliefs. We would realise that we are not justified in holding them (even if they make sense and even if they may be true). Moreover, if we could find such an evolutionary explanation it should not only make us agnostic with respect to the truth of our deeply cherished opinions. We should also, as is stressed by Richard Joyce, become agnostic with respect to the moral notions (concepts) we employ.19 But all his means no problem automatically for the moral realist. Quite to the contrary, such a suspicion about conventional morality is really the point of departure for the realist. Now, the suspicion about the moral notions will naturally lead to a close examination of them. And even if this examination somewhat unexpectedly leads to the conclusion that the moral beliefs we hold are indeed true, since we are to understand them as elliptic, making an implicit reference to the system of norms upheld in our society, this is only the beginning of further queries. For now it may transpire that the moral realist is not at all satisfied with conventional morality, nor with notions of right and wrong action used in her society. Then a kind of transcendental move is natural to expect. The words ‘right’ and ‘wrong’ take on a new, absolute sense, and critical moral views are put forward with the aid of these notions. And the debunking explanation of the content of common sense morality is now a reason to bracket these beliefs, in our pursuit for the truth. Here our cognitive and emotional capacities, provided for by evolution, are put to use. How, then, are we to think of these new critical moral views, partly in conflict with conventional morality? Well, to the extent we conceive of them as the result of progress in our moral thinking, it is impossible to just set them to one side. We treat them as true in an absolute sense. We take them to track moral facts.
Richard Joyce, Evolution of Morality, p. 181.
19
54
4 In Defence of Moral Realism
But can we not once again try to explain, and hence debunk, the content of these beliefs with reference to our biological evolutionary history? I think not; at lest with respect to some of them this would seem far-fetched to assume. This is true already of the idea that one should not torture an innocent child for no reason. It is even more obvious when it comes to the recent insight that the same is true, not only of children, but of all sentient creatures, human or non-human. We learn from sociobiology that we can be expected to behave altruistically towards those who share our genes. We also learn that we can be expected to have a capacity for cooperation, conditional altruism, and mutual accommodation of our interests. If such behaviour is called moral behaviour there exists an evolutionary explanation why it exists. However, it is hard to see why evolution should have put altogether impartial moral ideals into our heads. It is more reasonable to assume, I submit, that evolution has provided us with a capacity for understanding that allows us to grasp some basic moral truths. In the final analysis, we may also turn the plausibility of moral claims themselves against various nihilist or agnostic arguments in a manner adopted by A.C. Ewing, Ronald Dworkin and Thomas Nagel.20 But this move presupposes the truth of an objectivist analysis at least of our use of moral language, of course. If our moral opinions are mere emotions, the nihilist claim that there are no moral facts existing independently of our conceptualisation and thought does not contradict any one of our moral opinions. It might be tempting to argue that general moral facts such as the one that it is wrong to torture an innocent child for no reason cannot be grasped, since they lack location in space and time. To be in ‘contact’ with such a fact must be utterly different from, say, observing that there is a cat in front of you. And the content of the act of apprehension, to wit, that it is wrong to torture an innocent child, can thus not form part of any explanation of why you come to hold the belief. But this argument is not convincing. It is certainly true that the fact that it is wrong to torture an innocent child lacks location in space and time. But your contemplating this putative fact does take place at a certain place and time. And part of the explanation why you end up believing this to be a fact might be that it is a fact (together with the fact that you are an ordinary receptive and cleaver person). Compare when you come to believe that 5 + 7 = 12! Harman claims that this cannot be explained with reference to the fact that 5 + 7 = 12: Observation does not seem to play the role in mathematics it plays in physics. We do not and cannot perceive numbers, for example, since we cannot be in causal contact with them. We do not even understand what it would be like to be in causal contact with the number 12, say. Relations among numbers cannot have any more of an effect on our perceptual apparatus than moral facts can.21
This can be seen as the main thrust of Nagel’s book The Last Word. A similar move is made in Ronald Dworkin, “Objectivity and Truth: You’d Better Believe It”. And long before them, this kind of argument was used by A.C. Ewing in The Definition of Good, pp. 32–33. 21 Ibid., pp. 9–10. 20
4.5 Moral Knowledge
55
This does not make Harman a nihilist with respect to mathematics. But, on his account, mathematics is saved indirectly, so to speak. We need maths when we state our theories of physics. But a more straightforward argument is available to the mathematical realist (Platonist). At the right time and place in your early mathematical education, this truth, that 5 + 7 = 12, may well dawn upon you. And the truth of it forms part of a causal explanation of the origin of your belief. Karl Popper has noted this remarkable fact: … the human mind can see a physical body in the literal sense of ‘see’ in which the eyes participate in the process. It can also ‘see’ or ‘grasp’ an arithmetical or a geometrical object; a number, or a geometrical figure. But although in this sense ‘see’ and ‘grasp’ is used in a metaphorical way, it nevertheless denotes a real relationship between the mind and its intelligible object, the arithmetical or geometrical object; and the relationship is closely analogous to ‘seeing’ in the literal sense.22
This is how a Platonist about numbers should argue. This is how a consistent moral realist of a non-naturalist slant must argue, with reference to moral realities, as well. Finally, assuming that there are objective moral (normative facts), making our moral (normative) judgements true or false, how can we ever pretend to have knowledge about these facts? Is there any reason to believe that one opinion rather than another about these facts is any better founded? Can we have justified moral (normative) beliefs without using any faculty of moral perception or intuition ‘utterly different from our ordinary ways of knowing everything else’, to use Mackie’s words?
4.5 Moral Knowledge True and justified moral belief (knowledge) is possible. Let me here just indicate how I think we obtain justified moral (normative) beliefs. We can reach justified moral belief. When we do, and when our justified beliefs are true, we do possess moral knowledge. We go about in roughly the following way, using our ordinary emotional and intellectual capacities, produced by the evolutionary process. We grow up in a society where a certain moral code is taken for granted. At first we concur in taking it for granted. Then some of us come to a point where, from a detached perspective, we come to question what the code urges us to do. We become suspicious of common sense morality, not entirely, certainly, but partly. We find that there are many considered moral beliefs that we are hard put to give up, however, such as the one about not torturing innocent children for no reason. But there are also other beliefs, taught to us, that we find it hard to believe. If we are of a philosophical bent, we try to articulate speculative hypotheses explaining the truth of the
22
Objective Knowledge, p. 155.
56
4 In Defence of Moral Realism
judgements that strike us as plausible and explaining the falsity of those we find suspect. When articulating such speculative moral principles, we go for those that are simple, general, fruitful, and so forth. We select the one or the ones that we claim to give the best explanations of (the content of) our more particular judgements. We also put our particular judgements, our moral intuitions, to scrutiny. We search knowledge about how we may have obtained them. There is now much evidence from psychology and even from neuroscience, including the results of scans of our brains, to turn to. We submit our intuitions to cognitive psychotherapy and retain only those who survive our knowledge of how they may have come to exist, as still subjectively justified. We should rely mainly on those that have been generated by our cognitive capacities, I submit. We should rely on them, even when they run counter to our moral ‘gut feelings’; in particular we should do so if we can understand why it is rational to have these gut feelings, even if, now and then, they lead us astray. In the famous trolley cases, we should rely on our rational insight to the effect that it is better to save lives than to sacrifice lives, even when it means that we have to kill in order to do so. The gut feeling against active and physical killing is rational, however; it is rational in the sense that, in most cases, it is morally correct to rely on it. It is rational in the same sense that it is rational to fear all snakes, irrespective of whether they are poisonous or not. By fearing all we err on the right side and we stay alive. Hence we ought to cherish this kind of primitive attitude in our lives. And yet, in rare cases it is wrong to abide by it, so we should not count on it in an epistemic context. Our knowledge of its origin and function here has, and should have, a debunking function. It is not to be relied on in our pursuit for the moral truth. This is similar to our emotion with respect to snakes. If we want a correct classification of them, into those that are poisonous and those that are innocuous, we should not rely on our gut feelings (our primitive emotions). In this manner we take (the content of) our considered moral judgements or intuitions as evidence for the more general principles. It is crucial here, of course, that we have not simply consciously inferred our more particular beliefs from our general ones, supposed to explain them.23 The parallel to a scientific investigation is obvious. The process sometimes goes the other way round. We find that we have to let go of some more particular and considered moral judgements, in spite of the fact that they have survived cognitive psycho-therapy, since they conflict with principles that we have found have a strong support (they are supported by other particular considered judgements, as well as by methodological considerations; in addition to this, they may just feel intuitively right to us). When the process ends we have reached what John Rawls has famously
I elaborate on this idea in Understanding Ethics, Chapter 9. For literature about the results from scientific studies of the origin of moral intuitions, see for example Greene et al, ‘An fMRI Investigation of Emotional Engagement in Moral Judgment’, and Greene and Haidt, ‘How (and Where) Does Moral Judgment Work?’. The literature on this subject is rapidly growing.
23
4.5 Moral Knowledge
57
called a ‘reflective equilibrium’. And there seems to be genuine progress in this process, allowing us to think that we do come closer to the truth. Deontological ethics takes our gut feelings seriously. Utilitarianism takes them into account too, but gives them a different, secondary place, in a two level approach to ethics. In my opinion, this means that utilitarianism makes a theoretical advance over deontology. It explains the content of our considered intuitions better than deontology, it contradicts deontology, and it explains the relative success of deontology. In moral philosophy we are used to focus on disagreement, but there exists a growing common moral understanding, exemplified with the growing understanding that non-human sentient beings have a moral standing. I would even go so far as to claim that there is something close to a general recognition, that there must be some truth in utilitarianism, if utilitarianism is seen as more of a research program than a definite doctrine, and if it is acknowledged that there may be circumstances where utilitarianism may give the wrong answer (such as situations where its demands are overly demanding, where they mean that there is conflict with deontological constraints and individual rights, and so forth). In those cases where there are no such problematic aspects few would have any problems with adhering to the utilitarian solution of the moral problem. Few are prepared to argue that it is right to make the world a less happy place, unless there are any pressing reasons to do so. Those who are prepared to do so are simply wrong. What are we to say when we find that people, who have thought equally hard about these matters, as we have done, have come to different conclusions? Does this mean that we must give up our endeavour? Now, if they think differently about the torture example given above we are unperturbed by this finding. We set their opinions to one side; their moral opinions are just bizarre, and we try to find an explanation to do with some disorder or distortion within the dissenting person. What then if the dissenting party agrees with us about a hard core of ‘simple’ and obvious moral judgements, but disagrees completely in the ‘organisation’ of morality? This person believes in moral principles that we reject. And we find no obvious flaws in these people’s emotional or intellectual characteristics. They are at least as intelligent and sympathetic as we are. Does this mean that we give up our favoured moral principles, arguing that it would be far-fetched to assume that we, rather than those competent and decent people with whom we are in disagreement, should have reached moral knowledge?24 Obviously, we don’t, and we shouldn’t. To the extent that we come across such disagreement, where we cannot explain in any simple manner why it occurs, we become more humble. We realise that our own way of accounting for what we believe is a uniquely true morality may have gone wrong in many places. This is something we may come to find out in the future. However, each piece of our moral
For a discussion of this possible reaction, see George Sher, ‘But I Could Be Wrong’.
24
58
4 In Defence of Moral Realism
understanding has the argumentative support it has, so it would be irrational for us, given our epistemic goal (to be explained in Chapter 8), to wit, to believe true propositions and disbelieve false propositions, to give it up merely because other people have reached other results.
4.6 Conclusion I do not pretend to give any knock-down argument in this chapter for the thesis that objective moral facts or reasons exist, independently of our thoughts and actions. My main positive argument in defence of moral (normative) realism is the observation that it seems as though our making of some moral judgements, once we have been able to transcend conventional morality, is best explained with reference to moral facts. This is a rather week support. But it is the kind of support one can give for a thesis of this kind. And I do not think any stronger support is really needed. I tend myself to agree here with Thomas Nagel when he wants to show that pain and pleasure provides us with agent-neutral reasons, and starts his argument by saying: In arguing for this claim, I am somewhat handicapped by the fact that I find it selfevident.25
As was also noted above with reference to the same author, we may also turn the content of our moral beliefs – in form of particular judgements as well as in form of theories, and in form of the support we have gathered for them – against attempts to show a priori that they are false (moral nihilism) or that we must be agnostic with respect to them (because of a putative debunking explanation of why we hold them). Incidentally, I am prepared to go even further and claim that, unless moral facts did exist, independently of our way of conducting our lives and passing moral judgements, life would lose ‘meaning’, in an important sense.26 This final claim is of course no argument in defence of the truth of moral realism, or even an argument why it is rational to believe that it is true; or if it is after all a reason for a person who like me wants moral realism to be true in order to find meaning with life, to believe that moral realism is true, this reason has nothing to do with epistemic justification, of which more will be said in Chapter 8. However, the fact that moral realism can give meaning to our lives does indeed indicate why the discussion of it is of the utmost importance.
The View from Nowhere, pp. 159–160. I defend this claim in the concluding chapter of my book Moral Realism.
25 26
Chapter 5
Some Consequences of Moral Realism
5.1 Introduction In the previous chapter I indicated my reasons for my belief in moral (normative) realism. We have already seen that if moral realism is true, and, if there are some moral (normative) facts that have moral explanations, then there exist objective moral reasons. I will here defend some claims in addition to this one. One of them is that the moral nihilist doesn’t have access to moral reasons. This is perhaps not a very surprising consequence of moral nihilism. We have seen that it follows already from nihilism with respect to moral principles (particularism). Another consequence is that, somewhat more unexpectedly, the same goes for the moral naturalist. Finally I will discuss whether the notion of an objective moral reason, which I have defended, could help us to answer what Christine Korsgaard has called the normative question. I will argue that it can.
5.2 Moral Explanation In our lives, when faced with hard choices, we seek, if we take things seriously, practical guidance from moral theory. We try to find out what we ought to do. But, if we are of a reflective bent, we also want to understand why we ought to do what we ought to do. This means that we need to get the relevant facts right. This takes all sorts of skills. But this is only the beginning of our pursuit for an answer to our question. We also need to obtain a proper moral understanding of our situation. The only way to arrive at such an understanding is through a moral explanation of our obligations. We need an explanation of our actual obligation in terms of basic moral principles and relevant facts. If non-naturalistic moral realism is true, then such understanding is available to us. If there are true moral principles, the truth of which is independent of our conceptualisation and beliefs, then we can use them, together with relevant factual information, as premises of a deductive argument leading up to a conclusion about what it is we ought to do. In principle we can know then, not only what our obligations are, but also why we ought to do so and so.
T. Tännsjö, From Reasons to Norms, Library of Ethics and Applied Philosophy 22, DOI 10.1007/978-90-481-3285-0_5, © Springer Science + Business Media B.V. 2010
59
60
5 Some Consequences of Moral Realism
We know why we ought to do so and so once we have found a way of explaining, morally, our obligation. For example, if utilitarianism is true, then we can find an explanation why we ought, say, to give away much of our wealth, if we happen to belong to the rich part of the world population. By giving away much of our wealth we promote the general happiness. And this, according to utilitarianism, is what we ought to do. In morality just as we do in science we can produce nomological ‘covering law’ explanations. In morality we deduce our moral obligations from factual statements (by giving away much of our wealth we promote the general happiness) together with moral principles (we ought to promote the general happiness). If Kantianism is true, rather than utilitarianism, we deduce our obligations in a similar manner. Why do I ought to do F? Because I have promised to do F. And we ought to keep our promises. Kant also adds that the reason we ought to keep our promises is that by so doing, we abide by the Categorical Imperative.1 The Categorical Imperative is not itself a moral principle, however. It is rather a device that allows us to arrive at correct moral principles. So it does not in itself constitute a moral reason, or so I understand Kant. On this interpretation of Kan we can make best sense of his claim that we should act from duty. The immediate practical importance of this kind of moral understanding should be obvious. However, not only is it of interest because it allows us, in practical situations, to arrive at answers to the question what we ought to do. The model also provides the clue to the question how we can improve our moral understanding. By applying moral principles, in the manner indicated, to new situations, we may sometimes arrive at answers we find hard to accept. We can then move on to an investigation of other putative moral theories to see if some of these competing theories can explain the recalcitrant ‘data’ we have come up with. If some competing theory does, then we are allowed to say that these data are evidence for this competing moral theory. We have made an inference to the best explanation. This model of moral explanation, leading to (improved) moral understanding, is available to the moral realist, then. The moral realist has access, at least in principle, to objective moral reasons. However, if we take ‘nihilism’ to be the simple negation of the claim that there are moral truths obtaining independently of our beliefs and conceptualisation, the model is not available to the moral nihilist. Nor is moral explanation available to the moral naturalistic realist, either of the classical semantic variety, or of the new and fashionable Cornell variety. At least, that is what I will try to show.
5.3 Nihilism and Moral Explanation Moral nihilism comes in many varieties. One is the version put forward by Mackie, claiming that all positive moral assertions are false. On such a view, there is no room for moral explanation. If part of the explanans (the moral principle) of every
See his Groundwork of the Metaphysics of Morals.
1
5.3 Nihilism and Moral Explanation
61
moral explanation is false, all moral explanations must be spurious. The same is true, and even more obviously so, of the explanandum, of course. We cannot explain what does not obtain. But what about moral nihilists who adopt an expressivist semantic theory about moral language, can they consistently accept that there exist moral explanations? I think not. According to expressivists, when we use moral language we express emotions or attitudes, not propositions capable of being true or false. A moral opinion is an attitude. According to the expressivist, there are no moral facts. This is not to say that if expressivism is true, then we cannot use the word ‘true’ or ‘fact’. Even if expressivism is true we can say, for example, that it is true and even that it is a fact that murder is wrong. However, by so saying, given an expressivist understanding of what we are saying, we are only expressing an attitude with respect to murder.2 Certainly, we can say consistently if expressivism is true that what a person did was wrong ‘because’ it was an act of murder. And we may add that, as a matter of fact, murder is always wrong, and it makes sense from the point of view of expressivism. However, if expressivism is true, once again, by so doing we are just expressing an attitude. Which attitude? This is not entirely clear. What does it mean to have a negative attitude with respect to an abstract class of actions, such as murder? What does it mean, according to expressivism, when we claim that, if an action is an act of murder, then it is wrong? Or, what does it mean if expressivism is correct when a utilitarian claims that if an action doesn’t maximise the sum total of happiness in the universe, then it is wrong? Is it possible to have a positive attitude with respect to the class of all actions that maximise the sum-total of happiness in the universe? This boggles the mind. Could the expressivist say that, by saying that murder is wrong, I express the attitude: for all actions, x, if x is an act of murder, then boo for x? And, could the utilitarian claim be treated in a similar vein? I think not. What kind of an attitude is this? In particular, how are we to understand the use of the connectives ‘if’ and ‘then’ in a sentence such as this one? Certainly, we cannot understand them with reference to truth-values. Or, can we? But if we can, then we seemed to have moved all the way from expressivism into the moral objectivist camp again. This is of course the root of well-known Frege-Geach problem.3 Another possibility would be to take the expressivist to say that what I am expressing, when I say that murder is wrong, is a complex negative attitude with respect to accepting that an action is an act of murder and not having a negative attitude with respect to it. But even if this may be psychologically possible it
See C.L. Stevenson, Facts and Values, for a defence of this view of truth in ethics. The problem got its name from Peter Geach in his 1960 article ‘Ascriptivism’ where he attributes it to Frege.
2 3
62
5 Some Consequences of Moral Realism
doesn’t quite seem to capture what we want to say when we say that murder is wrong. We now express disapproval of not disapproving of murder, and that is not the same thing as disapproval of murder as such. And irrespective of how in detail we are to understand the expressivist it is difficult to see that, according to expressivism, a person who claims that if an action is an act of murder, then it is wrong, acknowledges that a certain action is an act of murder, but denies that it is wrong, makes any logical mistake. The expressivist may have recourse to a notion of consistency, but this notion of consistency has little to do with logic. May we say that such a person violates supervenience? I think not. From an expressivist point of view, and strictly speaking, supervenience makes no sense. Supervenience is the view that moral properties somehow supervene upon natural properties. If there are no moral properties, we cannot even formulate the supervenicence thesis. If the expressivist wants to speak about supervenience, he or she must conceive of it as some kind of psychological phenomenon. But then it is of little avail if we want to save ordinary logical relations between moral propositions. All this means that the expressivist lacks the notion of a genuine moral explanation. By saying that a certain action is wrong because it is an act of murder the expressivist doesn’t explain any obligation. What is going on here, according to the expressivist, is the mere expression of attitudes or emotions. According to expressivism, we do not derive the conclusion (the explanandum that the action is wrong) from the explanans (that the action is an act of murder and that all acts of murder are wrong) by the means of any logical inference. Moreover, when the expressivist cannot specify any explanation in particular, the expressivist cannot consistently say such things as that a moral explanation does exist why one should do so and so. For, according to the expressivist, there are no moral facts, no moral propositions capable of being true or false, and hence no moral explanations proper. The moral realist, on the other hand, can consistently say such things. The moral realist can claim consistently that there exists an explanation why one ought to perform a certain action, even if the moral realist doesn’t know which this explanation is (perhaps God or Mahatma Gandhi or Nelson Mandela has informed her about the existence of it, without spelling out its exact nature). The moral realist has access to, and can hence quantify over, objective moral reasons.
5.4 Quasi-realism Is there a middle position? What about quasi-realism? Can a quasi-realist accept that there are moral explanations? In order to be able to answer this question we need to understand what quasi-realism amounts to. In particular, it is of importance to understand how a quasi-realist conceives of truth in ethics. This is how Simon Blackburn has put the position:
5.4 Quasi-realism
63
The simplest suggestion is that we define a ‘best possible set of attitudes’, thought of as the limiting set which would result from taking all possible opportunities for improvement of attitude. Saying that an evaluative remark is true would be saying that it is a member of such a set, or is implied by such a set.4
However, this seems to be a characterisation of the meaning of evaluative remarks, not of the word ‘truth’. According to this characterisation, evaluative remarks are remarks about a certain ideal and hypothetical set of attitudes. But this means that quasi-realism collapses into (naturalistic) realism. To say that an action is wrong is then roughly to say that an ideal observer would dislike it. Note how similar Blackburn’s implicit semantic naturalism is to Michael Smith’s explicit semantic naturalism. According to Smith, to say of an action that it is right in circumstances C means to say that it possesses features ‘… we would want acts to have in C if we were fully rational, where these wants have the appropriate content’.5 Something similar seems to be going on in Chrispin Wright’s minimalism. He takes his minimalism to be capable of saving logic and truth in morality while giving up on realism. However, when the claim is made that truth is epistemically constrained within the moral domain, one must wonder how this is possible. Since there is just one notion of truth, and truth is not epistemically constrained in physics, the fact that it is in morality cannot have anything to do with the notion of truth as such. It must rather have everything to do with the meaning of moral utterances. So when Wright claims that, in morality, the notion of superassertability can play the role of truth, and when he claims that ‘… the morally true is that which can be morally justified and which then retains that justification no matter how refined or extensive an additional consideration is given to the matter’,6 then the only way of making sense of this, it seems to me, at any rate, is to take him to endorse a naturalistic (and hence realist) view of moral facts. I take it for granted that the criteria of justification can be specified in naturalistic terms. Otherwise Wright would be moving in a circle. It seems as though quasi-realism collapses into semantic moral naturalism. So the question we need to address is whether semantic moral naturalism is compatible with moral explanation. On this understanding of moral terms, it is obvious that moral explanation of the kind here discussed will not be available to us. For example, if ‘X ought to be done’ means the same as ‘X is conducive to general happiness’, then we cannot explain why we ought to perform a certain action with reference to the fact that it is conducive to general happiness. This would be as if we tried to explain why a man is not married with reference to the fact that he is a bachelor. The same goes for the idea of a ‘best possible set of attitudes’.
Blackburn, S., Spreading the Word. Groundings in the Philosophy of Language, p. 198. The Moral Problem, p. 185. 6 ‘Truth in Ethics’, p. 10. 4 5
64
5 Some Consequences of Moral Realism
5.5 Cornell Realism and Moral Explanation Some philosophers seem to slide into classical, semantic naturalism, as we have seen. In its classical semantic form, it amounts the claim that some crucial moral term such as ‘ought’ is synonymous with some natural term such as ‘is conducive to general happiness’, or it is claimed that the sentence ‘You ought to do X’ means the same as ‘X would be approved of by a person who had obtained the best possible set of attitudes’, and so forth. And this approach provides no room for moral explanation. Furthermore, it is vulnerable to Moore’s open question argument. On this view, it is no longer an open question whether an action that would be approved of by a person who had obtained the best possible set of attitudes was right. But this question seems to be open.7 When naturalism is defended in present discussions, however, it is usually defended in the Cornell style, where the assertion is not that moral terms and natural terms are synonymous, but where the claim is rather that even though they have different meaning, these terms refer to the same (natural) property. Such a move avoids the open question argument in its classical form. This is partly what has motivated the move from a semantic to an ontological version of naturalism. Moreover, should such a reduction of moral properties to natural ones succeed, this would grant morality a respectable place in our worldview. This is different, then, from an eliminative version of naturalism, where we give up the idea that there are moral properties and moral explanations. However, does such an ontological naturalist reductivist account, cast in terms of identity of properties, allow that we make moral explanations? I think not. Not only philosophers at Cornell hold this kind of ontologically naturalistic view; another important example is Tom Scanlon. Here I will focus on his statement of the view, since Scanlon does himself point at an important problem facing it. Scanlon claims that his theory, as well as utilitarianism and other moral theories, should be understood as conjectures about with which property the property of wrongness should be identified: That is to say, the property of moral wrongness can be identified with a certain normatively significant property which is shared by actions that are wrong and which accounts for their observed normative features …8
It seems clear that, on this account, we cannot explain why a certain action is wrong with reference to the fact that it has bad consequences, is a lie, and so forth. This is brought out very clearly with a comparison Scanlon himself makes, and which explains why I chose to focus on his statement of the kind of naturalism in question. On Scanlon’s view, the claim made by the utilitarian, that an action is wrong if it has an alternative with better consequences, should be understood in analogy with
I follow tradition and attribute this argument to Moore, even though it is doubtful whether he ever stated it. See Fred Feldman’s fascinating ‘The Open Question Argument: What it Isn’t and What It Is’. 8 Scanlon, T., What We Owe to Each Other, p. 12. 7
5.5 Cornell Realism and Moral Explanation
65
the claim that something is gold if it has a certain atomic structure. But on this account, just as we cannot explain why something is gold with reference to the atomic structure exhibited by it, we cannot explain why an action is wrong with reference to the fact that it has an alternative with better consequences. To block this move would be to distort utilitarianism, however. For, as we have seen, this kind of explanation, now rendered impossible, is exactly the kind of explanation the utilitarian is after. The same objection can be directed at all sorts of moral realism of the Cornell variety, granting that moral and natural terms differ in meaning (hereby dodging the criticism stemming from G.E. Moore’s ‘open question’ argument), while still insisting on the claim that moral properties are identical to some natural and familiar properties. In order to make clear what kind of explanation the utilitarian (or any moral theorist) is after, and in order to appreciate what we are missing out if we construe their theories as conjectures about the identity of natural and moral properties, it might be useful to return to the analogy with science. To explain why an action is wrong, say, with reference to the fact that it has an alternative with better consequences, or is a lie, is like explaining why a certain bird is black with reference to the fact that it is a raven. It is certainly true that a utilitarian is prepared to say that an action is wrong if and only if it has an alternative with better consequences, while we are not prepared to say that something is a raven if and only if it is black. However, even in the moral case, the explanatory relation obtains only in one direction. An action is wrong because it has a better alternative, but it does not have a better alternative because it is wrong. This is even more obvious with a claim such as Kant’s, that an action is wrong if it is a lie. The fact that an action has an alternative with better consequences or is a lie is what makes it wrong, but the fact that it is wrong does not make it an action, which has an alternative with better consequences or a lie. I have pointed to the finding that naturalism excludes moral explanation as a consequence of naturalism. However, it can also be turned into an argument against naturalism. For how do we establish a reduction of one property to another? How do we find out for example that water is identical to H2O? We scrutinise law-like connections, with ‘water’ and ‘H2O’ respectively in the antecedent, with a common consequent, I submit. We map the laws cast in terms of water onto laws cast in terms of H2O. If we succeed in doing so we have taken a first step towards reduction. We then assume that there are properties (in general). We observe, moreover, that H2O lends itself to further generalization, a kind of generalisation that was not possible with water. We then conjecture that water is identical to H2O. Is there a parallel to this in ethics? There may be some putative laws that can be mapped onto laws cast in the suggested new terminology, I admit. For example, if Sturgeon9 and I (see the previous chapter) and others are right when we claim that the fact that a certain action is right may in some circumstances explain why we believe it is right, then this law, with a moral term in its antecedent, could perhaps
See his ‘Moral Explanations’.
9
66
5 Some Consequences of Moral Realism
be mapped onto a naturalised version of itself. Say that utilitarianism is correct. Then, in some cases, the fact that an action maximises happiness can explain why we believe it is right. So far, so good, but the more important, and basic, normative principle to the effect that an action is right if and only if it maximises happiness, cannot be mapped onto any naturalised version of itself. Quite to the contrary, this normative (utilitarian) principle gets eliminated by the reduction. But a reduction eliminating rather than saving basic laws cannot be correct! So reductionism seems to be an unstable position. An outright eliminative naturalism does not face this problem of course. It bites the bullet and denies that there are any true moral principles. This flies in the face of our moral phenomenology, I submit, but it is a consistent position. It is against the reductivist ambition, that is the ambition to identify moral properties with natural ones in order to grant them a respectable place in our worldview, I think I have here given a sound argument. I mentioned another problem with Cornell realism. In our moral phenomenology we think there is a direction in moral principles, similar to the direction in a scientific law. The property mentioned in the antecedent gives rise to the property mentioned in the consequent. If we identify moral properties with natural ones, we cannot retain this asymmetry. This problem is perhaps not insurmountable, however. The Cornell realist can try to avoid it by arguing, not that moral properties and natural properties are identical (since no one has really performed the reduction in question yet, we cannot know for sure about this), but that moral properties are ‘realised’ or ‘constituted’ by, or being ‘homeostatic clusters’ of, many different natural properties; here we seem to have a ‘direction’ from the natural to the moral. But it is from clear why, if there are properties in the first place, there cannot exist disjunctive properties just as well as simple positive ones, and then, if Cornell realism is true, we will be able to identify ethical properties with disjunctive natural properties. And once again we are faced with the problem that the asymmetry in moral principles, as we know them from our moral phenomenology, is lost. Furthermore, even if we set this problem to one side, and even if different properties do make actions wrong in different situations – if the reduction without identification gets through – we still lose the possibility of explaining in these situations why these actions are wrong. If in one situation an action is wrong, and its wrongness is here constituted by its being a lie, while, in another situation, an action is wrong, and now its wrongness is constituted by its not maximising the sum-total of happiness, and if we cannot account for, in a general form, when an action is wrong (we cannot say what it is in the former situation that makes it wrong to tell lies and what it is in the latter situation that makes it wrong not to maximise the sum-total of happiness), we end up with a kind of moral particularism, rendering moral explanation impossible. For a person who believes that there are true moral principles, and for a person who believes that utilitarianism comes close to the truth, this, of course, is not a very attractive position.10
For further references, see D.O. Brink’s instructive, Moral Realism and the Foundations of Ethics.
10
5.6 The Normative Question
67
5.6 The Normative Question Moral reasons for action exist, we have seen, if they do exist, independently of whether we know about them or not. If this notion of a moral reason is accepted, does it preserve the ‘normativity’ of morals? Does this notion of a moral reason allow us to answer the question why we should do what we ought to do? This is how Christine Korsgaard has famously formulated what she has called ‘the normative question’: If I claim that you ought to face death rather than do a certain wrong action, I had better be prepared to back that claim up with an account of what makes the action wrong which is powerful enough to show that something worth dying for is at stake. But really this demand on moral theory is always there. Even when the claims of morality are not so dramatic, they are pervasive in our expectations of ourselves and each other. So these claims must be justified. That is the normative question.11
Of course, many thinkers have raised this question before her. Here is another example, Alan Gewirth, who asks: Why should one be moral, in the sense of accepting as supremely authoritative or obligatory for one’s actions the requirement of furthering or favorably considering the important interests of other persons, especially when these conflict with one’s own interests?12
Now, on the account here given, an answer to the normative question could be that, unless you sacrifice your life, many other people will get killed. And, in situations such as this one, you ought to act so as to sacrifice as few lives as possible. So you have a good moral reason to sacrifice your life. Of course, this answer could be questioned. Why, in situations such as this one, sacrifice as few lives as possible? Why not save one’s own life? One explanation may be that by sacrificing as few lives as possible you maximise the sumtotal of happiness in the universe (utilitarianism). This is why you ought to sacrifice your life. Is this kind of answering the question satisfactory? It might be felt that something is missing. How do I get from my acknowledging that sacrificing my life is what I ought to do, and from acknowledging why this is what I ought to do, to a reason to do it? As Henry Sidgwick famously put this point. Some people feel that: … the answer which we really want to the question ‘Why should I do it?’ is one which does not merely prove a certain action to be right, but also stirs in us a predominant inclination to do the action.13
Well, my answer to these people is that, this is how I have defined what it means to have a (moral) reason to perform an action. It means no more and no less than that
The Sources of Normativity. Alan Gewirth, Reason and Morality, p. 3. 13 The Methods of Ethics, p. 5. 11 12
68
5 Some Consequences of Moral Realism
there exists an explanation why I ought to perform the action in question. And I find no other way of conceiving of the quest for a reason – unless what I want is a motive for doing the action in question. Moreover, I share Sidgwick’s belief that … when a man seriously asks ‘why he should do’ anything, he commonly assumes in himself a determination to pursue whatever conduct may be shown by argument to be reasonable …14
5.7 Broome’s Answers Could we find any stronger connection between the belief that I ought to do X and my intention to do X? John Broome, who is not only a renowned economist and philosopher, but also a skilled sailor, has suggested some ways in which we can. According to his first proposal, the inference from ‘I ought to tack’ to ‘I shall tack’ is ‘… an example of correct normative practical reasoning’.15 However, he gives no positive arguments for this idea, and it is hard to make sense of it. Let us see what it might amount to. One idea would be to take the relation between ‘I ought to tack’ and ‘I shall tack’ to be one of implication, but then we immediately run into difficulties. For the inference from the premise to the conclusion cannot possibly be valid. I take it that a valid inference is such that, necessarily, if the premises of it are true, then the conclusion is true. But ‘I shall tack’ is the expression of an intention or a resolution, and hence not a proposition that can be either true or false. However, even if the inference is not valid, one may yet be tempted to believe that, necessarily, if someone really believes he ought to tack, then he will also form, or have formed, the intention to tack. Does this make sense? Well, it does if we take for granted the expressivist (prescriptivist) claim that normative opinions (such as the one that I ought to tack) are resolutions. However, this move is not open to Broome (or to myself), since we believe that there are (true, objective) moral reasons for action. A resolution cannot function as part of any moral explanation (neither as part of the explanans nor of the explanandum). Broome has thought further on the problem, however, and made another suggestion: we should think of the move from my believing that I ought to X to my forming the intention to X to be something required by rationality. And we may conceive of rationality as somehow normative.16 Are we then allowed to say that, if a person believes he ought to X, then he ought also to form the intention to X? This suggestion is not plausible, as Broome points out himself. It means that, a lazy person working in a concentration camp and
Ibid. ‘Normative Practical Reasoning’, Proceedings of the Aristotelian Society. 16 Wedberg lectures given at Stockholm University in Mars, 2004 (unpublished). 14 15
5.7 Broome’s Answers
69
believing that he ought to kill as many of the prisoners as possible, has an obligation to adopt the intention to do so. We must avoid a version of the requirement that allows this kind of detachment. Could we say then that it is the combination of the belief that one ought to do X and a lack of the corresponding intention that should be avoided? In the case of the person working in the concentration camp, what he ought to do is not to form the intention to kill as many prisoners as possible but, rather, to abstain from having both the belief that he ought to kill these people and abstaining from the intention to kill them. This is how Broome seems to think about the problem at present. But note that, from the point of view of rationality, there are then two ways of achieving consistency; either by giving up the belief that one should kill as many as possible or to be more ambitious in killing prisoners. And from the point of view of rationality, one of them is not superior to the other. Or, if one of them is superior, from the point of view of rationality, this must have something to do with some other requirement of rationality.17 The suggestion still lacks plausibility. We may of course stipulate such a use of the word ‘rational’, but if we do, it is hard to see why we should not sometimes allow ourselves to be, in this sense, ‘irrational’. And I can’t help feeling that this kind of stipulation would be very awkward. Suppose we find that a person is giving away all he owns to charity, living a miserable life. We ask ourselves why he is doing this. It would not be a plausible explanation to say that he does this because he believes that this is what he ought to do. Many people believe that this is something they ought to do and yet they don’t. Why is this person different? It is of little avail to offer for an explanation: He is rational. Quite to the contrary, some would say that the reason that they don’t give away everything they own has to do with the fact that they are rational. They know that they ought to give away everything they own, but they don’t want to do so, and they don’t to it, and by not doing it they are exhibiting an ordinary kind of instrumental rationality. Their actions, beliefs and desires fit each other in the manner we usually characterise as ‘rational’. So how should we explain that a person gives away everything he owns, if he does, if not with reference to his rationality? Well, I do not deny that this person too may be rational. But the explanation of his unusual behaviour lies elsewhere; it lies in his unusual sense of his moral duty. Suppose, however, that we want to insist on the view that it is irrational both to believe that one ought to do something and not form any intention to do it, and suppose one wants to insist that there is some connection between rationality and morality, at what price should one avoid this kind of relation between one’s beliefs and intentions? Would it be reasonable to torture one (just one, for a short moment) innocent child, say, if it meant that one would avoid one instance of this kind of ‘irrationality?’
I owe this qualification to a comment from John Broome (December 2008).
17
70
5 Some Consequences of Moral Realism
I think few would think so. But if not, why not? At what price should one avoid, for its own sake, this kind of ‘irrationality?’ When pressed, most people18 would answer this question in a manner that shows that, even if there is some value in itself in being rational, it is an utterly tiny value. Then there is little sense of making any fuss about it.
5.8 Korsgaard’s Answer What about Korsgaard herself, then, does she find another plausible way to answer the normative question? I think not. This is the answer to the question favoured by Korsgaard. According to Korsgaard, my reason to sacrifice my life may be that, by so doing, I avoid a fate worse than death – the fate of becoming alien to myself. This is how she puts her point: … morality can ask hard things of us, sometimes even that we should be prepared to sacrifice our lives in its name. This places a demanding condition on a successful answer to the normative question: it must show that sometimes doing the wrong thing is as bad or worse than death. And for most human beings on most occasions the only thing that could be as bad or worse than death is something that for us amounts to death – not being ourselves anymore.19
What does it mean to say that morality ‘requires’ of us that we sacrifice our life? I suppose it means that, as a matter of fact, this is what we ought to do. I agree, then, that a successful answer to the normative question must show that sometimes doing the wrong thing is as bad as, or worse than, death. A way of showing this can be to point out that, if I do the wrong thing, many people will get killed. Korsgaard prefers another answer to the question: by doing the wrong thing I would not be myself anymore. These seem to me to be competing explanations of my (assumed) obligation to sacrifice my life. Which explanation is the better one? My sympathy lies with the former explanation, referring to the fact that several lives are saved and presupposing that we ought to save as many lives as possible. However, be that as it may, it doesn’t seem to me that Korsgaard has been able to point to an explanation that is different in kind from this one. It is not quite clear what her answer amounts to, but it seems to presuppose some kind of ethical egoism. So we have to do research into normative ethics to find out whether I’m right when I say that we ought to save as many lives as possible or whether she is right when she proposes that we ought to preserve our own identity (even when it means that we have to sacrifice our life). We have to see which one,
Including John Broome who writes, in correspondence (May 2004), that he has ‘no formed views about the relation between rationality and normativity. He has made the same claim recently in conversation. 19 The Sources of Normativity. p. 16. 18
5.8 Korsgaard’s Answer
71
among these and, certainly, many other competing hypotheses, provides us with the best explanation of our judgement that, in the circumstances, we ought to sacrifice our life. However, Korsgaard seems to think that these answers to the normative question are not of the right kind. They are somehow off the mark: If the realist is not an intuitionist … he can go back and review the reasons why the action is right.… We need to remind ourselves that the action promotes pleasure, or meets a universalizability criterion, or fosters social life. But this answer appears to be off the mark. It addresses someone who has fallen into doubt about whether the action is really required by morality, not someone who has fallen into doubt about whether moral requirements are really normative.20
I don’t think these answers are off the mark. However, if moral realism is false, as Korsgaard believes, then they are not available, of course. Note further that, if these answers are off the mark, the same thing seems to be the case with Korsgaard’s own answer to the question. If it is off the mark to say that I ought to sacrifice my life since, if I don’t, I won’t foster social life, then, why is it not off the mark to say that I ought to sacrifice my life since, if I don’t, I will lose my (sense of) identity? I have to admit that I do not understand what further concern Korsgaard is here referring to. How is it possible to doubt whether ‘moral requirements are really normative’? What does it mean to say that morality requires of me that I sacrifice my life if it doesn’t mean that I ought to sacrifice my life? This statement is normative. At least given moral realism, this question is puzzling. The very idea that morality requires something from me is strange to a moral realist. With respect to conventional morality we may pose the question whether its requirements are valid or not. However, when I realise that I ought to sacrifice my life in order to save others this is not, given moral realism, something required by anyone or anything (such as a conventional system of rules) of me. This is an obligation I am under, this is something I am required to do, and there exists a perfect (true) explanation why I ought to do what I ought to do. I ought to save the others, since, in situations like this one, this is what people ought to do, simpliciter (and not ‘according to morality’ or anything of the kind). Admittedly, unless I am of an authoritarian bent, even if I know this, even if I know that I ought to sacrifice my life, I may want to know why I should do it. The fact that I ought to sacrifice my life is no reason for me to sacrifice my life. So there is room for a request for reasons. But then my notion of a ‘moral reason’ seems to fulfil its purpose. I can ask for a moral explanation why I ought to sacrifice my life. Of course, this is a somewhat open-ended request. Why ought I to sacrifice my life? Well, by doing so I save two lives. Why save two lives rather than one (my own)? Well, we ought to save as many lives as possible. Why save as many lives as possible? Well, in a situation such as this one, it means that you maximise the sum-total of happiness. Why maximise the sum total of happiness?
Ibid. p. l8.
20
72
5 Some Consequences of Moral Realism
Now it doesn’t seem appropriate to go on with this kind of quest. Yet, this does not mean that we have come to an end. What we now want are not further arguments of a more general kind, explaining the more particular ones, but evidence for the truth of the utilitarian principle. How does it compete with other putative moral principles? We need to dig deeply into normative ethics if we want to find out about this. We have to put the various putative moral hypotheses to test in applied ethics and in various different thought experiments. Furthermore, we need to learn more about the origin of the moral intuitions, against which we ‘test’ the different hypotheses in search for the best explanation of the intuitions that survive critical examination. I described this method of moral investigation in the previous chapter. All this makes sense to me; I suppose it makes little sense to Korsgaard, who doesn’t believe in moral realism. However, the claim she makes, that this kind of justification is ‘off the mark’, if available, is not correct. It is not available if moral realism is not true, of course, but if moral realism is true, and it is available, it does provide us with a perfectly comprehensible and adequate answer to the normative question. When Korsgaard claims that this kind of answer to the normative question is not adequate, what does she mean? Is perhaps Korsgaard’s point that we need an answer to the normative question such that, once we believe in it, we are motivated to do what we ought to do? Is her point that a moral reason must also (necessarily) function as a Humean motivating reason? But that kind of requirement is unreasonable. It is fine when a (true) moral reason has this function, in the sense that, when it is held, the person holding it happens to be motivated to act on it, but not something we can assume will always be the case. For, more often than not, we are not doing what morality requires of us (or, as I prefer to put the point, we are not doing what we ought to do). We know from experience that, most of the time, instead of doing what we ought to do, we succumb to our laziness, cowardice, egoism, or to convention, and our actions are simply indefensible. Most of us, at least sometimes, have some inclination to behave in a way that, from the point of view of morality, is not too bad, but this inclination is only one among many other ones and, certainly, for most of us, most of the time, not the strongest one. And there seem to exist people who lack it altogether. A similar response seems to be warranted to Alan Gewirth’s own attempt to answer the normative question. He tries to show that, at least all human rational agents have motivating Humean reasons to take other people’s interest into account; simply by performing actions ourselves, we accept a certain moral outlook: … evaluative and deontic judgements on the parts of agents are logically implicit in all action; and when these judgements are subjected to certain rational requirements, a certain normative moral principle logically follows from them.21
Alan Gewirth, ibid., pp. 25–26.
21
5.8 Korsgaard’s Answer
73
Few believe that Gewirth succeeds in establishing what he here claims. However, be that as it may, it is clear that, even if he did succeed, what he would have shown is not that there is a special type of reason beside moral and motivating Humean ones, as here defined, but merely that, contrary to what some have believed, we all (at least to the extent that we are rational agents) have motivating Humean reasons, in certain situations, to make sacrifices for others. Stephen Darwall makes a more cautious claim of the same nature. Like Korsgaard, he doesn’t believe that there are true norms that can explain our obligations. And yet, like Korsgaard and Gewirth, he doesn’t want to rest satisfied with the idea that the only kind of justification for an action that exists is Humean in nature. According to Darwall, a moral demand, such that the demand that we should sacrifice our lives, is justified for us if it represents what we would want if we were impartial, were behind a veil of ignorance regarding specific features and ends, but with full general information otherwise.22 This looks like an idealised version of a Humean reason for action, however. I can’t help feeling that this doesn’t really answer the normative question. Why should I act on the norms I would accept under certain ideal circumstances? This makes sense only if I have reasons to believe that what I would accept under ideal circumstances would be a true morality. But this means that what I would believe under ideal circumstances is really irrelevant to my quest for an answer to the normative question. I do not want to be informed, say, that under ideal circumstances I would believe in utilitarianism. In order to make a reference to utilitarianism when I answer the normative question, I want to know that utilitarianism gives a true account of morality. This means that I want evidence for utilitarianism, not information about situations in which I and other people would believe in it. Given moral realism, such evidence, for or against utilitarianism, exists. It exists in the inferences we make to the best explanation of our considered moral intuitions. If I realise that, in a situation, I ought to sacrifice my life, then it is natural if I ask for a reason to do so. Korsgaard is right about this. To the extent that there are things we ought to do, and to the extent that there exist explanations of our obligations, answers to this question exist; in Korsgaard’s words, it is possible to ‘justify’ the claim that I perform a certain action. The moral realist has an answer to the normative question. Neither a nihilist nor a ‘constructivist’ of Korsgaard’s and Darwall’s slant can answer it, however. What they present with reference to what we would want under certain ideal circumstances is a poor ersatz for an answer to it. However, even if an answer to the normative question can be given, given that there are true norms accessible to us, there is no way, pace certain suggestions made by John Broome, to reason from an obligation to an intention, and it is not irrational not to adopt the intention to do what one realises one ought to do. Nor is it irrational to do something because one wants to commit an immoral act.
Stephen L. Darwall, ‘Automist Internalism and the Justification of Morals’, p. 263.
22
74
5 Some Consequences of Moral Realism
If one doesn’t intend to do what one realises that one ought to do, then this means only that one is lacking Humean, motivational reasons to do what one ought to do. This does not mean that one is lacking moral reasons to do it. One’s moral reason to perform the action in question, the ‘justification’ – to use Korsgaard’s word – of the claim that one ought to perform it, is identical to the explanation of why one ought to perform it.
5.9 Conclusion In this chapter and the one that went before it I have set the stage for what is to come. I have noted that many people who have considered the question what types of practical reasons there are have come to the belief that there is such a thing as Humean reasons (for action). These reasons for action are dependent upon our individual desires or goals. Some people who have considered the question what types of practical reasons there are have also come to the belief that, besides Humean reasons for action, there exist objective moral reasons for action. I have sided with these latter people and roughly stated my arguments for doing so. And, irrespective of whether we believe or do not believe that objective normative (moral) reasons exist or not, we need the notion of such reasons – if only to render possible the rejection of their existence! However, I believe there are objective normative (moral) reasons, and the next few chapters rely on the assumption that there are such reasons. This assumption prepares the way for what is the main thrust of the rest of the argument of this book: my attempt to reject practical relativism by showing that there exists one, and exactly one, source of normativity. In the chapters that follow (Chapters 6–9) I deny that there is any such thing as prudential reasons proper, reasons from justice, reasons from aesthetics, proper reasons to believe, or proper reasons to feel; all these putative additional types of practical reasons can be ‘reduced’, either to Humean, or to moral reasons. I will now set out to show how such a reduction could take place.
Chapter 6
Reasons from Prudence and Rationality
6.1 Introduction It is obvious that Henry Sidgwick is a philosopher who shares my view that, besides Humean reasons, there exist also objective and categorical moral reasons, transcending the existing conventional common sense morality. He doesn’t use the term exactly as I do, of course, but he believes there are objective moral obligations, existing independently of our beliefs about them, and he does believe they could be explained – with reference to hedonistic utilitarianism (the principle of Rational Benevolence). However, it also seems as though Sidgwick believes that, beside these two types of practical reasons, there also exists a third one, which may come to conflict with our Humean and moral reasons for action. He speaks here of reasons of ‘rationality’, but I think we could just as well use the phrase ‘prudence’. In a similar vein, some philosophers have suggested that besides objective and categorical moral reasons there are also reasons to do with justice, aesthetics, epistemology or axiology, which must be fit into a full picture of all the obligations that exist. In this chapter I will first examine the idea that rationality or prudence may be a proper source of obligation, and in chapters to come I go on to discuss whether justice, aesthetics, epistemology or axiology, could provide us with such an additional sui generis source of normativity. My claim will be that there is no good reason to accept the kind of practical relativity, which is implied by the claim that there exist several different sources of normativity. There exists exactly one source of normativity. And this source is the one I speak of as the moral one. It should be noted that what I am discussing here are what could be called ‘heteronomous’ moral dilemmas, dilemmas arising out of putative conflicting sources of normativity. I will not discuss what one could call ‘autonomous’ moral dilemmas arising within morality. I do believe there are such dilemmas of a special kind. I may have an obligation to invite a friend and settle a dispute with her. At the same time I may have an obligation not to invite here since, if I do, I will quarrel with her rather than settling the dispute. I can satisfy each of these obligations but not both of them.
T. Tännsjö, From Reasons to Norms, Library of Ethics and Applied Philosophy 22, DOI 10.1007/978-90-481-3285-0_6, © Springer Science + Business Media B.V. 2010
75
76
6 Reasons from Prudence and Rationality
In order to defend this possibility I will have to reject the idea of agglomeration in deontic logic, but I am happy to do so. A moral realist should be sceptical with respect to deontic logic. If the properties of rightness, wrongness and obligatoriness are genuine ones, exhibited by concrete actions, we should be able to account for normative talk in first order predicate logic. The only addition we need to standard logic are the rules that if an action is right, then it is not wrong, if an action is wrong, then it is not right, and the rule that an action is wrong if and only if it is obligatory not to perform it. Perhaps we should add to this the idea that ‘ought’ implies ‘can’, though I am agnostic with respect to its status (is it really a conceptual truth?). Given this very slim version of deontic logic, we can make sense of the kind of autonomous moral conflict I have just described, where I have more obligations than I can fulfil, even though it is possible to fulfil each one of them.1 There may even exist situations where, whatever action I perform, I act wrongly. Derek Parfit has described a position in population ethics with this consequence.2 I have referred to it before in this book, in my discussion about strong and weak supervenience in Chapter 4. Only the fate of actual people (people who live or will as a matter of fact come to live) matters from a moral point of view. I can create either A or B. Each would lead a terrible life. If I create A I do the wrong thing (because A leads a terrible life). But if instead I would have created B, then B would have lead a terrible life, does not this mean that it was not wrong to create A? No, if I create A, then B is a merely possible person, not an actual person, so her possible suffering carries no normative weight. If instead I do create B, then, by a parallel reasoning, I do the wrong thing. All this is a consistent idea, implying a kind of autonomous moral dilemma, but, I would say, a rather weird one, violating what Wlodek Rabinowicz has called normative invariance.3 I will not go any further into any problems to do with autonomous dilemmas, I just note that such dilemmas may well exist. Here the interest is in the possibility that there are heteronomous ones. I will deny that heteronomous dilemmas exist.
6.2 Rationality and Prudence Sidgwick believed, with a qualification to be noted when we move on to our discussion of theories of justice, that utilitarianism (or Universal Benevolence) gives a true account of our objective moral reasons. Yet, for all that, he seems to feel that if, in 1 I defend the existence of this kind of dilemma in ‘Moral Conflict and Moral Realism’. But see Holly M. Smith ‘Moral Realism, Moral Conflict, and Compound Acts’ for an answer to my argument. 2 Reasons and Persons, p. 395. Parfit mistakenly claims that this leads to inconsistent (contradictory) value judgements. The correct observation is that value becomes relativised to possible worlds. 3 Rabinowicz formulated it in conversation. I did myself defend a moral view, intended to cater for vagueness, at variance with this requirement, in ‘The Morality of Collective Actions’. I see this view now as mistaken.
6.2 Rationality and Prudence
77
a situation, a person is required by morality (utilitarianism) to act in a manner that would hurt his or her own best interests, this person has a conflicting reason to see to his or her own best interest. He speaks here of reasons from ‘rationality’. It is clear that here he has a stronger notion of rationality in mind than the narrow and naturalistic notion I have employed in previous chapters. In order to keep them distinct we may here speak of ‘prudence’ instead of ‘rationality’. And Sidgwick is at a loss how to settle the conflict between morality and prudence. It seems as though we would countenance two conflicting sources of obligation in this field. Sidgwick himself used to speak of a ‘dualism’ of practical reason.4 But this would mean radical relativism. At least this is one way of understanding passages such as the following one: … even if a man admits the self-evidence of the principle of Rational Benevolence, he may still hold that his own happiness is an end which it is irrational for him to sacrifice to any other. (Methods of Ethics, p. 498)
Of course, this relativistic interpretation of the quoted passage is not the only one making sense of it. At least two other interpretations are possible. Here is one of them. Perhaps Sidgwick is just making the (externalist) point that, even if a person accepts utilitarianism and realises that he ought to do a certain action (since he realises that it maximises the sum-total of well-being in the universe), he may be reluctant to perform the action in question. He does not want to perform it since, for example, it seems to him to require too much of him, and, hence, he has no Humean reason to perform it. So he may therefore be tempted by some egoistic concern. But this is not a plausible interpretation, for Sidgwick is a moderate internalist. He believes that, if we hold a moral belief, we are normally inclined to act in conformance with the view in question. Moreover, to the extent that we lack such motivation (which we may occasionally do), this may be just as true in relation to egoism as it is in relation to utilitarianism. Even a person, Sidgwick admits, who accepts ethical egoism may find that, in a certain situation, he or she doesn’t want to do what the theory requires: As Butler urges, it is a matter of common experience that men indulge appetite or passion even when, in their own view, the indulgence is as clearly opposed to what they conceive to be their interest as it is to what they conceive to be their duty.5
So I do not think that Sidgwick’s point is that in some circumstances we just don’t want to do what we realise we ought to do. But then there is another possible alternative interpretation of the quoted passage: Perhaps Sidgwick is merely making the point that, even if it seems to him selfevident that utilitarianism gives a correct account of our moral obligations, even egoism has something going for it. Even egoism may strike us as self-evidently true, when we contemplate it in isolation, and then we have to admit that, even if
For example, in the preface to the second edition of Methods of Ethics, p. x. The Methods of Ethics, p. 36.
4 5
78
6 Reasons from Prudence and Rationality
not both utilitarianism and egoism can be true, each one of these theories has something going for it, and we do not know for certain which one of them is most plausible. If this interpretation of Sidgwick is correct, I have no quarrel with Sidgwick. What he is thus indicating is that a kind of epistemic relativism may be true. We may be equally justified in accepting two conflicting moral views. There is no way, a priori, to exclude this possibility. But in the circumstances, we are not completely justified in our belief in any of them. So perhaps the best strategy in this situation is not to say that each of them is self-evidently true but, rather, to suspend judgement altogether. Be that as it may, I doubt that this is a correct interpretation of the passage. If a man ‘admits the self-evidence’ of the principle of Rational Benevolence, then how can he in any way doubt that it is true? However, in the preface to the second edition of The Methods of Ethics Sidgwick makes the following claim: … as regards the two hedonistic principles, I do not hold the reasonableness of aiming at happiness generally with any stronger conviction than I do that of aiming at one’s own.6
Is the idea that he cannot choose between these principles? I think not. My impression is that he holds a very strong belief in both principles.7 Be that as it may, there seem to be several philosophers who argue in defence of the relativistic claim that there exist more than one source of objective and categorical practical reasons. Besides moral reasons, there are prudential ones, and perhaps there are also other types of practical reasons that can come into conflict with the moral ones, leading us into a kind of practical relativity. Here are some examples of this. In his book, Welfare, Happiness, and Ethics, Wayne Sumner assumes that a theory of welfare should satisfy two desiderata. One of them is a condition of normative adequacy. The other is a condition of descriptive adequacy. This is how he describes an approach based on the former condition: [this approach] begins by presupposing a particular theoretical framework and then goes on to test competing conceptions of welfare for their ability to play a predetermined role within that framework. The background theory could in principle be either explanatory or normative. In fact, however, philosophers who have argued from this direction have had a decided preference for normative theories, usually theories either of rationality or of morality. Let us imagine, then, some such framework with a particular niche to be occupied by welfare. If it is a theory of rationality it might take the form of prescribing that agents maximize their own welfare, while if it is a theory of morality it might require maximization of the general welfare, or the equal distribution of welfare, or maximization of the welfare of the worst off. Whatever the particular normative context, the only criterion of adequacy for a candidate conception will be its ability to play its designated role within the framework in question, which will be determined in turn by the viability of the framework when that conception is inserted into it.8
p. 10. For a fuller treatment of Sidgwick and his ‘dualism’, see Owen McLeod, “What is Sidgwick’s Dualism of Practical Reasons?” 8 L.W. Sumner, Welfare, Happiness, and Ethics, p. 8. 6 7
6.2 Rationality and Prudence
79
It should be stressed that Sumner himself does not rely exclusively on this condition. It must be tempered, he submits, with a regard also for descriptive adequacy. But this is beside the point in the present context. Here it is important to note only that Sumner takes it for granted that it is plausible to speak of two types of normative theories, a theory of rationality as well as a moral one. This seems to lead to normative relativism. However, Sumner goes on to speak of a difference in normative ‘context’. It is not clear to me what this means and to what extent it can ‘explain away’ the apparent relativism. Certainly, in most (all) situations both considerations from ‘rationality’ and ‘morality’ seem to be applicable. Even if I am the last surviving sentient being in the universe I can see to my own best interests, either because this is required by rationality (egoism) or by a concern for the sum-total of wellbeing (utilitarianism). In this situation the two doctrines are extensionally equivalent, but they are still both applicable. Derek Parfit, in a discussion of population ethics, discusses the view that there are lives that, from a prudential point of view, are worth living, while, from a moral point of view, they are not worth living (since they are below a critical level).9 And Wlodek Rabinowicz, in his discussion of the Priority View, assumes that, while, from the point of view of prudence, a certain prospect is preferable to another one, it may not be preferable from the point of view of morality.10 And this is true even if there exists just one person (where both doctrines, as we noted above, are still applicable). There is no need here to go into the details of any of these putative normative (prudential) theories, all of them problematic in many ways. The crucial thing here is the very claim that there are several such theories, providing us with different sources of normativity. This claim seems to be tantamount to acknowledging a kind of normative relativism. According to these views, a person may, in some circumstances, come to confront conflicting categorical norms. Or, should the idea be understood in a way that is normatively completely innocuous? Is the point only to describe two normative theories, without taking a stance towards them, and to point at certain implications of them? In that case, there is no conflict. We just have the idea that according to utilitarianism, I ought to sacrifice my life while, according to egoism, I ought to save it. The former and more daring interpretation is at least suggested by the way the problem is presented. I must admit that there are no clear and univocal statements of the view I want to attack, but it is very often suggested, so I take it for granted that there are in fact people who accept, or, are at least inclined to accept, what I am here denying. In my claim that prudence (or rationality) does not provide us with a normative source in its own right, I do not want to deny that there is a way of making sense of many claims about what prudence requires from us. My point is rather that unless such claims can be shown to make an implicit reference either to a Humean
9 Such a view is discussed in Reasons and Persons, Chapter 18. 10 W. Rabinowicz, ‘Prioritarianism and Uncertainty’.
80
6 Reasons from Prudence and Rationality
or to a moral reason, they are not comprehensible. However, many such claims can be seen, upon closer inspection, to make a reference to a Humean or a moral reason. When they do, the apparent normative relativism can be seen to be merely apparent. Or, if I stretch my imagination to the utmost, as I do in one section below where I introduce a version of moral relativism based on social constructivism, I can’t help believing that categorical normative claims, issued from the point of view of socially constructed universes, as there described, conflicting with real categorical norms, lack all normative authority. In this they are like truth in fiction. No one would try to refute a scientific theory with reference to what is stated in a science fiction novel. I now move on to try to establish these claims.
6.3 Person-Relative Practical Reasons Wlodek Rabinowicz has suggested the following line of argument11: Prudential reasons do exist along with moral reasons. The former, unlike the latter, are based on person-relative evaluations. If some outcome is good for a person then that person has a prudential reason to realise that outcome. Now, ‘good for’ and ‘better for’ are not reducible to ‘good’ or ‘better’ period (pace G.E. Moore). In particular, to say that an alternative x is good for P does not mean that the part of x that concerns P is good, period (or that x is good, period, if we abstract from the interests of everyone but P). One argument for the unacceptability of such reductions of ‘betterfor’ to ‘better’ would be that person-relative evaluations give rise to person-relative reasons. If x is better than y for P, than P has ipso facto a (prudential) reason to choose x over y, but nobody else need to have a reason to choose x over y. By contrast, impersonal evaluations give rise to person-neutral reasons. My first response to this objection is that I feel no need to cast normative arguments in evaluative terms (such as ‘good’, ‘good for’, ‘better’, and so forth).12 What we want, in the final analysis, are normative principles cast in clear empirical terms: An action is right (wrong), if, and only if, … (where an empirical criterion is given). I will return to this point in Chapter 9. So let me here focus on the objection when it is cast in terms of practical reasons rather than value. Now, I do not deny that a person may have agent-relative (or person-relative) reasons for managing his own affairs. However, to the extent that these reasons conflict with putative moral reasons, we have to make a choice. If there are true and
In correspondence. I take this for granted in my book Hedonistic Utilitarianism, and see no problem in developing a moral theory without having recourse to evaluative terms, and I discuss it also in ‘A concrete view of intrinsic value’. 11 12
6.3 Person-Relative Practical Reasons
81
unique answers to such questions, then the ‘prudential’ reason that I ought to manage my own affairs, and the ‘moral’ reason that I ought to maximise well-being in the universe, cannot both be true. Many people seem to share Rabinowicz’ view that if something benefits me, I have a (prudential) reason to bring it about. I do not deny that this may be so. However, a way (the only way) of cashing this out in normative terms, is to say that there seems to be at least something to ethical egoism. Of course, ethical egoism gives all of us agent-relative practical reasons (and agent-relative practical reasons only). Perhaps we do not want to say that egoism alone is the true moral principle. However, on this objection, a true account of our moral obligations, our obligations simpliciter, must concede something to egoism. Perhaps the idea put forward is something like the following. At least in situations where I can either benefit myself or someone else, and where it does not matter to the sum-total of well-being in the universe whether I benefit myself or this other person, and where this other person is not worse off than I am, and where I have not promised to benefit this other person, or wronged this other person, and so forth, then I ought to benefit myself. If this is true, then morality gives me an agent-relative reason to benefit myself. However, we are not here facing any conflict between prudence and morality. We have not come across different sources of our obligations. What we have found out is that the one and true morality must contain at least one agent-relative tenet. One clear example of the putative difference between prudential and moral reasons may be thought to exist, however, in Derek Parfit’s aforementioned idea about a ‘critical level’ of well-being. On this view, the addition of a person living a life, which is from the point of view of this person, worth living, may not, from the point of view of morality, add anything of importance to the universe. The reason that it does not add anything of importance from the point of view of the universe (and morality), is that it falls below a critical level of well-being, where life begins to matter from a moral point of view. If this view is correct, it means that a person who has good prudential reasons to go on living may still have no moral reasons to do so. Does this present us with a clear conflict between (basic) prudential and (basic) moral reasons? I think not. One way of conceiving of the conflict is as follows. The ‘prudential’ principle (go on with your life if it is true that you will stay above the level where life becomes worth experiencing; if you will fall below, kill yourself), and the ‘moral’ principle (even if you are above the level where life becomes worth living, you may just as well kill yourself, provided you are below the critical level) are inconsistent. That is why they yield conflicting recommendations. However, on this reading, we must conclude that at least one of them is false. No matter how we label them, they are in the same competition (and personally I would prefer to speak of both of them as competing putative moral principles). All principles inconsistent with what we ordinarily speak of as moral principles are moral principles on my stipulation. Or, is this too inclusive? On this terminological decision, what about moral nihilism? Is moral nihilism too a moral theory? Moral nihilism implies no positive normative judgement to the effect that any action is right, wrong, or obligatory, but it is certainly inconsistent with genuine
82
6 Reasons from Prudence and Rationality
moral views such as utilitarianism and Kantianism, so, on my terminological decision, I acknowledge even moral nihilism as a limiting case of a moral view.13 The same is true, it should be noted, also of a less radical version of moral nihilism, the view that there are no true moral principles, admitting that particular actions are right or wrong, a view typically held by so-called particularists. But that is also as it should be. Particularism is indeed a very general normative position. Does this mean that moral nihilism is self-defeating? It does not, since moral nihilism doesn’t imply any positive moral judgements. And the fact that moral metaphysics (of which the nihilist thesis forms part) cannot be strictly separated from normative ethics comes as no surprise. There is no first philosophy. Metaphysics is continuous with science, moral metaphysics is continuous with normative ethics; there is no difference in kind, merely in degree, to wit, in terms of generality of the questions pondered.14 However, are the dictates from prudence really inconsistent with the dictates from morality? Well, how we answer this question depends on our intuitions about deontic logic. To me they seem to imply contradictory judgements. Suppose a person below the critical level, but above the level where life becomes worth living, kills himself. Then, according to the critical level norm, he acted rightly (he did nothing wrong). However, according to the prudential norm, he acted wrongly. And, even, if we have seen, it is logically possible that a concrete and particular action is neither right nor wrong (if moral nihilism is true), it cannot be both right and wrong.
6.4 Indeterminate Cases Even if, in standard cases of conflict between requirements of morality and rationality, there is a unique answer, that is an answer to the question what the agent ought to do, there may perhaps exist cases where this is not settled. At least this is what has been suggested.15 I have the example from Jonas Olson. Suppose for instance that a person can save a stranger’s life by sacrificing a toe. We might then believe that this person clearly ought to save the stranger’s life since losing your life is clearly worse than losing a toe. However, suppose that another person can save the life of another stranger only by sacrificing both of his legs.
I owe this observation to Jens Johansson. This is consistent with my claim in The Relevance of Metaethics to Ethics to the effect that no normative conclusions follow from metaethical premisses since I there adopt a narrow view of metaethics, where metaethics is merely a theory about the meaning and role of ethical terms, not including moral metaphysics. 15 A similar idea is discussed in Derek Parfit’s book ms What Matters. See also David Philipps, ‘Sidgwick, Dualism and Indeterminacy in Practical Reason’, for an attempt to solve Sidgwick’s puzzle with the aid of an idea of indeterminacy. 13 14
6.4 Indeterminate Cases
83
Intuitively it is less clear that this person ought to save the stranger’s life since losing your life and losing both your legs are both great sacrifices. In this latter case there might be no clear or determinate answer to the question whether the person ought to save the stranger’s life. Let me now elaborate on this example. Suppose there are two principles available here, just like Sidgwick thought, one is the utilitarian one according to which, we assume, I ought, in both cases, to save the stranger, and then there is egoism or prudence or rationality, according to which I ought, in both cases, to save my own life. In one case it is clear that, absolutely speaking, and contrary to what Sidgwick thought, utilitarianism takes precedence. I ought indeed to sacrifice my toe and save the stranger. In the other case, however, there is no truth in the matter. There is no true answer to the question what I ought to do simpliciter. This conjecture, of course, is difficult to combine with anything Sidgwick could have thought about the situation. Is this even a possible situation? If we give up moral (normative) realism, it is a possible situation, of course. David Copp has famously defended this view: I will be defending the position that neither morality nor self-interest overrides the other, that there simply are verdicts and reasons of these different kinds, and that there is never an overall verdict as to which action is required simpliciter in situations where moral reasons and reasons of self-interest conflict.16
However, it has here been taken for granted that, as a rule, there is an answer to the question what I ought to do simpliciter. Is it then reasonable to assume that there are exceptions from this rule? This depends on how we understand moral realism, of course. I tend to believe that there cannot be ‘gaps’ like these in the moral reality. On some definitions of ‘realism’ this is an analytic truth,17 but that is not my point. Then it would take special arguments to show that moral realism is a realism of that kind. My point is that while a conventional morality, a morality of our construction, may well be incomplete in this sense, it is hard to make sense of a morality existing independently of our making, exhibiting ‘gaps’ like this. And, of course, if the kind of moral determinism I have been referring to is correct, then there are no gaps like this. Moreover, for those who feel that there are cases where neither utilitarianism nor prudence takes precedence, there is also an answer available, which is less metaphysically extravagant than the assumption that there are truth-value gaps in the moral reality. Why not say that, in these situations, the agent is permitted to do as he sees fit. Then we could explain this with reference to a combined principle, settling cases like this one.
David Copp, ‘The Ring of Gyges: Overridingness and the Unity of Reason’, pp. 86–87. Michael Dummett defends such a notion of realism, see his essay on realism in Truth and Other Enigmas. 16 17
84
6 Reasons from Prudence and Rationality
Or, finally, if we want to stick to the idea of indeterminacy, we could say that, in some cases such as these, one ought to save one’s legs while, in other cases, one ought to sacrifice one’s legs and save the stranger and, yet, there is no principled explanation when the one action is right rather than the other. One would like to have an explanation of why the one answer is right in some situations while the other answer is right in other situations, of course, but we cannot just assume that all particular normative facts have explanations. If examples such as these exist, which I doubt, then there are particular norms that defy moral explanation. The natural move then would be to have access to intuitionism of the kind famously defended by W.D. Ross. We see utilitarianism and prudence as principles giving us merely prima facie obligations in cases such as this one.
6.5 Different Sense of ‘Ought’ Let us stick to the example with the idea that there are two sources of normativity, one to do with prudence or rationality, and the other one to do with morality, such as seemed to be the fact with the egoistic view of prudence or rationality and the critical-level view of morality. Could we not conceive of the possibility of there being two such sources of normativity in the following manner: The egoistic theory of prudence or rationality, and the critical-level theory of morality, do issue, in some circumstances, conflicting obligations, but not inconsistent obligations (which would be absurd). And the explanation why these conflicting obligations are not inconsistent would be that the two views are using ‘ought’ in slightly different senses. On this notion, a person may be facing obligations in conflict, that is two obligations that cannot both be satisfied, such as, ‘You ought to stay alive’ and ‘You ought to kill yourself’. Yet these obligations do not contradict one another. We have here a case of practical relativism rendered possible by a kind of semantic relativism. According to this attempt to deal with the problem that normative relativism seems to lead to inconsistency we should realise that there is more than one sense of the word ‘ought’. Actually, as the prudential and the critical-level views were stated above, it is not quite true to say, as I just did, that the agent must be facing conflicting recommendations stemming from respective theory. For there is a way of satisfying both of them: to go on living. However, for the sake of the argument, let us state the moral idea of a critical level even more strongly: From a moral point of view, a person living below the critical level has a reason to kill himself. This is what he ought to do, once it is true that in the future he will come to live below the critical level. Then, if he is still living a life worth living, he cannot satisfy both requirements. Does this make sense? Are there two types of ‘ought’, a moral ‘ought’ and a prudential one? For a long while I thought this impossible, but has occurred to me that it might be a possibility after all. We can make sense of this idea if we take moral relativism seriously. We may then retain the idea that all statements of our obligations,
6.6 Taking Relativism Seriously
85
irrespective of whether they come to us in the form of morality, or etiquette, or what have you, are categorical and in possession of truth-values; they are either true or false. And yet, for all that, there is a sense in which a ‘transcendental’ (not socially constructed) ‘ought’ overrides all other normative considerations.
6.6 Taking Relativism Seriously There is indeed a way of making sense of the general idea that there are different ‘oughts’ with different meanings. And if there are different ‘oughts’, some of them could be prudential ones, while others are moral ones. I think here of a position developed and defended famously by Gilbert Harman and David Wong.18 According to Harman and Wong, when we pass moral judgements like, ‘This is what you ought to do’, these judgements are elliptic. There is an implicit reference made in them to a system of norms upheld in the society of the speaker (or upheld by a subculture to which the speaker belongs). This has the consequence that moral judgements have truth conditions. Yet, a problem with this naturalist semantic view is that the normativity of the statement seems to go away. We are back to: ‘According to morality, you ought to do this, but according to prudence, you ought not to do it’. This invites the categorical question: ‘But what is it that I ought to do?’ There is no answer to this question the moral nihilist may claim. There is a unique answer to this question, the moral realist may want to claim. But perhaps there are many answers to the question. This would truly be a kind of relativism. I have developed this possibility in detail elsewhere.19 Here is a rough sketch of the idea, which I see as an improvement upon Harman’s and Wong’s view, allowing both that two people, who pass conflicting judgements, may be right, and retaining the categorical form of their respective judgements. Neither of them makes judgements with any implicit reference to any system of norms. They both use their moral vocabulary in an absolute sense. An objectivist non-natural moral analysis (in the style of G.E. Moore or Henry Sidgwick) of what they say gives a correct representation of what they are doing. And yet, for all that, they pass conflicting judgements. The explanation why they can both, in an absolute sense, be ‘right’ in their judgements, is that they inhabit different moral (socially constructed) universes. So while it is true in the first person’s moral universe that a certain action is right, it is true in the second person’s moral universe that the very same action is wrong.
See, for example, Wong’s Moral Relativity and Harman’s contribution to Gilbert Harman and Judith Jarvis Thomson, Moral Relativism and Moral Objectivity. 19 See my ‘Moral Relativism’ for a further development of this idea. 18
86
6 Reasons from Prudence and Rationality
A moral universe consists of a system of common sense morality. We may compare common sense morality to grammar. And we should remember that it is possible to distinguish between descriptive and regulative grammar. In descriptive grammar we observe how language is actually used. We formulate hypotheses, and we try to find general answers to questions about language use. But we may also discuss a language from a regulative point of view. We may try to answer questions such as: is it correct to put certain words in a certain order? We may want to articulate general principles also in regulative grammar. However, these principles are not descriptive of the language in question. Rather, they answer the question: what makes a certain way of using language a proper or correct one? They provide reasons for answers to this question. In regulative grammar we take for granted that there are facts of the matter. There are right and wrong ways of using language. These are facts. These facts do not exist independently of us, of course. We may think of them as social constructions, that is as constituted by us. If we were to speak differently, then different facts would obtain. And yet, for all that, these facts are in a sense objective. We may be ignorant about them. By sound argument we may be set straight with respect to them. It is worth observing that there are two different kinds of answers to the question what ‘makes’ a certain use of language a proper or correct one. One (regulative) kind of answer does indeed stipulate a reason for a judgement about the case: it is wrong to put the words in a certain order because, in the language in question, the noun phrase must precede the verb phrase. This is the kind of reason stated in a principle or rule of regulative grammar. However, the same question, that is the question what ‘makes’ a certain use of language a proper or correct one, can also be understood as an ontological question about what constitutes right and wrong in a language. And then the answer must be along the following lines: the correctness of a certain way of using the words is constituted by the fact that this is how the words are actually used in the linguistic community in question. But this is not the end of the matter. It is also crucial how experts on grammar assess this way of using words. If they condone this way of using the words, this contributes to this way’s being a proper one. And, of course, one reason why the experts condone a certain way of using the words is that this is how the words are being used. But it is also true, to some extent, at least, that the fact that this way of using the words is condoned by the experts contributes to the explanation of why the words are used this way. In a similar vein we may think of (conventional common sense) morality (in a society). Conventional common sense morality is learnt by children in a manner similar to how they learn their mother tongue. They learn that it is right or wrong to perform certain actions. Why is it right or wrong to perform a certain action? Once again there exist two ways of understanding and answering the question. A certain rule can be given, providing a reason to perform, or not perform, the action in question. Or, the question may be understood as an ontological question as to what constitutes right and wrong action. Once again, the answer to the latter (ontological) question has something to do with both how people actually behave,
6.6 Taking Relativism Seriously
87
but also with how moral experts, or even people in general, tend to judge this kind of behaviour. These facts, in turn, may have, and certainly must have, some kind of natural explanation, for example one in terms of evolutionary biology. Socially constituted moral norms seem to come to us in the form of a moral universe. This means that we think of them, ideally, at any rate, as complete. We demand of the norms making up our moral universe that, in principle, they answer all moral questions (even though the answers may be hard to come by for us) in an unambiguous way. And the set of answers to these questions is the moral universe (a set of moral facts). Now, this is an ideal. In real life it may well turn out that there are moral questions that get no satisfactory solution. But then the typical reaction of a critical person is to try to develop such answers and have them accepted by others. This means that in the moral universe pursued by the moral reformist, who is bent on moulding common sense morality, not only answers to actual problems are sought for but also answers to hypothetical cases. For there seems to be no better way of testing alternative moral hypotheses, than to resort to thought-experiments. In general, if we want to find out about what constitutes right and wrong action in a certain society, we should focus more on how people justify their actions than on what they do. A crucial feature of morality is that we can use the morality a person explicitly adheres to when we want to criticise his or her actions, once they are at variance with this morality. But, at the same time, if everybody tend to perform a certain action, condemned by common sense morality, this may mean that, at the end of the day, this action is actually condoned by common sense morality. Common sense morality is in many ways a malleable and changing entity. And some moral reformists may well play a crucial role when common sense morality develops. Now, if moral nihilism is true (as we remember that Harman has argued), and the kind of relativism put forward by Harman, here improved upon and cast in objective moral language, is true, then this prepares the way for moral conflicts both within morality (between different socially constructed moralities) and heteronymous conflicts (between, say, morality and prudence). It may well seem reasonable to assume that someone may come to feel a certain conflict between prudence (egoism) and the demands of some other more conventional morality. Think of an American girl brought up in a small town in the Midwest, in a moral universe where it is taken for granted that the innocent human life, from conception, is sacred. She grows up, moves to Chicago where she takes up studies in economics. She belongs to a group of people fascinated with public choice theory, with a constructed morality where egoism (prudence) is just taken for granted. She gets pregnant, feels, however, that carrying the pregnancy to term will mean an obstacle to her studies, so without much ado she has an early abortion. She tries to explain her decision to her parents. She did the right thing, she claims, since the child would have meant an obstacle to her academic career. Her parents protest that it was wrong to kill an innocent human being. Given moral nihilism (with respect to a moral reality existing independently of our conceptualisation), this may seem as a genuine case of conflict. In an obvious sense the conflict is real. Even if she and her parents talk at
88
6 Reasons from Prudence and Rationality
cross-purposes – they use ‘ought’ in slightly different ways – their judgements conflict. There is no way for the girl to satisfy both the demands made by her parents and the demand from her own favoured morality. Each demand comes in a categorical form. So the enigma pointed at by Sidgwick seems here to surface. However, if moral realism of the kind I have defended is true, the enigma goes away. Or, so I am prepared to argue. Once we realise that moral universes are constructed by us, once we realise that, had we acted and judged differently, other moral (conventional) truths would have obtained, we want to know how our conventional moral universe ought to be constructed. If there is a truth in this matter, then all the ‘oughts’, in all the conflicting socially constructed moral universes, tend to loose their authority. They are not silenced, they are there indeed, and their categorical requirements may well be at odds with what, morally speaking, we ought to do. However, once we realise that they are at odds with what, absolutely and hence morally speaking we ought to do, they will not matter any more to our moral understanding. This is not to say that conventional morality (or normative grammar, or etiquette, or your own view of what is in your own best interest, or what have you, that can be seen as socially or individually constructed) cannot indirectly enter morality. It may well be that, as a rule of thumb, we ought to abide by conventional morality in the society to which we belong, or speak our own language in a proper fashion. However, when, in a certain situation, there is a conflict between the requirements from a morality that is not of our making, and a constructed normative claim of any sort, be it a requirement of grammar, etiquette or some constructed conventional morality, the moral claim has a special status. What we are facing here, as I said above, is like a conflict between truth in fiction and truth in reality. Truth in reality trumps truths in fiction – at least in reality. There is no way of showing this, I submit, but anyone who has seriously pondered a genuine moral question should know what it is I am here referring to.
6.7 A Further Stretch of Our Moral Ontological Imagination Let me make a further (final) attempt to stretch my moral ontological imagination. How can I be so sure that there are not genuine problems beside the one what I ought to do simpliciter? What if there are genuine problems apart from this one? One of them is what prudence requires of me. Another one is what morality requires of me, when ‘morality’ is not conceived of in the manner I do in this book (as the answer to the question what I ought to do simpliciter). What if there are competing versions of egoism, say, giving different answers to the question what prudence requires of us? What if there is a true answer to the question which one among these competing theories is the correct one? One could be cast in hedonistic terms, for example, while another is cast in terms of our interests. And one of the theories cast in terms of our interests may urge us, for example, to satisfy our present desires (in the sense of now for then) while another may urge us to be
6.7 A Further Stretch of Our Moral Ontological Imagination
89
temporally neutral (now for now and then for then)? Would we not then like to know which one of them gives the correct answer? This ‘possibility’ seems almost inconceivable. What kind of question is here given an answer? Furthermore, if these problems to do with prudence and morality respectively are genuine, how should we ever gain knowledge about them? We test our moral theories, I submit, against our considered intuitions with respect to what, in various different situations, we ought to do simpliciter. How are we to test competing theories of prudence? Or competing theories of morality, if they are not intended to answer the question what I ought to do simpliciter? I doubt that we have anything to rely on here but linguistic intuitions. But linguistic intuitions should not be capable of settling a normative issue. So I think we can safely assume that there is no such thing as a competing prudential source of normativity. And even if such a source existed, it would not be of much practical avail. Why bother with what we ought to do from the point of view of prudence (or morality), if we already know what to do, and why, simpliciter? It might be objected to this last claim that, in order to find out what we ought to do simpliciter, we need first to find out what is required of us from morality and prudence respectively, and then assign a certain weight to respective perspective. This seems to be how David Copp conceives of our situation. He defends, however, the nihilist idea that there is no way of arriving at an answer what we ought to do simpliciter. And the reason for this nihilism is as follows: The claim that a standard S has the property of supremacy is the claim that it is normatively the most important standard as assessed in terms of some other standard, R, which is the normatively most important standard. But only one standard could be normatively the most important.20
This sounds like sophistry, but perhaps the argument is sound. At least it may seem telling to a moral (ontological) nihilist who sees morality, prudence, and other sources of normativity as socially constructed and hence, on the face of it, on a par. Given that view, we may need a criterion to tell us which one among them overrides the rest, and then we seem to be in the predicament described by Copp.21 And yet, we can easily avoid it, if we accept that the norms simpliciter are both sui generis and authoritative; then we proceed in the following manner. We unpack the respective theories in question, answering the questions what we ought to do from the point of morality and from prudence. Suppose the right answer to the moral question is given by hedonistic utilitarianism and the right answer to the prudential question is given by hedonistic egoism. I now want to know what I ought to do simpliciter. Then I must examine hedonistic utilitarianism and hedonistic egoism (and other theories as well, I would like to add), in order to see which one, or which consistent combination of elements from each one of them, gives the right answer to this question.
‘The Ring of Gyges’, p. 103. Perhaps there is still a way of avoiding his conclusion; see Owen McLeod, O. “What is Sidgwick’s Dualism of Practical Reasons?” about this. 20 21
90
6 Reasons from Prudence and Rationality
But this we can do without any investigations into problems of prudence or morality proper; if we pursue such studies, and if such studies can be pursued in the first place, the only thing of normative relevance we reap from them are putative theoretical answers to what I have called the moral problem. We can access these theories without making any detour into prudence or morality (if morality is something different from the answer to this question).
6.8 Phronesis Let us leave the spectacular ontological ideas about putative genuine problems of prudence and morality and return to more mundane matters. The Greek root of the word ‘prudence’ is phronesis. In modern virtue ethics this term is often used to signify a certain skill in making moral distinctions, a skill in realising what, in a concrete and complicated situation, morality requires from us. However, if Sidgwick is right when he claims that utilitarianism gives a correct account of our moral obligations, then we will never know for sure what we ought to do. We cannot comprehend all the consequences of our actions, let alone can we survey what would have happened, had we acted differently. For epistemic reasons we seem to need, besides our criterion of right action, some kind of method of moral decisionmaking. And such a method is needed not just by utilitarians. Any moral theory paying any respect to the effects on well-being of all affected individuals, must be supplanted with a method of decision-making. And, as I have stressed in previous chapters, no plausible moral theory neglects well-being altogether. Our lack of knowledge renders the move to a method of decision-making necessary, then. Incidentally, this is an additional reason why Crispin Wright’s idea that truth is ‘epistemically constrained within the moral domain’, discussed in Chapter 5, is implausible. Does this mean that an omniscient deliberator would lack the need of a method of decision-making? Erik Carlson has made this claim: I believe that a reasonable theory should nevertheless be ‘action-guiding’ for an agent with complete knowledge of all morally relevant facts in the situation in question.22
This is too optimistic. It is true that lack of knowledge of relevant facts render necessary the use of a method of decision-making, but even an omniscient deliberator would need such a method. Not only epistemic considerations force us to adopt a method of moral decision-making. This has to do with the fact that, in many situations, it has bad consequences if we ponder in a systematic way what to do. My marriage would be ruined, for example, if all the time I pondered whether I should exchange my wife for someone better. A reasonable method of decision-making hence stops me from pondering this question and allows instead for spontaneity in
Erik Carlson, Consequentialism Reconsidered, p. 101.
22
6.8 Phronesis
91
my relations to my wife. Furthermore, in other situations we have to rely on habits rather than on deliberations, such as when we drive a car. Again, a reasonable method of decision-making stops me from deliberating in situations where, were I to do it, I would perform badly. All this means that a method of decision-making needs to be very complex, and it is likely that different people need different methods, in order to perform as well as possible. And all this means that, even if my claim is true, that there exists just one kind of moral question, to do with norms (‘What ought to be done and why?’) there is much room in our moral theorizing for considerations of motives, traits of character, and virtues. But inquiries into these matters belong mainly to the context of the development of methods of moral decision-making.23 Let us now stipulate that a prudent person is a person who has developed such a method. To simplify, by focusing only on the problem to do with lack of relevant knowledge, suppose the method informs the agent in a situation where it is acceptable to deliberate to account for the alternatives in the situation that are relevant to the decision at hand, to assess the possible outcomes of the various different alternatives, to make a rough evaluation of these, and then to maximise expected happiness. Could we not now say that, while the agent does not know which alternative to perform (i.e. which alternative maximises the sum-total of well-being in the universe), the agent still has a (prudential) reason to perform the action that is picked out by her favoured method of decision making? Could we not say that this is the action the agent ought to perform, since this is the action the decision-making procedure is picking out? Could we not say that it is prudent or rational to perform this alternative? But this alternative is perhaps not the right one to perform in the situation (according to utilitarianism, which we have assumed gives, at least in the circumstances, where no deontological constraints are at issue, a true account of our moral obligations). Does this mean that there are conflicting obligations here, one dictated by prudence (upon which the agent acts), and another one dictated by the true account of morality (utilitarianism), upon which the agent cannot act (since he or she does not know about it)? The agent has a moral reason not to perform the action and a prudential reason to perform it. I must confess that this is how I used to conceive of this kind of situation. I blush to admit that I even invented the notion of a ‘responsible decision’ to account for it.24 It seemed to me that, in a situation where it would be in accordance with my selected method of decision-making to perform a certain action, to perform it would be responsible of me. I had the idea that there are reasons to do with responsibility. I thought that I had here discovered a special source of obligation to do with
Some virtue ethicists attempt at formulating criteria of right and wrong action in terms of what a virtuous person does or would do. I criticise such approaches in my Understanding Ethics, but if such an attempt succeeds, then, of course, it is in the competition with utilitarianism, Kantianism, and so forth, for a correct answer to the normative question. 24 In my Hedonistic Utilitarianism, pp. 34–39. 23
92
6 Reasons from Prudence and Rationality
responsible decision-making. However, when pressed to explain what kind of reasons they were, or what kind of obligation it was to which I was referring, I was at a loss. I now believe I know why. There are no such reasons. There is no such source of obligation. I am not the only one, or even the first one, to have made this mistake. R.M. Hare has famously argued in defence of a two-level approach to moral problems, allowing us to rely on simple rules of thumb in certain concrete situations of practical decision-making. This is a plausible view. However, Hare goes wrong when he contemplates what implications it has for morality. This is what he says about such practical principles or rules of thumb: It is quite consistent even with act-utilitarianism … to admit the utility of inculcating such principles, so that people do, without thinking, what is likely to be optimific. To do what they do may turn out, in extraordinary cases, not to be right by the act-utilitarian criterion; but even by that criterion what they do is justified, because the criterion, regarded as a guide to actual choices at the time that they are made, can only require people to do what maximizes expectable utility …25
It is left open what it means to be in this sense ‘justified’. Certainly, what we are doing in the circumstances, if we follow the advice from the rule of thumb, is something we ought not to do. The action we perform is not ‘justified’ in the sense that it is right, then, let alone in the sense that there exists an explanation of why it is right (since it is wrong). Furthermore, even if we happen to believe that what we do is the right thing to do, we are not justified in this belief, since, more often than not, what our decision-method advices us to do is a wrong action. There are even cases where we know for sure that by abiding by the decision-method we perform the wrong action. An example of this would be Frank Jackson’s doctor Jill, who can prescribe either of three drugs, A, B, or C to her patient. She doesn’t know whether the correct diagnosis of the patient is X or Y (both possibilities are equally probable). If she gives A and the patient suffers from X then the patient will die. If the patient suffers from Y and is given A the patient is completely cured. The situation is reversed when it comes to the second medication, B. If the patient suffers from X and is given B the patient will recover completely. If the patient suffers from Y and is given B the patient will die. The third medication, C, is different. It will almost (but not completely) cure the patient, irrespective of whether the patient suffers from X or Y. To give C maximises expected happiness. Yet, we know for sure that it is wrong to give C to the patient. Either A or B is the right choice.26 All this is intriguing, but I would now prefer to give the following diagnosis of what is going on. Availing ourselves of the notions defined in this book we should describe the situation of the agent in the following way.
R.H. Hare, Moral Thinking, p. 17. See his ‘Decision-Theoretic Consequentialism and the Nearest and Dearest Objection’. I have simplified his example a bit. There are earlier versions of the example, in particular in Donald Regan’s book, Utilitarianism and Co-operation, pp. 264–65. 25 26
6.8 Phronesis
93
Suppose the agent has represented her choice as a choice between the two alternatives H1 and H2. Suppose that, if she performs H1, she acts in accordance with what morality requires (H1 maximises the sum-total of well-being in the universe). However, the agent does not know this. And, from the point of view of the decision-making procedure she has adopted, it seems as though H2 is the action she should perform, since, in the circumstances, H2 maximises the expected happiness in the universe. But does this really mean that, in a categorical sense, she ‘ought’ to perform H2? It does not. In the categorical sense she ‘ought’ to perform H1. And there is really nothing indicating that the categorical ‘ought’ is ambiguous here. However, it is certainly true that, if she wants consistently to abide by the decisionprocedure she has adopted (in the hope, I assume, that in the long run, if consistently abided by, this procedure will produce better outcomes than any other procedure she has been able to conjure up), then she ought to perform H2. In the circumstances, it is rational for her to do so. I say ‘in the hope that …’, since there seems to be no way to show that a decision-procedure really has this desired characteristic. Moreover, even if it could be shown that it has, this is of little avail when it comes to individual applications of it. Unless we have recourse to rule-utilitarianism, the relative success in the long run of a decision-procedure is no moral reason to abide by it in the individual case. Note also that this ‘ought’, in which she ought after all to stick to the procedure in the individual case, is a conditional or instrumental one. It signifies that there are Humean reasons for her to perform H2. It is rational for her to perform H2, in a purely naturalistic sense of the word. And these reasons are only there if this person wants to stick to the decision-method, in the situation. It does not signify that there are categorical reasons for her to perform H2. Certainly, if utilitarianism is true, and if H1 maximises the sum-total of wellbeing, there are no moral reasons for the agent to perform H2. She ought to perform H1 (even though she does not know this). There are no categorical prudential reasons, or reasons to do with responsible decision-making, conflicting with the moral ones. Moreover, in the individual case, if the person really ponders whether to abide by the decision-method or not, it may well happen that she finds that she doesn’t want to do so. Suppose following the decision-method means performing the action that maximises the expected happiness. In the situation where the agent is, this means taking a small risk that something terrible may happen. Suppose the agent doesn’t want to take this risk. And suppose a less risky alternative exits as well and tempts her to choose it, even thought it doesn’t maximise expected happiness. If this is the case, then she has no Humean reasons in the situation to use the method she has selected as her decision-method! So if, yet, she sticks to it, and performs the risky action in the circumstances, this has to do with the fact that this is how she has committed herself to act. Then there is no way of explaining her action and rationalising it (by her own lights). But, of course, once again, if the agent wants to stick to her favoured decision-making method, then this person has a Humean reason, even in the circumstances, to abide by it.
94
6 Reasons from Prudence and Rationality
This is all there is to say about the relation between criteria of right action and methods of decision-making.27
6.9 Conclusion Theories such as utilitarianism and egoism can be seen as giving competing, inconsistent, putative answers to the question what I ought to do simpliciter. Then, according to my stipulation, they are both moral theories and, since they are inconsistent, they cannot both be true. We have to make up our mind, then, as to which one of them is true or, if they are both false, with what they should be replaced. There is no place in this picture for any practical ‘dualism’. It is tempting to believe that, even if this may be so, another way of conceiving of utilitarianism and egoism is as practically incompatible but still not inconsistent theories, since they answer different questions. Utilitarianism is intended to answer the question what, morally speaking we ought to do, while egoism answers the question what, from the point of view of prudence, we ought to do. I have admitted that it is possible to make sense of this if we conceive of these theories as belonging to different socially constructed realities. In that case, however, the answers they provide have no authority. Why would we want to know what these theories inform us, if we already know the answer to the question I have called the moral one, to wit, the question what I ought to do simpliciter? This is the main result of the investigation in this chapter. Another and more peripheral one is that, even if it may be tempting to believe that, when a method of moral decision-making gives us instruction at variance with what is required from us from a correct criterion of right action, we face conflicting obligations, appearances are deceptive. We have Humean reasons to stick to our method of decisionmaking if we do so, and want to do so, in the hope that, consistently abided by, it will lead to a better outcome than if we were to abide consistently by some competing method of decision-making, but we have no objective moral reasons to perform the (wrong) actions we do perform.
The distinction between methods of decision-making and criteria of right-making characteristics was first clearly introduced by Eugene Bales in ‘Act-Utilitarianism: Account of Right-Making Characteristics or Decision-Making Procedure?’ For an excellent review of the problem of justifying the use of any such method in particular, see Holly M. Smith, ‘Making Moral Decisions’. 27
Chapter 7
Reasons from Justice and Aesthetics
7.1 Introduction Two further putative sources of normativity need to be examined: theories of justice and aesthetics. Certainly, those who advocate various different theories of justice often speak as though they contemplated a special realm of norms to do with distribution of goods. And they make claims such as, if a perfectly equal distribution were exchanged for an unequal one, then, even if this would mean an improvement for each individual, the change would be worse in one respect, that is with respect to justice. So there is a kind of normative ‘pull’ in the direction of equality, even though, in the envisaged circumstances, the exchange of equality for inequality may be morally desirable. And within aesthetics normative considerations enter in two different ways. There are philosophers who claim that an attitude of admiration or disinterested satisfaction is what it would be proper to adopt, when contemplating valuable art. Immanuel Kant made this claim, and it has been followed up by G.E. Moore and others. I will discuss it in Chapter 9. Another claim has been that, if something is beautiful, if something possesses aesthetic value, then, all other things equal, we have an obligation to produce it. This is the idea that will be discussed in the present chapter. However, in order to be plausible, this view needs the assumption that there is such a thing as aesthetic value. It must be shown that there are objects, natural or artefacts, of a kind such that we ought to construct tokens of them or, once they exist, keep them in existence. First I turn to theories of justice, however.
7.2 Theories of Justice At least on a superficial reading it may seem as though some philosophers take their theories of justice to provide us with a special source of normativity, which could possibly come to conflict with other sources, provided by theories of prudence or morality. This is not true of philosophers of a utilitarian bent, of course, such as Sidgwick. Utilitarians have indeed paid some interest to the notion of justice.
T. Tännsjö, From Reasons to Norms, Library of Ethics and Applied Philosophy 22, DOI 10.1007/978-90-481-3285-0_7, © Springer Science + Business Media B.V. 2010
95
96
7 Reasons from Justice and Aesthetics
In the first place, justice may be conceived of as a virtue, that is, as a desirable trait of character. In that case no utilitarian in general, or Sidgwick in particular, would have anything to object. However, according to the utilitarian, the explanation why a certain trait of character (in certain circumstances) is desirable, would have to be cast in terms of the consequences of adopting it, of course. Moreover, justice may also be an idea of what the basic structure of society should look like. Again a utilitarian would have no objection to adopt a view of justice. Once again, the idea would be that a basic structure of the society is just if, and only if, in the circumstances, there exists no alternative feasible basic structure the adoption of which would lead to better consequences. This would not mean that the utilitarian must acknowledge that there exists any additional source of normativity, then. A just basic structure is one we ought to establish, period. However, what if the utilitarian came to suspect that the maximisation of the sum-total of well-being is not the only thing that settles matters of rightness and wrongness? What if the utilitarian came to believe that, not only may the sum-total of well-being be of importance, but so may the distributional pattern as well? Could we not represent this as a case where the utilitarian suspects that, besides reasons to do with maximisation of well-being, there are also reasons to do with justice? This is a possible way of putting the point but hardly an illuminating one. Remember that Sidgwick felt this kind of tension. Upon consideration he came to accept that, if, in a situation, there are two ways of maximising the sum-total of wellbeing, one meaning that well-being be equally distributed and the other meaning that it be unequally distributed, the former way of action should be favoured over the latter.1 If we like, we may say that, from the point of view of maximisation, we may act either way while, from the point of view of equality, we ought to distribute wellbeing equally. But this is no answer to any normative question. Which way ought we to act? According to Sidgwick, we ought to distribute well-being equally. This is the only norm he is capable of arriving from this piece of reasoning. However, let us turn to a more recent advocate of a rather different conception or theory of justice, John Rawls. How does Rawls think of his theory of justice? In what way could we derive answers to normative questions from it? Could these answers come to conflict with norms derived from morality? It may seem as though John Rawls, in A Theory of Justice (in later writings he were less interested in basic moral problems and more in practical political matters) holds the view that there are basic normative problems to do with justice, different from moral ones. Principles of justice were thought to apply to institutions, and they were not to ‘… be confused with the principles which apply to individuals and their actions in particular circumstances’. (p. 7). And Brian Barry has made the claim that: If Rawls had achieved nothing else, he would be important for having taken seriously the idea that the subject of justice is what he calls ‘the basic structure of society’2
Methods of Ethics, p. 417. In his Justice as Impartiality, p. 214.
1 2
7.2 Theories of Justice
97
Does that mean that there are, according to Rawls, two kinds of basic principles, different in content, and hence capable of coming into conflict, if it were not for the fact that they have different fields of application? One set of principles concern institutions, another our individual actions. But this could not be exactly so. Also the principles concerning institutions, if they are to carry any normative import, must concern what individuals are supposed to do. They must be to the effect that such and such institutions should be established and sustained. Concerning other human actions, other principles are valid. On a superficial reading of Rawls, this may seem so. And Liam B. Murphy has tried to show that this is not only a superficial understanding of Rawls, but also a correct understanding of him.3 But this would mean, as Murphy points out, that we would have a hard time explaining why an individual should not through individual action try to ascertain the kind of justice prescribed by Rawls’ principles of justice, if there were no institutions available (in non-ideal conditions), to see to it that just conditions obtain. A more plausible interpretation of Rawls can be indicated along the following lines, however. Rawls does indeed hold that there is a set of principles telling us what kind of institutions to establish and sustain, but these principles are not basic ones, in the sense that, say utilitarianism and Kantianism, are basic. And it seems indeed as though he has a basic notion of obligation so to speak in the back-ground of his reasoning. If this diagnosis is correct, there is only one source of normativity in Rawls’ thinking; it is the source of both the requirements on individuals in general and on individuals in their dealings with institutions. I will now elaborate this possible interpretation. There is no need here to go into the exact content of Rawls’ theory with his two, lexically ordered, principles. Neither need we bother with any details in the contractual apparatus he utilises when he arrives at them. The important thing, in the present context, is the normative role he thinks the principles play. Their role, he claims, ‘is to be regarded as providing … a standard whereby the distributive aspects of the basic structure of society are to be assessed’.4 However, justice is but one among several virtues for the basic structure. All the virtues together with an idea about their relative weight when they conflict, according to Rawls, constitute a ‘social ideal’.5 What are we to make of this qualification? Does it imply that Rawls assumes that there are many, perhaps conflicting, sources of normativity? I do not think this is the most charitable interpretation. The most charitable interpretation would be something along the following lines. According to Rawls, while our obligation to adopt just institutions is categorical, there may exist situations, where several political options satisfy the requirement of justice. In those circumstances, we do not know which basic structure to adopt, unless we have
Liam B. Murphy, ‘Institutions and the Demands of Justice’. A Theory of Justice, p. 9. 5 Ibid. 3 4
98
7 Reasons from Justice and Aesthetics
worked out a complete social ideal. Once we have worked out this ideal we know how to answer the additional questions. But the answers to these questions do not conflict with any demands from justice. For the demands of justice trump other kinds of demands, to do with other virtues (such as efficiency) of the basic structure of society. However, the contractarian theory of ‘justice as fairness’ put forward by Rawls is thought of as merely a part of a complete ethical theory: Does this mean that the demands from justice may be at variance with our moral obligations? I think not, but here Rawls is facing a problem. How does his theory of justice fit into a complete ethical theory? According to Rawls, a full ethical theory must ‘include’ a theory of justice, along the lines developed by Rawls: I assume that any reasonably complete ethical theory must include principles for this fundamental problem [concerning the distributive principles for the basic structure of society] and that these principles, whatever they are, constitute its doctrine of justice.
Furthermore, according to Rawls, the contractual idea he uses to derive his theory of justice can be extended to the choice of ‘… more or less an entire ethical system, that is, to a system including principles for all the virtues and not only for justice’. And yet, this would not be the final word in ethical theory, … since it would seem to include only our relations with other persons and to leave out of account how we are to conduct ourselves toward animals and the rest of nature.6
The most charitable interpretation here is to take the obligations we can derive from Rawls’ contractual theory of justice to be conditional. On the assumption that, by our actions, we do not affect animals and the rest of nature, we have the obligations that follow from Rawls’ theory of justice, he claims. As a matter of fact, these obligations seem to be conditional even in other respects. The contractual model set up by Rawls can hardly handle problems of population ethics – or, at least, it has not been designed to do so by Rawls himself – so we have to assume, also, that the obligations we derive from the theory are conditional upon the assumption that, by our actions, we do not affect how many people will come to exist.7 What if these assumptions are not fulfilled? The consequence is that we cannot derive any norms from Rawls’ theory of justice – or, we can merely derive the hypothetical and not very illuminating conclusion that, if the provisos are fulfilled, then we ought to do so and so. To see this, imagine that by our actions we, members of the human race, do serious harm both to other animals and to nature. Suppose for the sake of the argument that some deep ecological thinkers, who do pay attention to animals and the rest of nature, and who claim that it would mean improvement if the human race disappeared from the world, are right. Suppose also that there exists a scientist who has invented a method of killing all human beings instantly and painlessly. She believes in the true morality and she puts her method into practice. The human race ceases to exist. Ibid., p. 17. I discuss this problem in ‘Why We Ought to Accept the Repugnant Conclusion’.
6 7
7.2 Theories of Justice
99
It is hard to find an example where Rawls’ idea of justice may seem to be violated more thoroughly. Everybody gets his or her right to physical integrity flouted by the scientist, who does the killing. However, appearances are deceptive. The scientist does the right thing, we have assumed. And her action does not violate Rawls’ theory of justice, precisely because of the conditional character of this theory. If the situation is like it has here been described, if the human race is a real threat to other animals and the rest of nature, then it might well be that it should be killed. And, since Rawls’ theory of justice is not a complete moral theory, it has nothing to say about this situation. We can derive normative information from it only in situations where the hypotheses upon which it is built are fulfilled. In the envisaged situation, they are not fulfilled. So it does not provide us with any reasons not to finish off the human race. Once again we must conclude that there are no conflicting sources of obligation. There are of course many other theories of justice, but it would be tedious, in the present context, to work through even all of the most influential ones. Some of them are, even on a superficial understanding of them, best conceived of as moral theories proper. This is true, for example, of Nozick’s theory of justice8 according to which an action is right (just), if, and only if, it does not flout any moral rights. On a superficial understanding of some other theories, such as Larry Temkin’s egalitarian theory of justice,9 it may seem as though there exist conflicting sources of obligation, one to do with justice, others to do with prudence, morality, and so forth. I alluded to this in the introduction to this chapter. This is how Temkin writes: I … believe that inequality is bad. But do I really think there is some respect in which a world where only some are blind is worse than one where all are? Yes. Does this mean I think it would be better if we blinded everybody? No. Equality is not all that matters.10
Does this mean that, according to Temkin, considerations of justice provide us with a special source of normativity? Does it mean that, on his account, norms to do with justice may come to conflict with norms to do with morality, and so forth? This would not be an illuminating way of putting his point. In the final analysis, there is only one normative (moral) problem, to wit, the problem of what ought to be done simpliciter. The correct answer to the question what ought to be done simpliciter may depend both on the sum-total of well-being produced by a certain action and the degree of inequality in which it results, of course. But there are not two different kinds of norms in conflict here. We merely meet with different incompatible suggestions as to how we should answer one and the same question: What ought to be done? According to Temkin, we ought to give some weight to considerations of equality, but not in a manner that forces us to accept a policy of ‘levelling down’. Temkin himself is not precise on how to do this, but there are many possible ways of achieving such a result, and they each result in normative principles in competition with utilitarianism. As developed in Anarchy, State, and Utopia. As developed in Inequality. 10 Ibid., p. 282. 8 9
100
7 Reasons from Justice and Aesthetics
What difference does it make if we speak of one or two (competing) sources of normativity, once we are prepared to allow them together to supply material for an answer to the question what ought be done? From an epistemic point of view, it makes all the difference in the world. If there were two different sources of normativity, then we could explore them, one at a time, independently of one another, and reach two preliminary, and sometimes incompatible, answers to the normative question. We would then be in a position to put them together into one single answer. However, if there is only one source of normativity, such an independent examination of any problem of justice proper, as distinguished from the ultimate normative question, is not possible. On Temkin’s view, equality (in the form in which he conceives of it) has some moral importance. If equality can be realised at no cost at all, it ought to be realised, he claims. And even, at some cost (in terms, say, of the sum-total of wellbeing), equality ought be realised, he seems to think. However, the price may be too high. In that case we ought to accept some inequality. All things considered, it is better that only some are blind than that all are blind. This has nothing to do with any plurality of norms pulling in different directions, then. If Temkin is right, inequality may, in some circumstances, be a wrong-making characteristic. In other circumstances inequality is morally acceptable. We may share or reject this view, without believing that there is more than one source of normativity.
7.3 Aesthetics The Cambridge philosopher G.E. Moore has famously defended the view that certain organic wholes have intrinsic value. He is addressing the question from the point of view of nature and culture alike. According to Moore, some pieces of art have intrinsic value, but so have certain ‘natural’ organic wholes. As a matter of fact, when giving his most often quoted example of an organic whole with (aesthetic) intrinsic value, he speaks of nature: Let us imagine one world exceedingly beautiful. Imagine it as beautiful as you can; put into it whatever on this earth you most admire – mountains, rivers, the sea; trees and sunsets, stars and moon. Imagine all these combined in the most exquisite proportions, so that no one thing jars against another, but each contributes to increase the beauty of the whole. And then imagine the ugliest world you can possibly conceive. Imagine it simply one heap of filth, containing everything that is most disgusting to us, for whatever reason, and the whole, as far as may be, without one redeeming feature. Such a pair of worlds we are entitled to compare.… The only thing we are not entitled to imagine is that any human being has, or ever, by any possibility, can, live in either, can ever see and enjoy the beauty of the one or hate the foulness of the other. Well, even so, supposing them quite apart from any possible contemplation by human beings; still, is it irrational to hold that it is better that the beautiful world should exist, than the one which is ugly? Would it not be well, in any case, to do what we could to produce it rather than the other? Certainly I cannot help thinking that it would: and I hope that some may agree with me in this extreme instance.11 Principia Ethica, pp. 83–84.
11
7.3 Aesthetics
101
We can hear this kind of argument echo in many modern and post-modern defences of deep ecology. I do not think that Arne Naess12 would reject an interpretation of his position along these lines. Moore’s argument is of course problematic. Could there be any such thing as a beauty that is not perceived? Well, the assumption that it could takes us some way into a problematic ontology, but perhaps not too far. So why not accept that the beauty is there, irrespective of whether it can be sensed or not? However, this only leads us to the next and crucial question. How can it be of any value, if no one is there to appreciate it? As I will explain in Chapter 9, I am not happy with putting normative points in terms of ‘value’. And in the present context it is clear that, instead of speaking of the intrinsic value of beauty, and of the ‘betterness’ of one world existing rather than the other, Moore could just as well have raised the question of whether beauty is something we ought to promote (he comes close to doing so when he speaks of it being ‘well’ to do what we could to produce the beautiful world rather than the ugly one). As a matter of fact, he often switches himself to this manner of speaking and I will stick to it. But the question remains. The fact that Moore himself cannot help thinking that beauty ought to be promoted does not show that it has value. Nor would it be of much avail if his hope that others will share his opinion were borne out by realities. Moore himself is not prepared to give any heavy normative weight to aesthetic objects. In this he differs from Arne Naess. The latter is prepared to say that, in the interest of the aesthetic existence of variation and richness in nature (with regard to many species) the human race should perhaps abdicate. He comes close to the kind of view I discussed in relation to John Rawls’ theory of justice in the previous section. Moore is more cautious, however. He believes that, in reality, there are no cases where aesthetic value gains the upper hand of other values. He goes on to say, switching in his typical manner from speaking of value to speaking of obligation, of a putative situation where we have to choose between other values and mere beauty: It is highly improbable, not to say, impossible, we should ever have such a choice before us. In any actual choice we should have to consider the possible effects of our action upon conscious beings, and among these possible effects there are always some, I think, which ought to be preferred to the existence of mere beauty.13
Be that as it may, what if there are aesthetic values of the kind indicated by Moore, no matter how heavy they weigh in a normative calculus? Does this mean that we have come across a new source of normativity? Are there special obligations to do with aesthetics, in addition to standard moral obligations? Does this mean that we may face conflicting duties, derived from different sources? It does certainly not. It is clear from Moore’s treatment of the subject that, on his notion of aesthetic values, these values are values of the same kind as other values. Cf. his Ecology, Community and Lifestyle: Outline of an Ecosophy. Ibid., p. 84.
12 13
102
7 Reasons from Justice and Aesthetics
This means that, if they are values at all, if they mean that we ought to do things in order to see to it that beautiful objects exist, then, this kind of obligation is in competition with other kinds of obligations.14 Aesthetics, then, on Moore’s moralistic notion of it, is part of ethics. And this seems to be true also of the conception of aesthetic value held by people like Arne Naess, who gives more weight to aesthetic values than Moore does. Naess does indeed see his environmentalist aesthetic theory as a proper part of morality.
7.4 Suppose I am Wrong In the previous chapter I considered the possibility that I was wrong in my insistence that there is no source of normativity to do with prudence. Even if it is somewhat repetitious, allow me to make the same concession in this chapter. Suppose there are problems in their own right to do with justice and aesthetics. Suppose there are true and false answers to the categorical questions what we ought to do, when put forward either in terms of a justice-ought, or an aesthetic one. What difference would this make? Suppose we have discovered that, in a situation, what we ought to do simpliciter is to make a rather happy person much happier, rather than making a less happy person slightly happier. Now two persons, who agree about what ought to be done simpliciter, disagree about whether doing this means that an injustice is done to the less happy person or not. One of them claims that, even if this is what has to be done, all things considered, it means that this person is treated in an unjust manner. Suppose the other person denies this and claims that, not only is this what should be done, it also means that justice is done. What can we make of their disagreement? I can’t help feeling that this must be a matter of conflicting linguistic intuitions, and nothing more than this. But linguistic intuitions should not be allowed to settle normative issues. However, suppose I am wrong about this, what practical importance would this have? One may think that if it is true that we are doing the less happy person an injustice by making the more happy person happier means that it was premature to conclude that we knew the answer to the question what ought to be done simpliciter. In order to reach an answer to that question we must know what to do from the point of view of justice, aesthetics, prudence, and so forth, and we must give relative weight to each normative perspective in order to reach an answer to the plain ought question as to what to do simpliciter. I have already rejected this line of argument (in the previous chapter). What we should do, under the envisaged circumstances, is rather to unpack the theories answering the respective questions, and see whether the considerations they take to
Moore himself gave up the idea that there are aesthetic values of this kind. See his Ethics, pp. 148 and 153. 14
7.5 Conclusion
103
be relevant have any weight when we want to answer the question what to do simpliciter. The person who claimed that it would be unjust to give a big benefit to a happy person rather than a small benefit to a less happy person, may be an advocate of some egalitarian or prioritarian normative theory. But such theories are already in competition with utilitarianism and other theories, when we are in search of a principled answer to the normative question: What ought to be done, simpliciter? We need not turn to theories of justice, prudence, or aesthetics, in order to get a hold of them.
7.5 Conclusion Neither theories of justice nor theories of the aesthetic value of objects, be they artefacts or natural ones, have been seen to provide us with any special source of normativity, capable of generating norms in conflict with morality. On the contrary, both theories of justice and aesthetics have been found to make, in some cases strong and I would say even bizarre, moral claims but yet, for all that, moral claims in their own right. We have to pause to think about these claims when we try to settle for an adequate moral (normative) theory. These claims are, in my manner of speaking, at any rate, truly ‘moral’ claims, capable of coming into conflict with (capable of being strictly speaking inconsistent with) claims made by some standard putative moral theories such as utilitarianism, egoism, moral rights theories, and so forth. If we find that it is not plausible that we should act on them, this move of ours does not leave them in the background, still speaking to us in a weak voice; it does indeed silence them. On the other hand, if we decide that one should act on them, they now form part of what we believe is the correct (or at least the best, so far) moral theory.
Chapter 8
Reasons to Believe
8.1 Introduction In a standard definition of ‘knowledge’, knowledge requires justified belief. I take this to mean that, in order for me to know that p, my belief that p must be justified, that is; I must be justified in holding it or, to cast the point in terms of reasons, I must have reasons to hold this belief. This is true irrespective of whether the belief is empirical, moral, or mathematical, or whatever, in its content. I speak here only of justification as a necessary condition of knowledge, in order to avoid going into any problems to do with the Gettier paradox. These problems are of no relevance to my argument in this chapter. However, what does it mean for a belief to be ‘justified’? This is a rarely discussed question in epistemology. At least it is rare that it is articulated in a clear manner. All sorts of theses are put forward, saying that under certain circumstances a believer is, or is not, justified in a certain belief. But in most cases it is far from clear what kind of claim is made. How are we to assess these various different claims? To borrow a distinction from R.M. Hare,1 we are presented with all sorts of criteria of epistemic justification, while the meaning of the term ‘epistemic justification’ is often left undefined. Or, in some cases, one suspects that a simple semantic naturalism is taken for granted: the criteria are thought of as straightforwardly giving the meaning of the term. In order to answer the question about meaning I will sketch some possible positions in the discussion and develop one of them, the one I find most fruitful. According to this position, epistemic justification can be conceived of in motivational, Humean terms. To be justified is to hold beliefs that it is rational to hold, given the rest of ones beliefs together with a desire which can be characterised as a desire to realise a specific epistemic goal. The kind of rationality we are here speaking of is familiar from standard causal theory of action, it is characterised by a ‘fit’ between what we do on the one hand and, on the other hand, what we believe and what we want to accomplish. The main thrust of my argument, however, is not that
1
The Language of Morals, Chapter 6.
T. Tännsjö, From Reasons to Norms, Library of Ethics and Applied Philosophy 22, DOI 10.1007/978-90-481-3285-0_8, © Springer Science + Business Media B.V. 2010
105
106
8 Reasons to Believe
this theory is the one that gives a ‘correct’ answer to the question what it means for a person to be ‘justified’ in a belief. My point is only to show that when answering this question, there is no need to resort to any normative notion of justification. It is not likely that, when we use the word ‘justified’ we presuppose that there are any epistemic norms that are sui generis and this is good news since, more importantly, no such norms exist. We can hence safely accept a naturalistic theory of epistemic justification and avoid both a normative notion and an error theory. It is far from clear what it means to ‘naturalise’ epistemology, of course, but, in a minimal sense, it means that epistemology is not seen, in its own right, as normative. Epistemology is not, in its own right, a source of normativity. In this (restricted) sense, at least, epistemology is ‘natural’, I will argue. Before going into the presentation of this position, however, it should be noted that many epistemologists seem to take up what we might call a semantic naturalistic view of epistemic justification. According to these thinkers, to be ‘justified’ in a certain belief means no more, and no less, than, say, having satisfied a certain requirement. At least many epistemologists give this impression when they claim that, in order to be epistemically justified in holding a certain belief, the content of this belief must be self-evident for the person at the time, or part of a coherent set of contents of beliefs held by the person, or arrived at in a reliable manner, and so forth. If this is a mere stipulation on their part I have nothing to object. Given such a stipulation it is also possible to stipulate a meaning of the word ‘knowledge’ in a standard manner (as justified true belief). And given these stipulations we may pose meaningful and interesting questions such as: Is it possible for us to gain knowledge (to acquire justified beliefs)? Is this possible in general? Is it possible in a certain field of inquiry, such as in ethics, mathematics, in inquiries into the past, or into the future, and so forth? And since these stipulations do not evoke any epistemic norms proper or sui generis, I need not question them. However, in many cases one gets the impression that these ‘theories’ of justification are not just intended as mere stipulations. Why then call them theories. So, if they are theories, what are they theories about? A natural understanding of them is to see them as semantic theories about what we mean when we claim that beliefs are or are not ‘justified’. Again I need not quarrel with these theories since, once again, no epistemic norms of a sui generis kind are here presupposed. However, I can’t help noting that this (simplistic) version of semantic naturalism is no more plausible in epistemology than in ethics. It is vulnerable to the open question argument: If such a definition were correct, the question ‘Am I justified in believing what is self-evident, or part of a coherent set of beliefs, or arrived at in a reliable manner?’ would not be open. But clearly, such questions are open. However, other thinkers have claimed that we should make a distinction between the meaning and criteria when we speak of epistemic justification. On this notion, a belief’s being arrived at in a reliable manner may be a criterion of its being justified (at least if the believer believes that it has been arrived at in a reliable manner), but it doesn’t constitute the meaning of ‘justification’. And the meaning of
8.1 Introduction
107
‘epistemically justified’ should be conceived of in normative terms. If I am epistemically justified in holding a certain belief then I am under an epistemic obligation to hold it (or, I am at least epistemically permitted to hold it). And, to the extent that I rely on all sorts of rules of thumb and methods of investigation when I adopt, hold on to, or revise my beliefs, I had better see to it that these rules and methods are epistemically justified, that is, that they are such that, epistemically speaking, I ought to use them. Recently it has been common to speak of ‘requirements of rationality’. Here it will be taken for granted that such a distinction between meaning and criteria makes sense. However, my claim will be that the best way of making sense of it is in terms of instrumental (Humean) rationality rather than in normative terms. Requirements of rationality are not normative. I make this semantic claim even if I need not really resort to it in my argument. My crucial claim is ontological and to the effect that there are no epistemic norms of a sui generis kind. This means that, if we cannot make sense of utterances about what we are and are not justified in believing, without resort to the idea that these utterances make reference to epistemic norms of a sui generis slant, we will end up with an error theory in epistemology. Since this is not an attractive position, let us see if there is a way of avoiding it.2 We should then try out the following possibility: A belief, or a set of rules of thumb and methods of investigation, is justified for a person at a time if, at the time in question, it is rational for the person to hold on to it, given the rest of this person’s beliefs and the desires held by the person. And, in particular, a person is epistemically justified in holding it if, given the rest of the beliefs the person is holding, and the epistemic goal of the person (to be explained below), it is rational for the person to hold on to it. We rely here on a standard notion of rationality, well known from the causal theory of action. It may be objected that we are never, or at least rarely, free to change our beliefs. Doxastic voluntarism is false. This was a claim that David Hume famously made: We may, therefore, conclude, that belief consists merely in a certain feeling or sentiment; in something, that depends not on the will, but must arise from certain determinate causes and principles, of which we are not masters.3
And Bernard Williams has made the even stronger claim that doxastic voluntarism is not only false, it is incoherent.4 I will not take any definite stand to these claims. I tend to believe that at least Hume is right, but, even if he is, we may speak of instrumental rationality of belief, if only indirectly. For, even if we rarely or never are free to believe otherwise from how we do, the instrumental notion of rationality is applicable to our methods of investigation, our skills at critical It leads to the awkward conclusion that no one is ever justified in accepting it (or any other proposition). Many people have commented on this, for example Terence Cuneo, The Normative Web, p. 117f, and Philip Stratton-Lake, ‘Introduction’, p. xxv. 3 David Hume, A Treatise of Huma Nature, p. 624. 4 See his ‘Deciding to Believe’ about this. 2
108
8 Reasons to Believe
thinking, and so forth; these we are capable of choosing or choosing to develop. But this means that, indirectly, we can say of a belief that it is rationally held by a person, meaning by this that it has been arrived at, and is kept in a rational manner, meaning only that it has been arrived at, or is kept, through the use of rational methods of investigation, and the rationality of these methods is assessed in relation to this person’s epistemic goal. In the sequel I will just take this proviso for granted and allow myself to speak of rational beliefs simpliciter. I will discuss a similar problem with respect to our desires and wants in another chapter (Chapter 9 on reasons to desire), where it is of more direct importance to my argument. We are not free to desire or want any more than we are free to believe, but we can foster desires and wants in ourselves indirectly, though our actions. Enough of this for now. Given this notion of an epistemic goal, epistemology gets ‘naturalised’. Epistemology is not conceived of as a source of normativity in its own right. There are no epistemic norms (or permissions) proper. This is the ontological part of the theory I defend. It has a semantic aspect as well. I claim that what is stated in the previous paragraph gives a rough answer to what it means for a belief to be ‘justified’ for a believer. When this claim is added to the negative ontological claim to the effect that there are no epistemic norms of a sui generis slant, we get a semantic naturalism that is not vulnerable to the open question argument; at least this is what I will argue. And we do steer clear of any error theory. It has been claimed that, if epistemology is in this way naturalised, then it cannot address, let alone answer, sceptical doubts of a Cartesian kind. Barry Stroud says, for example, with reference to Quine, who conceived of epistemology in a naturalised manner: … there still appears to be a question (and what looks like the most basic question we can ask about our knowledge of the world) which Quine’s epistemology does not and cannot answer.5
The problem Stroud has in mind is that of refuting scepticism by validating our belief that the world is not completely different from what it seems to be. According to Stroud, this is: … the traditional philosophical question of how we can ‘validate’ all our knowledge of the world and thereby know that the world matches up with the way it is perceived to be.6
I will not address the question of whether Quine does in an appropriate manner handle sceptical doubts, but I will try to show that, in principle, a naturalised epistemology, conceiving of epistemic rules in an instrumental, Humean, manner, can meet this challenge. But in order to show this I need to specify a kind of epistemic goal, in terms of which a belief can be justified for a person (at a time).
Barry Stroud, ‘The Significance of Naturalized Epistemology’, p. 460. Op. cit., p. 74.
5 6
8.2 The Epistemic Goal
109
8.2 The Epistemic Goal To the extent that we have reflected on our belief-forming activities, we may have come to hold thoughts of the following sort. When I observe something under perfect conditions, I should trust my senses. When I make an inference to the best explanation I should opt for accuracy, consistency, scope, simplicity, and fruitfulness, and so forth. If I find that I believe two contradictory propositions, I shall try to give up at least one of them. If I find that I believe p, and if I believe also that if p, then q, then I should believe q as well – or reconsider my belief, either in p, or in if p, then q. What kind of ‘should’ and ‘ought’ are operating in the present context? My conjecture is that they are instrumental. When we say or think things like this, we tacitly assume that there exists some goal that will be best served if we stick to these rules of thumb. What goal? Could it perhaps be any goal a person happens to have? This is not a plausible suggestion. Consider the following argument for moral realism put forward by Lars Bergström. Bergström argues that unless we believe in moral realism, our actual practise of trying to eliminate inconsistencies in our moral beliefs becomes pointless.7 We want to stick to this practice, he claims, and there is no empirical evidence against our common sense moral realism, so we are warranted in our belief in it. It is certainly true that we have no arguments for the truth or probable truth of this realism but yet, for all that, we have reasons to believe in it. This is too simplistic. To be sure, we have (Humean) reasons to believe in moral realism. But Bergström has not shown, or so I will argue, that these reasons are epistemic reasons. We can have reasons to believe all sorts of propositions (unless we do we may be sent to the gallows), but these reasons are, I submit, even if relevant to our practical purposes, irrelevant from an epistemic point of view (and therefore irrelevant to a rejection of scepticism). We have to distinguish between epistemic reasons (capable of silencing sceptical doubts) and other reasons for belief. One may think that our epistemic goal should be to hold as many true beliefs as possible, while avoiding holding false beliefs. However, this way of putting the goal runs into difficulties. How do we count our beliefs? So, perhaps, less ambitiously, this may be a (very) rough indication how the distinction could be made. First of all, a person, S, has epistemic reasons to believe a proposition, p, if his or her belief in p is (subjectively) rational in view of his or her epistemic goal. And, secondly, the epistemic goal of a person is the interest this person has in holding the belief the content of it is true, and in not holding it if it is false. This notion of a reason to believe is no less instrumental than the one presupposed by Bergström. However, the goal with reference to which our beliefs get rationalised is narrower. William James made an influential and often quoted statement of this goal: There are two ways of looking at our duty in the matter of opinion – ways entirely different, and yet ways about whose difference the theory of knowledge seems hitherto to have Cf. his ‘Outline of an Argument for Moral Realism’.
7
110
8 Reasons to Believe
shown very little concern. We must know the truth: and we must avoid error – these are our first and great commandments as would-be knowers; but they are not two ways of stating an identical commandment, they are two separable laws.8
In James’ statement of the epistemic goal it is clear that he takes it to provide us with what I have here called Humean reasons for belief. However, James overstates the first norm in one respect. What he is after is knowledge. I take knowledge to imply not only true belief but justification as well. As I will explain below, I don’t think that justification is part of our epistemic goal. Epistemic justification of our beliefs is not what we seek but what, if we are lucky, we arrive at. What we are after is rather to believe, than to know, what is true – just as we want to avoid error. Both norms are concerned with belief. There is a kind of restriction built into the two norms. We ought to adopt them ‘as’ would-be knowers. Why is this kind of restriction built into them? For the following reason, I submit. When these norms operate, we conceive of ourselves as reasonable persons who will go on from where we happen to be at the moment indefinitely with our intellectual pursuit unperturbed by all sorts of contingencies in life. In each case, when we ponder whether to add or retain a certain belief, or to add or retain its negation, or to suspend judgement the goal is operative. This may not be a realistic view. It might well be that, if we obtain a true understanding of our situation, we will become utterly depressed. At least it has been said that a statistic correlation exists between self-esteem and realism with respect to how we are conceived by others. Only depressed people have a realistic view of how they are looked upon by others. This may be more than a mere statistical correlation. It may be the case that realism in this respect breeds depression. We may even come to kill ourselves, if we find out the true story of our situation with respect to other people. So if we arrive at true beliefs about our situation, this might be the end of our intellectual endeavour. Does this mean that we have reasons not to try to find out more about our actual situation? It certainly does, but these reasons are not relevant to our epistemic goal. They have nothing to do with epistemic justification. Even if our finding out the truth of our situation means that we will kill ourselves, and gain no more true beliefs whatever, we may be epistemically justified in our beliefs about our situation, should we happen to find out about it. Unless we conceive of epistemic justification in this manner, epistemic justification will not silence sceptical doubts. A way of putting this may be to take the goal to be rather institutional, than personal. This is what Isaac Levi has to say about the goal, where it has also been restricted to a narrower field of investigation to do with prediction and estimation: The proposal to be made here is that, when efforts at prediction and estimation are not reducible to practical decision-making, the conclusions are to be judged in the light of two desiderata: truth and relief from agnosticism, and that the conclusions are to be judged in this way regardless of the scientist’s private motives and values.9
The Will to Believe and Other Essays in Popular Philosophy. Isaac Levi, Gambling with Truth, p. 58.
8 9
8.2 The Epistemic Goal
111
It should be recognised, then, that our epistemic goal is usually only one goal among many. It operates, in an incremental manner, when, qua ‘would-be knower’, I want to make up my mind whether to add a certain new belief to the ones I already hold, or whether to reject it, or whether to suspend my judgement. Once again, our goal is not to maximise our possession of true beliefs and to minimise our possession of false beliefs. So, when, as a young person, someone makes up his or her mind as to whether to study philosophy or physics, the epistemic goal is of little importance. It is reasonable to assume that, if I take up physics, I will, in some sense, come to hold many more true beliefs, and fewer false ones, than I would do if I took up philosophy. However, this is not relevant to my epistemic goal, which takes an incremental form. So it is nothing strange with a person who cares more about philosophical truth than about physical truth. If I am more interested in philosophy, I may go for philosophy, rather than physics. When I pursue philosophy, however, the epistemic goal is operative. Philosophical beliefs may be hard to come by, but when I adopt them, I want each of them to be true, and I want to avoid error. It is a sad fact that I must often suspend my judgement. This is not problematic, since I do not compare what I happen to learn while studying philosophy with the beliefs I would have obtained, had I instead studied physics. I compare each of the beliefs I adopt when studying philosophy with its negation, and with the possibility, in relation to each of them, of suspending my judgement. These comparisons, and these comparisons only, are relevant from the point of view of my epistemic goal. The epistemic goal, then, has built into it a kind of myopia. Isaac Levi has stressed this aspect of it, with respect to avoidance of error, in the following words: Avoidance of error is an invariant feature of the proximate goals of inquiries concerned with the improvement of knowledge. That is to say, in contemplating revisions of knowledge, X should desire to avoid error in making the revision immediately contemplated. Speculation as to the prospects of that revision leading to the avoidance or importation of error in further revisions which might take place later on are not to be taken into account. On my view, X should not be concerned as to whether the revisions he is about to make will or will not import error in the long run. He should be concerned with whether the revision he is about to make will immediately import error or not. But once he has made the revision and his conception of error has changed, he will no longer be concerned with avoidance of error according to his initial view. His new conception of error will take over and guide revisions at the next stage.10
This myopia must be given up, however, when we consider the epistemic goal and tries to decide what weight to give to the two goals, of believing what is true, and not to believe what is false, in our application of them. As Hempel has shown, a simplistic version of them may invite us to believe in a hypothesis, as soon as the probability given to it by available evidence is more than 0.5.11 For all sorts of strategic
Isaac Levi, The Enterprise of Knowledge. An Essay on Knowledge, Credal Probability, and chance, p. 23. 11 Carl G. Hempel, Aspects of Scientific Explanation, p. 77. 10
112
8 Reasons to Believe
reasons, this is not good policy. I ought to be much more cautious. The pursuit of truth is not just a matter of adding one belief to another. Sometimes when I add a belief to my existing web of beliefs, I become prepared to use it in turn as evidence. This is more like being prepared to follow a track than to gather stuff in a store. But then I had better be rather sure that, before I embark upon a certain track, I go in the right direction. The need to avoid false belief is much more pressing, then, than gaining new ones. Actually, this also answers another worry often voiced against an instrumental, Humean, view of epistemic justification. This is the worry that there is an obvious difference between practical and theoretical reasoning, which makes it hard to see how to reduce theoretical reasoning to practical reasoning, to wit, the fact that arbitrary choice among equally desirable alternatives is necessary in practical reasoning but forbidden in theoretical reasoning. As a matter of fact, it is necessary in theoretical reasoning as well. If my option is to accept h, or to suspend my belief, and if the evidence in support of h is very strong, but not strong enough to render belief in h epistemically obligatory, I may well be in a situation where an arbitrary choice is rational. When is the evidence sufficient to warrant assent, then? As was noted already by Hempel,12 and as is still the case, there seems to be no simple algorithm available here. So I do not pretend to having stated this epistemic goal in any entirely sharp manner, and neither did James, of course. The goal has only been put forward and adumbrated in a vague form. However, this is as it should be, since the kind of justification we are here thinking of is instrumental and hence subjective. The goal can be stated no more clearly than the concerned individual has actually stated it for him or herself.13 Note also that the exact form of the goal may vary between individuals. While some may stress (optimistically) the importance of believing what is true, others may be more eager (pessimistically) to avoid error. Since people have different interests, this would give a somewhat individual twist to the epistemic goal of people. All this renders plausible a kind of weak epistemic relativism. Some people, having studied the theory of belief revision, may have adopted very sophisticated and highly idiosyncratic ideals in this field (this is certainly true of Isaac Levi, for example), while the layperson in this field (like the present author) takes up a more relaxed view. Moreover, while some people are what we may call principled instrumentalists, since, for pragmatic reasons, they care only about the observational consequences of theories (they don’t give a damn about the truth about abstract matters), others have adopted a more thoroughgoing epistemic goal so, while, temporarily, for ‘tactical’ reasons, they may treat abstract theories as mere instruments,
Ibid. But, of course, the idea of an epistemic goal could be worked out in more detail and in a more sophisticated manner. For such a development of it, see Isac Levi, ibid. 12 13
8.2 The Epistemic Goal
113
they hope to find, in the final analysis, a true explanation of the very essence of our universe. But why should people adopt the kind of epistemic goal I have indicated, or at least roughly adumbrated? No one is under any obligation to adopt it. However, we all do. Why is this? As a matter of fact, we have little choice here but accepting the goal. The evolution of our species has indeed made the choice for us. If ever there were creatures who did not implicitly share this goal, or a goal of roughly this kind, they must have been swept away and, therefore, their genes have not multiplied. The obvious reason for this is that, whatever more specific goals we may set ourselves in life, we will not be successful in implementing them unless we are able at least to take up a fairly realistic view of our close environment. It might be objected to the epistemic goal here stated, that it is too crude. Why only accept three possible epistemic attitudes, to accept a given proposition, to reject it, and to suspend judgement with respect to it? This seems to be a true account of how we adopt epistemic attitudes, however. This does not mean that, when we assess a certain proposition from a different point of view, we cannot adopt a more sophisticated stance with respect to it. If I am invited to jump from a certain height, it may be true that I have to suspend my judgement with respect to the proposition that, if I jump, I will be hurt. I may feel that, for all I know, I may jump and be safe. However, if it seems likely to me that I will be hurt, it may be a good idea not to jump. Or, at least, it may be a good idea not to jump unless by jumping I may gain something. Here the idea of subjective probability comes in place. We may measure it in terms of our willingness to take chances. And, if we wish, we may even say that we have a certain degree of belief in a proposition, a degree that varies from 0 to 1. However, even in a case where it is plausible to say of me that, to a degree of 50%, I believe that p (I’m prepared to bet on the hypothesis), we should note that this is merely a metaphor. In the circumstances, I have really suspended my judgement. What if I truly believe that p, does this indicate that I ought to accept to bet on p, at any odds? It does not. If I believe that p, then, in my pursuit of truth, I am prepared to conditionalise new beliefs on my belief that p. However, even if I am certain that p (that, say, the person in front of me is my wife), I may still be reluctant to accept a wager offering me 1 dollar if she really is my wife, but which means that all my children will be killed if she is not. In the first place, the very presence of this wager may raise my suspicion. Perhaps the woman in front of me isn’t my wife, after all. My new concerns may mean that I loose my epistemic justification in my belief that the woman in front of me really is my wife. Perhaps she has been exchanged for a clever impostor? But even if this doesn’t happen, even if I stick to my conviction that she is my wife, I need not accept the wager. I may turn it down as disgusting. This need not mean that I suspend my judgement with respect to p. I may well stick to my belief that p. Furthermore, I may well be epistemically justified in my belief that p. And yet, this does in no way mean that I feel tempted to accept the wager. I’m simply not the kind of person who bets on hypotheses of this kind.
114
8 Reasons to Believe
8.3 Coherence and Justification In modern epistemology it is often taken for granted that coherence is a sign of justification. And according to the belief that coherence is a ‘guide to truth’ coherence and justification go together also on the notion of justification here defined. If coherence is a guide to truth then, if I find that a certain belief I hold coheres with the rest of my beliefs, I may take this as a further argument for the truth of this belief. Then, given my epistemic goal, it is reasonable for me to adopt it. However, even if there exist attempts to show that coherence leads to truth it would be premature to say that these attempts have been altogether successful. On the more plausible understanding of coherentism, Davidson,14 Tersman,15 and others to the contrary notwithstanding, coherence may not, objectively speaking, be a guide to truth after all. But even if it is not, coherence is a sign of justification where justification is thought of in instrumental, Humean, terms, and in terms of an epistemic goal of the kind here described. For a subjective relation exists between coherence and truth and a new metaphor for it will be provided: justification (in a belief, for a person) is a matter of how much truth the person believes to be invested in this belief. Even on this new metaphor coherence does connect (subjectively) to (objective) truth and it does provide all the justification we need for our beliefs. If the notion of justification here put forward is taken for granted, the following seems to follow: If p is a member of a set of all the beliefs B, of a person S, then, the more p coheres with B, the more justified is S in his belief in p. Or, if there are some incoherent members of B that are in no way relevant to the truth of p, this should not mean that S is not justified in his belief in p. One way of handling this would be to require that p should cohere with a – conservatively – revised version of B. Given this proviso, we may say that the more p coheres with B, the more justified is S in his belief in p. And this is so, even if the fact that p coheres with B does not provide S with any further argument for the truth or probable truth of p – not even if S knows about this coherence. The justification we find for our beliefs, when they form part of a coherent set, has to do with the simple fact that, the more coherent our web of beliefs is, the higher the price of giving up any single member of the web. The justification is pragmatic. And the price is specified with reference to our epistemic goal. When I give up a belief, which fits well with the rest of my beliefs, many beliefs have to go together with it. But this means that I have to give up what I believe are true beliefs. It might be wondered whether we really need a justification for holding on to the beliefs we have. It might be tempting to argue that we certainly need reasons for changing our beliefs, but that we need no reasons to hold on to the ones we happen
See his ‘The Coherence Theory of Truth and Knowledge’, pp. 307–319, where he argues that most of our beliefs must be true. 15 In his seminal Reflective Equilibrium, Chapter 5, Tersman elaborates on Davidson’s conjecture and connects it more clearly to coherentism, trying to show that those beliefs that best cohere with our system are most likely to be true. 14
8.3 Coherence and Justification
115
to hold? This is the line taken up by Isaac Levi16 but here I disagree with him. On his view, we need no justification for holding on to our beliefs, only for changing them. I think there is no difference between the two cases. However, it may seem that the difference between his view and mine is merely terminological. We should distinguish clearly here between, on the one hand, my justification for holding (on to) a belief and, on the other hand, the arguments I may possess for the truth of (the content of) the same belief. When I firmly hold a belief I need no arguments for the truth of it (all arguments must start with some premises that are taken for granted). However, this does not mean that I am not justified in holding on to it. This means only that, given my epistemic goal, and the rest of my beliefs (including a certain degree of self confidence to be explained below) it is rational for me to hold on to the belief (after all, given that I have the beliefs I have I must think that, had I not had these beliefs, I would have missed out on some truths about the world). It might be suspected that this view of justification leads to incorrigibilsm.17 And, in a sense, it does. We may say, correctly, that, once I hold a certain belief, I’m rationally incorrigible; given my other beliefs, and my epistemic goal, it would be irrational of me to give it up. However, rational incorrgibilism does not imply actual incorrigibilism. We often change our views. This is how it can happen. Suppose I hold a theory T. Then I observe that p. I’m struck by the belief that p. On my view, I have reasons to stick to p (since I believe that p). It is also rational for me to stick to T (since I believe that T, and since T coheres with many of my beliefs, we may add). But suppose p is inconsistent with T and suppose I realise this. Now I’m in trouble. I have to give up one of p and T (or both, and suspend judgement). If T, as I have assumed, is well entrenched, then my first temptation is to give up p. However, if I believe that my observation that p is of a kind that is likely to recur, I may find it more tempting, in the final analysis, to see what I can do about T.18 Perhaps I can find a way of revising T into another theory T*. I may hope that this other theory, T*, will, not only be consistent with p, but explain p as well as the relative success of T (in the way described by Popper).19 After all, I have inductive evidence to the effect that such things have happened in the past. Suppose I succeed in my intellectual endeavour. I succeed in articulating T*. And T* turns out to be the kind of bold conjecture I had hoped I could find. It is now rational for me to give up T and adopt T*. I’m now in an epistemic situation in which it is rational for me to stick to T* and p. So, even if the view on justification I defend implies a kind of (rational) incorrigibilism (and a certain degree of intellectual rigidity), it doesn’t imply actual incorrigibilism. We are often struck by new beliefs. This is neither rational nor irrational; it is something that simply happens. And once it happens, it prepares the ground for reasoned revisions of our beliefs.
According to a correspondence in relation to this chapter. This is a worry raised in correspondence by Isaac Levi. 18 This point has been elaborated on by Gilbert Harman in his book Change in View. 19 See his Objective Knowledge, p. 16. 16 17
116
8 Reasons to Believe
It is sometimes said by coherentists that beliefs, and beliefs only, can justify beliefs. Donald Davidson famously claims, for example: What distinguishes a coherence theory is simply the claim that nothing can count as a reason for holding a belief except another belief.20
But on the view here put forward this is not true. Beliefs cannot all by themselves justify beliefs; they owe their justificatory power also to a desire held by the believer (to live up to a certain epistemic goal). Beliefs and desires together justify (the adoption, the maintenance, or the revision of) beliefs in the same manner that beliefs and desires rationally justify (other) actions (by the agent’s own lights). Note that, on this notion of justification, coherence is not something we seek. Coherence is something we, hopefully, arrive at, when we pursue the truth and seek explanations for (the content of) our beliefs. Once we have arrived at coherence, however, we can’t help but finding ourselves justified in each and any one of our beliefs. And this is the kind of answer to sceptical doubts we need. It is certainly not possible to prove that the Cartesian sceptical challenge, that there are no tables and chairs around us, is false. But we have extremely good reasons to believe there are tables and chairs around us. And, more generally, we have extremely good reasons to hold on to our belief, to use Stroud’s way of putting it, that ‘the world matches up with the way it is perceived to be’.21 Much truth is invested in this belief. Moreover, to the extent that this belief happens to be true, we do possess knowledge.
8.4 Is Epistemic Justification a Feasible Goal? I have sketched what it means for a person to be epistemically justified in adopting a belief, p. It means that, from the point of view of his or her epistemic goal, and given the rest of her beliefs, adopting the belief that p makes sense, that it is ‘rational’, in the standard sense we know from standard causal theory of action. Is this a feasible goal? To show that it is I need to show that a person can have the goal in question. Furthermore, I must show that a person can believe about his or her addition of a certain belief, that it furthers the goal. Take the first question first. There is nothing strange or incomprehensible in the statement of the epistemic goal. We can all understand roughly what it is that James is after, when he states his two norms. And the clarification I have made of the goal makes it even more feasible. The clarification gives the individual some license in the weighing of the two norms (seek true belief while avoiding error). The assumption that we should consider ourselves as indefinitely engaged in a purely intellectual endeavour, within the field where we are for the moment operating, renders things easier as well. It is clear both that people can have, and do have, an epistemic goal of the kind adumbrated by James. Donald Davidson, ‘A Coherence Theory of Truth and Knowledge’, ibid., p. 310. Op. cit., p. 74.
20 21
8.4 Is Epistemic Justification a Feasible Goal?
117
The second question may seem more troublesome. Can I really have complex beliefs of the kind here required? I have to believe that, if I adopt p, or stick to p, or abide by a certain rule of thumb (such as: when you find that two of your beliefs are contradictory, give up at least one of them!), I further my epistemic goal. Can and do we form beliefs such as this one? Take first the most simple case. The mere fact that I believe p is a reason for me to stick to my belief. For, when I believe p, I believe that p is true. And, being reasonably self-confident, I must also believe that, if I believe that p is true, then my belief that p is true. I do not want to give it up. But could it not pay in the long run for me to give up a true belief? No, this is hard to believe, once we have taken for granted that my intellectual pursuit will continue indefinitely. This means, as we have seen, that the possibility that my true belief in p will, for some strange reason, cause my death (say), is irrelevant to my epistemic goal (as here defined). But could it not be argued against this kind of intellectual conservatism that, if I had had different beliefs, then I would have been justified in these different beliefs instead, so it should in no way be problematic for me to give up my favoured beliefs? The mere fact that I believe p cannot be a reason why I should believe p. No, this argument is mistaken. It is certainly true that there is a possible world where I hold very different beliefs from the ones I actually hold, and where I am justified in holding them instead. I could have been an omniscient God, for example. However, the closest world in which I do not hold a certain belief, which I do hold in the actual world, is not a world where I am God. It is rather a world where, carelessly, I have made some kind of intellectual mistake. This is how I must view matters if I am reasonably self-confident (and not omnipotent), at any rate. I must feel satisfied then, that I hold each one of the beliefs I do hold. For, had I not held on to it, I must think, then I would alas have been missing a true belief. And would I have held the negation of it, I would have had a false belief. The evaluation is made from the point of the view of the beliefs one actually holds. There is no other point of view accessible to me. Suppose I believe that there is a cup in front of me. I stick to this belief. And I’m pleased that I hold it since, in the closest possible world where I don’t, I have missed out on reality (in this very respect). I don’t deny that there is a possible world where there is no cup in front of me, and where I don’t believe there is one. But such a possible world is ‘further away’ from the actual one – if this spatial analogy is in order – than the one where the cup is where it is in the actual world, but where I happen not to notice this. All this means that in a sense I have to believe that I am infallible. But the kind of infallibilism that we must necessarily exhibit is of an innocuous variety. It is consistent with the acknowledgement that earlier I held, in the same manner, other beliefs, and, in the future, it is likely that I have given up some of my (now) cherished beliefs. In fact, inductive reasons tell me that the latter is probably the case. So I may well believe that my entire set of beliefs includes some false ones. However, in relation to each and any one of my beliefs, I must believe that it is correct. Not to believe this is not to hold the belief in question. And if a particular
118
8 Reasons to Believe
belief coheres with the rest of my beliefs, that is, if much truth is vested in it, I have very good epistemic Humean reasons to hold on to it. However, this argument takes care only of our existing beliefs, as we hold on to them. What are we to say of the case where we try to expand our web of beliefs? Even if, for the reasons just given, we are justified in sticking to a belief we already hold, on what grounds can we adopt new ones? How can we be justified in coming to hold new beliefs? As I have indicated, in order to change our views, we must make new experiences, changing our epistemic situation. We can seek such experiences, of course. When we do then, typically, we do so with reference to certain epistemic rules of thumb or methods of investigation. But how do we assess such rules of thumb or methods of investigation from the point of view of our epistemic goal? Is it true of us that we can hold, and we do hold, beliefs about the efficiency of these rules and methods with respect to our epistemic goal? One could think that, unless we take our criteria of theory choice, say, to be basic and sui generis, we would end up in scepticism. For we know very little about what way of selecting explanatory hypotheses is the most efficient way with respect to truth. To counter this objection we should observe that the epistemic goal (try to believe what is true and not to believe what is false!) has to be understood in a subjective sense. And, as a matter of fact, we have all sorts of beliefs in this field. Moreover, we succeed in making all sorts of theory choice all the time. Note that theory choice is not something that scientists do exclusively, but a commonplace experience in our mundane lives. We select among competing explanations when, in our dealings with people we meet, we practice ‘folk psychology’; we select among competing explanations when we mend our car engine, and so forth. So even if we are hard put to articulate the criteria we use (just as hard put as are most practising scientists), we do find ways to make our choices. And even if our beliefs in this field are extremely unsophisticated (unless we have studied the philosophy of science thoroughly), we do possess such beliefs, and we succeed in putting them to action.
8.5 The Sui Generis Approach22 We have seen that an instrumental, Humean, account of epistemic reasons is both possible and viable and that it does address the problem of scepticism. But are there any positive arguments in defence of the position that we should yet, for all that, take epistemic norms to be basic and irreducible, that is, to be normative sources in their own right? According to the alternative, sui generis account, when we say that a person is justified in holding a certain belief, we express a genuine evaluation in its own right. This section owes a lot to Lars Bergström’s highly instructive ‘A Defense of Quinean Naturalism’.
22
8.5 The Sui Generis Approach
119
There is a set of epistemic norms, requiring of us to believe certain things (in certain situations), to disbelieve other things (in other situations), or to suspend our judgement (in yet other circumstances). These norms include among them, on this account, such things as cannons of inductive logic to the effect that, when making an inference to the ‘best’ explanation, we should opt for accuracy, consistency, scope, simplicity, and fruitfulness, and so forth. These criteria of a ‘good’ explanation, then, are not taken in any instrumental sense. These criteria are taken to be both normative, in a categorical sense, and basic and irreducible to other types of norms or judgements. In particular, it would be a mistake to accept an instrumental account of epistemic justification of the kind just adumbrated. Such a view has indeed been defended explicitly, most notably by Jaegwon Kim, who claims that a main task of Western epistemology since the days of Descartes, Locke and Hume has been to ‘identify the criteria by which we ought to regulate acceptance and rejection of beliefs’.23 Kim says ‘that justification is a central concept of our epistemological tradition, that justification, as it is understood in this tradition, is a normative concept, and in consequence that epistemology itself is a normative inquiry whose principal aim is a systematic study of the conditions of justified belief’.24 And indeed: ‘Epistemology is a normative discipline as much as, and in the same sense as, normative ethics’.25 Is there anything to be said in defence of such an approach? I kind of doubt it. Let’s see how Kim argues. Kim’s main reason for claiming that epistemology is normative is that ‘justification’ is a normative concept. His argument for this claim is as follows: But justification manifestly is normative. If a belief is justified for us, then it is permissible and reasonable, from the epistemic point of view, for us to hold it, and it would be epistemically irresponsible to hold beliefs that contradict it.26
Kim is certainly right when insisting that, ‘from an epistemic point of view’, we have the kind of obligations and responsibilities he mentions. But this does not show that we are here facing any norms in their own right. It only means that, if we want to stick to our epistemic goal, we had better believe so and so. If he were right in his claim that epistemology presents us with a source of normativity in its own right, then this would have some very strange implications. If I am correct when I hold that we have all sorts of opinions about how to go about when selecting theories, having our epistemic goal before our eyes, how strange would it not be if, when we wanted to opt for a certain theory, in the belief that the theory gives us a true picture of the world, we found out that, from the point of view of certain epistemic norms in their own right, we should not opt for this theory! Perhaps such a thing could take place within a scientific community that has come to stand on ceremony. But no one would get away with it in a real life situation. ‘What is “Naturalized Epistemology”?’, p. 33. My italics. Op. cit., p. 35. 25 Ibid. 26 Op. cit. 23 24
120
8 Reasons to Believe
There we would not allow it to happen. But then we should not allow it to happen within any scientific community either. Once we contemplate the mere possibility that basic epistemic norms may conflict with our interest in holding true (and eschewing false) beliefs, we realise how utterly strange the idea is, that there could exist such basic epistemic norms. Or, should we say that there is extensional equivalence between various putative normative epistemic requirements and the fulfilment of the epistemic goals I have defined, and should we say that this is so for the simple reason that epistemic norms require from us that we adopt the epistemic goal?27 If this is so, the argument I have just given has no bite, of course. But if the epistemic norm plays no role other than the technical norm defined by the epistemic goal, why believe that it exists in the first place? However, it might seem like at least some criteria actually articulated and put to use in both science and ordinary life cannot really be indicators of truth (not even subjectively speaking) of the chosen theories. At least this seems to be true of the criterion of simplicity. For why should we believe that a simple theory is any closer to the truth than a complex one? Could this be construed as an epistemic norm with no relation to the epistemic goal I have defined? And if this is so, does it not then indicate at least that the criterion of simplicity states a basic norm in its own right? Or, that the criterion of simplicity must be jettisoned? Now, if we do not want to jettison the criterion of simplicity, and we don’t want to do that, I submit, then we should at least play down the role of it. In particular, it would not be a sound strategy to buy simplicity at the cost of truth. However, even if we do not want to exchange truth for simplicity, there may be a place for the criterion after all. First of all, we may use the criterion, not when we assess a hypothesis, but merely when, in the first place, we single it out for testing. Quine hints at this view of the role of simplicity when he writes: Simplicity is not a desideratum on a par with conformity to observation. Observation serves to test hypotheses after adoption, simplicity prompts their adoption for testing.28
However, the distinction between adoption of hypotheses for testing and testing of them, even if clear in principle, is often blurred in practise. Quine notes this and continues his argument in the following way: … decisive observation is commonly long delayed or impossible; and, insofar at least, simplicity is final arbiter.29
This indicates that we should sometimes select one hypothesis over another on the ground that, while both are empirically adequate, that is, consistent with our observational data, one of them is simpler. However, even if we thus restrict the use of the criterion of simplicity, and abide by it, I do not think that we must take it as sui generis. Sten Arndt has suggested this. W.V. Quine, Word and Object (New York: Wiley, 1960) p. 20. 29 Ibid. 27 28
8.6 Is Justification a Blessing?
121
But why abide by it, even in this restricted way, if it is not sui generis? A possible answer to this question could be simply that, as a matter of fact, we want our explanations to be simple. This would mean to say that our epistemic goal is more complex than I have indicated above. When we select hypotheses, we are not only interested in truth, but in simplicity as well. Another answer would be to note that, even if simplicity is not as such seen as any sign of truth, we might well believe that our option for simple theories will pay in the long run. By adopting (now) the simplest explanation of available data, we think we will become more successful in the long run in our pursuit of truth. It goes without saying that our judgement concerning our future prospects is made by our present lights. But note that, if, for reasons to do with simplicity, I opt for theory T rather than theory T*, where both T and T* are empirically adequate, then, on the notion of justification explained here, I should not really invest any belief in (the truth of) T. I should rather take up an instrumentalist stance with respect to T. I should be prepared to use T (as an instrument) when trying to arrive at further beliefs about the world, but I should not believe that it is true. Given my epistemic goal, it is reasonable for me to suspend my judgement with respect to the theory I have chosen. But this is as it should be. The use of requirement of simplicity in our choice of explanatory hypotheses is indeed what has been thought to render instrumentalism a necessary and reasonable position in the first place.
8.6 Is Justification a Blessing? Note once again that justification is something one gains involuntarily, in one’s pursuit of the truth. It is nothing sought for as such. This is similar to the hedonistic paradox. We seek perfection, not happiness, but once we succeed, we become happy. In a similar vein, we seek evidence for (the truth of the content of) various different beliefs, or explanation of putative facts, not justification. Justification of our beliefs is rather something we cannot avoid obtaining, once we consistently and successfully go on with our search of the truth, evidence, and explanation. The more coherent one’s web of beliefs becomes, the more justified one becomes in holding on to each one of them. But if justification is not something one seeks, but merely something one happens to gain, then it might be natural to ask whether justification is always a blessing. This question becomes the more pertinent if it is true, as I have claimed, that justification brings with it a kind of rigidity (rational incorrigibilism). I can reflect on the fact that when I pursue truth my beliefs become more and more closely knit together. If I do reflect on this fact, I may be struck by the following problem. From the point of view of my epistemic goal, more broadly conceived of, may not the pursuit of truth turn out to be counterproductive? What if two persons both hold the belief p, where for one person, Mary, p is as an isolated belief, with few connections to other beliefs while, for another person, Erik, p coheres with the rest of his beliefs.
122
8 Reasons to Believe
We have seen that this mean that Erik is rigid in a way that Mary is not, and may not this, from the point of view of Erik’s epistemic goal, put him at a disadvantage? Suppose that both Mary and Erik, who believe that p is the case, are confronted with a source of knowledge which they consider reliable, and this source tells them that, alas, non-p is the fact, does not that mean that it is easier for Mary than for Erik to adopt non-p instead of p, and does not this mean that, from the point of view of the epistemic goal, while coherence may sometimes be a blessing it can also be at variance with the epistemic goal itself? In the circumstances, and from the point of view of the epistemic goal, more broadly conceived of, isn’t it better to be like Mary than like Erik? Moreover, if all my beliefs are closely knit together, I stand a risk of having them fall together. Would it not be better, from the point of view of my epistemic goal, to have more isolated beliefs? Then I do not risk so much when I gain new knowledge.30 I would have none of this. It is certainly true that, in the circumstances, it is easier for Mary than for Erik to accept non-p. However, when Erik contemplates his situation, it is reasonable for him to think that it is a good thing that his belief in p coheres with a lot of other beliefs and that it is difficult for him to give it up. This means that his thinking is systematic. And systematic thinking is probably a necessary condition for a realistic view of the world, especially for a deep understanding of the world. Furthermore, a radical change in many of Erik’s beliefs, even if the price for it is high, or so he may be likely to think, may well be worthwhile. Certainly, this kind of change will not take place unless Erik comes to hold some new belief (because, say, he observes that q, where q is inconsistent with the rest of his beliefs and with p in particular). This event taking place has nothing to do with rationality, of course. It just happens, if it does. However, once it has taken place, Erik is in a position where he holds inconsistent beliefs, realises this and, hence, needs to go through a deliberate revisions of his beliefs. After he has made the revision he may come to hold a more realistic view of the world. It is hard to give up well-entrenched beliefs, and it is irrational to do so if they form a coherent set; however, one’s epistemic situation may change, and revisions of one’s beliefs may become necessary. In particular it is reasonable for a person like Erik, as we have seen, to give up those hypotheses that explained p in the hope that he will end up with, what Karl Popper has called, new bold conjectures which not only explain not-p, but contradict earlier theories while explaining their relative success.
8.7 Conclusion We have seen that it is possible to conceive of epistemic justification in simplistic semantic naturalistic terms. Then there are no epistemic norms proper, and no source of normativity to be found within epistemology. This semantic naturalism is no more I owe this point, partly to a misunderstanding of, and partly to a correct understanding of, an objection raised in correspondence by Erik Olsson. 30
8.7 Conclusion
123
plausible in epistemology than it is in ethics, however. And it is possible to distinguish between the meaning and the criteria of justification. Then we may be tempted to say that we are justified in holding a certain belief if we ought to do so. And we may go on to specify the circumstances under which we ought to believe what. I have no objection to taking this latter tack, but I have insisted that, if we do, we should conceive of the ‘ought’ in question as instrumental. The saying that I ‘ought’ to believe that p is to be explained in instrumental, Humean, terms. It should be explained with reference to a goal. But it should not be explained with reference to just any goal I happen to have. That would be too crude. It should rather be explained with reference to my epistemic goal. Unless we thus restrict our reference, we will have no proper answer to sceptical doubts. Thus conceived of, epistemology is still ‘naturalised’, in the weak sense. In this sense, the point is that epistemology is not as such normative. Epistemology is no source of normativity. This is a version of epistemic semantic naturalism too, if not a simplistic one, so one might wonder if it is not too vulnerable to the open question argument. It is not. If, given the rest of my beliefs, and my epistemic goal, it is rational for me to believe p (to believe p furthers my goal, given the rest of my beliefs), then it is not an open question for me whether to believe p. To the extent that I am rational I will believe p. What distinguishes epistemic reasons to believe from other Humean reasons to believe, then, is the specific epistemic goal, to wit, our attempt to believe what is true and not to believe what is false. On this notion we can see also that coherence goes together with epistemic justification. If a belief I hold coheres well with the rest of my beliefs, then, subjectively speaking, I must acknowledge that there is much truth invested in it. I should therefore be unwilling to give it up. The fact that there exists such a thing as instrumental epistemic justification does not mean that there are no other possible goals, of the kind discussed for example by Pascal, capable of generating other reasons for belief. However, these other goals (to avoid ending up in Hell, say), having nothing to do with a pursuit of truth, are completely irrelevant from the point of view of epistemology. What if, for a person at a time, there is a conflict between reasons to believe and reasons to disbelieve a certain proposition, where the reasons to believe have to do with some general goal in life of this person, while the reasons to disbelieve have to do with the particular epistemic goal of the person in question? What is this person to believe? What reasons should take precedence? Well, to the extent that this person has an overarching goal of being rational, he or she has a Humean reason to act on the reason that is strongest. But what ought the person to believe – categorically speaking? This is a moral question. If, for example, utilitarianism is on the right track, the person has a conclusive moral reason to adopt the belief the adoption of which leads to the best consequences. However, also on this notion it is probable that, as a matter of habit, in ordinary circumstances, most people have good moral (utilitarian) reasons to develop a character where the epistemic goal of believing what is true and disbelieving what is false, forms a part.
Chapter 9
Reasons to Desire
9.1 Introduction We are all lovers of the good (by our own lights), and we tend to admire what we find beautiful. Are we under any obligation to do so? We may have instrumental Humean reasons to love what we find good and to admire what we think is beautiful, of course, but have we here identified, not only some interests that people often hold, but also a special source of obligation? Are we objectively required to love what we find good and to admire what we find beautiful? Or, are we even required to love what is good (in itself) and to admire what is beautiful, irrespective of what we may think of it? Or, if we don’t, do we at least all share a common emotional goal, like the epistemic goal discussed in the previous chapter, in terms of which our likes and dislikes can be rationalised? In the present chapter I will discuss and reject all these ideas. In passing I will reject the very notion of intrinsic (or final) value, as a special category, additional to the notion of an obligation, in sound moral thinking. I will start by discussing the most radical claim, to wit, the idea that we ought to take up certain emotions with respect to things of intrinsic or final value. The idea is often expressed in terms of ‘fittingness’. According to this idea, certain emotions are fitting with respect to certain phenomena.
9.2 Fittingness The idea that some sorts of attitudes or emotions are fitting in certain circumstances, while others are unfitting, was probably born within aesthetics. This is how Immanuel Kant used to argue. Taste is the faculty for judging an object or kind of representation through a satisfaction or dissatisfaction without any interest. The object of such a satisfaction is called beautiful.1
Critique of the Power of Judgment, p. 96.
1
T. Tännsjö, From Reasons to Norms, Library of Ethics and Applied Philosophy 22, DOI 10.1007/978-90-481-3285-0_9, © Springer Science + Business Media B.V. 2010
125
126
9 Reasons to Desire
But while this claim in Kant’s philosophy looks like a mere stipulation, a definition of beauty, a more substantial claim is taken up by G.E. Moore who, as we remember from Chapter 7, defended the view that some natural objects and some artefacts are intrinsically valuable; we ought to preserve or produce them, for their own sake. To this view he adds the view that while we contemplate beauty, we ought to take up an appropriate emotion; I suppose he would say that the appropriate attitude is admiration, or, perhaps, more in line with Kant, a disinterested pleasure or joy. This is his way of putting the point: It is not sufficient that a man should merely see the beautiful qualities in a picture and know that they are beautiful, in order that we may give his state of mind the highest praise. We require that he should also appreciate the beauty of that which he sees and which he knows to be beautiful – that he should feel and see its beauty. And by these expressions we certainly mean that he should have an appropriate emotion towards the beautiful qualities which he cognises.2
In Moore’s moral philosophy this is just a special case of something’s forming an organic whole. It is not only that we should praise him who holds a fitting aesthetic attitude; it is good in itself that he does, according to Moore. And he does extend the view to more traditional examples in moral philosophy, outside aesthetics. According to Moore, it is not a good thing when someone takes pleasure in the pains of another innocent person. Since pleasure, according to Moore, has positive intrinsic value (in a context-independent manner), the pleasure as such contributes positively to the whole of which it is a part, even if the whole is constituted by someone taking pleasure in the pain of an innocent creature. However, its presence makes the whole worse. It detracts more value from the whole than it adds to it. In general, according to Moore, the world would be better without joyful contemplation of things which are themselves either evil or ugly: The world would be better even if the evil or ugly things themselves remained in the world while the pleasure taken in them would vanish: In order to illustrate the nature of evils of this class, I may take two instances – cruelty and lasciviousness. That these are great intrinsic evils, we may, I think, easily assure ourselves, by imagining the state of a man, whose mind is solely occupied by either of these passions, in their worst form. If we then consider what judgement we should pass upon a universe which consisted solely of minds thus occupied, without the smallest hope that there would ever exist in it the smallest consciousness of any object other than those proper to these passions, or any feeling directed to any such object. I think we cannot avoid the conclusion that the existence of such a universe would be a far worse evil than the existence of none at all. But, if this be so, it follows that these two vicious states are not only, as is commonly admitted, bad as means, but also bad in themselves.3
Personally, I cannot help thinking that this idea is silly. If it were taken seriously, one would have to claim that, of two situations, where, in the first one a person suffers terribly, but where no one takes pleasure in his suffering and where, in the other one, the person suffers slightly less, but where someone takes pleasure in his Principia Ethica, pp. 189–190. Ibid., p. 209.
2 3
9.2 Fittingness
127
suffering, the former one should be preferred to the latter one. I think at least that we should reject this special instance of what has been held to be an improper attitude to an emotion. And I think many people are, as a matter of fact, prepared to agree with my judgement on this point. I once tried out the following case on my freshmen students, before they took up philosophical studies. Suppose we have failed to cure a sadist.4 She is dangerous and we have thus decided to detain her for the rest of her life. She is terminally ill. She asks her warden of a last service. Could he not provide her with the public records from The Hague War Crimes tribunal? The warden sees no harm in this. He discloses the protocols to the sadist, who treats them as pornography and revels in them. Before she dies she thanks the warden; he has given meaning to the last part of her life, she claims. Did the warden act wrongly? Provided there were no bad side-effects, my intuition is that he didn’t. Among my freshmen students, only five thought that what the warden did was wrong, while 46 thought he did the right thing (three did not know.). However, I may be wrong about this. I may also be wrong about the value of people having the appropriate aesthetic feelings. These differences of opinion seem to belong to ethics, properly speaking, however. We are here just confronting possible counter-examples to hedonistic utilitarianism, it seems. And this is certainly the way Moore himself conceives of the problem. And if this is so, we are not here confronting any new or special source of normativity, besides the one with which we are already familiar from morality. But the idea that certain feelings are appropriate, while others are inappropriate, has also been taken up in a different and more general manner. Some philosophers have thought that if something is good in itself, if it is such that we ought to promote its existence, then ipso facto we also have object given reasons to desire it (or develop a taste for it). This claim is thought of as giving an analysis of the notion of intrinsic or final value. This analysis of the notion of (intrinsic or final) value then seems to confront us with yet another putative source of normativity. This tradition dates at least back to Brentano.5 According to Brentano, an object is valuable to the extent that it requires, or is a suitable object of, a positive attitude. A.C. Ewing has followed up the tradition from Brentano and famously defined ‘good’ in the following way: … we might adopt a technical term and define ‘good’ as what ought to be the object of a pro attitude (The Definition of Good, pp. 148–9)
Here it is also explicitly sated that, in the quotation, the meaning of ‘ought’ should be understood in (C.D. Broad’s) terms of fittingness. According to Ewing,
4 This example is an elaboration of an example suggested by Ragnar Ohlsson in Morals Based on Needs, pp. 91–92. 5 Brentano F. (1889) The Origin of Our Knowledge of Right and Wrong.
128
9 Reasons to Desire
… it is more appropriate, or fitting, to feel disgust than pleasure at cruelty, more appropriate to desire reconciliation than revenge, to admire fidelity than clever cheating, to feel aesthetic emotions on contemplating great works of art than not to do so (ibid. p. 151).
If we cast this idea in terms of reasons – as Ewing himself did in his book Second Thoughts in Moral Philosophy from 1959, long before Thomas Nagel, Thomas Scanlon, and others, had initiated the present vogue of reason speak6 – we get the claim that we have reasons to feel disgust at cruelty, to desire reconciliation rather than revenge, and so forth. A consequence of this view, it should be noted, is that, the reason why we ought to approve of pleasure, say, and disapprove of cruelty, has to do with the nature of pleasure and cruelty respectively – not with the fact that the former is good and the latter bad. So perhaps we do not really need any notion of intrinsic or final value at all? Let me elaborate on this possibility.
9.3 Morality Without Values Thomas Scanlon has recently defended a view on values pointing in the direction of their elimination from moral theory – or, if not elimination so at least reduction to other entities. His ‘buck-passing’ analysis of value seems to me to point in the right direction, even if it doesn’t go far enough. I will first focus on the merits of the view, and below I will return to the fact that it doesn’t go far enough. According to Scanlon: Goodness is not a single substantive property which gives us reason to promote or prefer the things that have it. Rather, to call something good is to claim that it has other properties (different ones in different cases) which provide such reasons.7
To me this seems to be a good reason to give up talk about (intrinsic or final) value altogether, and just concentrate on reasons for action. At least, we should not rely on talk about value in our moral explanations. For example, if someone claims that we ought to maximise the sum-total of well-being and I ask her why we should do so, it would not be a good idea for this person to answer: Because well-being is good. This answer doesn’t add any information.8 In objection to this it may be tempting to say that, besides problems to do with how we should relate to what is valuable (in itself), we also need to know what is valuable in itself. A utilitarian and an ethical egoist disagree on one point, it may be tempting to say, to wit, they disagree on what we ought to do with whatever it is that has intrinsic or final value, but they may very well agree about what it is that See Ewing, Second Thoughts in Moral Philosophy, p. 63. Scanlon, T. M., What We Owe to Each Other, p. 11. 8 In ‘A concrete view of intrinsic value’ I defend such a view of value, and in Hedonistic Utilitarianism I take it for granted. Roger Crisp speaks, in Reasons and the Good of this line of argument as the ‘redundance argument’. 6 7
9.4 Values Without Normative Import
129
has intrinsic or final value. They may both claim, for example, that happiness, and happiness only, has intrinsic value. But this won’t do. We ought to resist this manner of speaking since it is misleading. There is no special problem to do with intrinsic or final value that these two thinkers can ponder in common. When they want to settle what it is that has intrinsic or final value they have to return to their own favoured theories. The egoist must ask: What is it that each agent should maximise in his or her own life? And the utilitarian must ask: What is it that each agent should maximise in the world? There is no common problem here about intrinsic or final value that they can confront and solve in common. To see this we should contemplate the following possibility. Suppose the most plausible version of utilitarianism is hedonistic utilitarianism, according to which we ought to maximise the sum-total of happiness in the world, while the most plausible version of egoism is egoism of a preferentialist slant, according to which each person ought to maximise desire satisfaction in his or her life. We now want to ponder what theory is correct. Could we do this while first pondering whether happiness or desire satisfaction is intrinsically good? Obviously not, unless we see this as a quest for the best normative theory. But the truth may be that the correct normative theory is a mixture of utilitarianism and egoism, according to which we ought to maximise happiness in the universe, unless this brings, in terms of desire satisfaction, terrible sacrifices on ourselves. In that case we may go for our own desire satisfaction rather than for general happiness. We could speak here of utilitarianism tempered with some agent-relative prerogatives. If this is the correct answer to our normative query, then there seems to be no answer to the question what possesses intrinsic value. Furthermore, once we have arrived at an answer to the basic normative question, that is, once we have found the correct normative theory, that is a correct criterion of right action, there is no need any more for any theory of intrinsic value. For I suppose that the correct normative theory can be stated in purely empirical (nonmoral) terms (in terms of happiness, desire satisfaction, or what have you).
9.4 Values Without Normative Import Could not value questions be questions in their own right, of no importance at all from a normative point of view? Well, this would not be at odds with my argument to the effect that there is only one source of normativity, but the idea sits ill with my view that we do not need to resort to talk about intrinsic or final value at all. It is inconsistent with my claim that there is just one kind of moral problem. So it might be of interest to look at some attempts to defend the idea of value problems as problems sui generis. There are few knock down arguments in philosophy, but one of them is the following. If we find that a normative principles tells us to prefer a world A, where we have added one person with a terrible life (worth not experiencing), to a world B,
130
9 Reasons to Desire
where many people leading happy lives, but lives below a certain ‘critical level’ have been added, then we should reject this principle. We should not accept what Gustaf Arrhenius has nicknamed the ‘sadistic’ conclusion.9 And yet, it seems that many people discussing problems in population ethics have accepted principles implying the sadistic conclusion.10 We discussed one such idea in Chapter 6, to the effect that people who lead good lives, but below a critical level, lead lives of negative moral value. Some of those who have defended such views have just not noticed this implication, or tried to downplay the importance of the implication, but at least one of them has explicitly recognised it – while arguing that, even if this implication is not acceptable in normative ethics, it may form a respectable part of ‘axiology’. Thus Erik Carlson argues that the sadistic conclusion may be objectionable, if seen as a normative principle. However, if seen as an axiological principle it is more plausible: One could accept SC/the Sadistic Conclusion/, as an axiological principle, without accepting that we ever ought to create unhappy people.11
If we do, according to Carlson, we should formulate the sadistic conclusion in the following manner: If a number of people with negative welfare are added to a population X, the resulting population may be better than the population resulting from adding to X a number of people with positive welfare.12
And he adds that, to his mind, ‘… this reformulation removes some of the repulsiveness of SC’.13 What are we to make of these claims? I take it that Carlson would admit that, if we have a choice between the A world with one terribly unhappy person added and the B world with many happy persons living just below the ‘critical level’ added to it, we ought to go for the B world. Still he argues that the A world may be ‘better’ than the B world. Has he found a sense of ‘betterness’ without any normative implications? Again, if he has, this does not threaten my claim that there is just one source of normativity. Putting that problem to one side, however, I cannot help feeling that he has not. When he speaks of ‘betterness’ he does not use the term with a sense different from the one we have in mind, when we use it as a place-holder for what it is we ought to maximise, distribute in one way or the other, and so forth, but with no sense at all. It is nonsense, strictly speaking, to claim that the B world is better while acknowledging that we ought to produce the A world, if we had a choice between them. If I am wrong about this, then Carlson owns us an explanation of the sense he has in mind. Arrhenius, G. Future Generations: A Challenge for Moral Theory. See for example Parfit, D. ‘Overpopulation and the Quality of Life’, Blackorby, C., Bossert, W., and Donaldson, D., ‘Critical-Level Utilitarianism and the Population-ethics Dilemma’ and John Broome, Weighing Goods. 11 ‘Mere Addition and Two Trilemmas of Population Ethics’, p. 302. 12 Ibid. 13 Ibid. 9
10
9.5 The Meaning of ‘Good’
131
If this line of argument is accepted, then, when we hear that we have reasons to feel or desire certain things, we had better try to find out the ‘ordinary’ normative implications of this. If we accept that we have these reasons to feel or desire certain things, what obligations follow from this? This is the crucial question that needs to be answered.
9.5 The Meaning of ‘Good’ I am tempted to argue that we should never use the word ‘intrinsic (or final) value’. Indeed, it might be tempting to give up all kind of talk about ‘value’. I realise, however, that this not a good (sic!) idea. It would be quixotic to try to do without the value terminology. However, we should think of intrinsic value, or value, or goodness, as a placeholder to be filled out with something more concrete and empirical, when we state our normative theory. And unless we succeed in stating this theory without having recourse to any evaluative terms the theory hasn’t been given any definite content whatever. We may say things such that ‘the earthquake in Lisbon was a bad thing’, of course, but this easily translates into the statement that ‘the earthquake in Lisbon caused a lot of human misery’ – with the pragmatic implicature that we have an obligation if possible to avoid human misery. Those who are not utilitarians can make the same judgement, but with a different implicature, of course. They had better to adopt a normative theory, which sticks to the case at hand, however. For I suppose no one makes the claim implying merely that the earthquake made Lisbon look ugly and smell badly. And, when we resort to talk about ‘expected value’ in decision theory, which is a natural thing to do, we should keep in mind that, even if this notion relates in one way or the other to some notion of rationality, this notion of rationality in its turn is a natural one, without any normative implications of its own. We may even accept, if we like, what G.E. Moore said in Principia Ethica, where he stated: ‘right’ does and can mean nothing but ‘cause of a[n] [intrinsically] good result’.14
and where he also made the claim that: [T]o assert that a certain line of conduct is, at a given time, absolutely right or obligatory is obviously to assert that more good or less evil will exist in the world, if it be adopted than if anything else be done instead.15
for, certainly, my claim that there exists just one source of normativity is compatible with the use of both a normative and an evaluative vocabulary – as long as we keep in mind that they can be exchanged for one another without any change of meaning. §89 p. 147. §17 p. 25.
14 15
132
9 Reasons to Desire
Personally, I tend to prefer to speak in terms of normativity, and to get rid of the evaluative words. This seems to me to be the most simple and straightforward way of accounting for morality. In particular, if we stick to the value terminology we run into difficulties when we want to account, not only for a utilitarian normative ethics like the one defended by G.E. Moore, but also for normative theories specifying agent-relative obligations. These difficulties are not insurmountable, however. We may, if we wish, introduce also agent-relative values. So perhaps I am merely expressing a personal preference when I insist that we should cast our moral theories in normative terms and then go on to search for clear empirical criteria of rightness, wrongness, and obligation. If values can be reduced to norms and norms to values, then there is little point in asking which of these two categories is the ‘basic’ one.16 It is noteworthy that G.E. Moore himself abandoned the idea that norms can be reduced to values. When he did, however, he came to endorse an utterly abstruse theory about the relations between these two realms of morality. His view in Ethics was that ‘duty’ and ‘expediency’ (i.e. ‘being conducive to the greatest possible amount of intrinsic goodness’) when predicated of actions are equivalent, but not identical in meaning.17 I suppose he must have meant that ‘X is expedient’ and ‘X is our duty to perform’ share truth values in all possible worlds, but he failed when he was challenged18 to explain why this is so, if it is not because of synonymy between these two expressions. I suppose that the kind of explanation one could give would be along the following lines: normative and evaluative terms differ in meaning in the same sense that being ‘equiangular’ and ‘equilateral’ do. I don’t find this illuminating, however. For our most immediate ways of finding out whether a triangle is equiangular or equilateral differ; with respect to norms and values, there exists one unique epistemic problem, or so I have argued, at any rate. Hence I believe Moore was closer to the mark in Principia than he was in later writings. His idea that norms should be reduced to value was unfortunate and lead him astray, moreover. It would have been better if he had argued the other way round. His idea that norms should be reduced to value lead him to the mistaken belief that egoism is an inconsistent view. He could have avoided this fallacy, of course, if he had also introduced a notion of agent-relative value, but all this seems to me to be to complicate things. From a heuristic point of view we had better see the normative terminology, with which we are all familiar, as basic, and to try to articulate authoritative criteria of right and wrong actions (simpliciter) in non-moral terms. Once we have accomplished this we have solved the moral problem. Having clarified these terminological issues, let us return to the idea put forward by Ewing to the effect that certain emotions are in place, as ‘fitting’, in relation to certain phenomena. And let us focus on some modern attempts to defend this view.
In my Moral Realism I opted instead for intrinsic value as the basic moral category. I did so for what I (now) believe were bad epistemic reasons. 17 Ethics, pp. 107–108. 18 C.f. his ‘A Reply to My Critics’ about this. 16
9.6 Recent Attempts to Defend the Idea That There Are Reasons to Prefer or Desire
133
9.6 Recent Attempts to Defend the Idea That There Are Reasons to Prefer or Desire Derek Parfit, once again and most recently, has defended the claim that, besides Humean reasons, there are not only moral reasons for action, but also reasons to prefer or desire certain things.19 On Parfit’s view, if something (pleasure, say) is what we ought to maximise (in the world, or in our own lives), then this means also that we have an object given reason to prefer or desire happiness. In this claim, as we saw above, Tim Scanlon concurs. And this is the point where Scanlon, in my opinion, is not radical enough in his reduction, through his buck-passing view, of values. To the extent that he reduces them to reasons to act, that is to obligations to act and to moral explanations of these obligations, his analysis seems to be to the point. But when he also speaks about obligations to feel he goes astray, or so I will argue. Does it even make sense to speak of obligations to feel? Even setting details to one side, I think not. To have pro attitudes, to prefer or to desire something is not typically to perform any action. The fact that ‘prefer’ and ‘desire’ are verbs tends to mislead us. Our preferences and our desires are typically things that just happen to us. It doesn’t make sense to say that we have an obligation to prefer something to something else or to feel anything in particular. This is due to the simple fact that ‘ought’ is only applicable to actions. In a previous chapter I referred to thinkers such as David Hume who have claimed that doxastic voluntarism is false (and even to thinkers who have claimed that it is an incoherent position, such as Bernard Williams). It is even clearer that desire voluntarism is false. It is sometimes claimed that, even though we cannot desire at will, our desires are at least sensitive or responsive to critical arguments. But this is not true. We are sometimes sensitive or responsive to rational argument and can, hence, indirectly, change the way we will or want, and cultivate new desires. We are sometimes free to act otherwise from how we do (in a compatibilist sense of the word ‘free’), but we are never free to feel any differently from how we actually do. Ewing is well aware of the fact that it is not up to us to feel or not feel something, in the manner that it is up to us to do or not to do something. This is actually why, in his explication of the word ‘ought’, he is using the term ‘fittingness’. The normative claim that one ought to have a certain pro attitude means only that such an attitude would be fitting or appropriate. But we may want to know which are the ‘ordinary’ normative implications of such a claim. Is the claim then that we ought to cultivate emotions that are appropriate or fitting? There are two aspects to this question. First of all, does the fact that I ought to cultivate a certain emotion mean that this emotion is valuable? This is hard to believe, at least without qualification. For the reason that a certain individual ought to develop a taste for, say, literary achievement, may well be that both she herself and other people become happy if this ‘Rationality and Reasons’, ibid.
19
134
9 Reasons to Desire
means that she makes literary attempts and succeeds in them. But this does not show that a taste for literary achievement is of value in itself, that is, if we avoid speaking of value, it doesn’t show that a taste for literary achievement is something that should be maximised in the world or even in an individual life. So unless some qualification can be given, the analysis is not plausible. Secondly, and even more importantly, does the fact that something is valuable in itself, that is such that we ought to promote its existence, create a reason for me to develop a taste for it? This is even harder to believe. It is certainly true that, to some extent, we can take action in order to ascertain that we will come to prefer or desire certain things. And we may have both Humean and moral reasons to undertake these actions and, consequently, develop these preferences. But suppose that happiness is what we ought to maximise in all our actions. Does this mean that we ought to try to see to it that happiness is something we prefer or desire? Yes, according to the authors who defend this analysis. But what if our developing a taste for happiness is counterproductive? What if our desiring happiness makes us unhappy? Well, we still have a reason to develop this desire, according to the authors who defend this analysis. However, if our desire will prevent us from achieving what we desire, we have a ‘conflicting’ reason not to try to develop it. At least this is how Derek Parfit would like to put things.20 This is not convincing. If there is any reason to develop a desire for happiness, even if this does not make us happy, we would like to know how strong this reason is. How much happiness ought we to forgo, in our search of a taste for happiness? None, it seems to me, if happiness is what we ought to maximise. In fact, the assumption that we have a reason to try to develop a taste for what is valuable seems to be open to a simple reductio. Suppose that hedonism is true and that happiness is the only thing we ought to promote in our actions. On the view under discussion this means also that we have a reason to try to develop a taste for happiness. But if this is so, then this fact seems in its turn to imply that also the attempted development of a taste for happiness is valuable in itself (contrary to our hedonistic assumption). In order to block this argument we would have to say that there are some things we ought to promote, in spite of the fact that they are not of any value in themselves (nor means to anything that is of value in itself). But this assumption seems altogether ad hoc. It is noteworthy that even if Ewing seems to have thought that, if something is intrinsically good, it is something which, on its own account, we ought to welcome, rejoice in if it exists, seek to produce if it does not exist – but only if other things are equal (ibid. p. 149). What are we to make of this qualification? A way of understanding the qualification is that we ought only to welcome or rejoice, and so forth, if the consequences of doing so do not make the world worse.
Derek Parfit, in correspondence.
20
9.7 The Simple View
135
To develop emotions that are fitting is important – but only to the extent that we need not pay any price for having them. If this interpretation is correct it seems that Ewing has himself trivialised his claim. And the theory seems still to be open to the charge that it leads to a reductio. There is of course a quite different way of understanding Ewing’s talk about certain emotions being ‘fitting’ or ‘appropriate’ (in a situation). This could be taken as a mere aesthetic judgement with no normative implications whatever. If I have understood Kant correctly (above), this is how he conceives of aesthetic enjoyment. And once again, this seems to trivialise the claim. Moreover, it makes it completely irrelevant to the discussion of ‘reasons’ for action as here conceived.
9.7 The Simple View It might be thought that much of the problems I have created for the buck-passing view, claiming that we ought to love what we ought to promote, has to do with the fact that I have taken this view to imply that we ought to develop, or cultivate, at taste for what we ought to promote.21 I have done so in order to make sense of the view. Furthermore, this interpretation is strongly suggested by some formulations made by its advocates. Thus it is certainly suggested by Derek Parfit when he writes that ‘state-given reasons to have some desire are better regarded as object-given reasons to want to have it, or to try to have it’.22 But perhaps formulations such as this one are mere slips. Perhaps the view should be understood more straightforwardly. The point is not that we ought to cultivate a taste for what we ought to promote, the point is that we ought to like or desire what we ought to promote. Is there a way of making sense of this? Well, ‘like’ and ‘desire’ are verbs, so it might seem as though our likes and desires are in our command. To like and to desire is to do something. Then, perhaps, we can have obligations to like and desire. Now, I have claimed that this does not make sense. ‘Ought’ implies ‘Can’ and, I have claimed, even though I can like and desire all sorts of things it is not the case that, when I like A, I could instead like B. Of course, it could have been the case that I had liked B instead, but this does not mean that, when, as a matter of fact, I don’t, it is in my power to do so. Suppose I am wrong about this, however. Suppose desire voluntarism is true. Suppose hedonistic utilitarianism is true, and suppose that we ought to maximise the sum-total of well-being in the universe. Then, on the buck-passing view under consideration, we ought to like and desire universal happiness. Suppose I don’t. Does it make sense to say of me that I ought to do it? On this assumption of desire
I build here on a critical comment on my argument put forward by Sven Nyholm (in correspondence). 22 ‘Rationality and Reasons’, p. 24. 21
136
9 Reasons to Desire
voluntarism, this is what utilitarianism then tells me to do, but only if, by doing so, I maximise the sum total of happiness in the universe. Let us assume, then, that we do control in the relevant sense our likings and desires. Then the advocates of the buck-passing view avoid many of the arguments I have put forward in this chapter. And yet, for all that, they do not end up with any plausible version of the buck-passing view. For now, what was intended to be a detached analysis of value-statements has been turned into a normative view of a highly dubious nature. To see this, let us return to the assumption that hedonistic utilitarianism is true. Let us furthermore assume that I do not like or desire universal happiness. Now, we may further assume, if I were to like or desire universal happiness, then there would be less happiness in the world. Ought I still to like or desire it?23 Well, this would go against the assumption that hedonistic utilitarianism is true. Let us give up that assumption then, and take this version of the buckpassing view seriously as a rival normative theory. We could think of a normative theory according to which we ought to maximise the sum-total of well-being, but only when we perform ‘ordinary’ actions. Our likings are governed by other norms. Then we must query: At what price should we be lovers of the good? At what price should we change our desires and likings, when the price has to be paid in terms of suffering? My guess is that most advocates of the buck-passing view would retreat to the claim that we should not be lovers of the good if there is a price in terms of wellbeing to be paid for our love. But then there is little bite in the theory any more. So I conclude that the buckpassing analysis of value, to the extent that it results in the claim that we should be lovers of the good, makes poor sense. Let me elaborate on this claim. First of all, it is difficult to understand the claim that we can be under any obligation to like and desire what we, as a matter of fact, do not like and desire – if this does not mean that we ought to cultivate likes and desires different from the ones we actually hold. Furthermore, even if this makes sense, there is no point in liking those things we ought to promote, unless our liking them is instrumental in our promotion of them. In particular, the claim that we ought to like what we ought to promote, if this means that we will have less of it, is morally speaking absurd. And the claim that we ought to like what we ought to promote, but only to the extent that this is not counter-productive, is ridiculous. Be all that as it may, however. The most important conclusion to be drawn from the discussion in this section is that, even if desire voluntarism is true, even if the buck-passing view in the version here discussed did make sense, which I doubt, and even if it were true, which again I doubt, it would be a normative theory in competition with other normative outlooks such as hedonistic utilitarianism. Contrary to what has been supposed in the discussion about it, it would not be a mere analysis of what it means to make evaluative claims.
This kind of objection has been known under the name ‘The Wrong Kind of Reasons’-objection. It is discussed in Jonas Olson, ‘Buck-Passing and the Wrong Kind of Reasons’. 23
9.8 Must We Will the Means to Our Ends
137
9.8 Must We Will the Means to Our Ends The kind of reasons to hold certain desires or preferences discussed thus far are of a rather ‘moralistic’ bent. The idea is that some phenomena are such that certain emotions are fitting with respect to them. This is an objective view of fittingness. Even if it doesn’t work, one might think that there are views of a more subjective nature that are more promising. One may wonder whether there are simple reasons from rationality to desire or prefer certain things – or, at least, to keep a certain order among ones preferences. One idea to this effect is the thought that one must want the means to what one desires. I will discuss it in the present section. Other ideas have to do with the relations between our preferences or wants. We ought to avoid cyclical preferences, it has been held. And there is a proper way for us to value prospects. I will discuss these ideas in the following sections. Finally it might be thought that, just as there exists a common and very natural epistemic goal for us, there might exist some kind of common emotional goal. If there is such a common emotional goal, we might perhaps be able to rationalise our emotions with reference to it. Let me start, however, with the very strong idea that there is a kind of logical relation obtaining between our wants and intentions. We have seen that a person could have Humean reasons to try to develop and cultivate certain desires or wants. He or she has such reasons to the extent that, by adopting these desires or wants, he or she furthers his or her existing wants. Furthermore, a person may have moral reasons to develop and cultivate certain desires or wants. These reasons explain why he or she ought to adopt these desires or wants. Besides these reasons to adopt desires or wants, there exist no other reasons to do so, or, so I have argued. However, once a person has a certain desire, or want, couldn’t this person be somehow compelled by logic alone, to develop certain other desires or wants? In particular, should we not say that, unless a person wills the means to his or her ends, this person makes some kind of logical mistake? Should we not say that there is a way of arguing from (expressions of) existing desires or wants – or perhaps intentions – to (expressions of) other desires or wants – or intentions? I have conceded that a rational person will tend to act on his or her strongest desires. But the notion of rationality, then, is naturalistic. This is how we tend to be, most of the time, at least to some extent. Evolution has moulded us this way. Often we do exhibit this kind of fit between what we desire, believe, and do. However, the query now is different. Some philosophers seem to think that we are inconsistent if we do not will the means to our ends. They seem to perceive a strict analogy with ordinary logical inferences here. Could that be right? When philosophers speak of ‘practical syllogisms’ they seem to indicate that there exists a kind of practical logic, a logic of intentions. This is a position suggested by Donald Davidson, for example in the following passage: Corresponding to the belief and attitude of a primary reason for action … we can always construct (with a little ingenuity) the premises of a syllogism from which it follows that the action has some (as Anscombe calls it) ‘desirability characteristic’.24 Essays on Actions and Events, p. 9.
24
138
9 Reasons to Desire
But Davidson does not elaborate this view. A philosopher who does elaborate and defend it, however, is John Broome.25
9.9 Broome on Practical Reasoning Broome wants to show how we can reason from one intention to another intention, from an end to the means to the attaining of the end, so to speak. According to Broome, there is a correct (valid) way of conducting such an argument. This seems to be an instance, then, of the idea that we ‘must’ will the means to our ends. What we have come across is a special kind of practical argument or a practical ‘syllogism’. In order to understand the argument put forward by Broome we must know what an intention is. Do intentions exist in their own right? Broome argues that practical syllogisms exist. These syllogisms come close to Aristotle’s idea of a practical syllogism. However, there is one important difference, according to Broome. An Aristotelian practical syllogism ends up with a conclusion in the form of an (external) action. On Broome’s view, we cannot deduce any such action from the premises of a practical syllogism, in the way Aristotle taught. However, we can derive an intention as the conclusion of a practical syllogism, according to Broome. So the question needs to be addressed: What is an intention? Do intentions exist? Some philosophers, such as most famously Donald Davidson, have denied that intentions exist, as an ontological category in its own right.26 An action can be intentional (under a certain description). There is no such thing as intentions, existing over and above actions, desires, and beliefs, however. I tend to sympathise with this way of putting things. However, we may perhaps say that, to form an intention is like making a resolution, or making a decision, or making a choice. But then it seems as (the forming of) an intention (or the making of a resolution, decision, or choice) is indeed (a kind of) action, even if not an ‘external’ one. According to Broome, there are logical relations obtaining between what we express, when we express our intentions. Here is the example he discusses: A person, named Leslie, is sailing. He conducts the following practical argument: 1. I am going to leave the next buoy to starboard 2. In order to leave the next buoy to starboard, I must tack, so 3. I shall tack Are (1)–(3) sentences, or are they something expressed by sentences? I suppose the latter interpretation must be the one intended by Broome. (1) and (3) may look like factual propositions, but appearances are deceptive; in fact, they are the kind of entities we express when we express our intentions, that is, they are intentions. The sentences used by Broome to communicate (1) and (3) are descriptive sentences, ‘Normative Practical Reasoning’. This claim was made in ‘Actions, Reasons and Causes’ (1963), but seems to have been given up later, at least from ‘Intending’ (1978), both reprinted in Essays on Actions and Events. 25 26
9.9 Broome on Practical Reasoning
139
but here these sentences are used in a special way, indicating that they do not express descriptive (autobiographical) propositions, but they express rather whatever it is that we express when we express intentions. What do we express, then, when we express our intentions? What is an intention? Well, we do certainly not express any concrete mental event or act (our resolution to do so and so), so intentions should not here be identified with concrete mental acts. This event (act) exists or it does not exist, at a certain place and time, that’s all there is to say about it. We cannot share with other people our intentions, conceived of in this concrete manner (any more than we can share our beliefs, conceived of in a similar, concrete manner). But I suppose we are here to understand intentions as something we can share with others. However, it seems as though there is something we do communicate, when we express our intentions. And there is a way of accounting for it. We may say that a certain person intends (or is determined, or has decided, and so forth) ‘to’ do so and so, and we seem to take for granted that anyone who understands the relevant language can ‘grasp’ the relevant intention. So I think we may safely admit that there is a kind of semantic content, which we communicate when we communicate our intentions. There is indeed an analogy here with a case where we express our beliefs. When we express a belief ‘that’ so and so, there is a certain proposition we can be said to express. When I express my belief that there is a cat on the mat, the content of my belief, what I express, is the proposition that there is a cat on the mat. This proposition can be grasped, in principle, by any one. In a similar vein, when I express an intention (a resolution or choice) ‘to’ do so and so, there seems to exist a semantic component that I communicate, something I can share with others. Let us take this for granted. Assume that there is a kind of semantic content there for us to express, when we express our intentions.27 Assume also that this content can be ‘grasped’ by anyone who understands the relevant language. But how are we to characterise it? The proposition that I will leave the next buoy to starboard is true, if, and only if, I will leave the next buoy to starboard. Is there a similar way of characterising intentions? I think not. It seems to me that we have no proper language to characterise what is expressed when intentions, resolutions, and the like are expressed. It might be misleading therefore to use the words used above, in the statement of the relevant argument. I can use a descriptive sentence to express my intention to do so and so (‘I will do so and so’. would normally do) but this sentence, when used to characterise what I expressed, gives the impression that I did express an autobiographical descriptive proposition. However, when I express an intention I do not communicate any descriptive proposition. There is a difference between communicating a belief and communicating an intention. Perhaps a better way of characterising what it is I express when I express an intention would be to have recourse to the imperative mood (even if it may sound In English we say that I believe ‘that’ so and so, but that I intend ‘to’ do so and so. In Swedish the same phrase (‘att,’ the indicator of the infinitive mood) is used for attributions of beliefs and intentions alike. 27
140
9 Reasons to Desire
a bit schizophrenic to address oneself in the second person). To have the intention (to be determined) to leave the next buoy to starboard, then, is tantamount, I suggest, to accepting, or assenting to, the imperative, ‘See to it that you leave the next buoy to starboard!’ directed at yourself. Of course, this is a metaphorical way of accounting for the content of our intentions. It should not be taken to imply that actually we ‘talk’ to ourselves. Yet, some may find the imperative mood strange in the present context. Donald Davidson, for one, has objected to the idea that forming intentions bears much similarity to commanding: … it is enough to discredit these theories to point out that promising and commanding, as we usually understand them are necessarily public performance, while forming an intention is not.28
I think it does make sense to conceive of my forming an intention as my assenting to the relevant command, however. At least there seems to me to exist no better way of accounting for the semantic content of an intention. The recourse to evaluative language favoured by Davidson in his characterisation of the content of intentions is no less problematic. He depends in his turn on his expressivist understanding of this kind of language, and I have already in Chapter 4 indicated why I am sceptical about the expressivist understanding of evaluative language and why I prefer an objectivists account. Moreover, even if an expressivist view of normative language were accepted, some account of what it is we express when we use evaluative language, if we do not express factual (normative) propositions, would have to be supplemented any way, in order to give flesh to the bare bones of expressivism, and, once again, the imperative mood suggests itself as the most promising candidate. This is how I would prefer to express the relevant argument, then. Let ‘!’ represent an operator on propositions, with the meaning: ‘see to it that …’ We may then represent the person Broome has described as a person who ‘accepts’ or ‘assents to’ the following two premises: 1. ! (You will leave the next buoy to starboard) 2. If you don’t tack, it is not the case that you will leave the next buoy to starboard but who rejects: 3. ! (You tack) Is it correct to say that this person makes a logical mistake? When Broome wants to show that the argument from (1) and (2) to (3), in his original statement of it, is valid, he claims that what we are here confronted with is a sequence of mental states that have propositions as their ‘contents’. As I have indicated above, this strikes me as misleading. When we characterise intentions, we certainly make a reference to a factual proposition, but it is improper to say that, when an intention is expressed, this proposition is part of the ‘content’ of what is expressed, However, the assumption that our intentions have a factual propositional Essays on Actions and Events, p. 90.
28
9.9 Broome on Practical Reasoning
141
content is crucial to Broome’s argument. This is how he explains the case. We assume, with Broome, that the name of the person who conducts the reasoning in question is Leslie. Then, according to Broome, the ‘content’ of Leslie’s reasoning as a whole can be represented by this sequence of propositions: 4. Leslie will leave the next buoy to starboard 5. In order to leave the next buoy to starboard, Leslie must tack 6. Leslie will tack This is how Broome comments on his example: A spectator on shore might run through reasoning of this form. If she has the first two beliefs, she would thereby acquire the third. She would be right to reason this way. The propositions that constitute the content of her reasoning stand in the relation of validity: if the first two are true, then so must the third be. The spectator believes the first two. That is to say, she takes them to be true. Since they cannot be true without the third’s also being true (and since this is obvious), she cannot rationally take them to be true without taking the third to be true. This is why her theoretical reasoning is correct.29
The observation that the argument from (4) and (5) to (6) is valid strikes me as obviously correct but completely irrelevant in the present context. If (4) and (6) are taken as descriptions of actions by Leslie, then, certainly, (6) follows from (4) and (5). However, this does not mean that the argument from (1) and (2) to (3) is valid. For (1) and (3) do not describe any actions. According to Broome, (1) is something expressed by a person who expresses an intention; then it is not a belief. To express an intention must be like expressing a choice or resolution, or so say I. Nothing follows logically from such an act of expression, of course. And since it is the expression of an intention, or a choice, or a resolution, rather than the expression of a proposition (the ‘normal’ content of a belief) we cannot specify the proposition expressed. In particular, the proposition expressed in (1) is not that Leslie will leave the next buoy to starboard. If no proposition whatever is expressed then, certainly, this proposition is not expressed. We see all this more clearly when we use the imperative mood, as I would prefer to do. Then we may say that, to have the intention to leave the next buoy to starboard is to assent to an imperative urging oneself to do so. Could we not say that when we express our intentions, we express both a proposition and our attitude to it (where the attitude is indicated by the ‘!’ in my representation of the argument)?30 We could compare this case to a case where we express our fear that a certain dog will attack us. In the latter case we may say that we express both the proposition that the dog might attack us and a certain uneasy feeling associated with this possibility. Could we say something similar about our expressions of intentions? I don’t think that’s a proper account. It is certainly true that, unless a person believes that she will do X, she cannot form the intention to do X. Perhaps we should say that, a person who expresses her intention to do X, implies pragmatically ‘Normative Practical Reasoning’. This has been suggested by Broome in correspondence.
29 30
142
9 Reasons to Desire
that she believes that she will do X. But she does not express this factual and autobiographical proposition. A simple argument shows why the relevant factual and autobiographical proposition is not expressed, but merely pragmatically implied, when an intention is expressed. For, certainly, Leslie’s not tacking (as a matter of fact) is consistent with (1) and (2) in the above argument. However, if (3) would imply that he tacks then, from true and consistent premises, we would have deduced a contradiction. Or, to give another simple argument. If, when I leave in the morning, I say to my wife that I intend to be home at 5 pm and, eventually arrive at 6 pm, then, certainly, she may say that I have deceived her. If she suspects that I knew that I would be late, she may complain that I didn’t keep my promise. But it would be strange of her to claim that I had lied to her. Therefore, when we express an intention we do not express any descriptive proposition whatever. It is correct to say that we need to refer to the propositions about Leslie’s leaving the next buoy to starboard and Leslie’s tacking when we characterise what is expressed by him, when he expresses his intentions. And, once again, we may concede that pragmatically Leslie implies that he believes he will leave the next buoy to starboard when, sincerely, he expresses his intention to do so. And we may concede the same thing about his tacking. However, this doesn’t mean that the factual autobiographical propositions, that he will leave the next buoy to starboard and that he will tack, are expressed by Leslie, when he expresses his intentions. They are not expressed by him. The upshot of this, I believe, is that the argument from (1) and (2) to (3) is not valid. The kind of practical reasoning defended by Broome doesn’t exist. All the reasoning that is involved in a case, where a person wants to do A, realises that, if he doesn’t do B, he won’t do A, and, consequently does B, is a standard factual argument. Such an argument can produce an action in the person pursuing it, if this person is, in a purely descriptive sense of the word, ‘rational’ (like you and me). This is all there is to say about this. Or, perhaps not all. It should be noted that, if I am a rational person, then I do act (intentionally) on my strongest desires. However, this is true because this is how the (purely descriptive) notion of rationality has been defined. This has nothing to do with any logic of intentions. But we may add that, this kind of rationality is natural for us humans. Unless we exhibit it, in most situations we have difficulties in realising any ends. What if I am wrong about all this? What if (3) does follow from (1) and (2)? What if there are logical relations obtaining even between imperatives, resolutions, and the like? Of course, many authors have argued that there exists a logic of imperatives, of some kind.31 I am sceptical of this claim but will not try to pursue this theme in the present context. Needless to say, those who defend the idea that there are
The number of such attempts is legion. The best version, to my knowledge, is Lars Bergström, Imperatives and Ethics. 31
9.10 The Value of Prospects
143
valid inferences involving imperatives must define the notion of a valid inference in a non-standard manner. They cannot rely on the idea that an inference is valid, if, and only if, necessarily, if the premises of it are true, then so is the conclusion. But perhaps, after all, they can find a kind of ersatz for the standard notion. Wouldn’t this mean that we have, after all, identified a special source of normativity, beside morality? No, it wouldn’t. For logic as such is no source of normativity. Even if such a logic of practical reasoning did exist, which I doubt, we would still have to ask the question: Why hold consistent intentions? Why be in this sense ‘logical?’ We can say that it is a ‘requirement’ of rationality, that we do hold consistent intentions (as well as beliefs). I have no objection to this, if ‘requirement’ here is understood in line with how it is understood when we claim that it is a requirement of being a bachelor that you are male and unmarried. However, if it means anything more, if in particular it carries normative weight, then the belief that there are requirements of rationality is false. I have admitted that in most situations we do hold consistent intentions, and I have noted that this is no coincidence. In a sense, this is ‘natural’ for us. This disposition to hold consistent intentions has been selected for, when our species has evolved. But this does not mean that, all of the time, this is how all of us must or will behave. And there is certainly no obligation for us to behave like this in general, let alone in any particular case.
9.10 The Value of Prospects Decision theory is sometimes conceived of as in some sense ‘normative’. We have found reasons to believe that it is not a source of normativity proper, however. But it might be thought that we share some good Humean reasons to stick to its verdicts. This idea too is a non-starter, however. First of all, it should be observed in the present context that decision theory dealing with decisions under risk seems to be something that lures many people into the belief that there is such a thing as value, irrespective of our obligations. Instead they want to connect the notion to some notion of rationality. It is customary to say things like the following: For all outcomes oi−1, oi, and oi+1, if oi is worse than oi−1, and oi+1 is worse than oi, there is some probability value p, 0 < p 1, such that oi =g (oi−1, p, oi+1). Here. “=g ” stands for “is equally good as” and “(oi−1, p, oi+1)” denotes a binary lottery in which outcome oi−1 is assigned probability p, while the remaining probability 1 − p is assigned to outcome oi+1. This is a “standard” “continuity” assumption. My query is not whether it is true or not but rather whether it makes sense. What does it mean so say that a certain lottery has a certain value? Can we cash this out in simple normative terms? When we say that oi is worse than oi−1, and oi+1 is worse than oi a way of caching this out in normative terms would be to say that, if we have a choice between oi and oi−1, then we ought to opt for oi−1, and so forth. Can the same idea be applied to
144
9 Reasons to Desire
lotteries? For example, does the claim that the value of oi equals the value of (oi−1, p, oi+1) amount to the idea that, if we have a choice between oi and the lottery (oi−1, p, oi+1), then we may choose between them as we see fit? Well, I would protest to this interpretation. It is an open question, I would say, which option the agent ought to make, oi or the lottery. This depends on which moral theory is the correct one. According to my beliefs about this, the actual outcome of the lottery would be decisive! If this is correct, then it is quite clear that he or she should not be indifferent between the options. For, certainly, one of the options brings with it better consequences than does the other one. I now set to one side the interest the agent may have in exposing him or herself to risk, this may be a positive or it may be a negative factor to take into account of in a full picture of the value of the options. Those who adhere to other moral theories will answer the question differently. But according to their favoured theories, there will be a definite answer to the question. And it would be rash to assume that the answer would conform to any answer given by a mainstream decision theorist. However, mainstream decision theorists rarely pretend to give answers to moral questions. But what questions do they address, then, if not moral questions? Can we say that what is at stake here is not (moral) obligation but rather a matter of rationality? Can we catch out the talk about value and expected value in terms of rationality? Can we say that, if a person wants to be rational, this person has better conform to continuity and be indifferent between oi and the lottery (oi−1, p, oi+1)? Well, if my argument in Chapter 6 was successful, we should be careful not to interpret the notion of rationality in its turn in normative terms. If we don’t, then, of course, we may say that it is rational to be indifferent between oi and the lottery (oi−1, p, oi+1). This is what it means to be rational, we should then add. We may also add that these options have the same expected utility (this is how p was selected). But if we do, we must bear in mind that this too is the result of a mere stipulation. And note once again, by taking up the rational attitude of indifference with respect to oi and the lottery (oi−1, p, oi+1), we are certainly doing something we ought not to do! At least this is the case if act utilitarianism is a correct moral view. Furthermore, in the hope that this strategy will show superior to other strategies that they can conjure up, some people (even, and perhaps in particular, utilitarians) may have a desire consistently to try to maximise expected utility (happiness). Other people may be reluctant to do so, however, in particular when they realise that by so doing they may sometimes come to risk a very large loss in order to make a very slight possible gain; no matter how slight the risk is that they will actually come to lose and no matter how probable (almost certain) the slight gain is, these people may feel that the risk is not worth taking. These are facts about life. These are facts about people. Both kinds of people seem to exist. Both can avail themselves of some notion of rationality, satisfying their own temperament, allowing them to describe themselves as rational and those of a different bent as irrational. But with reference to no such notion of rationality are they allowed to deduce any normative conclusion from the observation that one action or the other is ‘irrational’. This action may well be one required by morality. This is all there is to this kind of talk about value and expected value.
9.11 Ought We to Avoid Cyclical Preferences?
145
9.11 Ought We to Avoid Cyclical Preferences? What are we to say of a person who has cyclical preferences? She prefers X to Y, Y to Z, but Z to X? Is at least such a person irrational? We may stipulate that such a person is irrational, of course. However, as we saw in Chapter 6, this does not settle the matter of whether one ought to adopt such preferences. However, it might be thought that, all of us, at least most of the time, have some shared Humean reasons to avoid having cyclical preferences? Do we not all share some basic interests, which force us, if we want to have them satisfied, to avoid forming cyclical preferences? In fact, I think there is no such shared interest in avoiding cyclical preference. So there is really no need to claim that a person with cyclical preferences is irrational, and this should come as no surprise. Furthermore, it may well be a good idea to form cyclical preferences in some circumstances. There are indeed no shared basic interests that will in all circumstances be frustrated, if we make such a move. As a matter of fact, most people have some cyclical preferences, even if they do not know about this, and they do not know about it since it does not matter in their daily life. Certainly, if a person believes that X is better than, Y, Y better than Z, and Z better than X, implying that there is more of something in X than in Y and in Y than Z (more happiness, say), while believing that there is more of the same thing in Z than in X, then this may mean that she holds inconsistent beliefs. And to hold inconsistent beliefs conflicts with the epistemic goal most people have, not to believe what is not true. However, not all preferences are beliefs. Typically, I would say, preferences are not beliefs. Rather, preferences are hypothetical choices, or tendencies, dispositions to make certain choices. And to the extent that they are not beliefs, but rather hypothetical choices or dispositions, there is no need to insist that a rational person must hold transitive preferences. In order to be called ‘rational’ it is sufficient for a person, I stipulate, to act on her strongest preference whenever a uniquely strongest preference exists. If she has cyclical preferences, and faces all the options in the cycle, it doesn’t matter (to her rationality) on which one of them she acts. And if we want to explain her action, what is crucial to our explanation is just, in the circumstances, on which preference she did act, that is which preference caused the action in question. Once we know this, we can explain her action. And if we are not capable of answering this question (perhaps none of her preferences caused her ‘action’, she just ‘picked’ one of the alternatives),32 then we should say that, even if this is a rational person, this behaviour of hers was not perhaps an ‘action’, in the full sense of this word (where a rational explanation is possible). This does not mean then, that this person is not a rational person, and it does not mean that most of her actions defy rational explanation. In order to call her ‘rational’, I suggest, we should require that, if she has a uniquely strongest preference for one option in a situation, then she goes for this For an illuminating discussion about the difference between picking and choosing, see UllmannMargalit E. and Morgenbesser, S., ‘Picking and Choosing’. 32
146
9 Reasons to Desire
option. This means that, in cases where she doesn’t meet all the options in the cycle, our Humean explanation of her action, if she is rational, is straightforward. She choose X over Y, say, because she prefers X to Y, period. However, is it not stupid to hold cyclical preferences? Doesn’t the possession of cyclical preferences make one vulnerable to a Money Pump? If you have cyclical preferences, does it not mean that you are prepared to pay in order to get from X to Y, and from Y to Z, and then from Z to X once again? And do we not all share an interest in not being vulnerable to this kind of threat? No, having cyclical preferences does not mean that, necessarily, any of our interests will be frustrated. The obvious, and hence rather boring, answer to the Money Pump argument is that, when you are faced by a series of choices, like the one here indicated, you make only one move, or no move whatever. Moreover, if it is not likely that one will come to face any money pump, it might be a good idea always to act on the preferences one has. In many such cases one will not even notice that one has cyclical preferences in the first place. Having cyclical preferences may, as was indicated above, mean that in some situation one will come to face all the options in the cycle; but then the solution is to take a ‘pick’ rather than acting on the strongest preference; and if one is perfectly happy with doing so, then what’s so terrible with this? What if one doesn’t mind? What if one – proudly – says to oneself: ‘This is the kind of person I am!’ And one may add that one finds people who do not exhibit a similar trait rather boring. Why change one’s preferences, if it is possible to live happily with those one has?
9.12 Is There a Common Emotional Goal? We saw in our discussion about epistemology that there seems to exist a common human epistemic goal, which we cannot help sharing with others; whatever our special interests otherwise, we will be poor at satisfying them unless we pay respect to our epistemic goal. This goal can be characterised as follows: to believe what is true and not to believe what is false. We also noted in our discussion about rationality in Chapter 2 that, even if our notion of rationality is purely naturalistic (non-evaluative), there is no doubt that most people, most of the time, are somewhat rational, that is they exhibit a typical fit between what they desire, believe, and do. Again this has much to do with evolution, but also with the fact that, whatever interests we have, we will be poor at satisfying them, unless we are, by and large, rational. This fact renders it possible to explain some of our actions by rationalising them (by our own lights). Perhaps this is not the end of the matter. Perhaps there is also a common emotional goal? If there is then, even if the (subjectively) good and beautiful, say, are not sources of normativity in their own right, they may still be something we all have common Humean reasons to aspire to. We can then rationalise our desires with reference to this goal. I have repeatedly made references to this idea in this chapter. Now time has come to investigate it.
9.12 Is There a Common Emotional Goal?
147
Is there a common emotional goal? If there exists such a common goal, what would it look like? To fear what is dangerous, to like what is useful, to care for one’s own future and the future of one’s kin, are desires that come to mind. We may think that, to the extent that we hold these desires, we do so because of the kind of creatures we have evolved into. If we borrow from Aristotelian virtue ethics, we may perhaps say that a common emotional goal would be to hold such desires that contribute to a flourishing life, on our own part, and on the part of our kin. However, there are many desires that are difficult to explain in the same manner and to rationalise with reference to this goal. I like to play the clarinet and I am eager to do philosophy. Can we find an evolutionary explanation behind these desires? Probably we cannot. But this is perhaps not so problematic, since every one does not share these desires. And we need not assume that all our desires can be rationalised with reference to the same goal. It would be interesting enough, if some of our desires could be thus rationalised. However, there are also many desires that seem to be at variance with what is to be expected from people, knowing their evolutionary background. For example, some people choose to live in celibacy, and others sacrifice their lives and their close ones to abstract moral goals. They refuse to take part in atrocities, even at some risk to themselves and their kin. Or they decide to become suicide bombers. This is not inexplicable, but the explanation cannot be cast in evolutionary terms. The explanation has to make reference to some cultural facts crucial to the development of the characters of these people. The desires these people have cannot be rationalised with reference to the putative emotional goal we are here considering. It seems then that, not even emotions such as fear of what is dangerous and a concern for one’s own future and the future of one’s kin, are shared by all people. Moreover, in some cases where people have desires that are, from an evolutionary point of view, unexpected, we tend to admire them. We are not prepared to say of them that they are ‘irrational’. Think of Hugh Thompson, a 25-year-old helicopter pilot at My Lai, as Jonathan Glover describes him: Hugh Thompson thought about how the Nazis had massacred people by making them lie in ditches and shooting them. He saw civilians, including children, running to a bomb shelter pursued by a group of soldiers. He landed between the villagers and the soldiers and sent a radio signal for help. He ordered his crew to use their machine-gun against the American troops if they fired on the villagers. He went to the bomb shelter and persuaded the civilians to come out. He called the two helicopter gunships nearby for help in rescuing the civilians. When his request was queried, he cursed and pleaded, and said that his machine-gunner would fire on the infantry if they fired on the civilians.33
It is doubtful whether Thompson’s desires were of a kind that promoted his own flourishing. They could very easily have lead to his premature death. And yet we do admire him. One reason, consistent with sociobiology, why we have very different desires with different prospects for our own flourishing may be that a population can consist Humanity, pp. 62–63.
33
148
9 Reasons to Desire
of individuals exhibiting very different strategies, where the existing mixture of strategies constitute an evolutionary stable situation. It is difficult to find any analogy to this within epistemology. Could we revise the claim that there exists a common emotional goal and cast it merely in terms of a common interest in practical consistency? This is not possible with respect to all our emotions, of course, since the notion of consistency doesn’t seem to be applicable to some of them, such as feelings of happiness, fear, and hope, but it is with respect to others. We may speak about them as ‘pro’ and ‘con’ attitudes. Typically desires are of this kind. Such attitudes can come into practical conflict since they have satisfaction values. We saw in our discussion of Broome’s views on practical reasoning that there are no logical relations between intentions. However, one could perhaps have recourse to a kind of requirement of practical consistency, for example to the effect that one should not hold a pro- and a con-attitude to the same thing (in the same respect). And we may add other requirements to this one. For example, we may require that if I hold a pro-attitude to everything that is A, then, if x is A, I hold a pro-attitude to x. I will not go into detail (and complication) here. The very idea seems to be a non-starter. Note that in epistemology the requirement of consistency is derived. To the extent that we want to abide by it, it is because we realise that two inconsistent propositions cannot both be true. There is no parallel with this with respect to practical consistency. Moreover, even in an epistemic context we may sometimes accept, and ‘bracket’ some inconsistencies, for strategic reasons. By doing so we hope to gain more knowledge in the long run. There may be a parallel to this with regard to pro- and con-attitudes. To some extent we may want ourselves, and even more others to be, somewhat consistent. This makes us predictable. And this predictability may engender common solutions to some coordination problems. We may hope to reach consensus. And yet, in many situations, we will be able to reach common solutions better if we are not too rigid in this kind of consistency. So the prospects of finding a common goal here are indeed bleak. Even a clever thinker such as Allan Gibbard, who has done his best to show that there exists a common interest in practical consistency, must admit that such an ambition, fostered by an interest in coordination through consensus, may sometimes prove counter-productive: … even if consensus were all that mattered, responsiveness to demands for consistency would not always promote it. If some are especially consistent, that might drive them away from consensus with the rest. Jointly inconsistent norms may each have a strong natural appeal. When that is so, there may be more consensus to be gained by sweeping inconsistencies under the rug than by exposing them and trying to root them out.34
Wise Choices, Apt Feelings, p. 288. Gibbard here speaks of ‘norms’ but he holds an expressivist view of norms so, according to his analysis, these are indeed pro- and con-attitudes of a kind. Therefore, the consistency here discussed is indeed a practical one. 34
9.13 Conclusion
149
9.13 Conclusion There seem to exist no reasons in their own right to prefer or desire anything. Something might be of ‘value’, in the sense that, this is what we ought to realise through our actions. And yet, for all that, this does not mean that we have any reason as such to develop a taste for it. Contrary to what is taught in the tradition from Brentano, there is no need for us to stand at attention in front of a right-making characteristic. This is not to say, of course, that there is absolutely no way of making sense of the claim that we ought (have reasons) to prefer or desire what we ought to realise in the world. For example, the claim that we have reasons to develop a taste for what we ought to realise could be seen as a claim to the effect that we have Humean reasons to develop such a taste. To the extent that we believe that, unless we desire and try to realise what we actually ought to realise, we will not be good at achieving what we ought to realise, and to the extent that we want to do our moral duty, this means, then, that we have good Humean reasons for desiring and striving for what we ought to realise. However, what we know about the paradox of hedonism should warn us against rash generalisations here. The best way of achieving some of the good things in life may be not to bother about them at all. But even if there are no object given reasons to adopt any attitudes in particular to any phenomenon, irrespective of the consequences of doing so, it might be thought that there are subjective reasons to keep a certain order among the emotions one holds. For example, ought we to want the means to our end? We have seen that, even if there is some truth in the saying that, unless we want the means to our ends we are ‘irrational’, this does not mean that rationality as such provides us with a source of normativity; the notion of rationality employed here is a merely descriptive one. And, even if in most cases, it pays to be, in this sense, ‘rational’, it may sometimes pay not to. In particular, it doesn’t seem correct to say that, once I have formed a certain intention, realised that, in order to fulfil it, I must perform a certain action, but do not form any intention to perform the relevant action, I make any logical mistake. Are there emotions that it is ‘natural’ for us to have, are there emotions that can be rationalised with reference to a common emotional goal, in the manner we can rationalise our beliefs with reference to a common epistemic goal? No, we have seen that there doesn’t seem to exist any common emotional goal, similar to the epistemic goal we all share. Hence, there is no possibility of rationalising our emotions with reference to any such goal. Finally, ought we at least to keep track of our preferences, in order not to have cyclical preferences? No, there is no compelling reason to do so. The Money Pump argument is defective. Many people hold cyclical preferences and this is consistent with all sorts of idiosyncratic goals they happen have, as well as with the requirements of morality.
Chapter 10
Conclusion
10.1 Introduction I have defined two notions of practical reason in this book: one Humean and one moral notion of practical reason. I have attempted to show that upon closer examination, putative prudential reasons, reasons from rationality or justice or aesthetics, as well as putative reasons to believe or to feel or desire certain things, turn out to be, either moral or Humean reasons proper. Moral reasons are a source of normativity, Humean ones are not. So the main claim has been that there is but one source of normativity. I have done my best to show that this source of normativity is real. I have defended moral realism. In particular, I have used my notion of a moral reason to answer Mackie’s complaint that we cannot understand how moral properties are consequential upon non-moral ones, and I have turned the content of our moral beliefs, and the evidence we have for them, against various different debunking strategies, intended to rob our moral convictions of any justification. It may seem strange to claim that while moral reasons exist in an objective and sui generis way, other putative kinds of reasons don’t. After all, I have claimed that moral reasons are not vulnerable to the arguments from queerness, relativity, empiricism, and so forth. Then, why do I not make the same claim with respect to other kinds of reasons (from prudence, aesthetics, and so forth)? This objection is based on a misunderstanding. What invites the misunderstanding is the classification of certain absolute normative reasons as ‘moral’ ones. My claim is that there exists just one kind of normative reasons. The singularity of action gives rise to one singular question: what ought I to do? And the unique explanation of the unique answer to this question is my reason for doing what I ought to do. We are familiar with this kind of reason, I have claimed, and it is neither queer nor defective in any other respect. We may speak of it as ‘moral’, but this is merely a matter of convention. We could have called it something else. It is not the case that morality requires things of us. If it did, it would be perfectly sensible to ask if prudence or aesthetics didn’t require (other and sometimes conflicting) things of us as well. The truth is rather that certain things
T. Tännsjö, From Reasons to Norms, Library of Ethics and Applied Philosophy 22, DOI 10.1007/978-90-481-3285-0_10, © Springer Science + Business Media B.V. 2010
151
152
10 Conclusion
are required of us, period. There are things we ought to do simpliciter and there are explanations of these obligations. The ‘ought’ in question is what Owen McLeod and others have called a just plain one.1 Or, perhaps the claim that our calling the normative requirements we face ‘moral’ is merely a matter of convention is an overstatement. The reason we (I) call them ‘moral’ has something to do with their content, of course. I haven’t taken a stand on substantial normative issues in this book, but it is my firm belief that something close to utilitarianism gives a hint of the most plausible normative view. If this is correct, then it is quite natural to speak of the normative requirements we face as ‘moral’ ones. If, on the other hand, egoism would be on the right track then, of course, it would be more natural to claim that the sole source of normativity is prudence! In any case, the central claim of this book is that there is exactly one kind of absolute, categorical, authoritative and sui generis norms that face us human beings. When we realise this, I conjecture, we will see through the talk of many ‘sources’ of normativity. The word ‘source’ is merely a metaphor, and a metaphor that tends, alas, to lead us astray in our thinking about moral reasons. We may say, as I have often done, that there is just one ‘source’ of normativity, but it would perhaps be better not to use the metaphor at all. There are obligations facing us, and that is it. I will now sum up my discussion about various different putative sources of normativity – and stress again why I do not think they are real.
10.2 Prudence, Rationality With respect to prudential reasons, or reasons from ‘rationality’, I have tried to show that they collapse into moral reasons proper. Some people may hesitate to call agent-relative reasons to do merely with the well-being of the agent ‘moral’ reasons. And yet, for all that, since such reasons can conflict with moral reasons, we must acknowledge that they are of the same type. A similar treatment has been made of the idea that there are reasons from justice. We may say, of course, that, from the point of view of the general welfare, a certain sacrifice ought to be made while, from the point of view of the welfare of the agent, or justice, the sacrifice ought not to be made. However, this is no normative judgement whatever. And if, in the final analysis, it is true that these sacrifices need not be made, then this means that morality admits of some options. At least this is how I have used the word ‘morality’ in this book. I have not taken a stand on the issue of what moral theory – if any – is the correct one. However, according to most suggestions in this field, we have to acknowledge that a moral theory cannot as such inform us, in any manner that we can utilise,
See his ‘Just Plain “Ought”’, also for further references.
1
10.3 Justice and Aesthetics
153
what to do in concrete situations. Most moral theories are too abstract to allow us to decide in concrete situations what to do. We know that we must not violate any rights, that we ought to maximise the sum-total of well-being in the universe, and so forth, but we do not know how to go about when we do so. A standard approach to this problem is to adopt some kind of two-level model. In a ‘cool hour’ we opt for a method of decision-making such that, when we adhere to it, we do so in the hope that we should better fulfil the goal of our favoured moral theory, than if we tried in each case immediately to ‘apply’ it. Could we say that such a method of decision-making does itself function as a source of normativity? If, in a situation, unbeknownst to us, A is what is actually prescribed by our favoured moral theory while B is what we know is prescribed by our favoured method of moral decisionmaking, does then this mean that we have conflicting obligations? Does it mean that, from the point of view of morality we ought to perform A while from the point of view of our method of decision-making (or from the point of view of ‘prudence’) we ought to perform B? It means nothing of the sort. In the circumstances we ought to perform A (provided that our favoured moral theory is the true moral theory) and we have reasons to perform A (explaining why we ought to perform A). We have no moral reasons to perform B. We do have Humean reasons to perform B, however, to the extent that we want to abide consistently by the method of decisionmaking which we have adopted. So prudence, thus conceived of, is no source of normativity.
10.3 Justice and Aesthetics We sometimes hear that, from the point of view of justice, it would be better that all differences of well-being were levelled out, even if no one would gain in terms of well-being by such a move of levelling down, and yet, for all that, it might be wrong to undertake such a move. And we learn that, from the point of view of aesthetics it would be obligatory to preserve a certain kind of environment, even if the cost to sentient beings would be considerable, and yet, for all that, it might be wrong to make this kind of sacrifice. Does this mean that justice or aesthetics provides us with a source of normativity of its own? It does nothing of the kind. The most plausible way of understanding such expressions is to take them to be statements of a particular kind of moral concern. However, until we know how much – if anything – we should be prepared to sacrifice, in terms of well-being, in order to see justice done or abstract beauty brought into existence, we should not pretend that we have understood them. And once we have been informed about this, we know how to fit them into our moral theory. This means that, even if justice and aesthetics may be sources of normativity, they are in this respect part of, not complementary to, morality proper. Since reasons to do with justice or aesthetics can compete and conflict with ‘ordinary’ moral reasons, they should be counted among all the
154
10 Conclusion
putative moral reasons that we want to consider in normative ethics. They are in no way ‘special’, then.
10.4 Epistemology With respect to reasons to believe, these reasons are most naturally taken, in most cases, to be Humean ones, taking their point of departure in an epistemic goal to do with our wish to hold beliefs rather than to be agnostics, and to do with our wish to hold true beliefs, rather than false ones. This means that epistemology is conceived of as, in a weak sense of the word, ‘naturalised’. Epistemology is not, as such, any source of normativity. I have admitted that, possibly, some people, in some contexts, may take the activity of science as constitutive of certain epistemic rules (rules of methodology), existing within scientific communities in roughly the same way that there are socially constructed or constituted moral norms (existing within moral universes). However, I have claimed that this does not seem to be the standard view of such rules. In most cases it is, I submit, naturally to take them to be instrumental. Moreover, even if, in some contexts, some epistemic rules, such as rules for theory choice, are taken as constituted by the scientific practice, this does not mean, of course, that scientific facts are ‘social constructions’. Certainly, they are not. The belief that they are exemplifies madness. And it would be strange for me, I have claimed, to hold on to a certain methodological rule, even if it would be generally endorsed in the scientific community to which I belong, if it would turn out that (by my lights) this rule does not seem to serve my epistemic goal. If this were to happen, and if I were a reasonably rational person, then I would give up my tendency to stand on ceremony in scientific contexts. Or, I would have to admit that this epistemic rule, with no regard for truth, is a kind of very strange moral principle, urging me to select, say, a certain hypothesis, merely because this means adherence to the rule – even if by selecting this hypothesis there will be both a loss in terms of the sum-total of well-being in the universe, and in terms of my holding a less realistic picture of the world. I think most people, upon reflection, would come to resent the kind of morality where this principle forms a part.
10.5 What Are We to Believe? Since my claim to the effect that there is but one source of normativity is a negative ontological one, the truth of it cannot be strictly proved. However, a person who in the spirit of William James wants to hold true beliefs and avoid false ones, has good Humean reasons positively to reject (rather than suspend his or her
10.6 Reasons to Feel and Desire
155
judgement concerning) the idea that there are any objective sources of obligation outside morality. And, the same person ought to conclude that, within morality there exits at most one such source that is both objective and independent of our moral judgement.
10.6 Reasons to Feel and Desire With respect to the idea that there are certain reasons to feel and desire things, so popular in the tradition from Brentano, I have argued that this idea leads into paradoxical results, once it is taken to have any true normative content, urging us to develop a taste for those things that, through our actions, we ought to promote. There are many ways of conceiving of this idea without making the assumption that it has a normative content of this kind, of course. One could say, for example, that if something is valuable in itself, then we have an obligation to produce it, but no obligations whatever to develop a taste for it. But once this normative content is dropped, the idea that there is a peculiar source of normativity to do specifically with feeling, desiring, and so forth, vanishes into thin air. Can we then find a common emotional goal, given to us by evolution? We have seen that most people seem most of the time to be somewhat rational (to exhibit a typical fit between what they desire, believe and do). Unless they are in this way rational, they do not survive. Furthermore, we have seen that we all share a common epistemic goal. We want to believe what is true and we want not to believe what is false. This means that we tend to conform to certain ‘requirements of rationality’. Unless we pay some respect to our epistemic goal we will have difficulties in satisfying our other goals, whatever they are. The epistemic goal does not provide us with any sources of normativity, the requirements of rationality are not normative properly speaking, but we can rationalise our actions and beliefs with reference to them. So it is tempting to think that there is also a common emotional goal, forced upon us by evolution, with reference to which we can rationalise our more particular emotions. One could think here of the goal that one should hold emotions that make oneself and one’s kind flourishing. However, even if some emotions, such as fear of what is dangerous, and care for our own future, are only natural, and could hence be rationalised with reference to this goal, there are situations where we want to act contrary to this goal. And we may have good moral reasons to do so. In some circumstances, the right thing to do may be to sacrifice one’s life, and, for moral reasons, one may be prepared to do so. So it is not a good idea to try to rationalise our emotions with respect to any shared emotional goal. Upon closer inspection there is no such shared goal. Not even an interest in practical consistency seems to be shared by all of us, all the time. Having reached this conclusion we are now in a position to return to the anecdote told about the War Tribunal in Chapter 1. This is the ‘true’ story told without any use of any vague notion of reasons at all, and with the outcome added to satisfy our curiosity:
156
10 Conclusion
10.7 The War Tribunal – The True Story The Judge: You admit that you drove the villagers into the camp? The Accused: I do. The Judge: Did you realise that they would most probably be killed in the camp? The Accused: I don’t think I considered that possibility. I didn’t think I had reasons to bother with that kind of thought. After all, it was none of my business. The Judge: You say you didn’t think you had ‘reasons’ to bother with that kind of thought. What do you mean by that? The Accused: I suppose I mean that I did not want to think about it and hence didn’t think about it. The Judge: So what you are saying is only that there is an explanation why you did not think about it? The Accused: I suppose so. The Judge: But considering what you must have known about other camps, you must have had reasons to believe that the same thing might happen this time? The Accused: Now you use the word ‘reason’. What do you mean by it? The Judge: I suppose I mean something along the following lines. You must certainly have an interest in believing what is true and not believing what is false. Given this kind of interest, don’t you think it would have been rational for you to adopt the belief that the villagers might get killed? The Accused: What do you mean by ‘rational’? The Judge: I suppose I mean only that, had you adopted this belief, this action would have ‘fitted in’ with your interest in believing what is true and the rest of your beliefs. The Accused: This may be true, but, as I said, my interest in believing what was true about the situation wasn’t that strong in the circumstances. The Judge: If you realised what might happen to them, why didn’t you release them instead of driving them into the camp? The Accused: As I said, I’m not certain I realised that. Not at the time. Anyway, I had my reasons to do what I did. The Judge: What reasons? The Accused: Well, even though I wasn’t happy with the situation in my country, I thought there was little I could do about it. As you know, the Americans supported the government. And I knew that if I did not obey my orders, I might myself have had to face some sort of consequences.
10.7 The War Tribunal – The True Story
157
The Judge: It seems to me that you are just repeating the explanation why you acted as you did. But those ‘reasons’ you mention, are these reasons really good ones? You could have saved several lives, and the cost to you would probably not have been devastating. I take it from what happened to others in your situation that, at worst, you would have lost your job. The Accused: Why should I not have acted as I did? I think it was quite rational for me to do as I did. My job was dear to me. And more serious consequences, if I did not comply, were not totally out of the place. The Judge: Perhaps it was rational, in some sense, for you to do as you did. But certainly, you had moral reasons to abstain. The Accused: What do you mean by a moral reason? The Judge: I suppose I mean something that can explain why one has a certain obligation. Such as the fact that you could easily have saved the lives of the villagers. The Accused: I’m not certain that there are moral reasons of that kind. At least, I’m not certain that they exist independently of us. But you are the judge, you are on the winning side, your morality prevails now, so I suppose you are right. I concede that I must have had moral reasons not to do what I did. The Judge: Then how come you did it. The Accused: As I said, I did not want to sacrifice either me or my family. I thought it was only prudent of me to do what I did. Even if I had not lost my life if I had released those people, I would certainly have lost my job and my family would have suffered. The Judge: You speak here of prudence. That seems to me to be another way of speaking of egoism. But is egoism a true moral theory? I think not. You could have saved several lives. This must be more important than your and your family’s well-being. The Accused: I suppose so. At least according to the morality that now prevails … The Judge: For certainly, if you can save innocent lives without making terribly severe sacrifices, you ought to do so! The Accused: From the moral point of view that now prevails, I suppose that is what I ought to do. But from a legal point of view, I saw no reasons not to comply with the orders I was given. These orders accorded with the law at the time. And, as I have said, I thought it was only rational to comply. The Judge: But aren’t what we call ‘moral reasons’ what count in the final analysis? The Accused: Now and here they do, obviously. And you seem to believe there is one single, true morality, which is our sole source of normativity. I don’t share that belief, I’m sorry to say. I did certainly not think like that, then. Even if I had been aware that, from a moral point of view such as yours, what I did was wrong – but
158
10 Conclusion
I’m not certain I saw things like that, then – I don’t think it is fair to say that my action was irrational. I have explained to you why I did as I did. I had my reasons to do it. I’m not crazy. I’m no monster. I’m not a criminal. I didn’t kill anyone. There were reasons for me to do as I did. I had reasons to do what I did. The Judge: I didn’t say your act was irrational. I admit that an explanation can be given of what you did. I believe you have given an explanation of your action, which rationalises it, at least by your own lights. I do say, however, that you acted wrongly. Would you be prepared to do it again? The Accused: I don’t know. And I doubt that there is any truth about the rightness or wrongness of what I did, at least I doubt that there is one single truth about this. Be that as it may, if you had been in my shoes, do you think you would have acted any differently? The Judge: What I would have done is beside the point. And I believe there is a single true morality. It is obvious that you have violated it. Considering the kind of trial this is, morality matters, that is morality matters also to the legal aspect of your case. So I sentence you to imprisonment, for having committed a crime against humanity. The Accused: I don’t know about morality. I doubt that there is such a thing any more. I used to hold moral beliefs, but my entire moral universe has been destroyed. I pass no moral judgements any more. I have no power. I have no desire. I accept your verdict; as I matter of fact, you may just as well sentence me to death. I seem not to belong to this world any more.
Bibliography
Arrhenius G (2000) Future generations: a challenge for moral theory. University Printers, Uppsala Bales E (1971) Act-utilitarianism: account of right-making characteristics or decision-making procedure? Am Philos Q VII:256–265 Baron M (2006) Acting from duty (GMS I, 397-401). In: Kant’s morals. Walter de Gruyter Verlag, Berlin, pp 72–92 Barry B (1995) Justice as impartiality. Clarendon, Oxford Bergström L (1962) Imperatives and ethics. Stockholm University, Stockholm (mimeograph) Bergström L (1972) Outline of an argument for moral realism. Grazer Philos Stud 12(13):215–225 Bergström L (1996) Reflections on consequentialism. Theoria 62:74–94 Bergström L (2008) A defense of Quinean naturalism. In: Wrenn CB (ed) Naturalism, reference, and ontology. Peter Lang, New York Blackburn S (1984) Spreading the word. Groundings in the philosophy of language. Oxford University Press, Oxford Blackburn S (1985) Supervenience revisited. In: Hacking I (ed) Exercise in analysis: essays by students of Casimir Lewy. Cambridge University Press, Cambridge, pp 47–67 Blackorby C, Bossert W, Donaldson D (1997) Critical-level utilitarianism and the populationethics dilemma. Econ Philos 13:197–230 Brandt RB (1959) Ethical theory. Problems of normative and critical ethics. Prentice Hall, Englewood Cliffs, NJ Brandt RB (1979) The theory of the good and the right. Clarendon, Oxford Brentano F (1889) The origin of our knowledge of right and wrong. Routledge & Kegan Paul, London/Henley (1969) Brink DO (1989) Moral realism and the foundations of ethics. Cambridge University Press, New York Broome J (1991) Weighing goods. Equality, uncertainty, and time. Blackwell, Oxford Broome J (2001) Normative practical reasoning. Proc Aristotelian Soc 75(Suppl.):175–193 (Here quoted from a revised and expanded net version) Broome J (2004) Reasons. In: Pettit P, Scheffler S, Smith M, Wallace J (eds) Reason and value: essays on the moral philosophy of Joseph Raz. Oxford University Press, Oxford Carlson E (1995) Consequentialism reconsidered. Kluwer, Dordrecht Carlson E (1998) Mere addition and two trilemmas of population ethics. Econ Philos 14:283–306 Chang R (2004) Can desires provide reasons for action? In: Wallace RJ, Smith M, Scheffler S, Pettit P (eds) Reasons and values: essays on the moral philosophy of Joseph Raz. Oxford University Press, Oxford, pp 56–90 Cooper JM (1999) Reason and emotion. Essays in ancient moral psychology and ethical theory. Princeton University Press, Princeton, NJ Copp D (1997) The ring of Gyges: overridingness and the unity of reason. Soc Philos Policy 14(1):86–106
159
160
Bibliography
Crisp R (2005) Reasons and the good. Oxford University Press, Oxford Cuneo T (2007) The normative web. Oxford University Press, Oxford Dancy J (2000) Practical reality. Oxford University Press, Oxford Darwall SL (1983) Impartial reasons. Cornell University Press, Ithaca, NY Darwall SL (1990) Autonomist internalism and the justification of morals. Nous 24:257–268 Davidson D (ed) (1980) Mental events. In: Essays on actions and events. Oxford University Press, Oxford, pp 207–227 Davidson D (1980) Essays on actions and events. Oxford University Press, Oxford Davidson D (1986) The coherence theory of truth and knowledge. In: LePore E (ed) Truth and interpretation: perspectives on the philosophy of Donald Davidson. Blackwell, Oxford Dummet M (1978) Truth and other enigmas. Duckworth, London; Harvard University Press, Cambridge, MA Dworkin R (1996) Objectivity and truth: you’d better believe it. Philos Public Aff 25:87–139 Ewing AC (1947) The definition of good. Macmillan, New York Ewing AC (1959) Second thought in moral philosophy. Routledge & Kegan Paul, London Feldman F (2005) The open question argument: what it isn’t and what it is. Philos Issues 15, Normativity, 22–43 Gauthier D (1986) Morals by agreement. Clarendon, Oxford Geach P (1960) Ascriptivism. Philos Rev 69:221–225 Gewirth A (1978) Reason and morality. The University of Chicago Press, Chicago, IL Gibbard A (1990) Wise choices, apt feelings. Clarendon, Oxford Glover J (1999) Humanity. A moral history of the twentieth century. Jonathan Cape, London Greene J, Sommerville RB, Nystrom LE, Darley JM, Cohen JD (2001) An fMRI investigation of emotional engagement in moral judgment. Science 293:2105–2108 Greene JD, Haidt J (2002) How (and where) does moral judgment work? Trends Cogn Sci 6:517–523 Hare RM (1952) The language of morals. Oxford University Press, Oxford Hare RM (1972) Moral thinking. Clarendon, Oxford Harman G (1977) The nature of morality. An introduction to ethics. Oxford University Press, New York Harman G (1986) Change in view. MIT Press, Cambridge, MA Harman G, Thomson JJ (1996) Moral relativism and moral objectivity. Clarendon, Oxford Hempel CG (1965) Aspects of scientific explanation. The Free Press, New York Hurley P (2006) Does consequentialism make too many demands? Ethics 116:680–706 Hume D (1739–40/1978) A treatise of human nature (Selby-Bigge LA [ed] 2nd edition with text revised and variant readings by Nidditch PH). Clarendon, Oxford Jackson F (April 1991) Decision-theoretic consequentialism and the nearest and dearest objection. Ethics 101(3):461–482 Jackson F (1998) From metaphysics to ethics. Clarendon, Oxford James W (1956) The will to believe and other essays in popular philosophy. Dover, New York Joyce R (2006) Evolution of morality. MIT Press, Cambridge, MA Kahneman D, Tversky A (1984) Choices, values, and frames. Am Psychol 39:341–350 Kant I (1998) Groundwork of the metaphysics of morals. Cambridge University Press, Cambridge Kant I (2000) Critique of the power of judgement. Cambridge University Press, Cambridge Kim J (1994) What is “naturalized epistemology”? In: Kornblith H (ed.) Naturalizing epistemology, 2nd edn. MIT Press, Cambridge, MA, pp 33–55 Kolodny N (2005) Why be rational? Mind 114:509–563 Korsgaard C (1996) The sources of normativity. Cambridge University Press, Cambridge Kripke S (1982) Wittgenstein on rules and private language. Harvard University Press, Cambridge, MA Levi I (1967) Gambling with truth. Routledge & Kegan Paul, London Levi I (1980) The enterprise of knowledge. An essay on knowledge, credal probability, and chance. MIT Press, Cambridge, MA
Bibliography
161
Mackie J (1977) Ethics. Inventing right and wrong. Penguin, Harmondsworth McDowell J (1978) Are moral requirements hypothetical imperatives? Proc Aristotelian Soc 52(Suppl.):13–59 McLeod O (2000) What is Sidgwick’s dualism of practical reasons? Pac Philos Q 81:273–290 McLeod O (2001) Just plain “ought”. J Ethics 5:269–291 Mele AR (1988) Effective reasons and intrinsically motivated actions. Philos Phenomenol Res 48:723–731 Moore GE (1903/1965) Principia ethica. Cambridge University Press, Cambridge Moore GE (1963[1912]) Ethics. Oxford University Press, London Moore GE (1942) A reply to my critics. In: Schilpp PA (ed) The philosophy of G. E. Moore. Northwestern University press, Evanston, Il Murphy LB (1999) Institutions and the demands of justice. Philos Public Aff 27:251–291 Naess A (1989) Ecology, community and lifestyle: outline of an ecosophy. Cambridge University Press, Cambridge Nagel T (1970) The possibility of altruism. Princeton University Press, Princeton, NJ Nagel T (1986) The view from nowhere. Oxford University Press, Oxford Nagel T (1997) The last word. Oxford University Press, Oxford Nozick R (1974) Anarchy, state, and Utopia. Blackwell, Oxford Ohlsson R (1995) Morals based on needs. University Press of America, Lanham, ML Olson J, Svensson F (2005) Regimenting reasons. Theoria 71:203–214 Olson J (2004) Buck-passing and the wrong kind of reasons. Philos Q 54:295–300 Parfit D (1984) Reasons and persons. Clarendon, Oxford Parfit D (1986) Overpopulation and the quality of life. In: Singer P (ed) Applied ethics. Oxford University Press, Oxford, pp 145–164. Reprinted in Ryberg J, Tännsjö T (eds) The repugnant conclusion. Essays on population ethics. Kluwer, Dordrecht (2004) Parfit D (2001) Rationality and reasons. In: Egonsson D, et al (eds) Exploring practical philosophy. Ashgate, Aldershot, HA, p 20 Parfit D W what matters? Petersson B (2000) Belief & desire. The standard model of intentional action – critique and defence. Studies in Philosophy 9, Lund University, Lund Philipps D (1988) Sidgwick, dualism and indeterminacy in practical reason. Hist Philos Q 15:57–78 Popper KR (1972) Objective knowledge. Clarendon, Oxford Quine WV (1960) Word and object. Wiley, New York Quinn W (1993) Putting rationality in its place. In: Foot P (ed) Morality and action. Collected papers of Warren Quinn. Cambridge University Press, Cambridge, pp 228–255 Rabinowicz W (2001) Prioritarianism and uncertainty. In: Egonsson D, et al (eds) Exploring practical philosophy. Ashgate, Aldershot, HA Radcliffe E (1999) Hume on the generation of motives: why beliefs alone never motivate. Hume Stud 25:101–122 Raz J (1999a) Practical reasons and norms. Princeton University Press, Princeton, NJ Raz J (1999) Engaging reasons. Oxford University Press, Oxford Regan D (1980) Utilitarianism and co-operation. Clarendon, Oxford Ross WD (1930) The right and the good. Oxford University Press, Oxford Scanlon T (1998) What we owe to each other. Harvard University Press, Cambridge, MA Shafer-Landau R (2003) Moral realism: a defense. Oxford University Press, Oxford Schneewind J (1977) Sidgwick’s ethics and Victorian moral philosophy. Oxford University Press, Oxford Sher G (2001) But I could be wrong. In: Paul EF, Miller FD Jr, Paul J (eds) Moral knowledge. Cambridge University Press, Cambridge, pp 64–78 Sidgwick H (1907/1966) The methods of ethics. Dover, New York Sinnott-Armstrong W (2006) Moral skepticism. Oxford University Press, Oxford Smith HM (1988) Making moral decisions. Nous XXII:89–108
162
Bibliography
Smith HM (1986) Moral realism, moral conflicts, and compound acts. J Philos LXXXIII:341–345 Smith M (1994) The moral problem. Blackwell, Oxford Stevenson CL (1944) Ethics and language. Yale University Press, New Haven, CT Stevenson CL (1963) Facts and values. Yale University Press, New Haven, CT Stoutland F (2002) The belief-desire model of reasons for action. Uppsala Prints and Preprints in Philosophy, Uppsala Stratton-Lake P (2002) Introduction. In: Ross WD, Stratton-Lake P (eds) The right and the good. Clarendon, Oxford Stroud B (1981) The significance of naturalized epistemology. Midwest Stud Philos VI:455–471 Stroud S (1998) Overridingness and moral theory. Pac Philos Q 79:170–189 Sturgeon N (1985) Moral explanations. In: Copp D, Zimmerman D (eds) Morality, reason, and truth. Rowman and Allanheld, Totowa, NJ, pp 49–78 Sumner LW (1996) Welfare, happiness, and ethics. Clarendon, Oxford Tännsjö T (1976) The relevance of metaethics to ethics. Acta Universitatis Stockholmiensis, Almqvist & Wiksell International, Stockholm Tännsjö T (1985) Moral conflict, and moral realism. J Philos LXXXII:113–117 Tännsjö T (1987) The moral import of modal realism. Theoria 58:87–96 Tännsjö T (1989) The morality of collective actions. Philos Q 39:221–228 Tännsjö T (1990) Moral realism. Rowman & Littlefield, Savage, NJ Tännsjö T (1995) BlamelessWrongdoing. Ethics 106:120–127. Reprinted in Writing social justice in the arts and humanities. Roosevelt University, Prentice Hall, Englewood Cliffs, NJ Tännsjö T (1998) Hedonistic utilitarianism. Edinburgh University Press/Columbia University Press, Edinburgh Tännsjö T (1999) A concrete view of intrinsic value. J Value Inquiry 33:531–536. Reprinted in Rönnow-Rasmussen T, Zimmerman MJ (eds) Recent work on intrinsic value library of ethics and applied philosophy. Springer, Dordrecht (2005) Tännsjö T (2002) Why we ought to accept the repugnant conclusion. Utilitas X:339–359. Reprinted in Ryberg J, Tännsjö T (eds) The repugnant conclusion. Kluwer, Dordrecht (2004) Tännsjö T (2007) Moral relativism. Philos Stud 135:123–143 Tännsjö T (2006) Understanding through explanation in ethics. Theoria 72:178–213 Tännsjö T (2008) Understanding ethics, 2nd edn. Edinburgh University Press/Columbia University Press, Edinburgh Tännsjö T (2009) On deviant causal chains – no need for a general criterion. Analysis 69:469–473 Temkin L (1993) Inequality. Oxford University Press, Oxford Thomson JJ (2001) Goodness and advice. Princeton University Press, Princeton, NJ Ullmann-Margalit E, Morgenbesser S (1977) Picking and choosing. Soc Res 44:757–785 Williams B (1970) Deciding to believe. In: Kiefer HE, Munitz MK (eds) Language, belief, and metaphysics. SUNY Press, Albany, NY, pp 95–111. Williams B (1981) Moral luck. Cambridge University Press, Cambridge Wilson G (2007) Action. Stanford encyclopedia of philosophy. http://plato.stanford.edu/entries/ action/ Wong DB (1984) Moral relativity. University of California Press, Berkeley, CA Wright C (1995) Truth in ethics. Ratio (New Series) VIII:209–226
Index
A Akrasia, 15 Argument from queerness, 41, 43–50 Argument from relativity, 50–51 Authority of moral norms, 19, 38–40, 45, 80, 88, 94 B Beauty, 6–7, 100, 101, 126, 153 Bergström, Lars, 32, 34, 109, 118, 142, 159 Blackburn, Simon, 48, 62, 63, 159 Blameless wrongdoing (blameful right-doing) 31 Bold conjecture, 115, 122 Brandt, Richard, 24, 52, 159 Braun, Eva, 37 Brentano, Franz, 7, 127, 149, 155, 159 Broome, John, 4, 6, 27, 31, 68–70, 73, 130, 138–143, 148, 159 Buck-passing theory of value, 8, 128, 133, 135, 136, 161 C Carlson, Erik, 34, 90, 130, 159 Carnap, Rudolf, 4 Chang, Ruth, 35, 159 Charity, principle of charity, 36–37 Constructivism, 73, 80 Coherence, 7, 114–116, 122, 123 Cooper, John, 5, 159 Cornell realism, 49, 60, 64–66 Covering law model, 27 Critical level, 79, 81, 82, 84, 130
D Dancy, Jonathan, viii, 11, 32–35, 37–38, 160 Darwall, Stephen, 18, 73, 160 Davidson, Donald, 5, 12, 14, 20, 23, 36–37, 114, 116, 137–138, 140, 160 Deliberative reasons, 20, 38 Deontic logic 76, 82 Determinism (moral determinism), 29, 83 Deviant causal chains, 12, 162 Direction of fit, 13 E Emotivism, 42 Empiricism, 6, 52–55 Epistemic filter, 33 Epistemic goal, 7, 52, 58, 105, 107–123, 125, 137, 145, 146, 154, 155 Error theory, 43, 106–108 Etiquette, 39, 85, 88 Evolution, 7, 18, 53–55, 87, 113, 137, 146–148, 155 Expressivism, 42, 61, 62, 140 Externalism, 49 Ewing, A.C., 7, 35, 54, 120, 127, 128, 132–135, 160 F Fittingness, 125–128, 133, 137 Folk psychology, 4, 118 Frege-Geach problem, 61 G Gauthier, David, 24, 160 Gibbard, Allan, 42, 53, 148, 160
163
164 H Hare, R.M. 24, 92, 105, 160 Harman, Gilbert, 6, 42, 52–55, 85, 87, 115, 160 Hitler, Adolf, 21, 37 Hobbes, Thomas, 5 Hume, David 5, 11, 13, 16, 107, 119, 133, 160, 161 Hurley, P., ix, 160 I Incorrigibility, 115, 121 Idealisation, 23–25, 36, 73 Indeterminacy (normative), 82, 84 Internalism, 45, 49, 73 Intuitionism (moral) 31, 84 J Jackson, Frank, 48, 92, 160 James, William, 7, 109, 110, 112, 116, 154, 160 K Kahneman, Daniel, 17, 160 Kant, Immanuel, 7–8, 28, 39, 48, 60, 65, 82, 91, 95, 97, 125–126, 135, 159–160 Kolodny, Niko, 15, 160 Korsgaard, Christine, 8, 59, 67, 70–74, 160 Kripke, Saul, 37, 160 L Levi, Isaac, 7, 110–112, 115, 160 Locke, John, 8, 119 M Mackie, J., 6, 41–47, 50–53, 55, 151, 160 Mele, Alfred, 14, 31, 161–162 Method of decision-making, 90, 91, 94, 153 Minimalism, 63 Moore, G.E., vii, 6–7, 35, 42, 64–65, 80, 85, 95, 100–102, 126–127, 131–132, 161 Moral explanation, x, xi, 8, 29, 48, 59–66, 68, 71, 84, 128, 133, 162 Myopia (epistemic), 111 N Nagel, T., viii, 35, 54, 58, 128, 161 Naturalism, 8, 48–49, 63–65, 105–106, 108, 118, 122–123, 159
Index Nihilism (moral), 13, 18, 58–60, 81–82, 87, 89 Constraint (the normative constraint) 37 Normative question (the normative qeustion), 8, 67, 70–73, 91 O Olson, Jonas 20, 27, 38, 82, 136, 161 Organic whole, 100, 126 ‘Ought’ implies ‘can’, 33, 76, 135 Overridingness of moral norms, ix, 25, 39, 40, 44–45, 83, 85, 89 P Parfit, Derek, viii, 6–7, 42, 51, 76, 79, 81–82, 130, 133–135, 161 Particularism (moral) ix, 29, 32, 48, 59, 66, 82 Petersson, Björn 14, 161 Phronesis, 90–91, 93 Plato, 44, 55, 162 Practical reasoning, 68, 112, 138–139, 141–143, 148 Prescriptive properties (inherently prescriptive properties), 43–46 Prima facie reasons and obligations, 29–32, 84 Priority view, 39, 79 Prospects (value of) 80, 111, 121, 137, 143 Pro tanto reasons, 31 Prudence, viii, ix, 5, 6, 8, 75–94, 99, 102, 103, 151–153, 157 Psychosis, 19 Q Quasi-realism, 62–63 Quine, W.V. 6, 108, 118, 120, 159, 161 Quinn, Warren, 34, 161 R Rabinowicz, Wlodek, 6–7, 27, 76, 79–81, 161 Radcliffe, Elisabeth, 13, 161 Rational Animal, 20 Reflective equilibrium, 57, 114 Repugnant conclusion, 51, 98 Requirements of rationality, 107, 143, 155 Rawls, John, 6, 56, 96–99 Raz, Joseph, viii, 34, 159, 161 Ross, W.D., 31–32, 84, 161–162
Index S Satisfaction-conditions (of objective norms), 44–46 Scanlon, Thomas, viii, 34, 64, 128, 133, 161 Shafer-Landau, Russ, 41, 161 Schneewind, J., ix, 161 Sidgwick, Henry, 6, 8, 35, 42, 67–68, 75–78, 82–83, 85, 89–90, 95–96, 161 Sinnott-Armstrong, Walter, 43, 161 Stevenson, Charles, 42, 53, 61, 162 Stoutland, Frederick, 34–35, 162 Stroud, S., ix, 108, 116, 162 Sumner, Wayne, 6, 78–79, 162 Sui generis, 41–42, 47–48, 106–107, 118–121, 129, 151–152 Superassertability, 63 Supervenience, 41, 47–49, 62, 76, 159 Svensson, Frans, 20, 38, 161 T Tännsjö, T., xi, 161, 162 Temkin, Larry, 6, 99–100, 162
165 Thomson, Judth Jarvis, 16–18, 27, 35, 42, 85, 160, 162 Thourette’s Syndrom, 22–23 To-be-pursuedness (intrinsic), 44, 46 Tversky, Amos, 17, 160 U Ultra repugnant conclusion, 51 Unexplained explainer, 29 V Value (final or intrinsic) as a place-holder, 7, 39, 41, 80, 100–101, 125–129, 131–132, 162 Williams, Bernard, 23, 30, 107, 133, 162 Voluntarism (doxastic), 107, 133, 135–136 Voluntarism (desire) 133–136 W Wong, David, 42, 85, 162
