From Psychology to Morality: Essays in Ethical Naturalism 0190878592, 9780190878597

Table of contents :
Contents
Acknowledgments
Sources
Introduction: Naturalism in Ethics
1 Concepts of Emotions in Modern Philosophy and Psychology
2 William James and the Rise of the Scientific Study of Emotion
3 Freud
4 Psychopathic Resentment
5 Reactive Attitudes Revisited: A Modest Revision
6 Is Empathy Requiredfor Making Moral Judgments?
7 Williams on Practical Reason
8 Sidgwick’s Conception of Ethics
9 Moral Ideals
10 The Emotional Significance of Punishment
11 Punishment and Proportionality: Part 1
12 Punishment and Proportionality: Part 2
Index
150-162.pdf
150
152
153
154
155
156
157
158
159
160
161
162

Citation preview

i

From Psychology to Morality

ii

iii

John Deigh

From Psychology to Morality Essays in Ethical Naturalism

1

iv

1 Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and certain other countries. Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America. © Oxford University Press 2018 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by license, or under terms agreed with the appropriate reproduction rights organization. Inquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above. You must not circulate this work in any other form and you must impose this same condition on any acquirer. CIP data is on file at the Library of Congress ISBN 978–​0–​19–​087859–​7 1 3 5 7 9 8 6 4 2 Printed by Sheridan Books, Inc., United States of America

v

In loving memory of Sarah Lynn Hill 1948–​2014

vi

vi

Contents

Acknowledgments  Sources  Introduction: Naturalism in Ethics 

ix xi 1



one Concepts of Emotions in Modern Philosophy and Psychology  13



two William James and the Rise of the Scientific Study of Emotion 



three Freud  four Psychopathic Resentment  five Reactive Attitudes Revisited: A Modest Revision  six Is Empathy Required for Making Moral Judgments? 

38 59 83 101 129



seven Williams on Practical Reason 

150



eight Sidgwick’s Conception of Ethics 

179

vi

viii Contents



nine Moral Ideals  ten The Emotional Significance of Punishment 

195 219



eleven Punishment and Proportionality: Part 1 

231



twelve Punishment and Proportionality: Part 2 

249

Index 

271

ix

Acknowledgments

I tried out most of the essays in this volume before various audiences at conferences, meetings of professional organizations, departmental colloquia, and research institute workshops. I am grateful for the interest that was shown and for the many helpful comments I received at these events. I have benefited too from the input of friends, colleagues, and students who read one or more of these essays in draft. I wish, particularly, to thank Daniel Brudney, Brad Cokelet, David Dolinko, Kathleen Higgins, Robert Kane, Herbert Morris, Martha Nussbaum, Larry Sager, Michael Stocker, and Paul Woodruff. My discussions with them over the years about my ideas have been immensely valuable. Similarly, years of discussions with Jonathan Adler and Robert Solomon, and a memorable seminar on emotions that Solomon and I jointly taught, were enormously helpful to me. Unfortunately, their untimely deaths prevent me from also thanking them on this occasion. The three essays on punishment with which the collection concludes came about from my participation in a working group on the problems of reintegration into society that confront felons who have served significant time in prison. The group was sponsored by the Institute for Criminal Justice Ethics at John Jay College, and I am grateful to Jonathan Jacobs, Director of the

ix

x

x Acknowledgments

Institute, and the other members of the group for the rich and illuminating discussions we had about these problems. I wish to thank Peter Ohlin, my editor at Oxford University Press, for his interest and support as well as his help in putting this collection together. Abigail Johnson and Isla Ng also helped with the Press’s editing and production of the collection, and I  thank them too for their assistance. Sarah Shamburg assisted me in preparing the manuscript for production and securing the needed permissions for reprinting the previously published essays. I am very grateful to her for her help. I am grateful too for the institutional support in the form of research leaves, funds, and assistance that I received from the University of Texas at Austin. I had the good fortune early in my career to study with Herbert Morris, who supervised my Ph.D. dissertation. Many of the seminal ideas in my work came from our conversations and his guidance during the years I  worked on the dissertation. Since then he has continued to be a friend and mentor whose encouragement and support, comments and criticism, I treasure. Not long ago he introduced me to the depth of thought in Poussin’s paintings. My choice of Spring as cover art for this collection was made in tribute to a wonderful friendship of more than forty years.

xi

Sources

1. “Concepts of Emotions in Modern Philosophy and Psychology” was first published in The Oxford Handbook of Philosophy of Emotion, Peter Goldie, ed. (Oxford: Oxford University Press, 2010). I presented a version of it at the Fifth Annual Royal Conference on Ethics at the University of Texas at Austin in 2008. 2. “William James and the Rise of the Scientific Study of Emotion” is a contribution to a special section on William James’s theory of emotions in Emotion Review 8 (2016): 56–​61. It is reprinted by permission of Sage Publications Ltd. 3. “Freud” is a considerably revised and expanded version of the article with the same title in A Companion to Continental Philosophy, Simon Critchley and William Schroeder, eds. (Oxford: Blackwell Publishers, 1998). Parts are reprinted by permission of Wiley-​Blackwell. 4. “Psychopathic Resentment” first appeared in Self-​Evaluation:  Affective and Social Grounds of Intentionality, Anita Konzelman Ziv, Keith Leher, and Hans Bernhard Schmid, eds. (Dordrecht: Springer, 2011) and was based on my contribution to a conference on self-​evaluation at the University of Basel in 2009. It is reprinted with permission of Springer.

xi

xi

xii Sources

5. “Reactive Attitudes Revisited: A Modest Revision” is a revised version of my contribution to Morality and the Emotions, Carla Bagnoli, ed. (Oxford: Oxford University Press, 2011). It originated as a paper I  gave at the “Emotions in Context” conference at the University of Chicago Law School in 2008. 6. “Is Empathy Required for Making Moral Judgments?” appeared in Forms of Fellow Feeling: Empathy, Sympathy, Concern and Moral Agency, Neil Roughley and Thomas Schramme, eds. (Cambridge: Cambridge University Press, 2018). It was my contribution to a conference on fellow feeling at the University of Essen in 2013. It is reprinted by permission of Cambridge University Press. 7. “Williams on Practical Reason” has not been previously published. An earlier version was my contribution to a memorial conference for Bernard Williams at the Jean Beer Blumenfeld Center for Ethics at Georgia State University in 2006. 8. “Sidgwick’s Conception of Ethics” was first published in Utilitas 16 (2004): 168–​183. It originated as my contribution to an American Philosophical Association symposium honoring the work of Jerome Schneewind, which took place at the Pacific Division Meeting in 2003. It is reprinted by permission of Cambridge University Press. 9. “Moral Ideals” appeared in Rationality, Rules, and Ideals: Critical Essays on Bernard Gert’s Moral Theory, Walter Sinnott-​Armstrong and Robert Audi, eds. (Lanham, Md.: Rowman and Littlefield, 2002). I presented part 1 of this essay at a conference on Bernard Gert’s Moral Theory at Dartmouth College in 1999 and later at an author-​meets-​critics session at a meeting of the American Society for Value Inquiry in 2001. It is reprinted by permission of Rowman and Littlefield. 10. “The Emotional Significance of Punishment” was published in Emotion Review 8 (2016):  56–​61. It is my contribution to a special issue on law and emotions edited by Susan Bandes and Terry Maroney. It is reprinted by permission of Sage Publications Ltd. 11. “Punishment and Proportionality: Part 1” was published in Criminal Justice Ethics 33 (2014):  185–​199. It is a product of my participation in the working group on prison practices and conditions in liberal democracies organized by the Institute for Criminal Justice Ethics at John Jay College, CUNY. It is reprinted by permission of Cambridge University Press. 12. “Punishment and Proportionality: Part 2” appeared in Criminal Justice Ethics 35 (2016): 20–​36. Like essay 11, it is a product of my participation in the working group on prison practices and conditions in liberal democracies organized by

xi

Sources xiii

the Institute for Criminal Justice Ethics at John Jay College, CUNY. The last part is taken from my contribution to the Fred Berger Prize symposium on Kit Wellman’s article “The Rights Forfeiture Theory of Punishment” at the American Philosophical Association’s Pacific Division meeting in 2015. It is reprinted by permission of Cambridge University Press.

xvi

xv

From Psychology to Morality

xvi

1

Introduction: Naturalism in Ethics

Naturalism has a long tradition in ethics. Arguably, it goes back to the beginnings of moral philosophy when Protagoras held forth at Callias’s salon on the origins of justice and conscience in human beings. The speech he delivered, notwithstanding its fabular character, was a piece of natural history about the emergence in human life of cities and the practices that kept them from breaking up and returning their citizens to an existence of scattered and vulnerable families.1 Its pivotal thesis was that all citizens of a city were equipped with justice and conscience just as other animals were equipped with great strength or great mobility. These traits were therefore unlike the skills of craftsmen, each of which only a small number of specially trained citizens possessed. Indeed, that they were possessed by all or nearly all citizens was crucial to the city’s stability and cohesion. Socrates, in asking Protagoras to back up his promise to young men who would put themselves under his tutelage that he would teach them how to excel at being citizens, had expressed skepticism about the possibility of teaching such virtue. He was certain that no one had expertise in looking after a city comparable to the expertise of skilled craftsmen, and he was equally certain that only those 1. Plato, Protagoras 320c10–​328d3.

1

2

2 Introduction

with such expertise could teach a craft. His mistake, according to Protagoras, was due to his looking in the wrong place for teachers of virtue. Because all or nearly all citizens were equipped with justice and conscience and all or nearly all exhibited these traits in their daily lives, any citizen, with the exception of the few who had turned against the city, were suitable teachers. In fact, Protagoras maintained, one often found ordinary as well as distinguished citizens in their engagement with the young teaching them virtue as naturally as they taught them Greek. Naturalism reached its zenith in ancient philosophy with Aristotle’s ethical writings. Aristotle’s program in ethics was to answer through the study of human psychology the questions that Socrates had asked about how one should live and about the nature of virtue. A deep understanding of the parts and workings of the human soul, Aristotle believed, would show what a good life consisted in and how through the development and exercise of the intellectual and moral virtues it could be achieved. To be sure, Plato too undertook to answer the same Socratic questions through a study of human psychology. He too thought that our coming to understand the parts and workings of the human soul would show us which life was the best for human beings and how one could acquire the virtues necessary to live it. And Plato, needless to say, was no naturalist. The difference between the two, however, is that Plato, unlike Aristotle, attributed to the rational part of the soul powers that gave it access to an intelligible world whose constituents were immutable and eternal and that made the soul itself immortal. Aristotle opposed Plato on these points. While he held that the rational part of the soul distinguished human beings from other animals, he understood the human soul to function entirely within the natural world. Its rational part belonged to the natural world no less than the parts that he identified as common to all animals. His answers to Socrates’s questions thus came entirely from a study of natural phenomena. The naturalist tradition in ethics went into hibernation during the Middle Ages. Its dormancy was due to the dominance of Christian thought throughout the period. Christian thought had merged with Neoplatonism in the last century of the Roman Empire, and this merger, abetted by Augustine’s making the doctrine of free will one of its fixtures, effectively excluded naturalism from moral philosophy for the millennium that followed. Indeed, the doctrine of free will, which Augustine had advanced to absolve God of any responsibility for human sin, was the most important contributor to the silencing of naturalism during the medieval period. The rediscovery of Aristotle’s ethical treatises in the thirteenth century initiated some movement towards a change in this situation. The great influence of these treatises on

3

Introduction 3

Aquinas led to his promoting natural reason as a source of knowledge independent of revelation, and his doing so in turn prepared the way for the tradition’s revival. But it took more than three centuries of philosophical study after Aquinas’s death and the rise of modern science to bring it about. The most important early modern contributors to the tradition’s revival were Hobbes and Spinoza. Neither man was a friend of Platonic forms or their progeny, and both opposed the doctrine of free will that had long since become Christian dogma. Hobbes thought the very notion of a free will was incoherent. Spinoza thought free will was an illusion. Both pursued ethics—​the science of virtue and vice, as Hobbes called it—​on the basis of the naturalist-​cum-​egoist theories of human psychology that they expounded. As a consequence of their radical views, both were targets of vigorous attacks by the defenders of religious orthodoxy, Judaic as well as Christian in Spinoza’s case. Philosophical resistance to Hobbes’s thought was also fierce. Spinoza, by contrast, spent years in the philosophical wilderness until late in the eighteenth century, when several prominent German thinkers began to take him seriously. In Britain, the impact of Hobbes’s philosophy was immediate. The debates over egoism that his ethics ignited produced two generations of rationalist, particularly Platonist, responses. In time, though, it also produced a new set of defenders of naturalism in ethics. The greatest of these was Hume, whose Treatise of Human Nature surpassed in the depth and subtlety of its theory of human psychology all previous works in British moral philosophy. Hume, like Hobbes, rejected the orthodox Christian doctrine of freedom of the will. At the same time, he went beyond Hobbes in explaining how human actions could be effects of antecedent causes compatibly with their being freely chosen. He then turned to expounding a theory of ethics based on his naturalist psychology. This exposition too went beyond Hobbes, not only in its rejection of egoism, but also in its account of justice as a virtue whose origins lay in the natural history of human beings. Hobbes had built his account of justice on the idea of the natural condition of humankind, but he did not understand this idea historically. The natural condition of humankind, in Hobbes theory, was the condition human beings would be in if they lived without benefit of a government, and such a condition could be the result of a civil war, as Hobbes feared, as well as the original condition of savages. Human behavior, in his view, would be the same in either case. Hume’s use of natural history to explain the origins of justice cancelled the advantage rationalist opponents of naturalism had seemingly held given the difficulty of reducing justice to either self-​ interest or benevolence. Nonetheless, naturalists were not alone in using natural history to support

4

4 Introduction

their theories. It did not belong exclusively to their toolkit. The outstanding example of its use outside of naturalism is Rousseau’s account in his Discourse on the Origin of Inequality of the birth and growth of human injustice and cruelty. Rousseau agreed with Hume in taking justice to have originated in the conventions of property. Unlike Hume, however, who explained these conventions as the consequence of cooperative schemes that human beings invented for their mutual advantage, Rousseau explained them as the consequence of consent to the appropriation of land by those who worked it. This difference between the two views is crucial. For Rousseau, in making consent to being excluded from the use of land by another’s declaring it his own the origin of justice, implied that justice and consequently injustice derived from the operations of the human will. Of course, if Rousseau had thought that a man’s will was itself determined by natural forces, his view would not have been crucially different from Hume’s. But Rousseau held exactly the opposite view: I see nothing in any animal but an ingenious machine . . . I perceive exactly the same things in the human machine, with this difference, that in the operations of the brute, nature is the sole agent, whereas man has some share of his own operation, in his character as a free agent. The one chooses and refuses by instinct, the other from an act of free will.2

Rousseau’s project was to explain the wretchedness of the human condition as entirely the work of men. He reiterated his theme in the opening sentence of Emile, “Everything is good as it leaves the hands of the Author of things; everything degenerates in the hands of man.”3 It is hard to think of a better coda for Augustine’s doctrine. The brilliance of Rousseau’s natural history lay in its developmental story of how humans began their sojourn on earth as innocent, solitary beings with simple desires for the necessities of self-​preservation and then, as the result of their entering into stable and complex social relations of increasing extent and complexity, acquired new desires that eventually made them capable of both moral virtue and moral vice, with a decided tendency towards the latter. What is striking about this history is its implication that human beings create the social structures that make it possible for them to achieve moral virtue or sink into moral depravity. Morality, in other words, on this history, is an artifact of human culture, and therefore human beings, rather than possess a 2.  Jean-​Jacques Rousseau, Discourse on the Origin of Inequality, in The Social Contract and Discourses, G. D. H. Cole, trans. (New York: E. P. Dutton & Co., 1959), p. 207. 3.  Jean-​Jacques Rousseau, Emile, Allan Bloom, trans. (New York: Basic Books, 1978), p. 37.

5

Introduction 5

conscience and moral sensibilities innately, acquire them through acculturation. While this view may have had earlier exponents, no one had expounded it more forcefully or imaginatively than Rousseau. Of course, the view itself poses no threat to naturalism in ethics. After all, it is open to defenders of such naturalism simply to disagree with Rousseau about the origins of morality. They could, as against Rousseau, explain conscience and moral sensibilities as being among our inherited powers and dispositions. No doubt, many followers of Darwin would explain them in this way. Nonetheless, a growing disenchantment with teleological explanations of human nature, accelerated by the appearance and acceptance of Darwin’s theory, increased the attractiveness of Rousseau’s thesis that morality is wholly an artifact of human culture as an alternative to the supposition that conscience and moral sensibilities are part of man’s natural endowment. It remained then for thinkers in the naturalist tradition to construct theories of the development of conscience and moral sensibilities that substantiated Rousseau’s thesis. The two most influential theories of this sort were Nietzsche’s and Freud’s. Both men expounded thoroughly naturalist accounts of the workings of the human mind. Their understanding of these workings, however, was decidedly different from that of the associationist psychology that Hume had expounded and bequeathed to nineteenth-​century British empiricism. Nietzsche and Freud based their accounts of the mind’s workings, not on principles of association, but rather on the dynamics of unconscious conflict among instinctual forces, and they used these dynamics to explain the formation in human beings of a conscience and moral sensibilities. Nietzsche weaved his explanations into the natural histories he offered in On the Genealogy of Morals. Freud made his explanations central to the individual psychology he developed and then fit them to a historical account of the emergence of conscience in human beings. Both took up Rousseau’s theme of representing the emergence of conscience as involving a transformation of human nature. Of the two theories, Freud’s had the farthest reach in its influence on subsequent thought. And to the extent that the view of morality as an artifact of human culture was widely accepted, his theory did more to advance naturalism in ethics than any other theory in psychology or social science put forward in the last century. The essays collected in this volume belong to the naturalist tradition whose course I have just sketched. The first ones are studies in psychology that bear on questions of ethics. Before one can take up the central question of ethics concerning how to live, one must come to some understanding of the place of desires and emotions in our lives and how reason works to

6

6 Introduction

direct their promptings. On a naturalist understanding of human nature, desires and emotions are the ultimate springs of our action, while reason provides us the intelligence we need to act well. The first essay is a critical survey of the major concepts of emotion in modern philosophy and psychology from the older, empiricist concepts of the eighteenth and nineteenth centuries to the present. The discussion highlights the importance of the theories of William James and Sigmund Freud first in overthrowing these older concepts and then in shaping later conceptualizations of the phenomena. The next two essays focus on James’s and Freud’s theories respectively. The essay on James explains how his concept of emotion derived from Hume’s distinction in the Treatise between calm and violent passions and then argues for the unsuitability of James’s concept for ethics and, indeed, for any field in which emotions are conceived of as motives. I conclude that recent attempts in neuroscience and philosophy to resurrect James’s theory are dead-​ends. The essay on Freud has two movements. In the first I trace the development of his thought from his original theoretical work on the neuroses exhibited in hysteria and obsessive behavior to his construction of a general theory of human personality and its disorders. In the second, I explain the philosophical import of his theory as a contribution to the naturalist tradition in Western thought and how it made significant advances over its predecessors. The two essays that follow deal obliquely with the doctrine of free will. Their aim is to vindicate the strategy for dealing with the doctrine that P. F. Strawson pursued in “Freedom and Resentment,” his landmark address to the British Academy. Strawson’s strategy was to give a naturalist account of our practice of holding people morally responsible for their actions. Such an account, if successful, would negate the proposition that a person is morally responsible for an action only if the action resulted from his or her exercising a free will, and Strawson used a study of certain emotions that he called “reactive attitudes” to show how one could give such an account. He thus supplied naturalism with a powerful argument against one of the main bulwarks of the doctrine. His address has had enormous impact on subsequent philosophical treatments of the question of free will. Many philosophers, writing under the influence of his argument, have drawn their accounts of moral responsibility from considerations of the reactive attitudes. Yet despite this impact, his argument has won few converts to naturalism. The reason is that it has been poorly understood. The misunderstanding is twofold. First, philosophers who have taken themselves to be following Strawson commonly mischaracterize the way he understood resentment. Second, the same philosophers have failed to grasp the crucial role of naturalism in his argument, and, consequently, their

7

Introduction 7

accounts of moral responsibility do not threaten views on which moral responsibility implies an exercise of free will or at least do not threaten such views in the way Strawson saw his argument as doing. “Psychopathic Resentment” corrects the first misunderstanding. “Reactive Attitudes Revisited” corrects the second. The sixth and seventh essays in the collection shift the focus to naturalist accounts of moral judgment and practical reason. The sixth essay examines the place of our capacity for empathy in making moral judgments. Hume’s ethics is commonly interpreted as taking empathy—​ what Hume called “sympathy”—​to be a necessary factor in our making moral judgments, and its role in his theory supports his thesis that moral judgments are inherently motivational. Against Hume, I argue for a more capacious understanding of moral judgment. This more capacious understanding fits Piaget’s account of how moral judgment develops in children, and Piaget’s account, I further argue, offers a better framework than Hume’s for investigating the motivational character of moral judgments. The seventh essay is a partial defense of Bernard Williams’s view that the source of an agent’s reasons for action is, in every case, some element in the agent’s set of desires, emotions, evaluative dispositions, inclinations, and the like. Williams’s view is defensible when one takes reasons for action as facts an agent considers when deliberating about what to do, given an Aristotelian notion of deliberation. It is indefensible when one takes reasons for action to be facts an agent considers when determining whether doing the action would be or was justified. Much of the controversy that Williams’s view stirred, including some of the most virulent objections to it, is due to confusion of these two different ways of understanding reasons for action. After explaining the confusion, tracing its origins to the beginnings of the philosophical literature on reasons for action in Anglo-​American philosophy, and dispensing with the objections that were based on it, I defend Williams’s view against objections, Christine Korsgaard’s, in particular, that were not based on it. The eighth essay also focuses on practical reason. Sidgwick’s conception of ethics is its specific topic. Sidgwick, in his masterwork The Methods of Ethics, proposed to treat each of the main theories of ethics as being rooted in a rational procedure for determining what one ought to do. In short, he conceived of ethics as the study of practical reason. He took egoism, utilitarianism, and dogmatic intuitionism as the main theories to be examined with the aim of showing that they could be unified. Such unification requires that the theories agree on the sense in which ‘ought’ is to be understood in moral imperatives of the form A ought to φ, where A is the name of an

8

8 Introduction

individual and φ is a verb of action. When egoism is formulated as a naturalist theory, however, the sense ‘ought’ has in these moral imperatives is different from the sense it has in the moral imperatives of dogmatic intuitionism. The reason is that on such a formulation, on Hobbes’s formulation, for instance, the starting assumption is that a brute desire for one’s own’s happiness predominates among the moving forces in one’s life. Consequently, on this formulation, moral imperatives of the form A ought to φ are hypothetical in Kant’s sense. By contrast, on dogmatic intuitionism, the sense ‘ought’ has in moral imperatives of this form is categorical in that, unlike hypothetical imperatives, the conditions of their validity do not include A’s having a desire that A would satisfy by φing. To be sure, a naturalist formulation of egoism was not Sidgwick’s preferred formulation. He preferred a rationalist formulation. Its starting assumption was that reason prescribed one’s own happiness as the ultimate end of action. But he allowed the possibility of a naturalist formulation. The upshot of his doing so, as I argue, is that his project of unifying the three theories was foredoomed. The ninth essay is a contribution to a volume of critical essays on Bernard Gert’s moral theory. It is chiefly a critical discussion of Gert’s treatment of moral duties and moral ideals. The central aim of Gert’s theory is to represent morality as a set of precepts that all rational and impartial persons would favor everyone’s following. Construction of such theories was common in ethics in the latter half of the twentieth century, and Gert’s construction is an especially clear and uncomplicated example of this type of theory. Gert divided the set of moral precepts into rules and ideals, though in truth he could have called them all rules and divided them into those that were meant to constrain action and be enforceable by punishment and those that were meant to promote action without the threat of punishment. He thus conceived of morality as a set of rules regulating behavior. This conception too was common to the type of theory his exemplified. The thrust of my criticism is that Gert’s is an impoverished conception of morality. Morality is as much concerned with our inner life as it is with regulating behavior. It contains, for instance, a duty of benevolence as well as duties not to kill and to keep one’s promises. Gert thinks the very notion of a duty of benevolence is problematic. Moral duties, he declares, correspond to rules that either forbid or require that certain actions be done, and requiring general benevolence would mean, he then argues, that a person would have to act benevolently all of the time towards everyone, which is impossible. If, however, one takes morality to guide the will or prescribe dispositions and not merely to regulate behavior, the problem Gert alleges does not arise. On this conception of morality, the duty of benevolence is the

9

Introduction 9

duty to maintain a good will or a benevolent disposition towards all, which means that what is required to fulfill it is a commitment of the will and not an action. Such a commitment, I argue, is certainly possible. I conclude by sketching an account of moral ideals on which they are understood as models or exemplars and not rules. The collection closes with three papers on punishment. The standard treatment of the topic in philosophy has for decades taken legal punishment as the paradigm case. Accordingly, philosophers writing on punishment have commonly understood its infliction as an institutional response to a breach of an institution’s rules. My treatment of the topic in these three papers rests, instead, on a broader understanding of punishment. On this understanding, punishment is as much at home in a natural context as an institutional one. The broader context is that of a stable social group. Outside of an institutional context, punishment is pain or loss inflicted by a member of the group on another member who is equally or less powerful in retaliation for the latter’s having displeased or angered the former. Within an institutional context, it is pain or loss inflicted by an agency formally empowered to enforce the rules of the institution in retaliation for their having been violated. In either case, it works to preserve the group’s social order, which is either natural or civil according as the relations of power among the members are unmediated or mediated by rules. This understanding covers not only legal punishment but also punishment that one country inflicts on another in retaliation for the latter’s having attacked its citizens or property, punishment that mob bosses inflict on underlings in retaliation for the latter’s having betrayed them to another mob or to some government authority, and punishment that husbands or wives inflict on their spouses and parents inflict on their children in retaliation for being angered by their spouses or children. All three papers take as their point of departure this broader understanding of punishment. The first of the three, “The Emotional Significance of Punishment,” explains how punishment works to preserve a group’s social order. I identify three functions that it has in this respect. It deters those inclined to act in ways that disrupt the order from doing so; it strengthens the disposition to maintain the order of those who are not so inclined by assuring them that disruptive acts are not tolerated; and it provides a means by which its recipients can repair the social relations they had ruptured. The explanation of the last function, punishment’s reparative function, proceeds from an account of the dynamics of these social relations that reprises the discussion of reactive attitudes in the collection’s earlier essays. Punishment has these three functions, not by design, but by virtue of how retaliatory behavior within a

01

10 Introduction

social group between members of equal power or by more powerful members towards less powerful ones works to maintain the group’s stability. The final two essays deal with questions of justice with regard to making the severity of punishment in the penal law match the gravity of the offense for which it is inflicted. The requirement of justice that the severity of punishment match the gravity of the offense is a requirement of proportionality. The first of the two essays examines whether a legislator’s or a legislative assembly’s determining the severity of punishment for an offense with the aim of optimally fulfilling the deterrent function of punishment would meet this requirement. This way of determining the severity of a punishment for a given offense, I  argue, yields harsher punishment than that of aiming at preserving the social order, and for this reason it tends to yield punishments that are disproportionately severe. The second of the final two essays takes up the strict retributivist’s way of determining the severity of a punishment for a given offense. This way of determining such severity, I argue, is morally problematic. It represents the view that justice requires retaliation for its own sake, and such a view contravenes principles of humanity basic to Western ethics, specifically, our general duty of benevolence towards human beings. At the same time, punishment is not similarly problematic when it is inflicted for a larger social end than mere retaliation, when it is inflicted, for instance, to promote public safety or preserve social order. Oxford University Press solicited two scholars to review this collection. Both made several good suggestions for improving its presentation. One of them, noticing the affinity of some of my views to Hume’s, suggested that I identify the kind of naturalism the collection’s essays represent as a version of sentimentalism. This suggestion, however, overstates the affinity of my views to Hume’s. Hume’s naturalism combines sentimentalism with an opposition to rationalism on the question of reason’s power to motivate action. I  share Hume’s opposition to rationalism. The seventh essay, the essay on Williams’s account of practical reason, makes clear my acceptance of his view that reason lacks the power to motivate action on its own. I  do not, however, share his subscription to sentimentalism. Yet given the tendency of scholars in a field to place a set of ideas in a familiar niche so as to better understand their complexity, it may be helpful to explain the separability of these two aspects of Hume’s naturalism. Sentimentalism is a thesis about how human beings tell the difference between virtue and vice. Specifically, it is the thesis that our affective responses to people’s behavior rather than our cognitions of its character are what ultimately determine whether we find it virtuous or vicious. “Morality,” Hume said in summarizing his sentimentalism, “is more properly felt than

11

Introduction 11

judg’d of.”4 Hume rested this thesis on his account of what he called indirect passions, the sentiments of approbation and disapprobation, in particular. These passions are built into the human frame, which means that as soon as a child’s affective capacities have developed enough to experience them, the child is conscious of virtue and vice. They are, in this respect, like visual impressions of color. That is, a child becomes conscious of different colors as soon as its visual powers are sufficiently developed to be capable of discriminating among them. Hume, indeed, affirmed the analogy between colors and virtues, colors and vices.5 One can reject this analogy, though, and therefore sentimentalism while at the same time opposing, as Hume did, rationalism on the question of reason’s power to motivate action. Thus the two aspects of his naturalism come apart. In fact, I do not accept the analogy. I do not believe it can be sustained. My discussion in the first essay of problems with the perceptual model of emotions explains why.6 The gist of the discussion is that one cannot tell the difference between virtue and vice unless one understands what would make a feature of a person’s character a virtue and what would make it a vice. It is not sufficient, contrary to Hume, that the motives one discerns behind people’s actions please or displease one when one views them disinterestedly. One must also have the concepts of benefit and harm so as to understand how actions that spring from such motives tend to benefit or harm the agent or others. By contrast, one does not need to understand anything about colors to see them. The failure of the analogy is evident in the difference between how one teaches a child the words for different colors and how one teaches it the words for different values. On the account of the latter teaching that I favor, it serves to retarget the child’s desires and emotions away from some of the objects they have as inherited dispositions and towards others that did not previously arouse them. Such teaching corresponds to the training of desires Aristotle takes to be necessary to one’s acquiring the moral virtues. The account, then, fits a naturalist theory of ethics on which our telling the difference between virtue and vice is a matter of cognition and not feeling compatibly with reason’s lacking the power to initiate action. 4.  David Hume, A Treatise of Human Nature, 2nd ed., L. A. Selby-​Bigge, ed. with rev. by P. H. Nidditch (Oxford: Oxford University Press, 1978), p. 470. 5.  See David Hume, Enquiries concerning Human Understanding and concerning Principles of Morals, 3rd ed., L. A. Selby-​Bigge, ed. with rev. by P. H. Nidditch (Oxford: Clarendon Press, 1975), p. 294. 6.  I give a more extended critique of Hume’s analogy in an earlier essay “Emotions and Values,” which appears in my collection Emotions, Values, and the Law (Oxford:  Oxford University Press, 2008).

21

12 Introduction

At the same time, the account yields an explanation of Hume’s thesis that morality is essentially practical. This result may strike some as untenable for any naturalist theory on which our telling the difference between virtue and vice is a matter of cognition. One cannot, so it is commonly thought, side with Hume in opposing rationalism on the question of reason’s power to initiate action without having either to reject Hume’s thesis or to deny, as he did, that we know the difference between virtue and vice through cognition. Yet the choice between these two options is a false dilemma. One can defend a naturalist account of ethical evaluation without being forced to decide between them. Words for values enter our vocabulary when we are young through teaching intended to reinforce or adjust the targets of the desires and emotions the liability to which we inherit from our simian ancestors. The aim of the teaching is the same as that of teaching us to live well as civilized human beings and members of the human society into which we are born. Our grasp and use of these words continue to guide us long after childhood when they have ceased to refer only to things that spark desire and arouse emotion either innately or through teaching. Though we learned the difference between danger and safety in learning what to fear and what inspires confidence, our ability to recognize things and situations as dangerous or safe outstrips our disposition to feel fear or confidence—​to be afraid or confident—​in the presence of those things once we have mastered the words and learned how to navigate our lives. We know that a gun if loaded or a car if speeding at high velocity is dangerous, and we thus take care in how we handle the gun or drive the car. We do this without having to be afraid of the former’s discharging or the latter’s crashing. Nor do we think, when we recognize the danger, that it would be rational or fitting to feel fear or that we have reason to feel it or ought to feel it. The thought guides our action, though the danger is not an object of fear. Philosophers who require a definition of the word ‘danger’, a verbal account of its meaning, that explains why applying the word to things and objects guides action are destined to be forever trapped in Wittgenstein’s fly bottle. I believe the account of ethical evaluation I give shows the way out.

13

1

Concepts of Emotions in Modern Philosophy and Psychology

Two major themes characterize the study of emotions in modern philosophy and psychology. One is the identification of emotions with feelings. The other is the treatment of emotions as intentional states of mind, that is, states of mind that are directed at or towards some object. Each theme corresponds to a different concept of emotions. Accordingly, the study has divided, for the most part, into two main lines of investigation. On one, emotions are conceived of as principally affective states. The concept on which this line proceeds is feeling-​centered. On the other, emotions are conceived of as principally cognitive states. The concept on which this line proceeds is thought-​ centered. Both concepts reflect revolutionary changes in the theoretical study of emotions that began to take place at the end of the nineteenth century and continued for several decades into the twentieth. The two main lines of investigation come out of objections to the concept of emotion that dominated philosophy and psychology in the eighteenth and nineteenth centuries. The revolution in our thinking about emotions begins with the abandonment of this older concept.

13

41

14  From Psychology to Morality

I The older concept is a fixture of empiricist psychology in the modern tradition. Empiricists of the eighteenth and nineteenth centuries saw the mind as a single field of thought and feeling, fully conscious and transparent to itself. Its chief inhabitants are distinct ideas and impressions. The latter include visual, auditory, and tactile impressions as well as other sensations of external things. The former are products of these impressions. They are either simple or complex, and if complex, then one can always understand them as combinations of simples. In Hume’s wonderful simile, impressions and ideas are “like players in a theater who successively make their appearance, pass, repass, glide away, and mingle in an infinite variety of postures and situations” (1978, p. 253). The class of impressions includes more, though, than sensations of external things. It also includes both localized and unlocalized feelings of pleasure and pain. The latter are the emotions. They are in Locke’s words “internal sensations” (1975, pp. 229–​230) and in Hume’s “secondary” or “reflective” impressions (1978, p. 275). The basic ones are simple impressions of pleasure or pain whose connection to ideas, physiological activity, and conduct is either that of cause to effect or conversely. Specifically, on Locke’s view, these internal sensations result from ideas of good and evil; on Hume’s their immediate causes are impressions of pleasure and pain along with, in some cases, ideas of external things. At the same time, on the empiricist’s view, we can conceive of these basic emotions in abstraction from their typical causes and effects, just as we can conceive of simple sensations of color or sound. Abstracted from their causes and effects, according to classical empiricism, emotions are discrete, episodic, and purely affective states of consciousness. The traditional empiricist concept of emotions came under a withering attack by William James. The attack was part of James’s broadside, in the famous ninth chapter of The Principles of Psychology, against traditional empiricism’s general understanding of mental states as either wholly simple states or complex states composed of these simples. Hume’s simile nicely presents James’s target. The empiricist understanding of mental states, James observed, presupposed that “sensations came to us pure and single” (1950, vol. 1, p. 233), that they recurred at different times in our experiences of the world, and that the great mass and constantly changing flow of thought in our minds was due entirely to the combination and recombination of these simple sensations and their corresponding simple ideas. None of these three suppositions, James argued, can be sustained. Sensory experience is no more made up of individual units of sensation than rivers are made up of individual drops of water. Our thought, James argued, naturally appears to us as

15

Concepts of Emotions in Modern Philosophy and Psychology  15

“sensibly continuous,” like a stream (1950, vol. 1, p. 237). It can no more be broken down into mental atoms whirling in a void than its correlative brain activity can be broken down into discrete, concatenated events. Secondly, this stream of thought is Heraclitan. One never, James insisted, had the same sensation twice. Indeed, sensations, according to James, are unrepeatable items of experience. Because each sensation corresponds to some modification of the brain, there could be no recurrence of the exact same sensation. For that to happen, the second occurrence would have to occur “in an unmodified brain,” which is “a physiological impossibility” (1950, vol. 1, pp. 232–​233). Thirdly, if, contrary to Locke’s supposition, there are no simple, recurrent sensations, then there are no corresponding simple, recurrent ideas either. Consequently, human thought does not break down into elementary units of feeling and thought. The continuous change in what we feel and think does not consist in the combining and recombining of the same set of simple, immutable sensations and ideas. We are led to think otherwise, James observed, by our habit of identifying sensations of sight, sound, taste, and so forth by their objects. We speak of the same sounds when we hear the same thing on different occasions, a musical note, say, or a bird’s chirp. Inattentive to variations in our sensory experience, we say that the sensations are the same when, in fact, there is no individual sensation on any of these occasions and therefore no relation of identity holding between sensations on different occasions. What is the same is the object of sensation. When we listen to the performance of a sonata, for example, we hear the same notes repeated at different intervals and in different arrangements, yet it would be wrong to think that each time we heard some note during the performance, we had a distinct auditory sensation that was separate from and a successor of an equally distinct auditory sensation we had in hearing the previous note and that was identical to an auditory sensation we had when we last heard this note. The experience of hearing the sonata does not consist of separate, individual sensations of sound experienced sequentially and repeated at different intervals. It is only the tendency to confuse these sensations with their objects that leads us carelessly to think of the experience in this way and to speak of having the same sensation when we hear the same note. We are not, of course, liable to confuse emotions with their objects, for we do not identify them with their objects. Fear, after all, is fear whether it is fear of spiders or fear of earthquakes. Nonetheless, emotions too are misconceived on the traditional empiricist understanding of them as either simple, recurrent states of mind or compounds of such states. “Pride and humility,” Hume tells us, at the very beginning of his discussion of these

61

16  From Psychology to Morality

passions, are “simple and uniform impressions” (1978, p. 277). So too, he says, are the other passions (though later, in book II of the Treatise, he identifies several passions, respect, contempt, and amorous love, as mixtures of two other passions (1978, pp. 389–​396)). The traditional empiricists’ misconception, in this case, is due, not to their mistaking emotions for their objects, but rather to their modeling emotions on external sensations. Thus, Locke, having begun his study with simple, external sensations, then introduces pleasures, pains, and passions, which arise with external sensations, either through impact on the body of some external object or with our thoughts of good and bad. The latter are emotions, and we experience them, Locke says, as internal sensations. Similarly, though more elegantly, Hume identifies external sensations as primary impressions, the simple ones being the starting points of his investigations, and then, when he takes up emotions, identifies them as secondary impressions. The understanding of emotions, in either case, therefore inherits the mistakes James criticized in the traditional empiricist treatment of external sensations. And James then went on in c­ hapter 25 of The Principles to level the same criticisms of the traditional empiricists’ understanding of emotions. “The trouble with the emotions in psychology,” he wrote, “is that they are regarded too much as absolutely individual things. So long as they are set down as so many eternal and sacred psychic entities, like the old immutable species in natural history, so long all that can be done with them is reverently to catalogue their separate characters, points, and effects” (1950, vol. 2, p. 449). James’s objection to the traditional empiricist concept of emotions was thus continuous with his objection to the traditional empiricist understanding of mental states generally. To identify each type of emotion as a distinctive, recurrent inner sensation or feeling, James argued, is to misunderstand the nature of emotional states of mind. At the same time, James accepted the traditional empiricist characterization of emotions as feelings. The feelings that he identified with emotions, however, were feelings produced by changes in the body. Emotions, according to James, were those feelings that arose as the result of the physiological and neurological changes that typically occurred as the result of the perception of some exciting fact. As James put it, “My theory . . . is that the bodily changes follow directly the perception of the exciting fact, and that our feeling of the same changes as they occur is the emotion” (1950, vol. 2, p.  449). By so identifying the emotion with these feelings, James reversed the commonsense order of events in an episode of emotion. For common sense has it that what excites the emotion is the perception of a fact, and the emotion then causes the bodily changes that express it. According to common sense, a person perceives a charging bear,

17

Concepts of Emotions in Modern Philosophy and Psychology  17

for example, feels fear, blanches, and runs, whereas on James’s account, the perception of the charging bear causes the person to blanch and run, and the feeling of these bodily movements is the fear. In effect, then, on James’s account, the emotion does not cause the behavior that expresses it or the behavior it is commonly said to motivate. Rather such behavior is the direct effect of the perception of some fact, and feeling the bodily changes the behavior consists in is the emotion. Emotions, on James’s account, are therefore epiphenomenal. They are the products of bodily changes, but they do not themselves cause any action.1

II James’s identification of emotions with bodily feelings is one of two major innovations behind the abandonment of the traditional empiricist concept of emotion. The other emerged in treatments of emotion that oppose the epiphenomenalism in James’s account. James, to be sure, regarded his reversing the commonsense account of the relation of emotions to their behavioral expressions as an advance in the study of emotions. But not every contributor to the study did. Some stuck with common sense. On the commonsense account, emotions are springs of action. They have motivational force. In reversing common sense, James implicitly denied that they had such force, and this consequence of his program is the rub. For it is not easy to give up citing emotions to explain actions.2 The traditional empiricist concept of emotions supported such explanations. On traditional empiricism, emotions are motives of actions. Specifically, they arise from thoughts or perceptions and immediately move their subjects to action. A good example is Hume’s account of the emotions 1.  In taking James’s rejection of the commonsense order of events in an episode of emotion as affirmation of the view of emotions as epiphenomenal, I do not mean to suggest that James held that all mental states were epiphenomenal. In The Principles of Psychology, ­chapter  5, James appears to reject the latter view on grounds of common sense. Hence, his express challenge to common sense in his theory of emotions implies that he is making an exception in this case to the rule of common sense that he invokes in rejecting epiphenomenalism as a general theory of the mind. See James, 1950, vol. 1, pp. 128–​144, esp. p. 144. 2.  Indeed, while James’s considered view was to deny that emotions had motivational force (lest they precede rather than succeed the bodily movements that express them), he too sometimes fell back to the commonsense understanding of emotions as springs of action. Thus, in the chapter on instincts, c­ hapter 24, he wrote, “Sympathy is an emotion . . . [s]‌ome of [whose] forms, that of mother with child, for example, are surely primitive, and not intelligent forecasts of board and lodging and other support to be reaped in old age. Danger to the child blindly and instantaneously stimulates the mother to actions of alarm and defense” (1950, vol. 2, p. 410).

81

18  From Psychology to Morality

he called “the direct passions.” The direct passions, on Hume’s account, include desire and aversion, joy and grief, hope and fear, anger and benevolence, among others. Like all secondary impressions in Hume’s psychology, they are inner feelings, each being qualitatively distinct from the others. At the same time, instances of each vary according as the feeling they consist in is more or less turbulent and as it is more or less forceful. Hume, that is, distinguished between the phenomenal character of an emotion and its motivational strength. Emotions are either violent or calm depending on how much agitation in the mind they entail. They are either strong or weak depending on the amount of force with which they motivate action. In other words, on Hume’s account, the degree to which an emotion is calm or violent is independent of the degree to which it is weak or strong (1978, p. 419). Consequently, a passion may be calm; it may produce little agitation in the mind; yet it may be stronger than some violent, co-​occurring passion and so move its subject to actions that are contrary to the actions the more violent passion is prompting. As Hume put it, “’Tis evident passions influence not the will in proportion to their violence, or the disorder they occasion in the temper; but on the contrary, that when a passion has once become a settled principle of action, and is the predominant inclination of the soul, it commonly produces no longer any sensible agitation” (1978, pp. 418–​419). Indeed, any passion, Hume observed, “may decay into so soft an emotion, as to become, in a manner, imperceptible” (1978, p. 276). Certain calm ones, in particular, “are more known by their effects than by the immediate feeling or sensation” (1978, p. 417). What I shall call “Hume’s doctrine of the calm passions” implies both that we can know of an emotion independently of its phenomenal character and that we can conceive of it as present and operative even though it lacks a phenomenal presence. The doctrine, consequently, suggests a different concept of emotion from that employed in traditional empiricism. It suggests a concept focused on the motivational rather than the affective character of emotions. Hume of course did not make this concept explicit. He could hardly have done so and kept quiet, as he did, about the seeming incoherence he introduced into his system in referring to an imperceptible impression. An unfelt feeling, after all, is a contradiction in terms. Hume, in other words, could not have maintained the doctrine of the calm passions unqualified without having to abandon his notion of an emotion as a secondary impression. More generally, he could not have maintained the doctrine without abandoning the framework of traditional empiricist psychology within which he worked. On that framework the mind is a field of consciousness, and its states, therefore, are essentially conscious. The doctrine of the calm passions, to the contrary, implies the possibility of

91

Concepts of Emotions in Modern Philosophy and Psychology  19

some emotions, though states of mind, occurring outside of this field. Some emotions may be, in a word, unconscious. The idea that emotions need not always be conscious states is the second major innovation behind the abandonment of the traditional empiricist understanding of emotion. While Hume’s doctrine of the calm passions suggests this innovation, it was not until Sigmund Freud developed his theory of the unconscious, 150  years later, that it was fully realized. Freud was not the first thinker to propose that there were unconscious mental states. He readily acknowledged predecessors. But those precedent proposals did not present much of a challenge to the traditional empiricist framework. Typically, what was proposed was that some thoughts and ideas existed just beyond the periphery of the mind, in a subconscious part, and if not easily retrievable, could become conscious nonetheless without effort. Examples were ideas corresponding to words that are “just on the tip of our tongues,” as we say, that we struggle to recall only to have them suddenly appear clearly in consciousness, and ideas that must occur to us during sleep since it is not uncommon for one, upon waking up in the morning, to see immediately the solution to a problem that had vexed one the night before. Freud, by contrast, held that some ideas and thoughts were deeply buried in an unconscious part of the mind and were blocked from being retrieved by repression. One’s mind, as it were, generated force to keep them from becoming conscious. Such repression was necessary because the repressed thoughts and ideas were themselves charged and energetic and would immediately rise to consciousness if there were no counterforce to keep them in check. Examples of such repressed ideas and thoughts were beliefs about being personally responsible for some traumatic event that occurred when one was very young. Such beliefs, while they may never become conscious, reveal themselves in dreams, self-​defeating behavior, even illness. That they manifest themselves in this way is Freud’s reason for understanding them as charged. Unconscious thoughts and ideas could not have the influence on people Freud saw in their dreams, irrational behavior, and illnesses that lack organic causes unless they were forceful in their own right. Freud called the tension between their force and that of repression “the dynamic unconscious.” The great originality of his theory of mind was due to this way of conceiving of unconscious thoughts and ideas. Freud took emotions too as capable of being repressed (Freud, 1915, 1923, 1924). They too could be unconscious. Consequently, he could not conceive of emotions as feelings. Nonetheless, he is commonly interpreted as having so conceived of them, notwithstanding the evident incoherence his taking emotions as feelings would, as in Hume’s case, introduce into his theory. (See, e.g., Lear, 1990, pp. 88–​90.) The passages that are commonly

02

20  From Psychology to Morality

cited to support this interpretation are not decisive, however, and, I would argue, are better read as supporting an interpretation on which Freud gives meaning to the term ‘unconscious emotion’ while acknowledging that the term ‘unconscious feeling’ is self-​contradictory (Freud, 1915; Deigh, 2001). Emotions may be unconscious, but feelings are necessarily conscious. Hence, Freud did not identify emotions with feelings. He identified them instead with states whose expression included feelings or that had the potential for such expression, a potential that was blocked from being realized by repression. Be this as it may, ideas of unconscious love, fear, hate, anger, and the unconscious sense of guilt are fixed points in Freud’s theory, and because of the theory’s great influence on twentieth-​century thinking about human psychology and culture, reference to them has become commonplace in our everyday thought and talk about the human mind. Freud’s theory, therefore, whatever Freud’s express opinion of the possibility of unconscious emotions was, offers a concept of them, a concept that is a major alternative to James’s. It remains then to define this concept. To begin with, Freud took emotions to be states of mind we are conscious of through the feelings that manifest them. As such, they are distinct from those feelings, and they may exist and operate in us even when we are unconscious of them. In addition, Freud sought, in the operations of unconscious emotions, explanations of a great range of phenomena of human life including odd behavior like parapraxes, unconventional sexual conduct such as fetishism and bestiality, inappropriate feeling or lack thereof, excessive fear, for example, or flat affect, and somatic illnesses that had no obvious organic cause. Such explanations were among the most distinctive explanations of psychoanalytic theory. They required, not only identification of an emotion that was either not manifest in what, if anything, the subject was feeling or, though manifest in the subject’s feeling, concerned with something other than the ostensible object of those feelings, but also identification of the person, thing, event, or state of affairs towards which the emotion was actually directed. The latter requirement indicates Freud’s special concern with the true objects of people’s emotions. While these were often people, things, events, etc. that produced the emotion, they were sometimes merely the products of the subject’s fantasies. As Freud put it, when he came to explain certain hysterical symptoms in his patients as the products of repressed pseudo-​memories of sexual trauma in early childhood, whether incidents of sexual abuse really occurred or were fantasized made no difference from the patient’s viewpoint. Because the fantasized events seemed to the patient as real as if they had actually occurred, the repressed pseudo-​memories and the terror they instilled in the latter case had the same effect on him as genuine

21

Concepts of Emotions in Modern Philosophy and Psychology  21

memories of real abuse. In effect, then, Freud, in giving these explanations, fixed on emotions as intentional states of mind. He saw, with each, how it was related to its object so as to give meaning or import to the feelings, behavior, and conditions that manifested it. For Freud, such intentionality was the core element of the concept of emotions he invented in introducing unconscious emotions into the theory of the mind. At the same time, Freud did not expressly consider what an emotion’s intentionality consisted in or what constituted its distinctive relation to its object. To introduce a concept of emotion that was a genuine alternative to the one James defined, it was sufficient for Freud to have made the intentionality of emotions and the import it gave to them the focus of his conception. Accordingly, on the concept he introduced, emotions are not, as distinct from the one James defined, epiphenomena. Rather they can be causes, and what they cause are actions. Secondly, being sometimes unconscious states, emotions are not, contrary to James, identical with feelings. Rather they are expressed by feelings. Thirdly, the feelings that express emotions are meaningful phenomena, whereas for James they are merely indices of bodily processes. Feelings of grief, shame, fear, or the like, on Freud’s view, have the same meaning or import as the emotions they express, and indeed that import is transmitted from the emotion to them. Grief, for example, has the import of loss, and in feeling grief one feels the loss over which one is grieving. Shame has the import of one’s appearing unworthy before others whose esteem one values, and in feeling shame one feels unworthy of them. On the concept James defined, by contrast, the feelings an emotion consists in, being nothing over and above the feelings of bodily changes, are not meaningful in this way. Their being feelings of grief, say, or shame is, rather, mere happenstance, a consequence of the concept that ultimately yielded telling empirical criticism.3

III The concept of emotion that Freud introduced thus stood in need of an account of the meaning that he located in an emotion’s intentionality. In the last third of the twentieth century, many philosophers and psychologists, often without recognizing their debt to Freud, converged on such an account.4 While the theories of emotion they advanced differed from each other in specifics, they 3.  Experimental work by Stanley Schacter and Jerome Singer (1962) is generally thought to have established that the bodily processes a person who is feeling some emotion undergoes are not by themselves sufficient to determine the type of emotion the person is feeling. 4.  Some, however, clearly recognized this debt. See Solomon 1976, pp. 180–​182.

2

22  From Psychology to Morality

agreed in taking some form of evaluative judgment as an essential element of an emotion and in citing the thought content of such a judgment to explain the intentionality of emotions. On these theories, for instance, judgments about good fortune are essential to joy, and judgments about undeserved misfortune are essential to pity. Accordingly, the import of the feelings expressing joy or pity derives from these judgments. Feelings of joy, therefore, concern something good that has happened to one, and feelings of pity concern undeserved misfortune that has befallen someone. The feelings that express emotions are, therefore, importantly different from feelings and sensations that merely register some physiological disturbance. The latter, being symptoms of bodily changes, do not concern anything. They have no import. When after sudden exertion, say, one is short of breath and feels weak or wobbly, the feeling is symptomatic of respiratory difficulty and nothing more. If, by contrast, upon a sudden attack of panic one is short of breath and feeling wobbly, the feeling is not just a symptom of respiratory difficulty. It concerns, rather, something of which one is intensely afraid and what determines the object of one’s fear is the judgment that one is in danger. The object of one’s fear, in other words, is what one thinks is endangering one. And the same pattern of analysis, so these theories hold, applies to emotions generally. Every emotion, that is, is necessarily about something, however vague or indeterminate, and what it is about is determined by the evaluative judgment it contains. Consequently, such a judgment is an essential element of the emotion. To take such a judgment as an essential element of an emotion, and not merely a common concomitant, is therefore to understand an emotion as essentially a cognitive state. This understanding, which now prevails among philosophers and psychologists who study emotions, represents a recovery of ideas about emotions that were prominent in the thought of the ancients. Indeed, Aristotle is often cited in the expositions of contemporary cognitivist theories of emotion as a source of their central thesis. And the boldest of them go so far as to endorse the ancient Stoic theory on which emotions are taken to be identical to evaluative judgments of a certain kind (Solomon, 1976; Nussbaum, 2001; see also Nussbaum, 1994, chs. 9–​13). Most of these theories, however, are less bold and give accounts of emotions that include, as essential elements, other things besides evaluative thought and articulate some complex relation among these elements. Agitation of the mind, autonomic behavior, and impulses to action are the usual additions. But even in these theories evaluative judgment is the primary element in the mix, for it is the element by which each emotion is principally identified. It is the element the theories principally use to define different types of emotion (Deigh, 1994, pp. 835–​842).

23

Concepts of Emotions in Modern Philosophy and Psychology  23

A theme that is common to these theories is that emotions, like other cognitive states, belong to intelligent thought and action. They are in this respect on a par with beliefs and other judgments, decisions, and resolutions. They are, that is, states that one can regard as having propositional content, which their subjects accept or affirm. Accordingly, one can treat them as warranted or unwarranted, justified or unjustified, by the circumstances in which they occur or the beliefs on which they are based. Thus fear would be warranted if its object evidently posed some threat to one and unwarranted if it evidently posed no threat. Likewise, anger would be justified if it were a response to a genuinely demeaning insult and unjustified if based on one’s mistaking an innocent remark for such an insult. In either case, the emotion is warranted or unwarranted, justified or unjustified, because the evaluative judgment in which the emotion consists, either in whole or in part, is warranted or unwarranted, justified or unjustified. In general, then, on these theories, an evaluative judgment is an essential component of an emotion. It is, moreover, the component by which one type of emotion differs from another. If you want to understand the difference between contempt and anger, say, then according to these theories the difference lies in the type of evaluative judgment that is essential to each emotion. When you have contempt for someone because he has behaved badly, you judge the person to be low or unworthy of your esteem in view of that behavior. When you are angry at someone because he has behaved badly, you judge that the person has injured or insulted you or someone close to you by so behaving. I will call this model of cognitivist theory “the standard model.” The standard model is subject to two powerful and related objections. First, sometimes one can experience an emotion towards something that one knows lacks the properties it must have for the emotion to be warranted. Consider, for example, the fear people typically experience when looking down from a precipice. They may know that they are perfectly safe and in no danger of falling, yet fear falling nevertheless. Similarly, common phobias such as snake and spider phobias supply examples of fear of an object the subject knows is harmless. Again, people sometimes feel disgust at foods they know are nutritious, benign, and perhaps even tasty.5 Defenders of cognitivist theories that fit the standard model must, to account for these examples, describe the subjects of these emotions as making contradictory judgments or holding contradictory beliefs. Yet familiarity with such experiences tells us that when, for example, one feels fear on looking down from a precipice, knowing that one is perfectly safe, one doesn’t judge or believe that one is in 5.  See Rozin and Fallon 1987.

42

24  From Psychology to Morality

any danger of falling. To react with fear, it is sufficient that one look down and see the steep drop below. The reaction is automatic, as it were. It does not depend on one’s making a judgment or forming a belief about one’s circumstances’ being dangerous. Cognitivist theories that fit the standard model cannot, then, satisfactorily account for such fear. The second objection to the standard model is that it cannot satisfactorily account for the emotions of nonhuman animals and human infants—​ beasts and babies. Like the first objection, this second objection identifies a class of emotions that resist being understood as consisting either wholly or in part in evaluative judgments of the kind that the standard model identifies as the essential cognitive element in emotions. The reason for this resistance is plain. Such judgments, like beliefs, are states of mind that imply acceptance or affirmation of propositions. Consequently, to have emotions requires being capable of grasping and affirming propositions. That is to say, one must have acquired a language. Since beasts never acquire a language and babies have yet to acquire one, the standard model cannot account for their emotions. The result is not entirely surprising. The most important antecedent of the standard model, the cognitivist theory of emotions that the ancient Greek and Roman Stoics advanced, denied that beasts and babies ever experienced emotions. Because beasts and babies lacked language, the Stoics argued, they were incapable of making evaluative judgments of the kind with which the Stoics identified emotions. But their view, like Descartes’ cognate view that only human beings have minds, has long since been reduced to a historical curiosity. Consequently, any theory of emotions must acknowledge that beasts and babies are capable of having emotions, and it must account for these as well as the emotions of mature human beings. Cognitivist theories that fit the standard model fail in this respect.

IV In view of these objections, recent defenders of cognitivist theories have dropped the standard model in favor of a broader account of the evaluative cognition that is essential to emotions (see Nussbaum, 2001 and Roberts, 2003). Such cognition, they argue, need not be an evaluative judgment of the kind that implies grasping and affirming a proposition. It may, instead, be a perception. Indeed, some defenders of cognitivist theories now argue that emotions are primarily forms of perception. The theories they defend fit what I’ll call “the perceptual model.” Since perceiving something does not imply that you make any judgment or form any belief about it, the first

25

Concepts of Emotions in Modern Philosophy and Psychology  25

objection does not threaten cognitivist theories that fit this model. When you look down from a precipice, knowing you are in no danger of falling, the fear you experience need not consist in or involve a belief or judgment that you are in danger of falling. You perceive danger in falling, but you do not believe or judge that you are in danger of falling. Hence, in feeling fear of falling you do not make contradictory judgments or hold contradictory beliefs. And the same goes for other examples that generate the first objection to the standard model. Does the perceptual model also escape the second objection? Can it account for the emotions of beasts and babies? Defenders of this model maintain that it can. Because perception is sometimes a more primitive form of cognition than the kind of evaluative judgment that the standard model identifies as essential to emotions, they argue, it is possible on the perceptual model for beasts and babies to have evaluative cognitions of the kind that is essential to emotions. Specifically, because not all perception requires propositional thought, attributing perceptions to beasts and babies does not presuppose that either have any linguistic capabilities. The fear a dog feels, say, when a man threatens him with a stick consists at least in part in the dog’s seeing the threatening motions his would-​be attacker makes with the stick as dangerous. The dog’s perception, so defenders of the perceptual model would argue, is like the perception we have when we look at a straight stick partially submerged in a deep pool of water. We see it as bent. Our so seeing the stick does not require any propositional thought on our part, and by the same token the dog’s seeing the man with the stick as dangerous does not require any propositional thought on the dog’s part. This argument for recognizing that the perceptual model is beyond the reach of the second objection would succeed if the point of that objection, when pressed against the standard model, were a point about the mode of cognition that on this standard model characterizes the evaluative judgments the model represents as essential to emotions. It would succeed, that is, if the reason the standard model fails to account for the emotions of beasts and babies is that the model takes such evaluative judgments to be states or acts of mind in which some proposition is affirmed or denied. Some perceptions, after all, are cognitive states in which the mind neither affirms nor denies a proposition. They are of a different mode of cognition. Hence, the second objection would have no force against the perceptual model since this alternative mode may characterize some of the perceptions the model represents as essential to emotions. But one could be making a different point in advancing the second objection. One could be making a point about the cognitive content of the evaluative judgments that the standard model represents

62

26  From Psychology to Morality

as essential to emotions. In that case, the argument has not succeeded in showing that the perceptual model is beyond the reach of this objection. For the cognitive content of a perception of someone, the man with the stick, say, as dangerous is the same as the cognitive content of a judgment that this man is dangerous. In either case, the concept of danger is deployed in a cognitive representation of the situation, and to have and deploy a concept of danger is to be capable of propositional thought. Hence, the perceptual model too runs afoul of the second objection. Defenders of the perceptual model will of course balk at this result. They will argue, in response, either that one does not need to have a concept of danger to see something as dangerous or that it is possible to have concepts even if one has no language or has yet to acquire one. Neither response, however, rescues the model from defeat by the second objection. Let us consider the second response first. By a concept, I understand, following the standard view in Anglo-​American philosophy, what words, particularly, substantival words, express when they are used in sentences that express propositions. They are, that is, what such words mean when they occur in meaningful sentences.6 Thus, the meaning of a word, its sense, as we sometimes say, is the concept it expresses. To have concepts independently of one’s having a language implies either of two possibilities. On the one hand, it may imply that one’s concepts are innate and that part of what takes place, when one learns language, is one’s matching words to them and constructing sentences that match the combinations of them one has created. On the other, it may imply that one first acquires concepts through sensory and affective experiences and later learns how they are encoded in language. The first possibility captures a view of language acquisition according to which it follows and results from the emergence and development of rational thought in children. Since animals other than human beings are not rational and never acquire language, the first possibility is not available to defenders of the perceptual model. The second possibility treats concepts as ideas in the mind that derive from one’s sensory and affective experience. On this possibility, the acquisition of such ideas is not exclusive to human beings. At the same time, this possibility entails a semantics for natural language according to which the meanings of our words are these ideas in our minds. While such a semantics was once widely held by classical empiricist philosophers and the positivist

6.  On this understanding of what concepts are, having the concept of x entails more than having the power to discriminate between x’s and other things. The point is as old as Plato’s early dialogues, but it sometimes needs to be brought forward because of the tendency among writers on topics in psychology to use the words ‘concept’ and ‘idea’ interchangeably.

2 7

Concepts of Emotions in Modern Philosophy and Psychology  27

philosophers whose theories of meaning dominated Anglo-​American philosophy in the first half of the twentieth century, powerful objections to these views that critics of positivism developed in the latter half of the twentieth century have shown this semantics to be untenable.7 Consequently, the second possibility is also a nonstarter. Defenders of the perceptual model must avoid attributing concepts to beasts and babies if they are to save the model from defeat by the second objection. We are left, then, with the alternative response, that one does not need to have a concept of danger to see something as dangerous. With regard to this response, the first thing to ask is what seeing something as dangerous amounts to. Specifically, is it like seeing a straight stick as bent when it is partially submerged in a pool of water? Or is it more like seeing an extended hand as an offer of friendship when you first make someone’s acquaintance? In the latter case, the perception entails an interpretation that consists in applying a concept, the concept of an offer of friendship, to a gesture. So if seeing something as dangerous is like this latter case, then the alternative response fails to save the perceptual model. In the former case, the perception, by contrast, arguably need not involve any application of a concept. In describing our perception as one of seeing a straight stick as bent, we describe how the stick appears to us in circumstances in which we know it would appear differently if we looked at it through a uniform medium. The expression ‘seeing as’ is apt in this case, not because we are interpreting how the object appears to us, but because we are perceiving the object through one of our sense modalities, vision, and the circumstances are ones in which we know that the way it appears to us visually is not the only way in which it can visually appear to someone. We are thus remarking a sensory property an object appears to have but may not in fact have. It is a property of a kind of which we are directly aware through one or more of our sense modalities. The question, then, is whether being dangerous is similarly a sensory prop­ erty of objects, whether it is a property of a kind of which we are directly aware through one of our sense modalities. If it were, then we would teach children what the word ‘dangerous’ meant in the same way that we teach them the meaning of words for sensory properties like ‘yellow’ and ‘sweet’. When one teaches a child words for such properties one assumes prior acquaintance with the properties. That is, one assumes that the child’s sense modalities have developed, work well, and been exercised. When one teaches a child color words, for instance, one finds objects that display the colors vividly, presents those objects to the child, 7.  See Putnam 1975.

82

28  From Psychology to Morality

and has it apply the different words to them. Other objects that are similar in color are also presented to the child, and it is taught to group them according to color. On the assumption, then, that the child can recognize the objects in each group as similar to each other, it is taught to apply the word to what each member in the group has in common with every other. Plainly, a different method is used to teach children what the word ‘dangerous’ means. Children are not presented with a collection of objects such as pencils with sharp points, rubber balls, plastic bags, matches, and billfolds and invited to group them according as they are dangerous or benign. One does not assume prior acquaintance with the property. That is, one does not assume that children can recognize, prior to their learning the meaning of the word ‘dangerous’, similarity among objects all of which are dangerous. Otherwise one would naturally teach them what the word means by teaching them to apply the word to what they already recognize as the property the objects have in common. Rather children are taught the meaning of the word by being told which things are dangerous and which are not. Teaching them the meaning and teaching them how to recognize the property are therefore one and the same. One assumes, that is, that the child is ignorant of which things have the property and which do not, that it cannot, without instruction, recognize a thing’s being or not being dangerous. Whether it’s strangers or matches or busy streets, a child is taught about the danger each poses, and the teaching typically includes some explanation of the harm that each can cause. Likewise, one teaches a child not to be afraid of things that initially frighten it when they are not dangerous. In so doing, one is not correcting a child’s misperception of danger. Rather, one is teaching the child that not everything scary is dangerous, and this teaching too may include a demonstration that these scary things do not cause harm. If being dangerous were a sensory property, such teaching would be unnecessary for getting children to recognize danger. A method of teaching like that of teaching children the meaning of color words would be sufficient. What then should we say about the dog that is afraid of the man threatening him with a stick? It is tempting, of course, to say that the dog sees the man as dangerous. After all, if a man threatened you with a stick, you would certainly see him as dangerous. But we need not suppose that the dog can perceive the property of being dangerous to understand the dog’s fear. We need only suppose that the dog sees the man as an imminent source of pain in that the dog anticipates pain from this man owing, say, to the man’s threatening behavior and the dog’s having previously experienced being violently struck with a stick. The previous experience and the perception of the man are jointly sufficient to explain the dog’s fear. For having once been

92

Concepts of Emotions in Modern Philosophy and Psychology  29

traumatized by pain, the dog will anticipate again feeling pain when he is in a similar situation, and consequently the occurrence of a similar situation is bound to trigger fear of the objects in that situation from which he anticipates pain. The mechanism is familiar. Think, for example, of the apprehension you experience when a nurse is about to draw blood or administer an anesthetic and you see in her hand a syringe with a long needle. We need not suppose any evaluative cognition to understand the inward shudder you feel as the nurse approaches. And the same is true of the dog. Cognitivist theories of the emotions on which evaluative cognition is essential to emotions have been the predominant theories for the past forty or so years. The attraction of these theories is the account of the intentionality of emotions they give. The account promises both a way of understanding the relation of an emotion to its object and the orientation towards the object that the emotion affords its subject. Yet they fall short of giving a satisfactory account of the intentionality of the emotions to which beasts and babies are liable. And since these are the first emotions that men and women who become capable of experiencing emotions whose intentionality does consist in evaluative cognition experience, it would seem that the soundest theory of emotions that proceeds from a concept of emotion on which intentionality is its core element should begin with an account of the intentionality of these primitive emotions and build on it an account of how they are transformed through moral development and education into emotions whose intentionality consists in evaluative cognitions (see Deigh, 2008).

V Recently, under the influence of work on emotions in neuroscience, there has been a return to conceiving of emotions as essentially feelings. Indeed, those in the forefront of this development have invoked James’s account as the source of their program.8 Accordingly, let us call them neo-​Jamesians. The neo-​Jamesians, like James, take the feelings in which emotions consist to be the feelings of the bodily changes that occur during an episode of emotion. Unlike James, however, the neo-​Jamesians hold that these feelings do not merely register those bodily changes. Rather they are also meaningful. The neo-​Jamesians, in other words, accept the fundamental idea on which the concept of an emotion that Freud introduced into psychology is constructed, the idea that emotions are intentional states. One can therefore 8.  See Damasio 1995 and Prinz 2004.

03

30  From Psychology to Morality

characterize their account of emotions as a hybrid produced by grafting this idea onto James’s account of emotions. This hybrid’s distinctive thesis, then, is that the feelings of bodily changes in which an emotion consists are intentional phenomena. Plainly, if the neo-​Jamesians can sustain this thesis, they will have made a significant advance on cognitivist theories. For they will have given a uniform account of the intentionality of emotions, an account that fits equally well the emotions of beasts and babies, on the one hand, and the emotions of men and women, on the other. They will have done so, that is, given that, on this hybrid, emotions, whether of beasts and babies or of men and women, consist in feelings of bodily changes and given that their intentionality has the same character in either case.9 They will have therefore given an account that eludes the problem that, as we saw, puts the cogency of cognitivist theories into doubt. At the same time, there is reason to wonder about the coherence of a hybrid that results from grafting the fundamental idea of the concept of emotion Freud introduced onto James’s account. To begin with, Freud introduced his concept out of an interest in citing unconscious emotions to explain a broad range of human behavior, feelings, and bodily conditions. His point was that an emotion need not be a conscious state to have influence in human lives, and, by identifying intentionality as the core element in our understanding of the nature of emotions, he secured that point. The point is obviously lost, however, when one takes the feelings of bodily changes to be the vehicle of an emotion’s intentionality. Indeed, neo-​Jamesians would appear to have to reject the possibility of unconscious emotions. Feelings of bodily changes cannot be unconscious. So one may wonder what the point is of their taking intentionality as an essential feature of emotions. Is their doing so just a case of amalgamating two otherwise independent features into a single conception? In addition, Freud’s use of a concept of emotion on which intentionality is the core element served to maintain the commonsense view of emotions as springs of action. To grasp an emotion’s intentionality is to understand how the emotion orients its subject towards the world or towards certain persons or things in it and what sorts of actions to expect in response to that orientation. James’s account is meant to oppose common sense on this point, for, on his account, the actions common sense takes as springing directly from an emotion are among the bodily movements the feeling of which the 9.  N.B. We can assume that their intentionality is the same in either case since it depends on the nature of these feelings and not on the nature of any cognitions and since in the lives of beasts and babies the feelings of bodily changes do not differ in kind from those in the lives of men and women.

31

Concepts of Emotions in Modern Philosophy and Psychology  31

emotion consists in and therefore cannot be explained by the emotion. So again one may wonder what the point is of the neo-​Jamesians’ taking intentionality to be an essential feature of an emotion if they mean to be faithful to James’s account and, particularly, to its treating emotions as epiphenomena. Taking emotions, instead, as states of motivation that are distinct from and can give rise to feelings would appear, in view, say, of Hume’s doctrine of the calm passions, to make better sense if one did not conceive of emotions as epiphenomena. Be this as it may, we need to examine the neo-​ Jamesian program as expounded by its adherents to see whether it makes sense and so progress within the philosophical study of emotions. A  useful starting place is Antonio Damasio’s theory of emotions in his influential book Descartes’ Error. Damasio draws his theory from his research in neuroscience on the brain. Damasio’s research yields a description of the neurological activity that occurs as we confront and respond to new and different situations. His main observation is of neurological activity that signals a need for adjustment or action that will promote one’s survival, correct for deviation in one’s functioning that threatens one’s health, say, or that will restore one to a condition of functioning well. The signals, in other words, enable us to monitor what is happening in and to our bodies and to make adjustments and take actions as necessary. We are the recipients of these signals, and we use them to regulate our welfare. In setting out his views on the nature of emotions, Damasio expressly claims James’s account as their principal antecedent. The feelings of bodily changes that James took emotions to consist in are, according to Damasio, a means by which we monitor what is happening in and to our bodies during an episode of emotion. They signal the need for us to make adjustments and to take action. Accordingly, Damasio characterizes them as perceptions of the body. In doing so, he renders them intentional phenomena. He therefore offers a way of understanding the intentionality of emotions on James’s account. And because of his thesis that these feelings are perceptions of the body, it is natural to take him as an adherent to the neo-​Jamesian program. Damasio, however, despite regarding himself as following James in giving his account of the emotions, does not. Specifically, he does not deploy James’s concept of emotion in his theory. Rather he deploys a concept on which emotions are distinct from the feelings to which they give rise. He deploys a different concept because he misreads James as identifying emotions with a bodily process, as “stripping emotion down to a process that involved the body” (1995, p.  129). Thus, having identified emotions with a bodily process, he takes the feelings of the corresponding bodily changes as the internal

23

32  From Psychology to Morality

perceptions of this bodily process. “If an emotion is a collection of changes in body state connected to particular mental images that have activated a specific brain system, the essence of feeling an emotion is the experience of such changes in juxtaposition to the mental images that initiated the cycle” (1995, p. 145). On his theory, then, feelings are distinct from emotions. His proclaimed descent, as a theorist of emotion, from James notwithstanding, Damasio bases his theory on a concept of emotion that is much closer to Freud’s than to James’s. And while the theory includes a thesis attributing intentionality to the feelings of emotion, the intentionality of these feelings, as Damasio understands it, does not correspond to the intentionality of the emotion that gives rise to them. This last point should be obvious. The intentionality of an emotion consists in the emotion’s being directed at or towards something. Typically, the object at which an emotion is directed is an object in the world, either real or imagined. When you fear an assailant who is threatening you with a knife, for instance, the object of your fear is the assailant. In experiencing such fear, your heart may race and your breathing may be labored. As a result, the experience may include your feeling your heart racing and your being short of breath. But these feelings, though part of your experience of fear, don’t change what the object of your fear is. It is still the assailant, in this case, and not your racing heart or your shortness of breath. If it were either of the latter, then we would expect the fear to move you to look for some heart medicine or an inhaler rather than to move you to flee or hide from the assailant. That it does not, that it moves you to flee or hide from the assailant, is sufficient to show that, whether or not we regard the feelings to which the emotion gives rise as perceptions of the body, they do not exhibit the intentionality of the emotion. Philosophers who are firmly in the neo-​Jamesian camp have recognized that Damasio’s thesis that the feelings of an emotion are internal perceptions of bodily changes fails on its own to yield a satisfactory account of the intentionality of emotions. Remaining faithful to James in taking emotions to consist in feelings of bodily changes, they have offered different accounts of the intentionality of these feelings. Here is an example due to Jesse Prinz (2004, pp.  52–​78; see also Robinson, 2006). The feelings of bodily changes that emotions, on James’s account, consist in, besides being perceptions of those changes, are also representations of properties related to their subjects’ vital interests. Feelings of grief, for instance, represent loss; feelings of shame represent unworthiness of others’ esteem; feelings of pity represent undeserved suffering; and so forth. That the feelings emotions consist in represent these properties is what explains their intentionality. The explanation requires some spelling out, however, to be grasped.

3

Concepts of Emotions in Modern Philosophy and Psychology  33

Accordingly, emotions are said to be mental representations of these properties in a special sense of the term ‘representation’. On this sense, a mental state S represents something O (an object, a property, a state of affairs, etc.) just in case two conditions hold. First, O must be the sort of thing that reliably causes mental states of the type to which S belongs. That is, in the jargon of the theory of perception in psychology, S must carry information about O.  Examples of one thing’s carrying information about another are drooping leaves on a plant, which because they are a sign of dehydration, carry information about the plant’s lack of water, and wet pavement on city streets, which because it is a sign of rain, carry information about its having rained. Consequently, using this piece of jargon, we can restate the first condition as follows: S must carry information about O. Second, it must be possible for S to be misleading. Specifically, it must be possible for S to carry information about O that is erroneous because the cause of S is not O or anything similar to O.  For this second condition to be met it is necessary that S be a mental state of a type instances of which regularly further some vital interest of their subject in virtue of their being caused by things like O, for in that case we can understand S as serving a function in the life of its subject and of its subject’s being misled by S if it is not caused by O or anything like O. Thus, sour smells carry information about their source. In the case of eggs, for instance, the information the smell carries would be that the eggs are rotten. Yet because sulfur has the same smell, it would, in circumstances in which one could easily attribute it to rotten eggs, carry erroneous information. It is clear that emotions, on James’s account of them, satisfy these two conditions. First, the feelings an emotion consists in carry information about the object whose perception excites the bodily changes that, on James’s account, produce those feelings. The feelings carry information about this object since objects of this type reliably cause such feelings when they are perceived by their subject. For example, because acting wrongly reliably causes feelings of guilt in someone with a conscience, the feelings carry information about his action. Second, it is possible for the feelings an emotion consists in to carry erroneous information, for they may have other causes than those on whose reliability their carrying information depends. A person brought up in a strict household, for example, may feel guilt over leaving food on his dinner plate because his parents, out of misplaced concern for waste, enforced a rule about eating every last morsel of food served one at dinner. In this case his guilt feelings are misleading because the information they carry is that his action is wrong, and there is nothing wrong with leaving some food on one’s plate if one has eaten enough to satisfy one’s

43

34  From Psychology to Morality

hunger. Feelings of guilt, then, as illustrated by this example are mental representations. They represent wrongful action and sometimes misrepresent such action. And more generally, emotions are mental representations. They represent properties related to their subjects’ vital interests that are detectable through the feelings the emotions consist in. In this way, we can identify emotions with feelings of bodily changes and at the same time understand their intentionality. Indeed, the feelings count as noncognitive evaluations. Or so the neo-​Jamesians argue. Yet this account of the intentionality of emotions, like Damasio’s, misidentifies what it aims to explain. Emotions, to repeat, are intentional states in virtue of being directed at or towards something in the world, real or imagined. Take again the example of fear of an assailant who is threatening you with a knife. The assailant is the object of your fear, and the intentionality of the emotion consists in its orienting you towards him. On the alternative neo-​Jamesian account we are considering, by contrast, the fear consists in feelings that represent danger. Nothing in this relation of representation, however, implies that these feelings are directed at or towards any object in the world, real or imagined. Nothing in the relation implies that they orient you towards the threat you are facing. In particular, even though the assailant is the object about which the feelings carry information in the sense of ‘information’ borrowed from the theory of perception, nothing in the relation implies that the feelings are directed towards him. For their carrying information about him just means that he, being dangerous, is a reliable cause of such feelings, and his being a reliable cause of the feelings implies nothing about the feelings being directed at or towards him. Obviously nothing changes if the information they carry about the object is erroneous, if for instance you mistook a deliveryman with a cell phone for an assailant wielding a knife. We have no more reason to regard the feelings in this case to be directed at or towards their cause than we had in the case of correct information. The feelings, in either case, represent danger exactly as the sound made by an alarm that is part of a home security system represents a home invasion. Just as the alarm’s sound is not directed at or towards whoever sets off the alarm, so the feelings are not directed at or towards the assailant or anyone, like the deliveryman, mistaken for one. The upshot is that the feelings do not explain whatever action you take in response to the threat you face. Since they merely represent danger and are not directed at the threat, they do not guide you in your dealings with it. In particular, they are not an evaluation you make in reaction to it. To identify them as such is to mischaracterize them. They no more count as an evaluation of something than the sound of a home security alarm counts as an

35

Concepts of Emotions in Modern Philosophy and Psychology  35

evaluation of something. They are the effects of bodily changes and have no direct role in explaining further actions that follow upon those changes. Like James’s account, then, emotions on this neo-​Jamesian alternative are epiphenomenal. The alternative’s failure to explain the intentionality of emotions simply draws attention to this fact. It draws attention, that is, to the fact that what it does explain is entirely consistent with James view and adds nothing significant to it.

VI A rather different proposal for using feelings to understand the intentionality of emotions is due to Peter Goldie (2000). Like the neo-​Jamesians, Goldie takes feelings to be an essential element of the experience of emotion. At the same time, he does not identify these feelings with feelings of bodily changes. Hence, his proposal lies outside of James’s theory. Goldie’s thesis is that an experience of emotion is an intentional state that consists in part in feelings towards the emotion’s object. These feelings are themselves intentional, and for that reason they are not identical with the feelings produced by the bodily changes that occur during an experience of emotion. Further, being intentional, these feelings are integral to the intentionality of the emotion that they partly constitute. One cannot, Goldie holds, treat them in abstraction from the latter as if they were a separable part. Attempts to separate the two by, say, distinguishing between the phenomenal or subjective character of the emotion’s object and its cognitive or objective character are misguided. While it is easy to suppose first that one can understand the intentionality of an emotion by specifying the emotion’s object and those features of it by virtue of which the emotion is an appropriate response and to suppose second that the features are conceivable independently of any feelings, Goldie denies this supposition. He holds, instead, that the intentionality includes as well how these features, as features of the object, appear to the subject. Such an appearance, Goldie maintains, is colored by the feelings towards the object that are an essential element of the experience. Clearly, Goldie’s proposal represents a different challenge from that of the neo-​Jamesian’s to the concept of emotion that Freud introduced. On Freud’s concept, the emotion’s intentionality is distinct from any of the emotion’s associated feelings. It thus makes it possible to include unconscious emotions in a theory of the mind, for it allows one to deny that emotions are identical with or consist partly in feelings. But if the intentionality of emotions is itself

63

36  From Psychology to Morality

determined by feelings the subject has towards the object, then Freud’s concept must be abandoned. On Goldie’s proposal, Freud’s concept gives at best an incomplete account of an emotion’s intentionality. The challenge, it should be evident, is not that Freud’s concept excludes the possibility of there being feelings towards an object, though of course to maintain the concept one must understand such feelings as manifestations of an emotion rather than as constituents of it. The concept no more excludes the possibility of such feelings than it excludes the possibility of behavior towards an object. And nothing in Freud’s concept brings the latter into doubt. Nothing in it, that is, raises any doubts about our common description of behavior that expresses an emotion as behavior towards the emotion’s object. Thus, when someone laughs at a joke, his laughter expresses his amusement at the joke. Similarly, when he shakes his fist at some antagonist, his behavior expresses his anger towards that antagonist. So too, when feelings express an emotion, they may be felt towards the object of the emotion. They too may be intentional on Freud’s concept. The challenge is to the implicit separation, on Freud’s concept, of the intentionality of the emotion from the feelings that go into an experience of emotion. Yet the thesis that the features of an emotion’s object by virtue of which one can understand the emotion as an appropriate response to the object must include their phenomenal character appears to be too strong even if one sets aside the possibility of unconscious emotions. To be sure, some emotions are such that one cannot understand them as appropriate responses to their object without including among the features of the objects in virtue of which they are appropriate responses features that are phenomenal or subjective. Horror, for example, is an emotion the object of which is typically something grisly or gruesome, and there is no understanding of what is grisly or gruesome apart from how grisly or gruesome things make one feel. Similarly, disgust at something that is foul or offensive to the senses, disgust at scum one suddenly finds oneself swimming in, is an emotion whose object would not be well understood if how scum feels when one comes into contact with it were omitted from the features that make it an appropriate object of disgust. But the intentionality of emotions whose objects are more abstract, disappointment, say, at not being hired for a job one had applied for and was keen on getting or embarrassment at confusing one’s guest with his brother seems fully understandable without bringing in any phenomenal features. Indeed, it is unclear how either event could be better understood as the object of disappointment or embarrassment in these cases. What this observation suggests, then, is that Goldie’s proposal does not so much challenge Freud’s concept as remind us that some and perhaps many emotions, while

37

Concepts of Emotions in Modern Philosophy and Psychology  37

distinct from the feelings that express them, cannot be understood apart from those feelings.

Bibliogra p hy Damasio, A. (1995). Descartes’ Error. New York: Avon Books. Deigh, J. (1994). “Cognitivism in the Theory of Emotions,” Ethics 104: 824–​854; reprinted in Emotions, Values, and the Law. ———​(2001). “Emotions:  The Legacy of James and Freud,” International Journal of Psychoanalysis 82: 1247–​1256; reprinted in Emotions, Values, and the Law. ———​ (2008). Emotions, Values, and the Law. New York: Oxford University Press. Freud, S. (1915). The unconscious. The Standard Edition of the Complete Psychological Works of Sigmund Freud (S.E.), vol. 14, ed. and tr. by J. Strachey. London: Hogarth Press, 1981. ———​ (1923). The Ego and the Id. S. E. 19. ———​(1924). The economic problem of masochism. S.E. 19. Goldie, P. (2000). The Emotions:  A Philosophical Exploration. Oxford:  Oxford University Press. Hume, D. (1978). A Treatise of Human Nature, ed. L. A. Selby-​Bigge. Oxford: Clarendon. James, W. (1950). The Principles of Psychology, vols. 1 & 2. New York: Dover. Lear, J. (1990). Love and its Place in Nature: A Philosophical Interpretation of Freudian Psychoanalysis. New York: Farrar, Strauss and Giroux. Locke, J. (1975). An Essay Concerning Human Understanding, ed. P. H. Nidditch. Oxford: Clarendon. Nussbaum, M. (1994). The Therapy of Desire: Theory and Practice in Hellenistic Ethics Princeton, NJ: Princeton University Press. ———​ (2001). Upheavals of Thought: The Intelligence of Emotions. Cambridge: Cambridge University Press. Prinz, J. (2004). Gut Reactions:  A Perceptual Theory of Emotion. Oxford:  Oxford University Press. Putnam, H. (1975). “The Meaning of Meaning” in Mind, Language and Reality: Philosophical Papers, vol. 2. Cambridge: Cambridge University Press. Roberts, R. (2003). Emotions:  An Aid in Moral Psychology. Cambridge:  Cambridge University Press. Robinson, J (2006). Deeper Than Reason: Emotion and Its Role in Literature, Music and Art. Oxford: Oxford University Press. Rozin, P. and Fallon, A. (1987). “A Perspective on Disgust,” Psychological Review 94: 23–​41. Schacter, S. and Singer, J. (1962). “Cognition, Social, and Physiological Determinants of Emotional State,” Psychological Review 69: 379–​399. Solomon, R. (1976). The Passions. New York: Doubleday.

83

2

William James and the Rise of the Scientific Study of Emotion

I When the philosophical study of emotions made the turn to cognitivism forty years ago, philosophers who initiated this turn typically cited William James’s identification of emotions with feelings of bodily changes as a clear example of the inadequacy of the then prevailing conception of emotions. James, they argued, in so identifying emotions, had missed the evaluative judgment that an emotion consists in or includes (Lyons, 1980; Pitcher, 1965; Solomon, 1976).1 By reducing the study to a physiological study he had, in their view, made altogether mysterious an emotion’s evaluative import, the sense of loss that sorrow signifies, the recognition of danger that fear implies, the judgment of inferiority or disadvantage that envy entails. And once the turn they initiated was complete, James’s view fell into neglect. Recently, however, a small but influential number of philosophers have revived interest in the view (Prinz, 2004; Robinson, 2005). A major inspiration for this revival is Antonio Damasio’s book Descartes’ Error (Damasio, 1994). Damasio, on the basis of research into brain-​ damaged people, 1.  For an account of this turn in the philosophical study of emotions, see Deigh (1994).

38

93

William James and the Rise of the Scientific Study of Emotion  39

proposed a theory of emotions and feelings on which physiological processes constituted “the mechanism essential to the understanding of [these phenomena]” (1994, p. 129). This understanding, Damasio maintained, was already present in James’s work. James, according to Damasio, anticipated by one hundred years the discoveries of late twentieth-​century neuroscience. Yet Damasio did not follow James in identifying emotions with feelings of bodily changes. Instead, he distinguished such feelings from emotions and took the latter to be the bodily changes themselves (1994, p. 145).2 Feelings of emotion, he declared, are internal perceptions of the bodily changes in which the emotion consists. One’s brain registers these changes, as it monitors bodily activity, and the feelings, being perceptions of bodily changes, are its registration of them. And like a thermostat that registers changes in temperature and adjusts the system’s output of hot and cold air accordingly, it then sets in motion the appropriate responses. In addition to this alteration in James’s view, Damasio criticized James for eliminating from emotions their evaluative import. They acquire this import, Damasio hypothesized, through learning and socialization. As a result, some emotions, by virtue of being felt, come to serve as signals of prospective benefits and harms that figure into one’s practical decisions. Accordingly, emotions that have undergone such education are, to use the term Damasio coined in formulating this hypothesis, “somatic markers” (1994, p. 173). The bodily changes they consist in serve to mark the opportunities and threats with which one’s situation presents one. What produces these somatic markers, then, is one’s evaluation of that situation. Such evaluations are the cognitive phenomena that James, because he tended to focus narrowly on emotions that result from the excitement of instincts, omitted from his account. Or so Damasio argued. The gist, then, of Damasio’s criticism is that James considered only “primary emotions,” emotions to which instincts alone make one liable, and failed to allow for “secondary emotions,” those the liability to which is the result of learning and socialization (1994, pp. 129–​139). These secondary emotions result directly from one’s evaluation of one’s situation, and in view of the “juxtaposition” of the two, such emotions acquire the cognitive character reflected in their role in deliberation and practical decision-​ making (1994, pp. 145–​146).

2.  “If an emotion is a collection of changes in body state connected to particular mental images that have activated a specific brain system, the essence of feeling an emotion is the experience of such changes in juxtaposition with the mental images that initiated the cycle” (italics removed; see also p. 129, where Damasio appears to misread James as “stripping emotion down to a process that involved the body”).

04

40  From Psychology to Morality

Damasio’s departures from James’s view set him apart from the philosophers who have revived interest in this view and who are commonly referred to as neo-​Jamesians. These philosophers avoid the incoherence Damasio courts in distinguishing emotions from feelings as he did. On the traditional division between mental and bodily phenomena, emotions belong to the former. They are states of mind, however one conceives of the mind, including as identical to the brain or the nervous system. Once one acknowledges this division in classifying the phenomena of sentient life, then saying or implying that anger, sadness, grief, joy, and the like are not mental states is like saying or implying that grass and shrubs are inorganic matter or that roosters, bulls, and stallions are female members of their species. Rather than follow Damasio’s unorthodox understanding of emotions, the neo-​Jamesians retain James’s identification of them with feelings of bodily changes. They also retain James’s exclusion of cognition from emotion. Damasio, though he criticized James for this exclusion, failed in fact to correct the error he attributed to James. Instead, he simply conjoined the relevant cognition to emotion, much as David Hume posited the constant conjunction of ideas and secondary impressions to explain his distinction between passions and sensations (Hume, 1978, pp. 275–​276). The conjunction Damasio posits implies only that cognitions accompany emotions and not that they are essential components of emotions. Thus his positing the conjunction is merely an ad hoc expedient, one that he thinks necessary to explaining the evaluative import of secondary emotions. The neo-​Jamesians, by contrast, deny any need for such an expedient. Their signature thesis is that some or all of the very feelings of bodily changes that an emotion consists in are themselves evaluations. Jesse Prinz, in particular, bases this thesis on Damasio’s account of such feelings as perceptions of bodily changes (Prinz, 2004, pp. 52–​60). Damasio’s account, Prinz believes, holds the key to understanding emotions as having evaluative import. Unlike Damasio, however, he thinks one can understand emotions as having such import without conjoining a cognition to the feelings in which emotions consist. No such added vehicle of evaluation is necessary, he argues. Nor, a fortiori, does he think one needs to identify emotions with cognitions or suppose that cognitions are essential components of emotions to explain this import. He rejects all such cognitivist theories. Contrary to these theories, as well as Damasio’s, he maintains that understanding the feelings in which emotions consist as perceptions of bodily changes is sufficient to explain it. Accordingly, to use the phrase he introduces to express the neo-​Jamesians’ signature thesis, emotions are “embodied appraisals” (2004,

41

William James and the Rise of the Scientific Study of Emotion  41

p. 78).3 The thesis, if sound, would yield a successful answer to the objection from an emotion’s evaluative import that many philosophers regard as a refutation of James’s view. Of course, it is one thing to label feelings of bodily changes “embodied appraisals” and another to establish that such feelings are appraisals.4 One could, after all, label a thermostat’s registrations of changes in room temperature “mechanical appraisals,” but they still would not be appraisals. The difficulty with labeling bodily feelings appraisals is that an appraisal is a type of judgment, and a feeling of bodily change is not (Deigh, 2008, pp. 83–​87). Bodily feeling and judging are distinct types of mental events. Indeed, the objection to James’s view that it misses the evaluative judgment an emotion consists in or includes would have had no force if bodily feelings and appraisals were not distinct. So the neo-​Jamesians’ labeling the former a form of appraisal just looks like verbal gerrymandering that invites confusion. Prinz, to be sure, does not merely assert that feelings of bodily changes are forms of appraisal. To the contrary, he advances a complicated argument for it.5 Here is a brief sketch of his argument. Emotions, he argues, being perceptions of bodily changes, have content analogous to the content of sense perceptions. This content includes the circumstances in the world that produce the changes just as the content of sensory perception includes the objects in the world that produce the sensory experiences. Prinz then characterizes this content as representing these circumstances in the way they produce the changes. Because different emotions, such as fear, joy, and sadness, have distinctive bodily expressions, the bodily changes of which they are perceptions are likewise distinctive. Hence, these changes represent circumstances of the sort that produce such changes, circumstances of danger, good fortune, and loss in the cases at hand. In other words, Prinz concludes, emotions are appraisals by virtue of being perceptions of these bodily changes.

3.  Jenefer Robinson, who does not tie her account as closely to Damasio’s, uses ‘affective appraisals’ and ‘noncognitive appraisals’ to express the thesis (see Robinson, 2005, pp. 28–​57). 4.  I use ‘appraisal’ and ‘evaluation’ interchangeably to mean a judgment of something as good or bad, either absolutely or relative to the subject’s interests. 5.  Robinson’s argument for the thesis, by contrast, is transparently verbal. Finding herself stuck between her subscription to the thesis that emotions entail appraisals and her acceptance of the evidence from experimental psychology and neuroscience that an emotion can occur without an intermediate or concomitant cognition, she invents the category of noncognitive appraisals to resolve her dilemma. But it takes more than an oxymoron to go through the horns of a dilemma.

24

42  From Psychology to Morality

Though Prinz’s argument for the thesis is more than mere wordplay, one should still have no trouble seeing the misnomer in it. The bodily changes, the perceptions of which are, in his view, emotions, represent the character of one’s circumstances in the same way a burglar alarm in a standard home security system represents, when triggered, invasions. The system detects invasions and sets off the alarm. But it does not thereby appraise the invader. Hearing the alarm, one thinks only invasion. There is no object to which one attends and attributes that invasion because at most an alarm calls attention to itself, it does not direct one’s attention to an object; that is, it contains nothing representing someone or some distinct thing to be appraised. The same goes for the bodily changes the perception of which Prinz takes as signaling danger, good fortune, loss, and so forth. On his account, emotions do not themselves include orientation towards an object. The account, for this reason, is defective (Hills, 2008). For typically, when one feels an emotion, one’s attention is directed at an object, and the appraisals that cognitive theories take the emotion to consist in or include are judgments of this object. Fear, for instance, is typically fear of some object—​an oncoming car, a bolt of lightning, an angry boss—​which in each case, according to these theories, one judges to be a threat to one’s life, limb, or well-​being generally. At the root of the failure of the neo-​Jamesians’ program is a misguided strategy for rehabilitating James’s view. To attempt a reconciliation of the view with the thesis that emotions have evaluative import is to miss the point of James’s defining emotions as feelings of bodily changes. It reflects, in other words, a basic misunderstanding of James’s view. The neo-​Jamesians are not alone in this error. The misunderstanding, I  suspect, was present even in the earliest criticisms the view received. In any case, the failure by both James’s critics and his champions to recognize how radical a change in our concept of emotion James was proposing is long-​standing. The change he proposed was such as to preclude emotions from having any direct influence on the will. Hence, whether or not they have evaluative import was of no consequence to the proposal’s viability. The irony, then, of the neo-​ Jamesian program is that its principal aim is to show how his definition of emotion yields an understanding of the phenomena that James had no interest in preserving. To see this, however, requires seeing James’s proposal as part of a larger change in the modern conception of emotions that took place over the course of 150 years. A history of this change is the history of how the study of emotions was transformed from a study integral to moral philosophy into a scientific study. It is in part a history of how the modern scientific study of emotions emerged out of the traditional philosophical study of the passions.

43

William James and the Rise of the Scientific Study of Emotion  43

The history begins with Hume’s (1978) contributions to moral philosophy in A Treatise of Human Nature, particularly his treatment in that work of the passions as the source of morals.6

II Hume divided his Treatise into three books. Passions are the subject of Book II, as indicated by its title, “Of the Passions.” Hume begins his study of the passions by reminding the reader of his taxonomy of mental states. All are perceptions. They divide into two types, impressions and ideas. Hume then subdivides the first type into original impressions and secondary ones. Original impressions are what he earlier, in Book I, defined as impressions of sensation; secondary impressions are, according to a companion definition, impressions of reflection. Original impressions arise from causes outside the mind. Secondary ones, by contrast, arise from prior impressions or ideas. “Of the first kind,” he writes, “are all the impressions of the senses, and all bodily pains and pleasures: Of the second are the passions, and other emotions resembling them” (1978, p. 275). This latter observation echoes an earlier statement, near the beginning of Book I, in which Hume identifies impressions of reflection as “passions, desires, and emotions” (1978, p. 8). Hume was the first philosopher to use the term ‘emotion’ liberally (Dixon, 2003, p. 104).7 It occurs often in Book II. He does not, however, use it as a name for a distinct species of secondary impression. Rather, like other eighteenth-​century writers of psychological treatises, he used several different terms interchangeably, including ‘passions’, ‘sentiments’, ‘affections’, and ‘propensities’, as names of the various states of mind he classified as secondary impressions. Traditionally, ‘passion’ was the term writers of such treatises chiefly used for such states, though it was also common for them to apply ‘passion’ to the more turbulent ones and ‘sentiment’ to the calmer ones. Hume, however, did not follow this practice. He treated ‘passion’ as a general term for any of the states that he also called affections, sentiments, emotions, and propensities. Hume’s use of ‘emotion’ was casual. He did not define it or use it systematically. Nonetheless, one can discern in certain passages in which it

6.  The two sections that follow are drawn from Deigh (2013). 7.  I put a word within single quotation marks to indicate that what I am referring to is the word itself and not what it denotes (see Quine, 1972, pp. 43–​44, for discussion of this convention and the confusions it serves to avoid).

4

44  From Psychology to Morality

occurs a meaning that is distinct from what he meant by ‘passion’, and his use of the term with this meaning bears importantly on the later emergence of the scientific study of emotions. Of course, whether Hume’s use of the term with this meaning contributed to how emotions came to be conceived as the science emerged and grew, or merely anticipated it, is hard to establish. But given Hume’s influence on later thinkers, it is reasonable to treat him as having planted the seed from which the later concept germinated. Hume’s use of the term with this meaning is evident in the opening section of Book II. Having reminded his reader of his taxonomy of mental states, he proceeded to draw two distinctions, that between calm and violent passions and that between direct and indirect ones. The first distinction depends on a passion’s turbulence, its agitation of the mind, and in explaining this phenomenon he used ‘emotion’ with a meaning distinct from the meaning with which he used the term ‘passion’. Thus he wrote: The reflective impressions may be divided into two kinds, viz. the calm and the violent. Of the first kind is the sense of beauty and deformity in action, composition, and external objects. Of the second are the passions of love and hatred, grief and joy, pride and humility. This division is far from being exact. The raptures of poetry and music frequently rise to the greatest height; while those other impressions, properly call’d passions, may decay into so soft an emotion as to become, in a manner, imperceptible. But as in general the passions are more violent than the emotions arising from beauty and deformity, these impressions have been commonly distinguished from each other. (1978, p. 276)

Much later in Book II, when Hume returned to the topic of calm passions, he observed that they are sometimes mistaken for reason: Now ’tis certain, there are certain calm desires and tendencies, which, tho’ they be real passions, produce little emotion in the mind, and are more known by their effects than by the immediate feeling or sensation. These desires are of two kinds; either certain instincts originally implanted in our natures, such as benevolence and resentment, the love of life, and kindness to children; or the general appetite to good and aversion to evil consider’d merely as such. When any of these passions are calm, and cause no disorder in the soul, they are very readily taken for the determinations of reason, and are suppos’d to proceed from the same faculty, with that, which judges of truth and falsehood. (1978, p. 417)

Yet passions of either kind, Hume immediately went on to note, are not necessarily calm: Besides these calm passions, which often determine the will, there are certain violent emotions of the same kind, which have likewise a great influence on that faculty. When I receive any injury from another, I often feel a violent passion of resentment, which makes me desire his evil and punishment, independent of

45

William James and the Rise of the Scientific Study of Emotion  45 all considerations of pleasure and advantage to myself. When I am immediately threatened with grievous ill, my fears, apprehensions, and aversions rise to a great height and produce a sensible emotion. (1978, pp. 417–​418)

Having thus pointed out that a passion’s influence on the will is not a function of how violent the passion is, Hume then distinguished between these two aspects of a passion: ’Tis evident passions influence not the will in proportion to their violence, or the disorder they occasion in the temper; but on the contrary, that when a passion has once become a settled principle of action, and is the predominant inclination of the soul, it commonly produces no longer any sensible agitation. As repeated custom and its own force have made every thing yield to it, it directs the actions and conduct without that opposition and emotion, which so naturally attend every gust of passion. We must, therefore, distinguish betwixt a calm and a weak passion; betwixt a violent and a strong one. (1978, pp. 418–​419)

On the evidence of these passages, we may conclude that Hume understood the difference between a calm passion and a violent one to consist in the amount of emotion with which a passion occurs. The more emotion, the more violent the passion. Calm passions are passions that occur with so little emotion as to be “in a manner, imperceptible”; violent ones are passions that occur with enough emotion to create considerable agitation in the mind. At the same time, we can infer from Hume’s observation that calm passions no less than violent ones influence the will, that the amount of emotion with which a passion occurs is independent of the passion’s motivational force. Accordingly, Hume used ‘emotion’ as a term for that feature of a passion that characterizes how violent the passion is and understood this feature to be distinct from the passion’s being a motive or spring of action. He understood it, that is, as a phenomenal property of passions, which as such implies nothing about their motivational strength or power to produce action. Hume, however, did not consistently use ‘emotion’ in this way. He also used it as a common name for secondary impressions, and indeed for secondary impressions for which he also used the term ‘passion’. Note, for instance, his referring in the third of the four passages above to “violent emotions” which are of the same kind as certain calm passions. This usage complicates interpretation of Hume because it is hard to square with the first way in which he used the term. Indeed, on a superficial reading, the two ways appear to clash with each other, for traditionally passions are understood to be motives. Consequently, it appears that Hume, by using ‘emotion’ in this second way, implied—​contrary to what the first way in which he used it implies—​that emotion has motivational force. The appearance, though, is misleading. Hume’s concept of a passion is not the traditional one. Passions, as Hume conceived of them, are not necessarily motives. Indeed, some passions, according to Hume, are not motives. So his two ways of using ‘emotion’ do not clash with each other.

64

46  From Psychology to Morality

Still, they are independent of each other. Neither, that is, derives from the other.8 Nor did he try to avoid confusion by restricting his use of ‘emotion’ as a common name for secondary impressions to secondary impressions that were not motives. To the contrary, he readily used it to denote passions that were motives. A good example occurs in the course of his well-​known argument for the inertness of reason. “’Tis obvious,” he wrote, “that when we have the prospect of pain or pleasure from any object, we feel a consequent emotion of aversion or propensity, and are carry’d to avoid or embrace what will give us this uneasiness or satisfaction” (1978, p. 414). In the face of passages like this one, then, we must conclude that Hume used ‘emotion’ with two different meanings without ever explicitly distinguishing them. The potential for confusion was, if anything, exacerbated by Hume’s departure from the traditional concept of a passion. Hume’s departure from this concept is a consequence of the second of the two distinctions among the passions that he drew in the opening section of Book II. This is his distinction between direct and indirect passions. Direct passions are motives of action. They arise immediately from pleasure and pain or from the prospect of pleasure and pain and, in either case, produce a volition to embrace what gives pleasure or avoid what gives pain.9 In addition, Hume identified certain direct passions that “arise from a natural impulse or instinct, which is perfectly unaccountable” (1978, p.  439). These include hunger and other bodily appetites, the desire for revenge or punishment of one’s enemies, and the desire for the happiness of those whom one loves. Like direct passions that arise immediately from pleasure and pain, these too produce volitions to secure their objects. Indirect passions, by contrast, do not. For this reason, they are not in themselves motives. At the same time, some give rise to direct passions and therefore generate motives. These are

8.  Nor is there any evidence that Hume meant to be using ‘emotion’ metonymically when he used it as a name for secondary impressions. 9.  Hume first characterizes direct passions as “aris[ing] immediately from good or evil, pain or pleasure” (1978, p.  276; see also p.  399). Sometimes, though, he speaks of them as arising from the prospect of pleasure or pain (see 1978, p. 414). He also describes volitions as impressions that arise immediately from pleasure and pain, while at the same time excluding them from being passions: “Of all the immediate effects of pain and pleasure, there is none more remarkable than the will” (1978, p. 399; see also p. 574). Later, however, he implies that the will is determined by passions:  “Beside these calm passions, which often determine the will . . .” (1978, p. 417). I believe the latter thesis best fits his argument and have interpreted him accordingly, that is, by presenting as his view that volitions are produced by direct passions. Note that in either case, Hume excludes volitions from the category of passions: “tho’, properly speaking, [the will] be not comprehended among the passions, yet . . . we shall here make it the subject of our enquiry” (1978, p. 399).

4 7

William James and the Rise of the Scientific Study of Emotion  47

chiefly love and hatred, which give rise to a desire for the happiness of one’s beloved, which Hume identified with benevolence, and to a desire for the misery of one’s enemies, which Hume identified with anger. Other indirect passions, however, do not give rise to direct passions. These are chiefly pride and humility. They are neither motives nor producers of motives. Hume set out this difference between the two sets of indirect passions in the section of the Treatise on benevolence and anger: The passions of love and hatred are always followed by, or rather, conjoin’d with benevolence and anger. ’Tis this conjunction, which chiefly distinguishes these affections from pride and humility. For pride and humility are pure emotions in the soul, unattended with any desire and not immediately exciting to action. But love and hatred are not compleated within themselves, nor rest in that emotion, which they produce, but carry the mind to something farther. (1978, p. 367)

The traditional concept of a passion, in modern philosophy, is well illustrated in Descartes’ (1989) The Passions of the Soul. Descartes held that all passions give rise to volitions, or as he put it, in keeping with his view that the will is free, they all dispose the will to act as they bid. Thus, in Article 40, Descartes wrote, The principal effect of all the passions in men is that they incite and dispose their soul to will the things for which they prepare their body, so that the sensation of fear incites it to flee, that of boldness to will to do battle, and all the rest. (Descartes, 1989)

What Descartes described as the principal effect of all passions, a volition, is for Hume the principal effect of direct passions only. By distinguishing indirect passions from direct passions by virtue of their not giving rise to volitions, Hume therefore not only departed from the traditional concept of a passion but rejected as well the understanding of a passion as an inclination to act that, through an exercise of will, a volition, one either follows or resists. Because Hume regarded volitions as secondary impressions,10 he rejected Descartes’ model of the will as acting on a passion and endorsed instead a mechanical model in which a so-​called act of will, a volition, is a secondary impression produced by a direct passion. Even more instructive, perhaps, for seeing the significance of Hume’s departure from the traditional concept is Thomas Reid’s objection, in his Essays on the Active Powers of the Human Mind (1969), to Hume’s notion of a calm passion. Reid objected in particular to Hume’s thesis that calm passions are

10.  “The impressions, which arise from good and evil most naturally, and with the least preparation are the direct passions of desire and aversion, grief and joy, hope and fear, along with volition” (Hume, 1978, p. 438).

84

48  From Psychology to Morality

“readily taken for the determinations of reason” (Hume, 1978, p. 417). Hume had advanced this thesis to explain how the great majority of philosophers, both ancient and modern, had come to the false view of reason and passion as combatants that oppose each other in struggles to determine the will. Reason, Hume had argued, is the slave of passion. Being powerless to oppose passion, it is necessarily under its rule. He then, to supplement this argument, gave his explanation of how philosophers had come to the false view of the combat between reason and passion. Both the argument and the explanation, Reid rejoined, were based on an abuse of language. Passions, Reid observed, are, in the most general sense of the term, “principles of action” that are also commonly referred to as appetites and affections (Reid, 1969, p. 70). They are the propensities to action that humans share with brutes and must therefore be understood to spring from the irrational part of the human frame. Yet ‘passion’, Reid maintained, is more commonly used with a more specific sense that distinguishes passions from other natural desires and affections (Reid, 1969, pp.  177–​178). In this more specific sense, its ordinary sense, passions are principles of action whose strength agitates the body and clouds the understanding. Indeed, their strength as motives of action corresponds to the violence with which they occur, for the stronger the passion the more the body is uncontrollable and the mind is clouded. As Reid saw things, Hume’s talk of calm passions, passions that prompt action without any sensible agitation, is oxymoronic. A defender of Hume could, of course, grant Reid’s point about common usage without conceding any significant error in Hume’s position. Hume’s position depends on there being propensities to action that influence the will and produce action without any sensible agitation. Since Reid allowed that there were such propensities, since he allowed that natural desires and affections could be calm in Hume’s sense, his criticism of Hume for his sham use of ‘passion’ appears to be no threat to Hume’s thesis that these calm affections and desires are commonly mistaken for determinations of reason. But Reid’s criticism goes deeper into Hume’s thought than its linguistic surface. It strikes rather at Hume’s distinction between the motivational strength of a passion and its violence. The distinction underpins Hume’s thesis, and Reid, in assuming a correspondence between the strength of a passion and its violence, thus rejected the distinction. In Reid’s view, one could not, pace Hume, mistake in a man who resolutely resists temptation a passion for the determinations of his reason, for if the man’s will to resist temptation derived from a passion, the violence of that passion would be greater than that of the passion it defeated. Its presence and operation would, therefore, be unmistakable. And if one were to suppose instead that the man resisted the passion

94

William James and the Rise of the Scientific Study of Emotion  49

in order to secure a good that he desired and his desire for that good was calm, then its being calm, Reid held, would mean that it neither weakened the man’s control over his body nor clouded his judgment. Consequently, his resistance would have to be due to a determination of reason, for the desire, being calm, would have been too weak to check the passion he resisted. Hume’s idea that such resistance could be the work of a calm passion was, in other words—​given a correspondence between the violence of a passion or other affection and its motivational strength—​incoherent. Reid, as a result of his rejecting Hume’s distinction between the strength of a passion and its violence, denied in effect that the amount of emotion with which a passion occurred was a feature of a passion distinct from its power to produce action. Reid, then, because he held to the traditional concept of passion, did not allow the possibility of a passion’s not being a motive. It was for him inconceivable. For Hume, by contrast, it was certainly conceivable. The distinction between the amount of emotion in a passion and a passion’s power to produce action guaranteed its possibility. After all, one need only think of a turbulent state of mind that had no tendency to produce action to conceive of such a passion. And Hume, in distinguishing indirect passions from direct ones, did just that. His account of pride and humility as passions that were “pure emotions” fit the bill exactly (Hume, 1978, p. 367). Reid, as we have seen, treated motivational strength as inseparable from turbulence. The term ‘emotion’ appears nowhere in his chapter on the passions. Hume’s characterization of pride and humility as pure emotions thus reflects his distinction between the violence of a passion and its strength. It reflects, in other words, his use of ‘emotion’ to denote that feature of a passion that characterizes how violent it is together with his understanding of this feature as independent of a passion’s power to produce action. By virtue of this understanding, Hume’s concept of a passion departed from the traditional one. Later writers who took up Hume’s other use of ‘emotion’, his use of it as a general name for secondary impressions, did so without also departing, as Hume did, from the traditional concept of a passion (e.g., Bain, 1875, pp. 383–​392; Brown, 1822, I, p. 251).11 Generally, then, while Hume’s use of ‘emotion’, as a name for secondary impressions, including passions, emerged

11.  Bain, in introducing the category of states of feeling that are neutral in the sense of being neither pleasurable nor painful—​surprise is his example—​may seem to be departing from the traditional concept of passions since he takes their neutrality to imply that they are not motives to action. But his explanation of them as fixing one’s attention on an object so as to require effort to turn one’s attention elsewhere shows that he still conceives them as states that influence the will (see Bain, 1875, pp. 13–​14, 390).

05

50  From Psychology to Morality

as the prevailing one in the nineteenth century, his use of the term as a name for a feature of a passion that is distinct from that of the passion’s motivational strength did not acquire a similar following. The notion it expressed disappeared for much of the century. Not until William James published his work in psychology did it—​or rather something like it—​reappear and come firmly and distinctly into view.12

III To understand the lasting significance of James’s work on the emotions, one must attend to writings of his other than the chapter on emotions in The Principles of Psychology (1890) or the earlier, more widely read, article “What is an Emotion?” (James, 1884). One must attend, above all, to the famous ninth chapter of The Principles (Deigh, 2001). This chapter contains James’s attack on the conceptual scheme for mental states characteristic of classical British empiricism, the scheme of Locke’s psychology and Hume’s, along with that of each of the many minor philosophers who belong to this tradition. The revolutionary import of James’s theory of emotions depends on his attack on this conceptual scheme. In following this scheme, Locke, Hume, and the many empiricist philosophers they influenced divided all mental states into two or three main types according to their origins. Thus Locke divided all mental states into sensory ideas, ideas of imagination, and abstract ideas, and Hume, as I noted above, divided all of them into impressions of sensation, ideas, and impressions of reflection. They then treated the states in each of these types as either simple or complex, and if complex, then analyzable into simple states. And finally they construed mental states as occurring in the mind singly, discretely, and serially. Hume’s wonderful description of the mind as “a kind of theatre, where several perceptions successively make their appearance; pass, re-​pass, glide away, and mingle in an infinite variety of postures and situations” nicely captures the classical empiricist conception of mental states (Hume, 1978, p. 253). James rejected this conception as both false to our experience, when carefully considered, and inconsistent with any understanding of consciousness as a product of neurophysiological processes. Conscious experience, which fills our waking lives, does not break down into sequences of individual, recurring impressions and ideas, “mental atoms or molecules” as James liked to call them (James, 1890, I, p. 230). In a seeming rejoinder to 12.  C.  G. Lange, independently of James’s publication, published a similar account of emotion around the same time (see James & Lange, 1922).

51

William James and the Rise of the Scientific Study of Emotion  51

Hume, James declared, “A permanently existing ‘idea’ . . . which makes its appearance before the foot-​lights of consciousness at periodical intervals is as mythological an entity as the Jack of Spades” (James, 1890, I, p. 236).13 James then applied this attack to the empiricists’ program for studying emotions. The traditional program of British empiricism was that of a taxonomic science. It consisted in setting out a general definition of the phenomena to be studied, dividing those phenomena into their several species, dividing those species into subspecies, and so on. James’s reaction to this enterprise was sharp. He wrote, The trouble with the emotions in psychology is that they are regarded too much as absolutely individual things. So long as they are set down as so many eternal and sacred psychic entities, like the old immutable species in natural history, so long all that can be done with them is reverently to catalogue their separate characters, points, and effects. (1890, II, p. 449)

Such treatment of the subject, James observed, has led to flat, profitless descriptions of a seemingly endless variety of emotions: The mere description of the objects, circumstances, and varieties of the different species of emotion may go to any length. Their internal shadings merge endlessly into each other, and have been partly commemorated in language as for example by such synonyms as hatred, antipathy, animosity, resentment, dislike, aversion, malice, spite, revenge, abhorrence, etc. etc . . .. But there are limits to the profitable elaboration of the obvious, and the result of all this flux is that the merely descriptive literature of the subject, from Descartes downwards, is one of the most tedious parts of psychology. (James, 1892, p. 374)14

Scientific psychology, James declared, to get to the deeper levels of understanding that “all truly scientific work” achieves, must abandon its treatment of emotions as recurring types to be differentiated and catalogued according to their observable features and typical circumstances (James, 1890, II, pp.  448–​449). It must, instead, treat the variability of emotion as what is to be explained: But if we regard them as products of more general causes (as ‘species’ are now regarded as products of heredity and variation), the mere distinguishing and cataloguing becomes of subsidiary importance. (1890, II, p. 449)

13.  That James has Hume in mind is then made explicit on the next page. “[A]‌necessary consequence of the belief in permanent self-​identical psychic facts that absent themselves and recur periodically is the Humian [sic] doctrine that our thought is composed of separate independent parts and is not a sensibly continuous stream . . .. [T]his doctrine entirely misrepresents the natural appearances” (1890, I, p. 237). 14. The passage is a more pithy restatement of the same points in James (1890, II, pp. 447–​448).

25

52  From Psychology to Morality

James, as this passage makes clear, aimed at reorienting the scientific study of emotions away from taxonomy and towards genesis. Progress in this study, he believed, required investigating the origins of emotional variety, and to investigate this phenomenon psychologists must study the different external conditions and events to which emotions are reactions, as well as the bodily processes that mediate those reactions. To this end, James redefined emotions as the feelings of bodily changes, as they occur, which changes directly follow upon perception of an exciting fact (1890, II, p. 449). Plainly, he did not put forward this definition to establish a category of things whose “separate characters, points, and effects” can be catalogued. He put it forward so as to turn our attention in the study of emotions to the general causes of the bodily changes that occur in episodes of emotion. While the feelings of those changes, as they occur, are what ‘emotion’, as he defined the term, means, they are not, he held, isolable, recurring units of consciousness whose nature and composition is the subject of scientific study. They are, rather, like rapids and eddies in a river, to be understood as disturbances and agitations in an unbroken stream of thought, which one studies by examining the forces and conditions that produce such changes in the flow. The study of emotions, in other words, in James’s view, is the study of the causes of bodily changes that are made manifest in the mind through turbulent or stirring feelings. Because these bodily changes include the movements of muscles and limbs behind the voluntary behavior of which passions, on the traditional concept of them, are motives, James’s definition of an emotion departs from this concept. James, to be sure, was aware of his definition’s unorthodoxy. Indeed, he regarded its being contrary to common sense on this point as its hallmark. Immediately, after giving his definition, he wrote, Common sense says, we lose our fortune, are sorry and weep; we meet a bear, are frightened and run; we are insulted by a rival, are angry and strike. The hypothesis here to be defended says that this order of sequence is incorrect, that the one mental state is not immediately induced by the other, that the bodily manifestations must first be interposed between, and that the more rational statement is that we feel sorry because we cry, angry because we strike, afraid because we tremble, and not that we cry, strike, or tremble, because we are sorry, angry, or fearful, as the case may be. (1890, II, pp. 449–​450)

Two things are noteworthy about this passage. First, it shows that James, by defining an emotion as the feelings of bodily changes, not only took an emotion to be the sensible agitation in the soul that Hume had identified as the determinant of a passion’s violence, but also understood that, as such, an emotion was not a motive of action. To this extent, his definition recovers the notion Hume had in mind in using ‘emotion’ to denote the feature of

53

William James and the Rise of the Scientific Study of Emotion  53

passions that characterized how violent they are. Second, it spells out the chief implication of James’s definition, its incompatibility with the classical empiricist taxonomy of mental states. For by pointing out that, contrary to common sense, an emotion is not the direct consequence of a perception or thought but rather arises only after the interposition of those bodily changes that a perception or thought excites, James jettisoned one of the principal categories of mental states on the classical empiricist taxonomy. On that taxonomy, emotions, passions, affections, and the like all belong to the category of states of mind that arise from antecedent sensory or intellectual states. For Locke these were internal sensations; for Hume, they were secondary or reflective impressions. James’s definition in effect removes such states from the study of psychology since its implication is that physiological processes and not prior states of mind are the immediate cause of the feelings with which we identify emotions, passions, affections, and the like. Consequently, his definition differs from the notion Hume had in mind in that it does not represent an emotion as a feature of some mental state that may also have, independently of that feature, the power to produce action. In James’s account of human psychology, there are no such mental states. The question, then, given that James had removed passions, as traditionally conceived, from human psychology, is how he accounted for human motives. The answer is found in the chapter in the Principles on instincts, Chapter xxiii, which immediately precedes the one on emotions. This chapter and the one immediately preceding it—​Chapter xxii, “The Production of Movement”—​concern the bodily actions the feelings of which James identified as emotions. Many of these bodily actions are autonomic. Shortness of breath, a palpitating heart, and shivering are examples. James discussed such actions in Chapter xxii. He then dealt with voluntary actions in Chapter xxiii. Each instinct, according to James, is an innate tendency to act, a tendency built into the nervous system of human beings and other animals as a result of natural selection.15 Each such tendency, each instinct, operates through an impulse to action (1890, II, p.  385). While instincts originally operate blindly, which is to say, reflexively, in response to certain sensory stimuli, they become in animals capable of foresight the source of voluntary actions through repeated experiences of their operation. An animal’s familiarity with its circumstances enables it to anticipate the impulses they excite 15.  James, in the chapter on instinct, does not explain instincts as the result of natural selection, but treats them as innate and the result of adaptation. In the last chapter of Principles, however, in the section “The Origin of Instincts,” he argues for the superiority of Darwin’s theory of natural selection to Lamarck’s theory of the inheritance of acquired traits (see 1890, II, pp. 678–​688).

45

54  From Psychology to Morality

and to encourage or resist them by thinking of objects that strengthen or counter them. In this way, the animal gains some measure of control over its muscular movements and so its actions. Correspondingly, while the earliest operations of an instinct, being entirely reflexive, produce actions of whose ends the animal is incognizant, later operations produce actions whose ends it cognizes and, to the extent the actions fall within its control, it consciously pursues. Accordingly, instincts qualify as motives of these actions. As principles of or propensities to action that humans share with other animals, they replace passions in James’s account of human psychology. James’s discussion of the instincts special to human beings confirms this observation. Human beings, James maintained, possess an enormous variety of instincts. He included among them tendencies to attack what opposes one, to flee what threatens one, and to comfort one’s kin and conspecifics. In each case, one perceives some exciting object, and the perception then mobilizes one to action. James characterized the operations of these instincts as anger, fear, and sympathy. Later in the discussion, he took up the human tendency to acquire or appropriate things that please and noted how this tendency can turn into the impulse to harm others who possess things that one covets. He characterized such impulses as those of envy and jealousy. Finally, he added the tendency of parents, mothers in particular, to cleanse and feed their young, and characterized the operations of this instinct as parental love. Each of these characterizations identifies an instinct by the type of emotion that the action it tends to produce typically expresses. These characterizations suggest a correspondence between the feelings of bodily changes that the operations of certain instincts produce and the different types of emotion that theorists of the mind since Plato had distinguished and studied. If James had affirmed this correspondence, then his distinction between emotions and the instincts whose operations produced them would have mapped on to Hume’s distinction between a passion’s degree of violence and its motivational strength or power to produce action. James, however, denied it. Specifically, he denied that one’s feeling an emotion of one of these types—​fear, anger, joy, or the like—​necessarily resulted from one’s undergoing a bodily movement or syndrome of bodily movements that the operations of the corresponding instinct produced (1890, II, p. 442). Indeed, he denied that it necessarily resulted from one’s undergoing a bodily movement or syndrome of bodily movements of any determinate type. Because of the indefinite variability among people in their reflex responses to the perception of the same exciting object, James argued, no determinate type of bodily change or syndrome of bodily changes corresponds to any of these types of emotion. A frightening object may induce flight in

5

William James and the Rise of the Scientific Study of Emotion  55

one person and paralysis in another. Triumph may cause one person to jump and another to prance. And a similar variability among people holds of autonomic responses as well: [T]‌he moment the genesis of an emotion is accounted for, as the arousal by an object of a lot of reflex acts which are forthwith felt, we immediately see why there is no limit to the number of possible different emotions which may exist, and why the emotions of different individuals may vary indefinitely, both as to their constitution and as to objects that call them forth. (1890, II, p. 454)

It follows, therefore, in James’s view, that motives of voluntary action originate in instincts that can have no more than a statistical association with the emotions to which he referred homonymously. James’s separation of emotion from the springs of action was thus even greater than the one Hume implied in using ‘emotion’ to capture the feature of a passion that characterized how violent it was as distinct from how powerful a motive it was. Hume, in so using ‘emotion’, identified a distinctive feature of passions, their sensible turbulence or agitation, which, departing from the traditional concept of passion, he took to be independent of a passion’s motivational strength. He did not, however, treat this feature as a special object of psychological study. His purpose in identifying it as a feature distinct from a passion’s motivational strength was to secure his notion of a calm passion whose operations were commonly mistaken for determinations of reason. Passions were the object of his study. James, by contrast, went further. Not only did he identify emotion with sensible turbulence or agitation that had no motivational potential and was therefore independent of whatever had motivational strength, but he also, by removing from the study of psychology the type of mental state to which passions, on the classical empiricist taxonomy, belonged, made such turbulence—​turbulence directly produced by bodily changes and not prior mental states—​the object of psychological study. A major consequence, then, of his applying his attack on the classical empiricist conceptual scheme for mental states to the empiricist program for studying emotions, was to redefine the study of emotions as the study of such states of arousal. His redefinition represents a significant change in how emotions, as the object of scientific study, are conceived. He thus brought to completion the transformation of the study from one in moral philosophy to one in experimental psychology.

IV In the traditional concept of passion, a passion influences the will by virtue of a cognition that guides action through a representation of something good to

65

56  From Psychology to Morality

be realized or bad to be avoided by doing that action. Modern philosophers, from Descartes forward, typically explained this cognition as arising in conjunction with the passion. Their predecessors, the Scholastics, typically explained it as an essential element of the passion. In either case, it is the vehicle of evaluation that, together with or as an element of the passion, moves the subject accordingly in a direction to which the will either consents or sets itself in opposition. As such, it is built into the very way passions are understood as motives of voluntary action. Since to give a theory of passions entails giving a theory of these motives, and since one could not give such a theory without distinguishing at some level of differentiation the variety of motives to which human beings are subject, the theory perforce must include a taxonomy whose types are defined at least partly by the different types of evaluation that accompany or give essential form to a passion. Hence, taxonomic studies of the sort characteristic of the writings on emotion that James pronounced “one of the most tedious parts of psychology” are integral to the study of emotion on a concept of the phenomenon that is or derives from the traditional concept of passion (1890, II, p. 448).16 While James may have been fully warranted in deploring the amount of minutiae found in these writings, his criticism of their aim of expounding a taxonomy of emotions masks how radical the change is he introduced in how the phenomenon is conceived. His criticism, then, of these taxonomic studies as superficial, and at best secondary to a proper scientific study of emotions, is due to his conceiving emotions as states of arousal only. It is due, that is, to his excluding evaluation from his conception of the phenomenon. Having explained motives as originating independently of emotions and as merely correlated statistically with them, owing to the overlap of the class of bodily movements the feelings of which an emotion consists in and the class of bodily movements that initiate voluntary action, he rendered pointless any conception of emotions as entailing evaluations. Hence, to construe feelings of bodily changes as vehicles of evaluation is likewise pointless if it is not done for the purpose of assigning these feelings a role in the production of voluntary action. Yet to identify emotions with such feelings, and at the same time assign them such a role so as to preserve the concept of an emotion as a motive, is to miss completely James’s point in identifying emotions with feelings of bodily changes. It leads, moreover, to theories of emotion of the very sort James meant to 16.  James’s dismissal continues with even harsher criticism. In the next sentence, he wrote, “And not only is it tedious, but you feel that its subdivisions are to a great extent either fictitious or unimportant, and that its pretences to accuracy are a sham” (1890, II, p. 448).

57

William James and the Rise of the Scientific Study of Emotion  57

discredit. The neo-​Jamesians, who put forward such theories in the interest of rehabilitating James’s theory, thus profoundly misunderstand it. The question that confronts anyone who wishes to revive James’s theory is whether his proposal to treat emotions as states of arousal that are independent of states of motivation represents an advance or a dead end in the scientific study of emotions. It would represent an advance only if giving up our understanding of anger, fear, envy, and the like as states of mind that both contain feelings and serve as motives were feasible. Perhaps it seemed feasible to James because he took introspection to be as reliable a method of observation in science as unaided perception of external things. Introspection, he thought, was the primary method of gathering the facts about the mind that psychology studies. It is, he said, the “first and foremost” form of observation on which psychologists rely (1890, I, p.  185). And, he later added, though it is “difficult and fallible . . . the difficulty is simply that of all observation of whatever kind” (1890, I, p. 191, italics removed). Such confidence in introspection undoubtedly led him to assume that the subjects of emotion could identify what emotion they were feeling even if they could not describe it fully. Hence, he could believe that, owing to the reliability of introspection as a method of observation, one can directly identify a state of arousal as fear, anger, sorrow, joy, or the like without applying the criteria specified in a taxonomy of such states and therefore one can rely on this method to study emotions as exclusively states of arousal. Introspection, however, has long since lost its place in psychology as the primary method of observation. It fell as experimental psychologists came to adopt canons of scientific investigation which require that the facts gathered as evidence for or against a hypothesis be intersubjectively confirmable. Without confidence in introspection as a reliable method of observation, then, it is hard to see how giving up our understanding of anger, fear, and the like as states of mind that both contain feelings and serve as motives could be sustained. The prospects, in other words, for reviving James’s theory of emotions are dim.

References Bain, A. (1875). The emotions and the will (3rd ed.). New York, NY: D. Appleton. Brown, T. (1822). Lectures on the philosophy of the human mind. Andover, UK:  Mark Newman. Damasio, A. (1994). Descartes’ error: Emotion, reason, and the human brain. New York, NY: Avon. Deigh, J. (1994). Cognitivism in the theory of emotions. Ethics, 104, 824–​854. (Reprinted in Deigh, 2008.)

85

58  From Psychology to Morality Deigh, J. (2001). Emotions:  The legacy of James and Freud. International Journal of Psychoanalysis, 82, 1247–​1256. (Reprinted in Deigh, 2008.) Deigh, J. (2008). Emotions, values, and the law. Oxford, UK: Oxford University Press. Deigh, J. (2013). The emergence of emotion as an object of scientific study. In J. Deigh (Ed.), On emotions:  Philosophical essays (pp. 215–​ 232). Oxford, UK:  Oxford University Press. Descartes, R. (1989). The passions of the soul (S. Voss, Trans.). Indianapolis, IN: Hackett. Dixon, T. (2003). From passions to emotions: The creation of a secular psychological category. Cambridge, UK: Cambridge University Press. Hills, D. (2008). Response to Gut reactions. Philosophy and Phenomenological Research, 76, 720–​728. Hume, D. (1978). A treatise of human nature (2nd ed., L. A. Selby-​Bigge Ed.). Oxford, UK: Oxford University Press. James, W. (1884). What is an emotion? Mind, 9, 186–​205. James, W. (1890). The principles of psychology. New York, NY: Henry Holt. James, W. (1892). Psychology: Briefer course. New York, NY: Henry Holt. James, W., & Lange, C. G. (1922). The emotions (K. Dunlap, Ed.). Baltimore, MD: Williams and Wilkins. Lyons, W. (1980). Emotions. Cambridge, UK: Cambridge University Press. Pitcher, G. (1965). Emotion. Mind, 74, 326–​346. Prinz, J. (2004). Gut reactions:  A perceptual theory of emotions. Oxford, UK:  Oxford University Press. Quine, W. V. O. (1972). The methods of logic (3rd ed.). New York, NY: Holt, Rinehart, and Winston. Reid, T. (1969). Essays on the active powers of the human mind. Cambridge, MA: MIT Press. Robinson, J. (2005). Deeper than reason: Emotion and its role in literature, music, and art. Oxford, UK: Clarendon Press. Solomon, R. C. (1976). The passions: The myth and nature of human emotion. Garden City, NJ: Anchor Press/​Doubleday.

95

3

Freud

Sigmund Freud created a theory of psychology that has had a more profound influence on our thinking about human life and human culture than any other theory produced in the twentieth century. Freud presented his theory as the product of scientific work. He did not offer it as part of a philosophical system and did not advance philosophical arguments to defend it. Rather he based it on evidence he gathered from the observations he made as a physician specializing in nervous disorders, and he enlarged and modified it on the basis of further observations and reflections. Consequently, to understand its philosophical import, one must first explain its origination and subsequent development and identify its central ideas. I will preface my exposition of his theory with a brief account of Freud’s early life.

1.  Freud’s Early Life Freud grew up in Vienna, where his family had moved when he was four. He was born 6 May 1856 in Freiburg, Moravia, which is now part of the Czech Republic. The family lived there for three years, followed by a year in Leipzig, before moving to Vienna. Freud’s father, a wool merchant, was

59

06

60  From Psychology to Morality

twice married. His second marriage at age forty was to Freud’s mother. She was twenty. Freud was their first born. Over the next ten years the couple had seven more children, five girls and two boys. Freud also had two half-​ brothers from his father’s first marriage. They were close to his mother in age, and one of them was married with a son. Freud thus entered the world already an uncle. The family lived in what Freud characterized as limited circumstances, though in so characterizing them he could have just as well been describing the limited opportunities open to Jews in late nineteenth-​ century Vienna as the family’s limited finances. Freud’s schooling in Vienna was at the Gymnasium, where he excelled. He remembered questions about human concerns piquing his curiosity more than questions about natural objects.1 When Freud was seventeen, he entered the University of Vienna as a medical student. He was drawn to medicine, not from a desire to become a physician, but rather from an interest in science. The coursework of his first two years reflected this interest. It consisted of a steady diet of lectures in physics, chemistry, biology, physiology, botany, anatomy, and so forth, though he did manage during this period to take a few courses in philosophy from Franz Brentano. Midway through his third year he began to do research in zoology but by the end of the fourth he had switched to physiology. His imagination and commitment to science had been captured by Ernst Brücke, the director of the physiology laboratory, and he remained Brücke’s student and a member of his Institute for the next five years. During this time he focused in his research on the nervous system and, en passant, received his MD. While there can be no doubt as to the success of his research, he had no realistic hope of advancing to any of the higher positions within the Institute. In 1882 he left and began preparing for a career as a practicing physician. His expertise in neurophysiology equipped him for specializing in nervous diseases. He became a resident at Vienna’s General Hospital, where he concentrated on brain anatomy and neurology while also acquiring the broad knowledge and skill necessary for his new career. At the end of his residency, in the summer of 1885, he won a six-​month travel grant for the purpose of studying with Jean Martin Charcot at his clinic in Paris. Charcot, one of Europe’s leading neurologists, was at that time lecturing on hysteria and demonstrating its treatment through hypnosis. The impact of these lectures and demonstrations on Freud’s thinking cannot be overstated. On the one hand, Charcot presented a symptomatology of hysteria that, in Freud’s view, 1.  “An Autobiographical Study,” SE v. 20, pp. 4–​5.

61

Freud 61

established it as a true neurosis, for previously it had been widely considered to be merely a grab-​bag diagnostic category, equally applicable to faking as to genuine nervous disorders. On the other, seeing Charcot use hypnotic suggestion to induce as well as remove the physical symptoms of hysteria—​ anesthesia, convulsions, paralysis, etc.—​convinced Freud of the importance of psychological factors in the etiology of the neurosis, at least in certain of its forms. Thus, the idea of psychoneurosis took hold in his mind and became the inspiration for the subsequent theoretical and clinical work that constitutes his extraordinary intellectual legacy.

2.  The Development of Freud’s Theory The heart of that legacy is the theory and practice of psychoanalysis. It took Freud a long time, roughly a dozen years after he returned to Vienna from Paris and started a private practice, to arrive at its fundamental principles and techniques. To relate how he did so would require recounting the many twists and turns his thinking took as he wrestled with ideas for which there was very little sympathy in Vienna’s medical community. Suffice it to say, he developed the theory of psychoanalysis from successive attempts to fit the hypothesis of psychoneurosis to both the clinical observations he made from the cases he treated and the understanding of neurophysiology he retained from his studies and research at the university and the hospital. And he developed the therapeutic techniques of psychoanalysis through discovery of ways to elicit from his patients material that provided insight into the ideas and impulses at the root of their illnesses and also ways to relieve his patients of their suffering by making them aware of those ideas and impulses and helping them to work through them. Studies in Hysteria, a book that he wrote with his friend and older colleague Josef Breuer and published in 1895, contains his early theoretical ideas and the case histories from which they were drawn. It was his first book on psychopathology. The full flowering of his psychoanalytic researches in this formative period appeared four years later in his next book, his masterwork, The Interpretation of Dreams. The last chapter of this work lays out the basic structural features of psychoanalytic theory. Although Freud significantly revised these in later works, the essential ideas remained constant throughout his subsequent writings. Thus, to begin with, Freud divided the mind into conscious and unconscious parts. On this scheme, the workings of the latter explain many of the thoughts and feelings that occur in the former and the behavior that manifests them. The principal explanations center on conflict between the mind’s conscious

26

62  From Psychology to Morality

and unconscious parts. The unconscious part contains thoughts and wishes that would be painful and distressing if one were conscious of them. Indeed, their having become unconscious is itself due to the necessity of ridding oneself of the pain and distress they cause. The memory of a traumatic experience, for example, or the wish to see a loved one dead can be so upsetting that, to regain peace of mind, one must not only banish it from consciousness but also block it from returning. One must exclude the painful memory or wish from the field of conscious experience and then repress it. In this way one gains an immunity from its pain and distress, though renewed exertions of repressive force are necessary to keep up the immunity. They are necessary, according to Freud, because unconscious thoughts and wishes are themselves invested with psychic energy in consequence of which they would force themselves into consciousness if left unimpeded. The conflict between the mind’s conscious and unconscious parts is therefore to be understood as a clash of these psychic forces. On the one side is the force with which unconscious thoughts and wishes are invested, and on the other is the force that is exercised in repressing those thoughts and wishes by the conscious part of the mind, more exactly, the agency that Freud assigned to it and that he called the ego. And this clash continues as long as the repression is successful. The principal explanations of psychoanalytic theory consist then in citing the dynamics of this conflict and following out its effects. Because of the force with which unconscious thoughts and wishes are invested, the ego’s mastery of them is never total. They do not, in other words, completely succumb to the repressive force the ego applies but rather continue to exert pressure within the mind. Being blocked from becoming conscious, they exercise influence in other ways. The pressure they continue to exert finds alternative outlets, and both their force and their content, albeit in distorted and disguised forms, are communicated through these outlets. As a result, they affect their subject’s conscious thought and behavior while remaining unknown to him. The effects of their influence include emotions with inappropriate objects, abnormally strong desires and urges, dreams and delusions, and other disturbances of thought and behavior. The symptoms of the psychoneuroses Freud treated are prime examples. Dreams too, which Freud took to be disturbances of thought that occurred when the ego’s defenses were temporarily weakened during sleep, are examples of first importance in the development of his theory. In either case the phenomenon is traced back through associations and ideational transformations to its origins in the subject’s unconscious thoughts and wishes, and accordingly what seems pointless or incoherent in a person’s life is shown to be meaningful in the context of that life. The triumph of these explanations is that they make such

6 3

Freud 63

familiar but disturbing phenomena of human life as compulsions, obsessive thoughts, excessive fears, and dreams, which were once regarded as the products of alien spiritual forces or the expressions of mysterious somatic disorders, intelligible as effects of the human struggle to escape pain and achieve happiness. Freud, having anchored his theory with the explanations of these disturbing phenomena, proceeded to extend its reach. In The Psychopathology of Everyday Life and Jokes and Their Relation to the Unconscious, two books he published soon after the appearance of The Interpretation of Dreams, he offered distinctively psychoanalytic explanations of other familiar but more innocuous types of human behavior. These included accidents, mistakes, slips of the tongue, acts of forgetfulness, jokes, and comical behavior. Surveying a broad range of examples, he illustrated the workings of the unconscious in the production of these actions. Later he considered religious rituals and presented a psychoanalytic account of these on which they were understood as analogous, in both their genesis and their forms of expression, to the thoughts and behavior characteristic of obsessional neurosis. And then, in 1913, he published Totem and Taboo, his groundbreaking work on the origins of religion and morality in human society. Using anthropological studies of aboriginal cultures, Freud proposed an account of the primitive religious practices and moral norms of these cultures according to which they represented resolutions of conflict between parents and children that each person, as a child, had internalized and repressed. Freud then suggested that this account could be extended to the religious practices and ethics of his own, Judeo-​Christian society, that they too represented resolutions of such conflicts. The Psychopathology of Everyday Life, with its deeper explanations of human actions that to common sense seemed like chance events or the results of mere lapses in thought, had attracted favorable interest from a gen­ eral, educated public. Totem and Taboo, by contrast, with its unflattering account of the origins of religion and morality, added to Freud’s notoriety as a purveyor of shocking and dangerous ideas. Freud had originally gained such notoriety from the hypotheses about human sexuality that he ultimately adopted in developing his theory of the psychoneuroses. He had come to the conclusion from his clinical work with hysterics that sexual trauma was at least partly responsible for his patients’ condition, and as he delved more deeply into the material these patients provided, he became increasingly persuaded that this trauma occurred in early childhood. He initially formed the hypothesis that the patients, as small children, had been the victims of sexual abuse and that the memory of this experience, now repressed, was at the root of their illness. As alarming as

46

64  From Psychology to Morality

this hypothesis was, he replaced it with an even more alarming one, that the trauma, in many cases, resulted from fantasies the patient had as a small child. After all, Freud noted, whether the illness was due to repressed childhood memories or repressed childhood fantasies would not affect the character of its manifestations in the adult patient. It would, however, say something about the character of the child’s sexuality, and this implication is what made the new hypothesis even more alarming. For the fantasies that, on the new hypothesis, Freud attributed to small children, unlike the memories that on the old one he attributed to them, expressed sexual desires. In other words, where the old hypothesis merely implied that small children were sometimes the objects of another’s sexual interests, the new one implied that such children themselves had sexual interests and experienced sexual impulses. To the members of a society whose moral view of human nature required belief in the sexual innocence of the prepubescent child, Freud’s new hypothesis was nothing short of scandalous. Freud expounded and defended this new hypothesis and its auxiliaries in Three Essays on the Theory of Sexuality, which appeared in 1905. Next to the basic structural features of psychoanalytic theory outlined above, these hypotheses are its most significant elements. The structural features identify the theory as a representative of psychological hedonism, and as such it descends from theories of human motivation that originated in the thought of the ancient Greeks. By using his hypotheses about human sexuality to fill in the structure that these features defined, Freud was then able to advance this tradition in a remarkably fertile way. For one thing, filling in the structure with these hypotheses enabled him to organize the theory around common, age-​specific patterns of experience in infancy and early childhood. Thus the content of the unconscious thoughts and wishes to which distinctively psychoanalytic explanations appealed could be systematically determined in relation to the infant’s or child’s experiences of sexual gratification and frustration at different stages of its development and, moreover, in relation to the different aims and objects of the infant’s or child’s sexual interest at those stages. For another, it enabled Freud to locate the psychic energy with which those thoughts and wishes were invested in a single source, namely the sexual instinct or libido, and consequently to unify psychoanalytic explanations under the general idea of tracing the transmission and discharge of this energy under conditions of repression. Nowhere did these developments in his theory prove more fruitful than in the elaboration of his famous idea of the Oedipus complex. The hypothesis of infantile sexuality created for Freud a problem that did not arise on the traditional view of children as asexual until puberty. Both the dramatic physical

6 5

Freud 65

and psychological changes that occur at puberty and the absence in adults of any memory of their having experienced prepubescent sexual impulses gave powerful support to this traditional view. Since Freud too regarded puberty as a time of sexual awakening and recognized the absence of such memories in adults, he faced the problem of how to square these phenomena with his hypothesis. He did so by assuming a period of sexual latency in childhood at the onset of which all memory of the prior impulses and experiences of infantile sexuality was repressed. This period ended with puberty, though the repression with which it began continued. The assumption fitted his hypothesis to the problematic phenomena, and the question then was what could bring this period of sexual latency about. What, that is, could bring about repression of such strength and scope as to block from coming into consciousness memories, thoughts, and wishes that have accumulated over several years and that carry the vital force of the libido? Clearly, to explain such repression required very frightful fantasies and wishes, and the Oedipus complex, with its fantasies and wishes concerning incest and parricide, met this requirement. Accordingly, Freud placed the Oedipus complex in the final stage of infantile sexuality. At this stage the child, on the simplest exposition of the Oedipus complex, sought sexual union with the parent of the opposite sex, and came to regard the parent of the same sex as a hated rival whom it wished to see dispatched. The emotional conflict within the child that these incestuous and parricidal wishes and their related fantasies reflected could not be sustained indefinitely, and its resolution, Freud proposed, came about through their sharp and lasting repression. Such mastery of the complex not only brought the period of infantile sexuality to an end and initiated that of sexual latency but also created the store of unconscious thoughts and wishes that were chiefly responsible for any psychoneuroses the child was liable to develop later if mastery of the complex was less than complete. The Oedipus complex, as Freud liked to say, was “the nucleus of the neuroses.”2 Two other principal outcomes that Freud attributed to the Oedipus complex signal a growing interest in general psychology. The first is the adult choice of a spouse and other lovers; and the psychoanalytic explanations Freud offered of such choices, while direct extensions of his theory of psychosexual development, also depended on the role parents have as their children’s first ideals. The second is the acquisition of moral sensibilities, specifically, a sense of guilt; and the psychoanalytic explanation Freud offered of this development depended on the child’s relations to its parents as authorities who set 2.  The Ego and the Id, SE v. 19, p. 32.

66

66  From Psychology to Morality

and enforced the bounds of right and wrong in its young life. In constructing these explanations, Freud increasingly focused on the processes by which individual ideals and moral precepts became internalized, and as he deepened his account of these processes, he found it necessary to revise the basic structural features of his theory. Specifically, he concluded that the representative of parental ideals and parental authority in the mind was neither entirely conscious nor entirely unconscious and thus was not happily located in either of the mind’s two parts. And concurrent with this conclusion, he realized that the ego itself was not entirely conscious and, indeed, that a satisfactory account of repression required that it be partly unconscious. The upshot of these points was a new model of the mind’s organization on which it was divided into three parts, ego, id, and superego. Indeed, Freud, from the perspective of this new model, characterized his original model as topographical and his new model as structural. The advent of the structural model, which received its fullest statement in Freud’s 1923 monograph The Ego and the Id, completed the shift in his interests from the study of the psychoneuroses and other disturbances of thought and behavior to the general study of human personality. In working out this new, structural model, Freud, for the reasons mentioned above, abandoned the idea of defining the mind’s parts according to the conscious or unconscious quality of their states and processes. Instead, he defined them by their functions. The ego’s functions corresponded to those it had on the old model, and the id’s corresponded to those of the unconscious. Thus the ego was the agent of such conscious activities as deliberation and decision-​making about how best to adapt to the environment, and also the unconscious action of repression, and the id functioned as the storehouse of repressed thoughts and wishes as well as the source of instinctual energy. The superego, then, which had no counterpart on the topographical model, was assigned the functions of conscience and maintaining the individual ideals to which the ego aspired. Repression remained a central dynamic in the explanations of psychoanalytic theory, but the three-​part division of the mind yielded a more complex account of this dynamic on which the ego, in effecting repression, acted under the direction and the pressure of the superego. This difference reflects the general thrust of the changes Freud made to the basic dynamical features of his theory:  conflict between the mind’s conscious and unconscious parts gave way, as the basis of psychoanalytic explanation, to more complex interactions among ego, id, and superego. At the same time as Freud was making these changes he was also revising his theory of the instincts. Initially, he had assumed two, those of sex and self-​preservation, and conceived of repression as a site of opposition between

6 7

Freud 67

them. That is, on his original conception of repression, the sexual instinct supplied the psychic energy with which unconscious thoughts and wishes were invested, and the self-​preservative instinct explained the activities of the ego in its efforts to block those thoughts and wishes from becoming conscious. Later, Freud saw that, on his way of understanding the instincts, the self-​preservative instinct might derive from the sexual one. For crucial to his understanding of the instincts, the sexual instinct in particular, is the idea that they do not have fixed objects. Indeed, their lability is necessary to Freud’s account of how, under repression, the psychic energy with which unconscious thoughts and wishes are invested is communicated and reinvested in new thoughts and wishes such as those that lie behind the symptoms of the psychoneuroses. The derivation, then, of the self-​preservative instinct from the sexual instinct followed directly once Freud realized that the ego itself could become invested with libidinal energy. Freud advanced this hypothesis of ego libido in his seminal paper of 1914, “On Narcissism: An Introduction,” and for the next several years he proceeded in his theoretical work on the assumption that libido was the sole instinct in human life. This assumption, however, clashed with his conception of repression as a site of opposing instincts, a conception he regarded as fundamental to his theory. He eventually resolved this difficulty with his postulation of a death instinct and his subsequent treatment of it as the source of the aggressiveness with which the superego pressed its demands on the ego. The postulation, by restoring dualism to Freud’s theory of the instincts, thus allowed him to give an account of repression that conformed to his fundamental conception of it. Freud’s abandonment of his topographical model of the mind also led him to rethink his account of anxiety. He had long distinguished between two types of anxiety, realistic and neurotic. The former, on his account, is a response to situations of real danger. These are situations in which the subject judges him-​or herself powerless to defeat some threat and recognizes that the injuries with which he or she is threatened would have a traumatic impact on his or her psyche. Such impact, Freud theorized, would correspond to the subject’s first traumatic experience, which Freud took to be that of being born. This first traumatic experience is one of severe, unpleasant respiratory and cardiac agitation and feelings of helplessness that leave a permanent impress in one’s psyche the recall of which is triggered by perceptions of real danger. Realistic anxiety, then, consists of similar but considerably less intense unpleasant sensations and feelings that one experiences in anticipation of danger. Accordingly, it is a fainter mnemonic reproduction of the first traumatic experience and serves to alert its subject to the dangerousness of his or her situation. Neurotic anxiety, on Freud’s account, was, as the

86

68  From Psychology to Morality

name suggests, anxiety associated with various neuroses. On Freud’s original account of such anxiety, while it was like realistic anxiety in its phenomenal and physiological character, it was unlike it in having little or no connection to situations of danger. Freud came to this conclusion about the difference between the two from observations he made of the anxiety his patients displayed or described during treatment. Some patients, he noted, suffered from a kind of free-​floating anxiety, anxiety that, as it were, attached itself inconstantly to different objects none of which posed any threat to the subject. Others related experiencing attacks of anxiety in the wake of prematurely terminated sexual activities, the more passionate the aborted activity the greater the resultant anxiety. Still others experienced anxiety when they were prevented from engaging in the obsessive or phobic behavior that was the chief symptom of their neurosis. Obsessive patients, for instance, suffered severe anxiety when prevented from indulging their obsession. Phobic patients suffered severe anxiety when confronted with the things that aroused their fear. Yet in neither case did the interference that provoked anxiety put them in any danger. Freud inferred from these and other cases that neurotic anxiety was a transformation of libido. More exactly, he inferred that sometimes, when libidinal energy is blocked from realizing its aim, as happens in repression, it is transformed into anxiety. It is, in these cases, a discharge that results from repression’s having failed to prevent traces of the libidinal instinct from reaching consciousness. Hence, his conclusion that neurotic anxiety was a different type from realistic anxiety. Freud kept to this account of anxiety until his last major theoretical work, Inhibitions, Symptoms and Anxiety. In that work, he reviewed, through the lens of his new, structural model, the evidence on which he had based his distinction between neurotic and realistic anxiety. As a result of this review, he significantly revised his understanding of the former. He had, however, already, in The Ego and the Id, laid the grounds for this revision. In that work’s last chapter, he examined the ego’s relations to the id and superego, noting first how the ego functions to gain control of the instinctual forces that emanate from these other parts of the mind and then considering its vulnerabilities in so functioning. “From the other point of view” (i.e., the point of view of the ego’s vulnerabilities), he wrote, “we see this same ego as a poor creature owing service to three masters and consequently menaced by three dangers: from the external world, from the libido of the id, and from the severity of the superego. Three kinds of anxiety correspond to these three dangers since anxiety is an expression of a retreat from danger.”3 With this observation Freud, in effect, 3.  Ibid., p. 56.

96

Freud 69

retracted his understanding of neurotic anxiety as different from realistic anxiety in its having little or no connection to danger. To the contrary, Freud here implied that neurotic anxiety was like realistic anxiety in being a signal of danger. Subsequently, in Inhibitions, Symptoms and Anxiety, he reaffirmed this point and went on to explain the repression behind neurotic anxiety as itself a response to what the ego took to be a situation of danger, danger that consisted in the threat of serious injury or deprivation the repressed libidinal demands portended owing to the threats the subject as a child imagined coming from its parents as a result of its Oedipal wishes. Freud, as the quotation from The Ego and the Id just given indicates, not only revised his understanding of neurotic anxiety, but also revised his general account of anxiety by adding a third type, moral anxiety. Anxiety of this third type signals the danger that the reproaches and beratings of conscience represent. It arises, therefore, only after the superego is formed, since conscience, being a function of the superego, is not acquired any earlier. In this respect, moral anxiety is like the sense of guilt, an affect to which one becomes liable only with the internalization of moral strictures that takes place through the formation of the superego. Freud relied on this point in distinguishing moral anxiety from neurotic anxiety. Both, as he observed, signal danger that arises in the child’s mind from its Oedipal wishes. And while the danger in either case is traceable to the serious injury or deprivation with which the child, when subject to the Oedipus complex, imagines being threatened, its locus is different in the two cases. In the case of neurotic anxiety, its locus is the libidinal impulse to engage in incestuous sexual intimacy, owing to the grave consequences of gratifying that impulse. Neurotic anxiety manifests fear of libido. In the case of moral anxiety, the locus of danger is conscience, owing to the punishment for disobedience with which conscience threatens the ego. Moral anxiety manifests fear of the superego. Accordingly, Freud observes, if the dissolution of the Oedipus complex is normal, the impulse to engage in incestuous sexual intimacy vanishes, and the child grows up free of the traumatic neuroses Freud attributed to the repression of the libidinal instinct and therefore free of neurotic anxiety. By contrast, the normal dissolution of the Oedipus complex no less than any pathological dissolution results in the formation of a superego. The superego, once formed, is in either case a permanent fixture of human psychology. Hence, the child becomes liable to moral anxiety whether or not it later develops some traumatic neurosis. A second way in which moral anxiety differs from neurotic anxiety is in the nature of the threat that excites it. In neurotic anxiety libidinal impulses represent a threat to the subject because of an association they have in his or her mind with serious injury or deprivation. They represent a threat just as a

07

70  From Psychology to Morality

drill in the hands of a dentist represents a threat to a patient who experiences realistic anxiety at the sight of the drill because of an association in his or her mind with intense pain from dental work experienced as a child. By contrast, in moral anxiety the reproaches of conscience not only represent a threat to the subject because of an association in his or her mind with serious injury or deprivation but are also in themselves palpably threatening in virtue of the severity with which conscience delivers them. The harshness of the superego in pressing its demands on the ego, which reflects its being the repository of the aggressive energy of the death instinct, makes it in itself an intimidating object. Moral anxiety is thus a mixture of anxiety due to the ego’s anticipation of punishment if it disobeys the superego and anxiety provoked by confrontation with a harsh and demanding overseer. Freud first postulated the death instinct in his 1920 monograph Beyond the Pleasure Principle, and then, in the final chapter of The Ego and the Id, he examined its role in the superego’s relations to the ego and the sense of guilt that characterized those relations. This chapter combines the fundamental changes Freud made to the mind’s organization with the revisions he made to his theory of the instincts, and its theoretical brilliance amply demonstrates the great fecundity of these changes. Their fecundity was, then, further confirmed in Inhibitions, Symptoms and Anxiety by the greater coherence and clarity he brought to his account of anxiety in reworking it in light of these changes The changes he made in this work were, with the exception of a modification to his account of the death instinct (discussed below), the last major changes he made to his theory.

3.  The Philosophical Import of Freud’s Theory Freud’s theory belongs to the tradition of naturalism in Western thought. The tradition stretches back to antiquity. Its founders are several pre-​Socratic philosophers, most notably the atomists, who advanced programs for understanding all phenomena as the products of natural forces. This tradition, after disappearing into the darkness of the early medieval period, began its re-​emergence with the rediscovery and acceptance of Aristotle in the twelfth century. It then fully resurged three centuries later with the rise of modern science. Accordingly, the naturalist ideal has become that of bringing all that is observable in nature within the scope of the natural sciences. Freud, who developed psychoanalysis as a natural science, repeatedly identified it as a contribution to the realization of this ideal. His theory, to be sure, was not the first in the modern period to explain the workings of the human mind as

71

Freud 71

entirely the products of natural forces. Both Hobbes and Hume, for instance, expounded comprehensive naturalistic theories of human psychology. But it surpassed these predecessors in its explanations of the phenomena that opposing traditions regularly advanced as proving the untenability of naturalism in psychology. Above all, it surpassed them in confronting the objections of Christian spiritualism, the dominant tradition in Western philosophy. Indeed, the apparent success of psychoanalysis in dispatching the doctrines of this tradition, doctrines about the essential mysteriousness to science of the elevated aspects of human thought and conduct, secured its claims to having advanced the naturalist ideal farther than any previous scientific theory of the mind. The advantages that Freud’s theory has over its predecessors within the tradition of naturalism derive from its structural features. First, its conception of the mind as divided into parts—​tension between which produces various thoughts, feelings, and behavior—​introduces dynamic relations within the mind that have no counterparts in earlier naturalistic theories. The introduction of these relations enabled Freud to give cogent explanations of inner conflicts—​the struggle against sin, say, or the combat of reason and passion—​that spiritualists since Plato had adduced as evidence against a wholly naturalistic understanding of human psychology. Secondly, the theory’s assumption of unconscious thoughts and wishes whose psychic force affects conscious thought and feeling gives depth to its explanations that is alien to theories like Hobbes’s and Hume’s. And the possibility of these deeper explanations enabled Freud to treat the civilized motives that people offered as explanations of their own behavior and that the opponents of naturalism regarded as revealing a dimension of human existence unaffected by natural forces as surface movements whose underlying causes sprang from the more primitive drives of animal life and thus qualified fully as forces of nature. Of these two advantageous features, the second is generally regarded as Freud’s most original contribution. Yet arguably the first is the bolder innovation. The bare assumption of unconscious thoughts and wishes had precedents in other psychological work, precedents that Freud readily acknowledged, and its germ is already evident in such devices as Hume’s notion of a calm passion. By contrast, the idea of incorporating into a naturalistic theory of psychology the conception of a divided mind the parts of which were sources of conflicting motivational forces was wholly new. Before Freud such a conception was the intellectual property of spiritualism. Earlier naturalist theories put forward conceptions of the mind as either an undivided field of mental states and activity or as divided into parts each of whose

27

72  From Psychology to Morality

workings never conflicted with those of the other parts.4 Freud, therefore, one might say, raided the spiritualist tradition’s domain of one of its most valuable possessions when he incorporated its way of conceiving the mind as divided into his theory. Before Freud, disputes between naturalism and spiritualism in the modern period regularly turned on the question of the mind’s structure, specifically whether it was unified or divided, for the tendency of naturalist thinkers like Hume was to identify the mind with a single, uniform field and to organize the great variety of mental phenomena in human life around a few simple principles that operated co-​ordinately in this field. With Freud’s innovation, the terms of this dispute changed radically. Of course, a conception of the mind as divided is not essential to Christian spiritualism. Paul, after all, is likely to have had no such conception in mind when he warned of the contrariety between spirit and flesh,5 and Descartes, who made spiritualism a cornerstone of his philosophy, can plausibly be read as adhering to Pauline doctrine in maintaining the unity of the mind and its absolute division from the body. Still, Descartes’s position was not that far removed from a conception of the mind as divided, his assertions to the contrary notwithstanding. For to preserve the dualism in human nature that is essential to Christian spiritualism, Descartes had to give a novel account of human weakness.6 On this account, such weakness involves direct conflict between mind and body in which the mind’s power to move the body yields to the automatic forces at work in the body’s animal movements, whereas on the standard account, it involves conflict internal to the mind in which the force of reason yields to that of desire or passion. And the obvious correspondence between these two accounts makes it clear that Descartes was in fact applying the model of a divided mind even while denying its truth. The irony is that one can see his account of human weakness, once the model he was applying becomes apparent, as containing its own version of unconscious thoughts and wishes. A more orthodox defense of Christian spiritualism in modern philosophy is found in Kant’s work, his ethics in particular. Kant divided the mind into

4.  Among the ancients, the former conception belonged to the theories of the atomists and Epicureans. Modern thinkers, like Hobbes and Hume, followed their lead. The latter conception is Aristotle’s. It did not, as a naturalist conception, survive the rise of modern natural science. Its teleological tenets ensured its obsolescence as a scientific theory. It has, however, survived within Christian spiritualism owing to the way Aquinas transformed it into a theory that Christian theology could embrace. 5.  Galatians 5:17. Paul is likely to have drawn on the Stoic conception of the mind as a unitary whole in attributing sinful inclinations to the flesh. 6. Descartes, The Passions of the Soul, pp. 44–​46.

73

Freud 73

two faculties, reason and desire. Typically, the two operate harmoniously, as they do, for instance, when reason shows one how best to satisfy some desire. Hence, they typically cooperate in the formation of plans and schemes for pursuing happiness. Sometimes, however, they oppose each other, and to exemplify their opposition Kant turned to occasions when one desires to do something that is forbidden by a moral principle whose validity is indubitable. Analysis of such conflicts, Kant argued, shows that the constraint on desire one experiences must be due to the force of reason as it operates in one’s recognition of the principle’s validity. Hence, Kant concluded, the principles of morality are categorical dictates of reason, and moral motivation is an expression of the practical force of reason in the lives of human beings. That this force sometimes opposed the force of desire and that neither was resolvable into the other meant that the two forces were absolutely distinct. And from this conclusion Kant’s division of the mind followed immediately.7 Kant then built upon this division a metaphysics of clearly spiritualist inspiration. Human beings, he maintained, belong to two worlds. On the one hand, they belong, in virtue of possessing the faculty of desire, to the natural world. The faculty comprises the natural appetites and passions that humans share with other animals, and its operations are determined by the forces of nature. On the other, human beings belong, in virtue of possessing reason, to the intelligible world. Because the operations of reason when it opposes desire are disconnected from natural appetites and passions, they occur spontaneously, undetermined by the forces of nature. Thus the faculty of reason implies a capacity for freedom that distinguishes human beings from other animals. In exercising this freedom, human beings act from principles whose validity their own reason affirms rather than on impulses whose stimulation occurs regardless of reason’s approval. Their choices, in other words, are determined by the legislative processes of their reason rather than by the external stimulation of their appetites and passions. As inhabitants of the natural world, humans live as integral parts of a system of nature. As inhabitants of the intelligible world, they live as independent members of a community of rational beings, a kingdom of ends. These metaphysical doctrines, which echo similar doctrines in the works of Plato and Augustine, secure Kant’s place in the spiritualist tradition of Western philosophy. As long as the exponents of naturalist theories conceived of the mind as undivided, their responses to Kant’s challenge were limited to various attempts to resolve moral motivation into natural desire. They then faced the following difficulty. Natural desires, the desire for warmth, the desire for 7. Kant, Critique of Practical Reason, pp. 28–​30.

47

74  From Psychology to Morality

food, the desire for companionship, and so forth, while they might differ in intensity or felt urgency, did not differ in rank or felt authority. Moral motivation, by contrast, when it clashed with some natural desire, was experienced as a constraint on the latter as if from a superior force. How, then, could one resolve moral motivation into natural desire without loss of its felt superiority? The question stymied naturalists who conceived of the mind as undivided, but it posed no similar difficulty for Freud. Freud’s mature theory especially, with its three-​part division of the mind and its revised accounts of anxiety and the instincts, could answer Kant’s challenge directly, and it did so quite powerfully. The answer consists in Freud’s explanation of how the phenomena of conscience ultimately derive from the same primitive drives that are the sources of natural desires. The experience of constraint on natural desire as if from a superior force belongs to these phenomena. Freud characterized all of them as manifestations of the sense of guilt. To have ‘a bad conscience’, to be troubled by one’s conscience, to suffer compunctions of conscience, all result from one’s having either done something bad or wished to do something bad. Each expresses tension between the ego and the superego, tension that is modeled on tension between subject and governing authority. Thus, Freud identified the sense of guilt generally with such tension. Accordingly, the features of moral constraint that Kant invoked in his argument for the irresolvability of moral motivation to natural desire are found in the sense of guilt. The felt superiority of moral motivation to natural desire is captured by the superego’s authority over the ego, and the opposition between moral motivation and natural desire is captured by the harshness with which the superego treats the ego. Freud then accounted for both of these features through his explanation of how the superego is formed, which is to say, his explanation of how the child masters the Oedipus complex, and his explanation of the superego’s subsequent persistence in treating the ego harshly. The first feature results directly from the internalization of parental authority that coincides with the introjection of the hated parent of the Oedipus complex. The second feature results from what happens to the hostility the child has towards this parent. This hostility, being directed at a figure who is loved as well as hated, is emotionally difficult to sustain. To resolve the difficulty the aggressive energy powering the hostility is redirected. Specifically, the child invests it in the internal authority, the superego, that is erected through introjection of the hated parent. The harshness of the superego in its treatment of the ego is thus primarily explained as an expression of the aggressive energy with which the superego in its formation was instilled. The strictness with which the hated parent, on which the superego is modeled,

75

Freud 75

treats the child is then at most a secondary explanation. Finally, the superego receives fresh investments of aggressive energy every time the child’s hostile impulses towards its parents or others who inherit their authority are thwarted. This account of the superego’s harshness has an obvious affinity to Nietzsche’s conjecture, advanced in the second essay of On the Genealogy of Morals, about the origins of bad conscience in man. Indeed, Freud’s overall account of how the superego is formed and how it draws its power from the aggressive instincts—​more exactly, the death instinct—​constitutes a naturalistic answer to Kant that was already budding in Nietzsche’s wry remark in that essay about the smell of cruelty in the Categorical Imperative.8 The success of Freud’s answer depends on the cogency of his explanation of how a child develops a sense of guilt. Specifically, it depends on the intelligibility of his making the formation of a superego a necessary condition of that development. Critics of Freud who have wanted to preserve Kant’s account of morality as being solely the work of reason have denied that Freud’s account of the formation of a superego explains or is needed to explain how human beings become liable to a sense of guilt. One such critic bases his criticism on Freud’s identification of a sense of guilt with moral anxiety.9 Anxiety, this critic argues, is no more a moral feeling when labeled ‘moral’ than when labeled ‘realistic’ or ‘neurotic’. It is in all cases a signal of a danger or threat to one’s well-​being, and as such it prompts self-​defensive, which is to say, self-​ interested or egoistic action and not moral action. Another critic holds that one can understand the child’s development of a sense of guilt independently of its forming a superego.10 In particular, he argues, one can understand it, compatibly with Kant’s ethics, as an intellectual process the materials for which figure in the formation of a superego but are present before it takes place. Either of these criticisms, if sound, would show Freud’s explanation of the sense of guilt to have fallen short of vindicating a naturalist answer to Kant’s challenge. If sound, the criticisms would show either that the explanation fails to account for the child’s acquisition of moral sensibilities or that it fails to anchor those sensibilities in natural desires as opposed to intellectual processes. Both criticisms, however, misfire. The first errs in overlooking the difference Freud highlighted between moral anxiety and the other two types. It reaches its conclusion by treating all three types of anxiety as equally signals of danger to the subject’s well-​being and then assuming that the only differences among them are in the sources 8.  Nietzsche, “Guilt, Bad Conscience, and the Like” in Genealogy of Morals, p. 65. 9.  See David Jones, “Freud’s Theory of Moral Conscience,” p. 51. 10.  See J. David Velleman, “A Rational Superego,” pp. 550–​558.

67

76  From Psychology to Morality

of the danger being signaled. But moral anxiety differs from the other two types in manifesting what Freud called “a need for punishment,”11 and this difference implies a difference in its motivational character as compared with the motivational character of realistic or neurotic anxiety. The danger a child who has not yet developed a superego knows it has placed itself in when it misbehaves is the punishment with which its parents threaten it for such behavior. The anxiety the child experiences as a signal of this danger is realistic, and the child is then motivated to avoid the danger by evading detection, by hiding from its parents its having misbehaved. Once the child develops a superego, however, it cannot avoid suffering pangs of conscience when it misbehaves, for it cannot hide its having misbehaved from its superego. To quiet its conscience and regain peace of mind requires submitting to punishment, seeking forgiveness, offering reparations, and the like. Hence Freud’s characterization of the sense of guilt as expressing a need for punishment. Moral anxiety is thus aptly understood as a moral emotion, for the actions it prompts entail acknowledgment either that one has acted badly and must submit to punishment (or doing an equivalent penance) to put things right or that one desires to act badly and must suppress that desire to keep peace with one’s conscience. The error in the second criticism is also due to the critic’s failure to grasp the import of Freud’s characterizing the sense of guilt as an expression of a need for punishment. Freud’s explanation of how a child develops a sense of guilt begins with the child’s initial love of its parents. This love entails the child’s identifying with its parents and, as a result, coming to idealize them. Such idealization means that the child sets its parents up as ideals, and the second critic uses this factor in Freud’s theory to make his case for the child’s developing a sense of guilt before it forms a superego. Once the case is spelled out, however, its shortcoming should, in light of the first criticism’s deficiency, be evident. The case consists in explaining the child’s development of a sense of guilt as resulting from a change in its understanding of its relations to its parents owing to its having set them up as ideals. The child’s parents, as ideals, are objects of its admiration, and consequently it regards them as exalted figures whose exercise of power in disciplining it is justified. The child sees its parents, in other words, as having moral authority and understands the pain or deprivation they inflict in enforcing their demands as punishment that is its due on account of its having disobeyed them. As the child acquires this 11.  “The Economic Problem of Masochism,” SE v. 19, p. 166.

7

Freud 77

understanding of its parents as having moral authority over it and of the pain or deprivation they threaten for misbehavior as punishment, the emotion that such misbehavior stirs undergoes a change as well. Before setting its parents up as ideals, the child views them as powerful figures whose threats represent the use of brute force to bring about compliance with their demands. Accordingly, they are objects of fear and anxiety, and the child experiences such fear and anxiety acutely when it misbehaves. When the child’s view of them changes and it now sees them as having moral authority over it and as exercising that authority in threatening punishment, the anxiety the child experiences when it misbehaves acquires a moral content. It is anxiety about having disobeyed moral authority and being thereby subject to punishment. Hence, the child develops a liability to moral anxiety and, by the same token, a sense of guilt. And because this entire process can take place independently of the formation of a superego, the latter is unnecessary to the development of a sense of guilt. Freud, therefore, so the second criticism concludes, failed to answer Kant’s challenge. Plainly, the criticism goes wrong in identifying as moral anxiety what Freud would have taken to be realistic anxiety. The point here is not the trivial one that ‘moral anxiety’ is a technical term in Freud’s theory that applies only to anxiety signaling danger coming from the superego. Rather it is the substantive one that anxiety that signals danger coming from an external source motivates one to avoid the danger or retreat to safety if possible, given that one cannot overpower and subdue its source. When anxiety signals danger coming from the superego, one is not so motivated, for it is neither possible to avoid the danger—​to hide from the superego or keep it from detecting one’s misbehavior—​nor possible to overcome and subdue its source. Instead, one finds relief from the anxiety, from one’s tormenting conscience, by submitting to punishment or doing some other form of penance that requires suffering. Accordingly, the anxiety manifests a need for punishment. What the second criticism characterizes as moral anxiety does not manifest such a need. To the contrary, it still has the motivational structure of anxiety that signals danger coming from an external source and accordingly moves the child to avoid the danger or retreat to safety, given that the child cannot overpower and subdue the source of the danger. Indeed, it has this motivational structure whether the child understands the danger as punishment at the hands of elevated figures who have moral authority or as harm inflicted by powerful figures who use brute force to induce compliance. Hence, the second criticism fails to show how a child could develop, independently of its forming a superego, a sense of guilt and a liability to moral anxiety as Freud conceived these states.

87

78  From Psychology to Morality

By replacing Freud’s conceptions of a sense of guilt and moral anxiety with different ones, it leaves Freud’s answer to Kant’s challenge untouched.

4.  A Modification at the Last Stage Freud did not explicitly affirm his idea of the sense of guilt as an expression of a need for punishment when he first introduced his structural model of the mind. At the same time, he did make use of it to explain the inordinate resistance to recovering their health that certain of his patients displayed during the course of treatment. It was, Freud observed, as if they needed to remain ill. This need, was evident in the satisfaction they got from continuing to suffer. Their continued suffering satisfied an unconscious sense of guilt, he concluded.12 It was as if it brought them relief from a burden of guilt they bore. Freud returned to this observation a year later in “The Economic Problem of Masochism.”13 In this paper he introduced the expression ‘need for punishment’ as a useful substitute for ‘unconscious sense of guilt’. The substitution, he thought, helped to illuminate what he meant by the latter. It helped, that is, to remove the air of self-​contradiction that the latter, particularly in its original German, ‘unbewusstes schuldgefüh’, carried.14 It also clarified how citing an unconscious sense of guilt served to explain ways in which patients seemingly invited gratuitous pain into their lives and harm to themselves. What Freud called moral masochism was a singular case of this phenomenon. Structurally, Freud pointed out, an unconscious sense of guilt was indistinguishable from a conscious one. Both expressed tension between the ego and superego. What interested him about the unconscious sense of guilt was the strength of its influence on his patients. The extent to which they suffered on account of it was an indication of the harshness with which the superego treated the ego. That harshness, while partly due to the superego’s being modeled on strict parental authority as the child experienced or imagined it, was largely the expression of the aggressive energy that was invested in the superego. Freud attributed this energy to the death instinct, once its 12.  “The Ego and the Id,” SE v. 19, pp. 49–​50. 13.  See “The Economic Problem of Masochism,” SE v. 19, p. 166. 14.  A literal translation of the German expression is ‘unconscious feeling of guilt’. Freud was acutely aware of the self-​contradiction apparent in talk of an unconscious feeling. Ibid. For his most extensive discussion of the general use in psycho​analysis of references to unconscious feelings and emotions see ‘The Unconscious’, section III, SE v. 15, pp. 177–​179. See also my essay “Emotions: The Legacy of James and Freud,” pp. 1249–​1252.

97

Freud 79

destructiveness had been diverted outward and targeted on external objects. Yet because he also held that the death instinct operated jointly, as a silent partner, with libidinal instincts, he had to suppose an end to the partnership as part of the dissolution of the Oedipus complex. The reason is that to explain how the dissolution brought a temporary end to the child’s sexuality and initiated a period of sexual latency required taking the continued love the child had for either parent to be no longer erotic, and such desexualization entailed that the death instinct as provider of the aggressive energy with which the superego was invested no longer operated jointly with libidinal instincts. The investment of aggressive energy in the superego at its formation was, as Freud said, due to “a defusion of instinct.”15 Freud then explained moral masochism as a condition of one’s getting satisfaction from punishment and suffering not only as a result of their satisfying one’s need for punishment (i.e., not only because they brought one relief from a burden of guilt that one bore unconsciously) but also as result of the suffering’s having an erotic character. The satisfaction, in other words, was libidinal as well as anxiety reducing. Moral masochism, Freud declared, was thus a significant piece of evidence of the fusion of the two instincts. Indeed, Freud, as far back as Beyond the Pleasure Principle, the work in which he introduced the hypothesis of a death instinct into his theory, took the phenomena of sadism and masochism as the prime evidence in support of this hypothesis.16 He had otherwise arrived at it theoretically. Accordingly, his thesis of the fusion of the two instincts, libido and death, directly followed from his attributing such evidential import to these phenomena. Similarly, given the primacy of sadism, his further thesis that the death instinct, in fusing with libido, became redirected outward and so converted into aggression that operated on external objects also followed. Freud was, of course, aware of the skepticism his hypothesis provoked, and his characterizing the death instinct as a silent partner in its cooperation with libido helped to keep these critics at bay. As long as he conceded the lack of direct evidence of the death instinct’s operating independently of libidinal instincts, he could preserve his hypothesis, even if only as a tentative one. When he took up the question of human aggression and the efforts of society to contain it in Civilization and Its Discontents, however, his position changed. He ceased to treat the hypothesis as tentative and lacking support from evidence independent of the erotic entanglements and disentanglements in how the death instinct 15.  “The Economic Problem of Masochism,” SE v. 19, p. 167. See also The Ego and the Id, pp. 54–​55. 16.  Beyond the Pleasure Principle, SE v. 18, pp. 53–​55.

08

80  From Psychology to Morality

operated. “I know,” he wrote at the point at which his position changed, “that in sadism and masochism we have always seen before us manifestations of the destructive instinct (directed outwards and inwards), strongly alloyed with erotism; but I  can no longer understand how we can have overlooked the ubiquity of non-​erotic aggressivity and destructiveness and can have failed to give it its due place in our interpretation of life.”17 This change allowed Freud to simplify his explanation of the origin of a conscience and to clarify his idea of the sense of guilt as an expression of a need for punishment. By treating the aggressive instincts as operating independently of the libido, Freud could explain the origin of a conscience without involving any erotic attachment that the child had to a parent. In particular, he could explain it without bringing in the standard Oedipus complex in which the child has incestuous desires for sexual relations with one of its parents and parricidal desires generated by jealousy of the other, who, as a rival, stands in the way of its satisfying those incestuous desires. Simply citing the child’s ambivalence towards its parents that consists in, on the one hand, the love it develops for them owing to its helplessness and dependence on them for protection and provision of its needs and, on the other, the hatred it develops towards them owing to their preventing it from satisfying its urges and strivings—​ including especially the hostile ones the hatred generates—​was sufficient. For such ambivalence is, like the ambivalence central to the Oedipus complex, unsustainable, and Freud could then cite the same mechanisms of identification and introjection to explain its resolution. At the same time, because the objects with which the mechanisms in this case deal are the parents as external authorities who have thwarted the child’s urges and strivings rather than the parent who as a sexual rival stands in the way of the child’s satisfying its Oedipal desires, Freud could understand the introjected figure as solely an internal authority to be obeyed rather than also an ideal to be looked up to. Accordingly, Freud could simply explain the figure’s harshness in its treatment of the ego as due initially to the investment in it of the store of vengeful hostility towards the parents that had built up in the child at the time of introjection. He did not need to explain it as also due to the way the child masters the Oedipus complex by identifying with and introjecting the powerful parent who, as sexual rival, stands in the way of its satisfying its Oedipal desires. Freud, then, as his attention was drawn increasingly to the contribution of the death instinct to the workings of conscience and the sense of guilt,

17.  Civilization and Its Discontents, SE v. 21, pp. 119–​120.

81

Freud 81

came to offer a different explanation of how the child, through the formation of a superego, develops a sense of guilt from the explanation he advanced when he introduced his new, structural model into his theory. This later explanation, unlike the original, is narrowly focused on the acquisition of a conscience, an authoritative organ of moral directives and criticisms that serve to check the aggressive drives in human beings whose destructive force threatens social peace and harmony. And it highlights more than the original how the destructive force of these drives is redirected from its outward orientation back onto the individual whose aggressive instincts are its source. Accordingly, one could characterize the later explanation as a substantial elaboration of Nietzsche’s conjecture about the origins of bad conscience that the original explanation contained embryonically. The elaboration is carried out in the last two chapters of Civilization and Its Discontents once Freud in the chapter preceding them came to acknowledge the great extent to which the death instinct operated independently of libidinal instincts.18 The view he articulated represents his most cogent answer to Kant’s challenge. It took Freud ten years from his introduction of the death instinct into his theory to reach this view. The question that he then faced, in light of his attributing to it, as an independently operating instinct, so much human aggression was how human beings could live together peaceably given this unbounded natural drive towards hostility. How did civilization sustain itself? The answer, Freud held, lay in the way civilization, conceived of as a natural process by which social relations among human beings are fostered and regulated, turned these aggressive instincts in human beings to its own advantage. It did this principally through the formation of the superego, which, as the internal agent of morality, as conscience, drew its power from those drives. The real work of the superego, Freud held, was therefore to thwart and deflect the aggressive instincts so that they did not realize their socially destructive potential. As counterpoint to Kant’s notion of conscience, this idea could not have been more apposite. Where Kant saw conscience as a form of intelligence that transported its possessor beyond the determinations of nature, Freud saw it as a device that nature implanted in human beings for her own purposes. And where Kant celebrated the transcendental freedom that the possession of a conscience brings, Freud regretted the unhappiness that its possession made inevitable, the unhappiness that nature made the lot of humankind as the price of being civilized.

18.  See “Remarks on Some Difficulties in Freud’s Theory of Moral Development,” essay 4 in my The Sources of Moral Agency for further discussion of these points.

28

82  From Psychology to Morality

The pessimism that runs through the argument of Civilization and Its Discontents reflects the greater role the death instinct had acquired in Freud’s theory. It marked as well the declining condition of his health and of the political situation in Europe. In 1938 the Nazis invaded Austria, and Freud was forced to leave Vienna. He went briefly to Paris and then to London, where he died the following year.

Bibliogra p h y Deigh, John. The Sources of Moral Agency:  Essays in Moral Psychology and Freudian Theory (Cambridge: Cambridge University Press, 1996). ———​. “Emotions: The Legacy of James and Freud,” International Journal of Psychoanalysis 82 (2001): 1247–​1256. Descartes, Rene. The Passions of the Soul (1649), tr. S. Voss (Indianapolis, IN: Hackett, 1989). Freud, Sigmund. Das Unbehagen in der Kultur (Wien: Internationaler Psychoanaltyscher Verlag, 1930). ———​. The Standard Edition of the Complete Psychological Works of Sigmund Freud (SE), ed. James Strachey (London: Hogarth Press, 1953–​1971). ———​. The Origins of Psychoanalysis:  Letters to Wilhelm Fleiss, Drafts and Notes, 1887–​1902, ed. Marie Bonaparte, Anna Freud, and Ernst Kris (New  York:  Basic Books, 1954). Hobbes, Thomas. Leviathan (1649), ed. Richard Tuck (Cambridge: Cambridge University Press, 1996). Hume, David. A Treatise of Human Nature (1739–​ 40), ed. L. A. Selby-​ Bigge (Oxford: Oxford University Press, 1888). Jones, David. “Freud’s Theory of Moral Conscience.” Philosophy 41 (1966): 34–​57. Jones, Ernest. The Life and Work of Sigmund Freud (New York: Basic Books, 1957). Kant, Immanuel. The Critique of Practical Reason (1788), tr. L. W. Beck (Indianapolis, IN: Bobbs-​Merrill, 1956). Neu, Jerome (ed.). The Cambridge Companion to Freud (Cambridge:  Cambridge University Press, 1991). Nietzsche, Fredrich. On the Genealogy of Morals (1887), tr. W. Kaufmann and R. J. Hollingdale (New York: Random House, 1967). Velleman, J. David. “A Rational Super​ego.” Philosophical Review 108 (1999): 529–​558. Wollheim, Richard. Sigmund Freud (Cambridge: Cambridge University Press, 1971).

83

4

Psychopathic Resentment

Hitchcock, in Shadow of a Doubt, introduced a new kind of villain into his films. His previous villains, while conventionally ruthless and sinister, were largely plot devices. His most successful earlier films were about espionage, reflecting the growing political tensions in Europe at the time, and the villains in those films were spies and assassins whose intrigues created the suspense for which Hitchcock had become famous. In Shadow of a Doubt, by contrast, the suspense comes from the villain himself: one Charles Oakley, masterfully played by Joseph Cotton. Oakley, to his sister and her family, is the worldly Uncle Charlie, a man who exudes boyish charm, displays impeccable manners, and is attentive to all in his company, especially the women. And they in turn adore him. He is also, as we see in the opening scenes, in flight from the law, and as we later learn, a serial killer who preys on widows and who has sought refuge in the small California town of Santa Rosa, where his sister and her family live. As the film progresses and Oakley comes under the suspicion of his niece, he becomes increasingly secretive, manipulative, and menacing towards her. Slowly we come to realize that this charming, debonair fellow connects with no one and has no conscience. He is Hitchcock’s first major attempt at exploring the chilling personality of a psychopath.

83

48

84  From Psychology to Morality

Hitchcock continued this exploration in Rope, which is loosely based on the infamous Leopold and Loeb case,1 and then in Strangers on a Train, which is based on Patricia Highsmith’s eponymous novel. The villain in the latter film, Bruno Anthony, is, if anything, an even more gripping example of a psychopathic killer. Robert Walker, who gives a brilliant performance as Bruno, endows the character with a naive egocentricity, like that of a child, and in doing so bids us to see his lack of conscience and inability to connect with others as products of a severely arrested moral development. Bruno, who has murdered the estranged wife of the film’s protagonist, Guy Haines, in the belief that he and Guy have entered into a compact in which in return for his killing Guy’s wife, Guy would kill Bruno’s father, becomes enraged when it becomes evident to him that Guy has no intention of killing his father. Feeling double-​crossed, Bruno seeks revenge. The last part of the film is then taken up with his repeatedly frustrated attempt to return to the scene of the murder so as to plant evidence against Guy and Guy’s attempt to stop him. Bruno’s displays of resentment and anger in this part of the film seem entirely in keeping with his personality. They match, in this respect, the resentment and bitterness Oakley exhibits in a scene in which he suddenly and vehemently denounces the world as a “foul sty” and belittles his niece as a know-​ nothing. It is, I  believe, no accident that Hitchcock made the capacity for these volatile, retributive emotions part of both villains’ personality profile. While the anger and resentment these villains exhibit seems entirely in place in Hitchcock’s portrayal of them, their having a capacity for these emotions may nonetheless seem incongruous to those who are familiar with what has become the standard view in Anglo-​American philosophy about their nature. Resentment, in particular, is now commonly understood as an emotion a person cannot experience unless he or she is capable of moral judgment, and indeed rather complex moral judgment. Stephen Darwall’s recent treatment of the emotion in his book The Second-​Person Standpoint is representative.2 Darwall writes, “Resentment  .  .  .  is felt in response to apparent injustice, as if from the victim’s point of view. We resent what we take to be violations against ourselves or those with whom we identify. If you resent someone’s treading on your foot or, even more, his rejecting your request or demand that he stop doing so, you feel as if he has violated a valid claim or demand and as if some claim-​exacting or responsibility-​seeking response by you, or on your behalf, is justified.”3 What makes such accounts 1.  The film is an adaptation of Patrick Hamilton’s play Rope. 2.  Darwall (2006). See also G. Watson (1988), and R. J. Wallace (1996). 3.  Darwall (2006, 68).

85

Psychopathic Resentment  85

of resentment seem ill suited to the resentment displayed by the villains in Hitchcock’s films is not so much the complexity of the judgment that the emotion is represented as containing as the moral orientation that they presuppose. It is not an orientation one would think was available to someone like Oakley or Bruno. It is not, one would think, the orientation of a psychopath. Yet it is essential to these accounts. On Darwall’s account, for instance, resentment is one of a set of emotions that he calls, following P. F. Strawson, reactive attitudes.4 The set also includes gratitude, moral indignation, guilt, and forgiveness, and to have the capacity for these emotions one must, according to Darwall, see oneself and others as members of a moral community in which every member has authority to address any other, to make obligation-​creating claims on him or her, and to hold that person responsible for failures to meet those obligations.5 Accordingly, the injustice to which or to the appearance of which Darwall characterizes resentment as a felt response consists in a failure to meet some such obligation. And similarly for the two other emotions, moral indignation and guilt, that Darwall takes to be felt responses to injustice. What differentiates each of these emotions from the other two, then, is the viewpoint from which the injustice is seen. Thus, when one feels resentment towards another, one takes oneself to have been treated unjustly by him or her; when one feels moral indignation towards another, one takes the person to have acted unjustly towards a third party on whose behalf one is taking this retributive attitude; and when one feels guilt, one takes oneself to have acted unjustly towards another. Consequently, on Darwall’s view, to be liable to any of these emotions is to be capable of making moral judgments that reflect a view of oneself as joined together with others in mutual relations defined by reciprocal rights and obligations.6 And this means, given the capacity of psychopathic killers like Hitchcock’s villains for resentment, that they too are capable of making such moral judgments. Yet the lack of a conscience and inability to connect with others that are principal marks of this type of personality imply that they are not.7 Explaining how such killers could have a capacity for resentment consistently with Darwall’s account of the emotion may not, however, seem all that problematic. For it requires explaining how they can regard themselves as victims of injustice though they lack a conscience and are unable to connect 4.  P. F. Strawson (1962). All references to the reprint. 5.  Darwall (2006, 67). 6.  Ibid., 66–​68. 7.  See R. D. Hare (2003).

68

86  From Psychology to Morality

with others, and one might give such an explanation by appealing to their egocentricity. After all, being egocentric does not obviously imply that one is incapable of judging that another has treated one unjustly, even if there is no question that it implies that one lacks a conscience and cannot connect with others. Accordingly, since the primary evidence for taking the lack of a conscience as a principal mark of psychopathy is that psychopaths exhibit no guilt or remorse for their crimes and since the absence of these emotions is also evidence of their inability to connect with others, it may suffice for understanding Darwall’s account as consistent with acknowledging the capacity of psychopaths for resentment to point out that a capacity for resentment requires only that a person be capable of making moral judgments from his viewpoint. Hence, that psychopaths are incapable of making moral judgments from others’ viewpoints need not be a problem on Darwall’s account. Or so one might argue. Of course, if one makes this argument, one must abandon the common characterization of psychopaths as incapable of moral judgment. But this departure from settled opinion may not seem that worrisome. As long as it is understood that psychopaths are incapable of taking the viewpoints of others and making moral judgments from those viewpoints, one might think that it is possible to attribute to them, consistently with accounts of resentment like Darwall’s, a limited capacity for moral judgment. Thinking this, however, would be a mistake. On such accounts, the capacity for moral judgment cannot be limited in this way. The reason is that the capacity for moral judgment that these accounts presuppose is a capacity for judgments that reflect an understanding of oneself as belonging to a moral community in which one’s relations to the other members are relations of reciprocal obligation. Because of this fact, one cannot, according to these accounts, judge, when another fails to meet an obligation he owes one, that he has acted unjustly towards one unless one is also able to judge, when one has failed to meet a similar obligation that one owes him, that one has acted unjustly towards him. Or to put the point in the language Darwall favors,8 the obligations that any member of a moral community has are reasons to do that which he is obligated to do, and therefore, if one sees oneself as a member of a moral community, whatever reasons one thinks these obligations give others to act are reasons one also recognizes oneself as having. Consequently, if one expects others to meet their obligations to one in view of these reasons and likewise demands of them that they make

8.  Darwall (2006, 74–​79).

8 7

Psychopathic Resentment  87

amends if they don’t, one must then see oneself, in view of the same reasons, as required to meet the obligations one has to others and be disposed to demand of oneself that one make amends if one doesn’t. Resentment, on Darwall’s account, that one feels towards someone who has not met his obligations to one implies both that one expects him to meet those obligations and that one demands of him that he make amends since he didn’t. And it does so because the moral judgment it contains reflects an understanding of oneself as a member of a moral community. So, on this account, resentment implies that one expects the same of oneself and is disposed to make similar demands on oneself if one fails to meet one’s obligations. Plainly, then, it is not possible, on this account, to be liable to resentment if one is incapable of taking the viewpoints of others. The resentment of which psychopaths are capable remains, therefore, inexplicable.9 No doubt, this problem will seem to many to be merely an issue of labeling. Surely, they will think, we can restrict the term ‘resentment’ to that species of anger that is moral in character and felt in response to one’s being treated unjustly or at least to what one takes to be unjust treatment. To say that it is moral in character is to say that the judgment about being treated unjustly it implies is a moral judgment. And it follows, then, that the anger towards others of which psychopaths are capable, while it may be similar to resentment, should not be called “resentment” since psychopaths are incapable of making genuine moral judgments. After all, the standard view in Anglo-​ American philosophy about the nature of resentment emerged in the process of our regimenting our terms for moral emotions so as to have a precise way of discussing the emotions distinctive of moral agency. Such regimentation is necessary because language is used loosely in everyday speech. The standard view, then, once one understands that it has come about through stipulation necessary for making discussions of the psychology of moral agents precise, is not threatened by the anger displayed by psychopaths towards those whom they see as having betrayed them. The problem, however, with this way of understanding the standard view is that it makes any representation of the liability to resentment as a mark of moral agency vacuous. This result is especially problematic for accounts of resentment like Darwall’s, for the apparent force of these accounts comes from their representing the liability as such a mark.10 Typically, their proponents put forward the liability to resentment and other emotions like guilt and 9.  For further discussion of what it means to say that psychopaths lack the capacity for moral judgment see Deigh (1995, 743–​763). 10.  Darwall (2006, 17).

8

88  From Psychology to Morality

indignation as an essential feature of being a moral agent so that the accounts of these emotions they propose are seen as illuminating the nature of moral agency. Such illumination, however, would be illusory if the emotions were defined in advance as the emotions of a moral agent. Hence, to be genuinely illuminating, an account of resentment, like accounts of other natural phenomena, must presuppose that one can identify the emotion prior to giving the account. For to give an illuminating account of some natural phenomenon one must first be directly acquainted with the phenomenon, which is to say, one must first know how to identify it by name. Having so identified the phenomenon, one then brings its nature to light by developing an account that fits it. One cannot, therefore, understand the liability to resentment as a mark of moral agency unless the attribution of resentment to psychopaths is a misidentification. Yet nothing supports taking such attributions to be misidentifications apart from the thesis in question, that the liability to resentment implies a capacity for moral judgment. While the source of the anger Oakley expresses when he denounces the world as a “foul sty” is obscure,11 it is clear that Bruno’s anger at Guy is due to his belief that Guy has double-​crossed him. He is not, therefore, merely furious with Guy for thwarting his plan to end his father’s life. His anger, that is, is more than the fury of someone whose efforts are frustrated by another’s refusal to do what he wants him to do. Such fury is no different from anger at some machine or beast for refusing to work as it should. The farmer who beats his mule mercilessly because it refuses to pull the plow may thereby show anger amounting to rage that is due to the mule’s obstinacy and the farmer’s frustration at being unable to get the mule to move, but because the farmer need not see the mule as having betrayed him, the farmer’s anger need not be vengeful or retributive. Bruno’s, by contrast, is, which is why we readily identify it as resentment. Bruno’s sense of being betrayed by Guy comes from an attachment he has formed to Guy. His fondness for Guy develops during the conversation they have on the train that sets the story in motion and continues afterwards. He thinks of Guy as his friend and expects that this new-​found friend will follow through on the agreement he imagines they have made. His attachment to Guy is childlike, to be sure, but even small children regard their playmates as friends. A psychopath is not necessarily a loner. Thus Bruno’s anger, when he realizes that Guy does not plan to follow through on the agreement he

11.  Hitchcock plants in the film’s dialogue the suggestion that the cause of Oakley’s antisocial personality is a childhood injury to the brain.

98

Psychopathic Resentment  89

imagines they have made and, indeed, is not his friend, is anger at what he feels to be a betrayal. It is resentment. The standard view in Anglo-​American philosophy about the nature of resentment is therefore mistaken. A capacity for moral judgment is not necessary to be capable of the emotion. Someone whose personality is entirely egocentric may be capable of it as well. Consequently, its place in human psychology is different from that of the other emotions that are commonly explained in conjunction with it as emotions that imply judgments about injustice and that differ from each other according to the viewpoint with respect to the injustice from which the emotion is experienced. Thus, think again of Darwall’s account of these emotions according to which resentment is a response to injustice experienced from the viewpoint of the victim, indignation a response to it experienced from the viewpoint of a third party, and guilt a response to it experienced from the viewpoint of the perpetrator. While the symmetry among these emotions that structures this account neatly fits the underlying thesis that each emotion is distinctive of moral agency, it is belied by the psychopath’s liability to resentment. Only indignation and guilt imply moral judgments. The question then is how resentment, which does not imply such judgment, is related to them. Ironically, the work that exponents of the standard view about the nature of resentment usually cite as its source, Strawson’s “Freedom and Resentment,” offers an answer. It is embedded in Strawson’s explanation of the role of these emotions in the dynamics of interpersonal relations. The emotions belong to a set whose members, as we’ve already noted, Strawson calls reactive attitudes, and the insight on which Strawson proceeds in this essay is that the dispute over whether determinism threatens the soundness of our practice of attributing moral responsibility to people for their actions can be resolved through a study of these emotions. Strawson’s explanation of their role in the dynamics of interpersonal relations consists mostly of his description of the conditions of their arousal and expiration, and to see the answer to our question that Strawson’s essay offers, the question, that is, about the relation of resentment to the other reactive attitudes, requires elaboration of his explanation. The explanation begins with some plain, natural facts—​what Strawson calls “commonplaces”—​about men and women as social beings. The facts he adduces are these. A central concern of people who live together in communities is what their intentions and attitudes are towards each other. They place great importance on these intentions and attitudes, and they interact with each other with an expectation that generally their fellows behave with goodwill towards them. The goodwill each expects of others consists in the others’ keeping, as

09

90  From Psychology to Morality

best they can, from harming him in their conduct towards or interactions with him and in their being responsive to harm they may nonetheless cause him by being solicitous of his injury and feelings. Such goodwill may be basic to human personality generally (though not universally), or it may derive from more basic emotions and primitive attachments. Either is compatible with the general fact that Strawson means to emphasize: that people are highly sensitive to and greatly concerned with whether others are well-​or ill-​disposed towards them. This sensitivity and concern gives rise in a person to certain attitudes and feelings when others—​particularly certain others—​behave in ways that manifest good or ill will towards him. Thus, while one naturally takes pleasure in benefits that come one’s way on account of another’s behav­ ior, one responds, not just with pleasure, but with gratitude too, when one sees that the other intended to benefit one through his actions and because of his affection, esteem, or goodwill towards one. Likewise, while the harm inflicted on one by another’s actions is disagreeable, one will feel, not just displeasure, but also resentment if one recognizes in his actions malice or ill will towards one. Gratitude and resentment are Strawson’s initial examples of reactive attitudes that result from one’s seeing in another’s actions evident good or ill will towards one. Other examples are love, forgiveness, and hurt feelings. He calls these personal reactive attitudes. One cannot overstate the importance to Strawson’s explanation of the distinction between being benefited or harmed inadvertently or accidentally and being benefited or harmed intentionally out of evident good or ill will towards one. It is essential to how he conceives of the personal reactive attitudes. They are, on his conception of them, emotional responses in which one discerns (or thinks one discerns) in the beneficial or harmful action of another the person’s good or ill will, regard or indifference towards one. Hence, they differ from one’s merely being pleased or delighted with the benefit, pained or aggravated by the harm. They differ in that the cognition implicit in the response in either case, the cognition by virtue of which the response is an intentional state, is an awareness of the attitude towards one that is expressed in the action to which one is responding. The ability to discern in actions that benefit or harm one the actor’s dispositions and attitudes towards one is, therefore, necessary to one’s being liable to the personal reactive attitudes. Indeed, the ability, by virtue of the feelings and attitudes it makes possible, is basic to human sociability. Strawson makes this point implicitly when he invites us to “think of the many different kinds of relationship which we can have with other people—​as sharers of common interest; as members of the same family; as colleagues; as friends; as lovers; as chance parties to

91

Psychopathic Resentment  91

enormous range of transactions and encounters” and to think, “in turn . . . of the kind of importance we attach to the attitudes and intentions towards us of those who stand in these relationships to us, and of the kinds of reactive attitudes to which we ourselves are prone.”12 A decade later, John Rawls, in A Theory of Justice, makes it explicitly. Concerning the three psychological laws that, on his theory of moral education, explain the transitions between the stages of moral growth in a child’s development of a sense of justice, Rawls writes, “[T]‌he three laws are not merely principles of association or of reinforcement. While they have a certain resemblance to these learning principles, they assert that the active sentiments of love and friendship, and even the sense of justice, arise from the manifest intention of other persons to act for our good. Because we recognize that they wish us well, we care for their well-​being in return. Thus we acquire attachments to persons and institutions according to how we perceive our good to be affected by them. The basic idea is one of reciprocity, a tendency to answer in kind . . .. [T]his tendency is a deep psychological fact. Without it our nature would be very different and fruitful social cooperation fragile if not impossible.”13 The depth of this fact (and so the ability to discern the dispositions and attitudes of others towards one) was observed two centuries earlier by Rousseau, whom Rawls acknowledges as his predecessor.14 For Rousseau, the fact has to do with the transition in the very young child from the solipsism characteristic of infancy and the corresponding regard for others as mere instruments of or threats to the satisfaction of its needs to its uniting with or setting itself in opposition to others and the corresponding regard for them as fellows or foes. “Every child,” Rousseau writes, “is attached to his nurse . . .. At first this attachment is purely mechanical. What fosters the well-​being of an individual attracts him; what harms him repels him. This is merely a blind instinct. What transforms this instinct into sentiment, attachment into love, aversion into hate, is the intention manifested to harm us or be useful to us. One is never passionate about insensible beings which merely follow the impulsion given to them. But those from whom one expects good or ill by their inner disposition, by their will—​those we see acting freely for us or against us—​inspire in us sentiments similar to those they manifest towards us. We seek what serves us, but we love what wants to serve us. We flee what harms us, but we hate what wants to harm us.”15 12.  Strawson (1962, 76). 13.  Rawls (1971, 494–​495). 14.  Rawls (1963, 281–​305). 15.  Rousseau (1979, 213).

29

92  From Psychology to Morality

While Rousseau takes account of both relationships of mutual affection and of mutual enmity as products of our awareness of and responsiveness to the attitudes of others towards us, Strawson concentrates on the former. In general, Strawson thinks of these relationships as based on the parties to them expecting and demanding “some degree of goodwill or regard” from each other.16 This is another of the plain, natural facts about us as social beings he adduces. The personal reactive attitudes occur, then, against the background of these expectations and demands. They and the responses they elicit in turn are evidence of them. Resentment is Strawson’s prime example. Thus, when one party in a relationship of mutual regard injures another and the injured party feels resentment in response, the offender may try to remove these feelings with expressions of regret or apologies and such explanations as “It was an accident,” “I didn’t mean to,” “I didn’t see you coming” or similarly “I couldn’t help it,” “I slipped,” or “I had no choice.” Each of these explanations is typically offered to assuage resentment by denying that the offending action manifested ill will. They are meant, that is, to restore the relationship to the prior state of mutual goodwill or regard by giving the injured reason to abandon his belief that the offender acted with indifference or malice towards him. The offender in making such pleas is in effect saying that he did not, despite his having committed the offense, lack the goodwill towards the person he injured expected of him. He is saying, in other words, that while he is an appropriate object of reactive attitudes particularly at the time he committed the offense, the resentment the injured feels is misplaced in this case. Strawson compares such pleas with pleas that are offered to assuage resentment by denying that one was even, at the time one committed the offense, an appropriate object of reactive attitudes. “I lost my head,” “I was only a child,” “It happened during a psychotic breakdown,” “I was in a complete trance” etc. Each of these explanations is meant to exempt the offender, at least at the time of the offense, from the class of people of whom one expects some degree of goodwill in their dealings with one. It is in effect a plea to look on the offender, not as a fellow participant in the kind of friendly or cooperative relations that presuppose mutual goodwill and regard, but rather objectively as someone to be managed, controlled, or handled. In taking this outlook towards someone, one ceases to see his actions as manifesting good or ill will towards one and consequently one loses one’s disposition to feel resentment, gratitude, or any of the other reactive attitudes towards him. One

16.  Strawson (1962, 76).

9 3

Psychopathic Resentment  93

may of course still be liable to some feelings towards him. That one looks on the offender objectively does not mean that one becomes incapable of feeling sympathy, fear, or irritation towards him. What it means, though, is that one no longer engages with him with an attitude of goodwill or regard and an expectation of its being reciprocated. While his actions can, then, provoke in one emotional responses, these responses do not belong to the class Strawson defines as that of the reactive attitudes. Hence, the objective view of someone, which pleas of this second type invite, is sharply opposed to the view one takes of another with whom one participates in an interpersonal relationship. Strawson makes this comparison between the two types of plea to bring out the difference between our viewing someone as a participant in an interpersonal relationship with us and our viewing him as an object whose behavior is the product of the external forces and conditions to which he is subject and who is to be managed or controlled accordingly. The latter is consistent with the view we would take of people and their actions if we were to accept the thesis of determinism and regard them in its light. We are disposed to take such a view of someone when we see him as immature or dysfunctional in a way that incapacitates him for normal interpersonal relationships. In that case, we abandon the view of him as a participant with us in such relationships and regard him, instead, with detachment of a sort that an objective view of him entails. Nor is this the only occasion on which we would view someone objectively. We also, typically, take such a view of people generally when we are making social policy or designing a program to provide certain services to a given population. And we sometimes take such a view when, because of the strain of our relations with someone, we need to put distance between him and us of the sort that a detached attitude towards him creates. In each of these cases, we maintain an objective view only temporarily or only towards certain individuals. We could not, Strawson argues, permanently hold such a view of all people, for doing so would require our doing something we cannot in fact do, namely, cease to have interpersonal relationships. Nor, to continue Strawson’s argument, would it be wise permanently to adopt the objective view even if we could, for to do so would mean giving up human life as we know it, and abandoning our humanity is an unreasonable sacrifice. Determinism, Strawson concludes, is therefore no threat to the personal reactive attitudes. We would not cease to be liable to them even if we did accept its truth, nor would we have good reason to suppress them. Whether Strawson’s argument, at this point, succeeds in vanquishing the bogey of determinism is not our concern, however. What we are after

49

94  From Psychology to Morality

is how Strawson explains the relation of resentment to the moral emotions of indignation and guilt. To make this explanation he adds to the class of reactive attitudes vicarious analogues of the personal reactive attitudes. These analogues are attitudes one takes towards the goodwill or ill will of another, not in reaction to the manifestation of that will in the person’s actions towards one, but in reaction to its manifestation in his actions towards others. Specifically, Strawson calls the vicarious analogue of resentment “indignation” or “disapproval.” Plainly, his use of ‘indignation’ and ‘disapproval’ is, as he notes, somewhat stipulative. In common speech, for instance, we are as comfortable with saying that one feels indignation at one’s own mistreatment as that one feels it at the mistreatment of another. But since his point is to have distinctive names for vicarious resentment and not to give an independent account of indignation or disapproval, the stipulative use of these terms is unproblematic. Strawson also remarks that the vicarious character of indignation, that one can conceive of it as “resentment [felt] on behalf of another, where one’s own interest and dignity are not involved,” makes the adjective ‘moral’ an apt modifier of the attitude.17 Strawson, in other words, takes indignation to be a moral attitude in virtue of its being impartial or disinterested. He does not, then, in conceiving of it as a third-​party analogue of resentment, suppose that to feel it in response to someone’s injuring another in a way that shows a lack of due regard for the victim one must have a special attachment to or specially identify with the victim. Rather he supposes that people’s expectations of and demands for a reasonable amount of goodwill towards each other that form the backdrop of interpersonal relationships generalize as they, as individuals, come to recognize that everyone reasonably expects and demands such goodwill and regard on the part of others towards himself or herself in his or her transactions with them. The recognition of this general fact, Strawson further supposes, naturally disposes a person to respond with indignation or disapproval when, say, the person witnesses an injury done to someone to whom he has no special attachment by another, who manifests attitudes towards the injured that fall short of the goodwill and regard for others that is generally expected of people in their transactions with others. Strawson, apart from a quick reference to human connection, does not say why it is natural for people’s expectations of and demands for a reasonable amount of goodwill from others towards themselves to generalize in this way. Nonetheless, one

17. Ibid., 84.

9 5

Psychopathic Resentment  95

could, to explain this development, appeal in the way Hume did to something like our general sympathy with humankind. Strawson, in fact, seems to be tacitly following Hume in this part of his essay. Hume too distinguished attitudes that reflect a disinterested view from those reflecting an interested one and took the former to be moral because they reflected this view and occurred in response to a perception of the quality of the will manifested in the action that is their object. Indeed, Hume took the distinction between the personal or interested view and the general or disinterested view to be crucial to the explanation of our praising some people for their moral virtue and censuring others for their moral depravity. Such praise and censure, Hume held, express the sentiments of moral approbation and blame that we experience in response to our perceiving behind the actions of those whom we praise and censure such motives as kindness and compassion, on the one hand, envy and malice, on the other, and for our perception of such motives to give rise to these sentiments we must perceive them from a general or disinterested view. We could not otherwise, Hume argues, explain our praising and censuring the heroes and villains of the ancient world, like the Roman senators who defended the republic at the time of its crumbling and the Roman emperors who came later and whose cruelty and injustice are infamous, since their actions can hardly affect our interests.18 What Strawson then adds to Hume’s account is the idea of certain moral feelings and attitudes being analogues of feelings and attitudes that reflect a personal or interested view. In addition to the vicarious reactive attitudes, Strawson introduces reflexive ones into the class of reactive attitudes. He calls these “self-​reactive attitudes.” They naturally arise, he suggests, as a result of one’s generalizing the expectations of and demands for goodwill on the part of others towards one and thereby coming to expect and demand such goodwill of others towards each other in their transactions with themselves. For once one has generalized these expectations and demands, one realizes that they apply to oneself as well, that one is no different from others in being someone of whom such goodwill in interpersonal relations is expected and demanded. Accordingly, one becomes liable to feeling obliged or bound to act as the goodwill towards others that is expected and demanded of one requires, to feeling compunction, and to feeling guilt or remorse if one fails to meet this requirement. These are what Strawson refers to as the self-​reactive attitudes. He mentions them, he says, for the sake of completing the picture.19 But he 18.  Hume (1975, 215–​218). 19.  Strawson (1962, 84).

69

96  From Psychology to Morality

also notes that an agent who was not liable to the self-​reactive attitudes would be abnormal in a way that supported, were he to injure another, a plea like the pleas of immaturity and derangement that invite taking an objective view of him and thus suppressing the resentment or indignation towards him that otherwise would be forthcoming.20 Strawson, having completed his picture of the reactive attitudes, returns to the dispute over whether determinism poses a threat to our practice of attributing moral responsibility, and to show that it doesn’t, he reprises the argument he gave to show that accepting the thesis of determinism would not affect our being liable to the personal reactive attitudes. The argument, he believes, applies mutatis mutandis to the vicarious reactive attitudes, since they are analogues of the former. At the same time, he suggests that we can directly see in the argument’s application to this subclass of the reactive attitudes that our accepting the thesis of determinism is no threat to our practice of attributing moral responsibility. The argument, as applied to the vicarious reactive attitudes, affords this observation, he argues, because our susceptibility to these attitudes, unlike our susceptibility to the personal reactive attitudes, is the same as our expecting and demanding of everyone in our community some degree of goodwill or regard towards others generally in his or her dealings with them, and this generalized, disinterested expectation and demand amounts to an understanding of everyone in the community as lying under moral obligations to act with some degree of goodwill towards others generally in his or her dealings with them and as therefore someone to whom moral responsibility for failures to meet these obligations and for harms that result from some such failures is, in the absence of an exculpatory plea, properly attributed. Attributions of moral responsibility, according to Strawson, require an understanding of people as moral agents and as members of a moral community, and the interpersonal relationships in which one participates are seen as falling among the relationships that a moral community comprises only when one regards people in one’s community from a general or disinterested standpoint. Strawson initially makes this point by supposing someone whose view of himself and others is entirely interested or personal.21 Such a person is liable to the personal reactive attitudes but not to the vicarious ones. He is someone, then, who expects and demands from others a degree of goodwill towards himself in their dealings with him but does not expect of or demand

20. Ibid., 86. 21. Ibid., 85.

9 7

Psychopathic Resentment  97

from others such goodwill or regard in their dealings with each other. Such a person, Strawson contends, would be a moral solipsist. He would not, that is, see the interpersonal relationships in which he participates as embedded within a moral community. He would not see the demands he makes on others that they act with a degree of goodwill or regard towards him as general requirements to act with that degree of goodwill towards others to which everyone is subject and that as such represent moral obligations under which everyone lies. The ideas of belonging to a moral community, of having in common with others reciprocal moral obligations to act with a degree of goodwill towards one another, and of being, as everyone else is, morally responsible for actions that affect others do not arise, then, until one takes up the standpoint of others in one’s community and their expectations and demands on those with whom they interact.22 These ideas do not arise, in other words, until one regards oneself and others from a general or disinterested standpoint. Here too Strawson may be meaning to follow Hume in supposing that perceptions and judgments of moral relations and moral conduct must come from a view we share with others. The intelligibility and practical force of the language of morals, Hume argued, presuppose a common point of view, and therefore we must vacate our respective interested or personal viewpoints, which are peculiar to each of us, and take up a general or disinterested one (or suppose that we have), which is common to all, to take in the moral character of people’s relationships and actions within them and to speak properly of it. Nonetheless, though Strawson may be meaning to follow Hume here, he departs, I think, from Hume in supposing that we get our ideas of moral obligation and moral responsibility directly from generalizing the expectations of and demands for goodwill towards ourselves that we make on others, which is to say, by vacating our personal view and taking up a general or disinterested one. For Hume does not think we come to have these ideas until we apply the generalized expectations and demands that we acquire in taking 22.  Strawson then reaffirms the point in considering how pleading, in response to injuring someone, that one was at the time incapacitated for normal participation in interpersonal relationships (“I was suffering hallucinations, undergoing a psychotic breakdown, in a trance, etc.”) affects our susceptibility to moral indignation. Thus, Strawson, after introducing the case by writing of seeing the agent in a different light, “as one whose picture of the world is an insane delusion; or as one whose behaviour . . . is unintelligible to us” etc., goes on, “[A]‌bstracting now from direct personal interest, we may express the facts with a new emphasis. We may say: to the extent to which the agent is seen in this light, he is not seen as one on whom demands and expectations lie in that particular way in which we think of them as lying when we speak of moral obligation; he is not, to that extent, seen as a morally responsible agent, as a term of moral relationships, as a member of the moral community.” Ibid., 86.

89

98  From Psychology to Morality

up the general view to ourselves in those circumstances in which we find ourselves lacking the natural motives by which we typically manifest goodwill or due regard in our dealings with others. Specifically, Hume thinks we acquire these ideas as a result of a kind of indignation with ourselves for lacking the motives of goodwill towards others that we come to demand of everyone including ourselves when we take up the general or disinterested view.23 That is, seeing that we lack the motive of benevolence, say, or fellow feeling that in our dealings with another would show us to bear goodwill towards him and at the same time expecting and demanding on his behalf that we so act, we reproach ourselves for our lack of benevolence and fellow feeling and are thereby moved by a sense of duty to do the action that we would have done had we been naturally moved by either to do it. It is this experience of self-​reproach for lacking the proper natural motive of goodwill and the consequent sense of duty to do what we would naturally do if we had that motive that gives us the idea of moral obligation, the idea that acts displaying goodwill or due regard for others in our dealings with them are obligatory even when we lack in ourselves the motives of benevolence or fellow feeling that would naturally produce them. And presumably, though Hume does not expressly say this, the same experience that gives us the idea of moral obligation gives us as well the idea of being morally responsible for doing or failing to do what we now understand as obligatory. Be this as it may, the important point is that Strawson, like Hume, holds that the acquisition of these ideas occurs only after one becomes liable to the vicarious reactive attitudes. Being liable to the personal reactive attitudes is by itself insufficient for having them. Let us distinguish, then, between Strawson’s original account of the reactive attitudes and that account modified to reflect Hume’s view of how we come to see ourselves as having moral obligations. I will refer to the latter as the modified Strawsonian account. Clearly, either account shows how someone might be liable to resentment though not to indignation. Such a person, while capable of experiencing personal reactive attitudes, would be incapable of experiencing the vicarious ones. To be capable of the former he must be able to form attachments to others, to regard them as friends, mates, pals, or the like, and thus to have some affective tie to them. Such attachments are possible, though, even if one is incapable of experiencing the vicarious reactive attitudes. For one can form an attachment to another without being able to understand the other’s interests sympathetically, and

23.  See Hume (1978, 479).

99

Psychopathic Resentment  99

on either account, to be capable of the vicarious reactive attitudes means that one can take an impartial view of others, a view in which one regards their interests sympathetically and thereby feels personal reactive attitudes on their behalf. A person who is incapable of taking such a viewpoint is thus incapable of experiencing the vicarious reactive attitudes. On the modified Strawsonian account, such a person is also incapable of experiencing the self-​reactive attitudes, for on this account, a person can experience such attitudes only if he is capable of experiencing the vicarious reactive attitudes. Being capable of experiencing the latter attitudes, in other words, is a necessary condition for being capable of experiencing the former. Hence, at least on the modified Strawsonian account, a thoroughly egocentric person, someone lacking a conscience and incapable of empathy, could nonetheless be liable to resentment. This account, in short, supports an understanding of psychopaths as capable of resentment.24 If psychopaths, by virtue of their lacking a conscience, are the paradigm of human agency that is incapable of moral judgment but otherwise rational, then the modified Strawsonian account yields the better explanation of this incapacity and thus yields a better understanding of the nature of moral agency. The two accounts differ in identifying the capacity for moral judgment with the acquisition of different liabilities to reactive attitudes. The modified Strawsonian account identifies it with the acquisition of a liability to the self-​reactive attitudes. Strawson’s original account, by contrast, identifies it with the acquisition of a liability to the vicarious reactive attitudes. On his account, acquiring the capacity for these attitudes is independent of one’s acquisition of a liability to the self-​reactive attitudes. Yet if the hallmark of the capacity for moral judgment is the ability to see oneself as others see one and to react accordingly, then Strawson’s original account, in identifying the capacity for moral judgment with the liability to experience the personal reactive attitudes vicariously, falls short. A general sympathy with others that gives rise to one’s being liable to the vicarious reactive attitudes is, in other words, insufficient for understanding oneself as joined with others in relations of reciprocal moral obligations and for understanding, in consequence, the moral responsibility that can attach to failures to meet those obligations. Hence, only with the development of the self-​understanding that comes

24.  Strawson’s characterization of the moral solipsist is that of someone who is capable of experiencing the personal reactive attitudes but not the vicarious ones, and Strawson allows that such a person might also be capable of the self-​reactive attitudes. Consequently, the moral solipsist, as Strawson characterizes him, does not fit the profile of a psychopath. At the same time, Strawson raises the possibility of a moral idiot, who does. See Strawson (1962, 85).

001

100  From Psychology to Morality

from being able to see oneself as others see one and of the liability to the self-​reactive attitudes that is consequent to it does one acquire a capacity for moral judgment. Such development marks a significant advance over the development of a capacity for experiencing vicarious attitudes.25 The modified Strawsonian account of the reactive attitudes captures this advance. By doing so it explains, as Strawson’s original account does not, the moral deficit that is so memorably displayed by Hitchcock’s villains in Shadow of a Doubt and Strangers on a Train.

References Darwall, S. 2006. The Second-​Person Standpoint: Morality, Respect, and Accountability. Cambridge, MA: Harvard University Press. Deigh, J. 1995. “Empathy and Universalizability.” Ethics 105:743–​763. Hare, R. D. 2003. Manual for the Revised Psychopathy Checklist, 2nd ed. Toronto, ON: Multi-​Health Systems, Inc. Hoffman, M. 2000. Empathy and Moral Development: Implications for Caring and Justice. Cambridge: Cambridge University Press. Hume, D. 1975. Enquiries concerning Human Understanding and concerning the Principles of Morals, edited by P. H. Nidditch, 3rd ed. Oxford: Clarendon Press. Hume, D. 1978. A Treatise of Human Nature, edited by P. H. Nidditch, 2nd ed. Oxford: Clarendon Press. Rawls, J. 1963. “The Sense of Justice.” Philosophical Review 72:281–​305. Rawls, J. 1971. A Theory of Justice. Cambridge, MA: Belknap Press. Rousseau, J. J. 1979. Emile or On Education. Translated by A. Bloom. New York: Basic Books. Strawson, P. F. 1962. “Freedom and Resentment.” Proceedings of the British Academy, 48, 1–​25. Reprinted in G. Watson (ed.), Free Will, 2nd ed. Oxford: Oxford University Press, 2003, 72–​93. Wallace, R. J. 1996. Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Watson, G. 1988. “Responsibility and the Limits of Evil:  Variations on a Strawsonian Theme.” In Responsibility, Character, and the Emotions, edited by F. Schoeman, 256–​286. Cambridge: Cambridge University Press.

25.  See Hoffman (2000, 93–​112).

0 11

5

Reactive Attitudes Revisited: A Modest Revision Commanding, questioning, recounting, chatting are as much a part of our natural history as walking, eating, drinking, playing. —​Ludwig Wittgenstein, Philosophical Investigations, §25

Few essays in contemporary1analytic philosophy have had as great an impact on the study of their topic as P. F. Strawson’s “Freedom and Resentment.”1 Even fewer have been as poorly understood.2 The innovation for which Strawson’s essay is famous is his use of the emotions of gratitude, resentment, moral indignation, love, and forgiveness, among others—​what Strawson calls the reactive attitudes—​to explain our practice of attributing moral responsibility to people. The brilliance of this innovation and the use to which Strawson put it has captured the imagination of many philosophers who work on the problems of moral responsibility, and it is now common for them either to follow Strawson in explaining moral responsibility by reference to the reactive attitudes or at least to acknowledge his explanation as consistent 1T . his essay is a revision of my “Reactive Attitudes Revisited” in Morality and the Emotions, Carla Bagnoli ed. (Oxford: Oxford University Press, 2011). The main revisions are to the material in section 1. 1.  Proceedings of the British Academy 48 (1962): 1–​25. Reprinted in Free Will, Gary Watson, ed., 2nd ed. (Oxford: Oxford University Press, 2003), pp. 72–​93. All references to the reprint. 2.  I mean this generally, of course, not universally. One exception, which has significantly influenced my understanding of Strawson’s essay, is Jonathan Bennett’s treatment of it in his “Accountability,” Philosophical Subjects: Essays Presented to P. F. Strawson, Zak van Straatan, ed. (Oxford: Clarendon Press, 1980), pp. 14–​47.

101

021

102  From Psychology to Morality

with theirs. Nonetheless, the agreement expressed in either case is typically agreement with Strawson’s theme only. Beyond affirming the importance of the reactive attitudes to our practice of attributing moral responsibility, these philosophers’ explanations have little in common with Strawson’s. One seldom, for instance, finds agreement with how he conceived of the reactive attitudes. Indeed, one often finds significant yet unrecognized disagreement. Such disagreement alone indicates the depth of the misunderstanding that now characterizes philosophical writing about Strawson’s essay. Above all, one finds in this writing that reactive attitudes are conceived of in ways that implicitly reverse Strawson’s understanding of their relation to the practice of attributing moral responsibility. This reversal of Strawson’s understanding is most apparent in recent, influential accounts of the practice according to which the attitudes disclose complex normative thought about obligations people have to one another and the authoritative claims they make on each other when these obligations are not fulfilled. Perhaps the two best known of these accounts are R. Jay Wallace’s, in his Responsibility and the Moral Sentiments,3 and Stephen Darwall’s, in his The Second-​Person Standpoint.4 On Wallace’s and Darwall’s accounts, the reactive attitudes are a consequence of the existence of the obligations and claims that the complex normative thought they disclose represents and therefore have merely a decorative place in those accounts. All the real work is done by the relations of obligation and authority represented in this thought. By contrast, on Strawson’s account, the attitudes do real work in explaining the practice. They are its material, and from it moral relations of a kind similar to those Wallace and Darwall presuppose in their accounts emerge. In what follows, I will first expound Strawson’s account with the aim of substantiating these last remarks. Substantiating them means showing that Strawson did not conceive of the reactive attitudes as implying complex normative thoughts. Rather he conceives of them (or at least some of them) as implying cognitions that need not even have any conceptual content and are thus, in this respect, much simpler than cognitions whose content incorporates the concepts of obligation and authority. Parts I and II contain this exposition. Once one grasps Strawson’s conception of the reactive attitudes, it will be apparent that Strawson saw the practice of attributing moral responsibility as emerging from these states and the interpersonal relationships to which 3. R. Jay Wallace, Responsibility and the Moral Sentiments (Cambridge, MA:  Harvard University Press, 1996). 4.  Stephen Darwall, The Second-​Person Standpoint: Morality, Respect, and Accountability (Cambridge, MA: Harvard University Press, 2006).

013

Reactive Attitudes Revisited  103

they are integral and not as necessarily embedded in those relationships. In part III of the paper I will contrast this account with the accounts of moral responsibility in Wallace’s and Darwall’s books. My chief aim will be to show that, unlike Strawson’s account, these accounts take the practice as already embedded in those relationships and, as a result, the reactive attitudes do no real work in these accounts in explaining the practice. What remains, when one removes from these accounts the appeals to the reactive attitudes, is just the traditional account of the practice that Strawson meant his account to supercede. In part IV, I take up an objection to my interpretation of Strawson’s conception of the reactive attitudes. I argue, in answer to this objection, that Strawson does not rest his conception of these attitudes on the assumption that their subjects and objects are moral agents. Rather his conception provides a way of seeing how we develop the idea of moral agency and its conditions from our susceptibility to these attitudes.

I Strawson, in his famous essay, takes up the modern question of the possibility of human freedom in a deterministic universe. Specifically, the question is whether determinism implies that men and women lack the freedom necessary to being morally responsible for their actions. Since our practices of praising good deeds and punishing bad ones presuppose our attributing moral responsibility to those whom we praise and punish, determinism thus raises questions about the justifiability and sustainability of these practices. Some philosophers, whom Strawson dubs “pessimists,” see no way to reconcile determinism with the freedom necessary to sustain these practices. To have such freedom, they think, is to be a prime mover. It is to have the power to move oneself independently of the impact of external events on oneself, and one cannot have such power if determinism is true. Other philosophers, whom Strawson dubs “optimists,” see no problem with reconciling determinism with the freedom necessary to sustain our praising good deeds and punishing bad ones, because they think the only freedom necessary for this purpose is the freedom implicit in our being capable of being influenced by praise and punishment. The point of our praising good deeds and punishing bad ones is to regulate behavior for social good, and accordingly we distinguish between behavior that is responsive to the positive feedback of praise and the negative sanctions of punishment and behavior on which praise or punishment can have no influence. We distinguish, that is, between the voluntary actions of men and women of sound mind and the actions of the

041

104  From Psychology to Morality

insane or of people acting in a panic or under such compulsion of fear as to be virtual automatons. The former are free in the only sense that matters to being morally responsible for their actions, and in that sense their freedom is not threatened by determinism. Strawson’s goal in the essay is to resolve the opposition between the pessimists and the optimists by finding a view both can accept. To achieve this resolution, Strawson thinks, requires finding an account of our practices of praising good deeds and punishing bad ones that the pessimists can accept even though it means abandoning their conception of men and women as prime movers. Recourse to that conception, Strawson suggests, is an overreaction to the threat that determinism poses to these practices. At the same time, Strawson agrees with the pessimists that the optimists’ way of explaining the freedom necessary to sustaining these practices misses something essential to our understanding of what justifies praise and punishment when they are justified. It misses our understanding of them as fitting responses to good and bad deeds, which is to say, it misses our understanding of them as justified in the first instance by being called for by such deeds and thus apart from the social benefits that follow from them. The pessimists’ dissatisfaction with the optimists’ explanation of the basis of praise and punishment is therefore, according to Strawson, well-​founded, for the explanation adduces facts that provide an inadequate basis of these practices. Facts about the socially beneficial influences of praise and punishment are not what make the bestowal of praise or infliction of punishment justified when either is justified. Consequently, Strawson argues, if one can adduce facts that provide an adequate basis of these practices, the pessimists will then believe their understanding of what is essential to justifying praise and punishment has been vindicated, and they will therefore no longer feel driven by determinism to a conception of men and women as prime movers. Strawson’s resolution of the opposition between the pessimists and the optimists thus depends on his coming up with a set of facts supporting our practices of praise and punishment that is different from those to which the optimists appeal. Moreover, they must support these practices while fitting them within a deterministic universe. Otherwise Strawson would not achieve his goal of finding a view acceptable to both pessimists and optimists. Hence, it would ill serve his program to put forth, as an alternative set of facts providing an adequate basis of praise and punishment, facts about, say, there being a government of the universe to which all humans are subject and in view of whose laws one can identify morally worthy actions as well as moral wrong ones. For though such facts would sit well with the pessimists, the optimists would surely need to be convinced that a deterministic universe

015

Reactive Attitudes Revisited  105

could have such a government built into it, that the responsiveness of men and women to moral law that such a government presupposed was a kind of activity consistent with determinism. And it is no part of Strawson’s program to develop arguments that would convince optimists on points like these. Rather Strawson looks to facts about human psychology and human behav­ ior that appear to be completely consistent with determinism. Specifically, he looks to what appear to him to be plain, natural facts about us humans as social beings. His turning to interpersonal relations and the reactive attitudes that are integral to them is thus meant to be a turn to a field of study whose defining phenomena are consistent with determinism. What recommends this field is its being a field whose study parallels the study of our practices of praise and punishment and is at the same time less fraught than the latter with uncertainties about the consistency of what is being studied with determinism. Strawson begins his statement of the plain, natural facts about men and women as social beings—​what he calls “commonplaces”—​with a fact he takes to be central to collective human life. It is that we humans attach great importance to the intentions and attitudes that others have towards us. These intentions and attitudes are a matter of major concern to us as we engage interpersonally with others. Our own feelings and conduct towards others are in large part reactions to what we take to be their intentions and attitudes towards us. We expect that they will generally act towards us with goodwill. This goodwill that we expect of them consists in their keeping, as best they can, from harming us in their conduct towards or interactions with us and in their being responsive to any harm they may nonetheless cause us to suffer by being solicitous of our injuries and feelings. Such goodwill may be basic to human personality generally (though not universally), or it may derive from more basic emotions and primitive attachments. Either is compatible with the general fact that Strawson means to highlight, which is that we are highly sensitive to and care deeply about whether others are well-​or ill-​disposed towards us. This sensitivity and concern makes us liable to the particular feelings and attitudes with which we respond when those with whom we interact behave in ways that manifest good or ill will towards us. Strawson calls these particular responses the personal reactive attitudes. Being liable to them implies a capacity to discern in the actions of others the intentions and attitudes they have towards us in so acting. Such a cognition distinguishes the personal reactive attitudes from other responses to behavior that benefits or harms us. Thus, while we naturally take pleasure in benefits that come our way on account of another’s behavior, we respond, not just with pleasure, but with

061

106  From Psychology to Morality

gratitude when we recognize in our benefactor’s action his meaning to benefit us through this action and because of his affection, esteem, or goodwill towards us. By the same token, while we naturally feel pain or distress when another harms us, we feel more than just pain or distress, but also resentment, when we recognize in his action malice or contemptuous indifference towards us. Gratitude and resentment are Strawson’s initial examples of personal reactive attitudes. His other examples are love, forgiveness, and hurt feelings. The distinction between the personal reactive attitudes and other responses to being benefited or harmed is crucial to Strawson’s account of the dynamics of interpersonal relations. It explains the importance within such relations of the difference between being benefited or harmed inadvertently and being benefited or harmed intentionally out of evident good or ill will towards one. Think here of the difference between receiving a windfall and receiving a gift, or again of the difference between suffering a mishap and suffering a beating. In the one case, one reacts with delight or dismay at one’s good or bad fortune but not with any particular warmth or chill towards the person who unwittingly or accidentally benefited or harmed one. When a stranger abandons a newspaper one is keen to read, one is certainly glad at the opportunity to read it but will typically give no thought to the stranger who left it. In the other case, one’s reaction is that of gratitude or resentment, thanks or enmity, towards the person who gave one the gift or the beating, and in so reacting one has an impulse to return in kind what one received or suffered. The impulse to reciprocate is at the heart of the dynamic that interests Strawson. It arises by virtue of one’s having discerned (or thought one discerned) in the beneficial or harmful action of another the person’s good or ill will, regard or indifference, towards one. And within the relationships in which it arises it serves either to strengthen the bonds between people when it moves them to return favors to their benefactors or to create conflicts and rifts among them when it moves them to retaliate against their injurers. The latter possibility then induces from injurers placative actions when they wish to preserve good relations with those whom they have injured, and these actions in turn serve to soften the hard feelings that the latter have towards their injurers and to repair the rifts the injurers’ harmful actions created. Plainly, then, such reciprocity would not be possible if we did not have the ability to discern in the actions of others their intentions and attitudes towards us. This ability, when coupled with how greatly those intentions and attitudes matter to us, is basic to human sociability. Strawson implies as much when he invites us to “think of the many different kinds of relationship which

017

Reactive Attitudes Revisited  107

we can have with other people—​as sharers of common interest; as members of the same family; as colleagues; as friends; as lovers; as chance parties to enormous range of transactions and encounters” and to think, “in turn . . . of the kind of importance we attach to the attitudes and intentions towards us of those who stand in these relationships to us, and of the kinds of reactive attitudes to which we ourselves are prone.”5 Each such relationship is harmonious to the extent that we have and exhibit goodwill towards each other. Demonstrations of it strengthen the trust that we put in those with whom we share relations to act with our welfare in mind. Signs of its absence weaken that trust. And in either case the response of greater or lesser trust is evidence of how much their intentions and attitudes matter to us. The personal reactive attitudes are thus inherent in relationships of mutual trust. They are hallmarks of human social life. An ancient conceit of Western thought is that this form of sociability is exclusive to human beings. Man alone, on this view, lives in societies that are held together by schemes of cooperation that sustain relations of trust and reciprocity compatibly with each member’s retaining his or her individuality. Plato weaved this view into the speech Protagoras makes at the beginning of the dialogue Plato named for this greatest of sophists. The gods, Protagoras tells his listeners, gave to the other animals physical assets, like strength and swiftness, that enable them to survive in a hostile world, but they neglected to give the same to man. To make up for their omission, they gave human beings justice and conscience, along with the tools of the practical arts, for men and women, Protagoras explains, need cities to survive, and they need justice and conscience to run their cities and to keep them from falling into civil strife.6 Similarly, Aristotle writes in book I of the Politics, “[T]‌hat man is more of a political animal than bees or any other gregarious animals is evident. Nature, as we often say, makes nothing in vain, and man is the only animal whom she has endowed with the gift of speech. And whereas mere voice is but an indication of pleasure or pain and is therefore found in other animals . . . , the power of speech is intended to set forth the expedient and inexpedient, and therefore likewise the just and the unjust. And it is a characteristic of man that he alone has any sense of good and evil, of just and unjust, and the like and the association of living beings who have this sense makes a family and a state.”7 5. Ibid., p. 76. 6. Plato, Protagoras, 320c7–​322c3. 7. Aristotle, Politics, 1253a7–​1253a18 in The Complete Works of Aristotle:  The Revised Oxford Translation, B. Jowett, trans. (Princeton: Princeton University Press, 2016).

081

108  From Psychology to Morality

The conceit recurs as well in the modern period. The statement of it Hobbes gives near the period’s outset is exemplary. Echoing Aristotle, Hobbes writes, “It is true that certain living creatures, as Bees and Ants, live sociably one with another, (which are therefore by Aristotle numbered amongst Politicall creatures;) and yet have no other direction, than their particular judgments and appetites; nor speech, whereby one of them can signifie to another, what he thinks expedient for the common benefit.”8 Human beings, by contrast, according to Hobbes, cannot live sociably one with another for very long without a scheme of cooperation that they mutually accept and that is centrally enforced. To substantiate this claim Hobbes goes on to list several properties that distinguish human relationships from those of other social animals. His list includes, among others, the competition for honor and dignity in which human beings are continually engaged and which, in the absence of a scheme of cooperation that is mutually accepted and centrally enforced, leads them inevitably into war. For Hobbes such competition shows the great importance human beings attach to the attitudes that others have towards them, and he reinforces the point by noting that, unlike human beings, other animals do not respond to being harmed with negative reactive attitudes. “[I]‌rrational creatures,” he observes, “cannot distinguish between Injury and Dammage, and therefore as long as they be at ease, they are not offended with their fellows.”9 One can find similar claims to the special character of human sociability in the writings of Hume and Rousseau.10 Recently, this traditional conceit has come under attack from naturalists who closely observe and record the social life of primates both in the wild and in captivity. No primatologist has pressed this attack more energetically than Frans de Waal. The springboard he uses to launch it is the theory in evolutionary biology of reciprocal altruism. The theory originates in the work of Robert Trivers. In his seminal article “Evolution of Reciprocal Altruism,” Trivers advances the thesis that natural selection would favor, in certain animal populations and habitats, the inheritance by the populations’ members of dispositions to behave in ways beneficial to other members, despite their not being kin or their kinship’s being too

8. Thomas Hobbes, Leviathan, Richard Tuck, ed. (Cambridge:  Cambridge University Press, 1996), p. 119. 9.  Ibid., p. 120. 10. See e.g., David Hume, A Treatise of Human Nature, P. H. Nidditch, ed., 2nd ed. (Oxford: Clarendon Press, 1978), p. 326 (bk. 3, pt. 1, sec. 2); Jean-​Jacques Rousseau, Social Contract (bk. 1, ch. 8)  in in The Social Contract and Discourses, G. D.  H. Cole, trans. (New York: E. P. Dutton & Co., 1959), pp. 18–​19. The passage in Hume should be read in light of his later remarks in bk. 3, pt. 2, sec. 2 about the origins of conventions of justice and property.

091

Reactive Attitudes Revisited  109

distant to explain this inheritance, when the members who engage in the behavior can expect its beneficiaries to return the favor at some later time.11 Such behavior qualifies as altruistic in the evolutionary biologists’ sense when it worsens the chances of reproduction of those who engaged in it. Accordingly, Trivers characterized it, when it qualified as altruistic, as reciprocal altruism, and the burden of his argument, then, was to show that natural selection could explain the disposition in the members of an animal population to such altruism. De Waal, inspired by Trivers’s theory, has marshaled evidence, both ethological and experimental, of reciprocal altruism in the social lives of chimpanzees. His case against the traditional conceit rests on this evidence. It shows, de Waal maintains, that chimpanzees form relationships of trust and reciprocity in which the personal reactive attitudes of gratitude and resentment inhere. De Waal’s evidence consists of observations he and his collaborator made of chimpanzee behavior that was altruistic in the biologists’ sense.12 Specifically, their observations were of colonies of chimpanzees whose members shared food with other members. Such sharing, being unilateral, qualified as altruistic since the sharers, in allowing other chimpanzees to take some of the food they possessed, gave up something of value without receiving there and then anything of value in return. De Waal determined that such sharing was neither indiscriminate nor always done with kin. Sometimes, that is, it was done selectively, and the beneficiary was not kin. On these occasions, the sharers did not allow every chimpanzee who gathered near them while they ate to take some of their food but did allow fellow chimpanzees who had earlier groomed them to take some of the food. Such sharing, given the sharer’s expectation of reciprocal benefits from the beneficiary as de Waal’s observations confirmed, thus qualified as reciprocal altruism. This result alone supports the view that chimpanzees can discriminate between those of their fellows who have previously behaved well towards them and those who have not. Such discrimination implies an ability to discern in the behavior of a fellow a favorable attitude towards one. It implies, that is, the ability to recognize another as being well-​disposed towards one, to recognize him or her as a friend. The significance of these implications should be clear. Because chimpanzees in natural habitats live together and communicate with each other without words or language, their having these abilities implies that 11.  Robert Trivers, “The Evolution of Reciprocal Altruism,” Quarterly Review of Biology 46 (1971): 35–​57. 12.  Kristin E. Bonnie and Frans B. M. de Waal, “Primate Social Reciprocity and the Origin of Gratitude” in The Psychology of Gratitude, Michael McCullough and Robert Emmons, eds. (Oxford: Oxford University Press, 2004), pp. 213–​229.

0 1

110  From Psychology to Morality

linguistic competence is not necessary to having them. It implies, in other words, that the cognitions that consist in one’s discerning in the behavior of another a favorable attitude towards one need not be propositional. Admittedly, de Waal does not report observing chimpanzees spontaneously expressing gratitude at the time a fellow chimpanzee does them a good deed. His evidence, then, may be incomplete. The absence of such spontaneous expressions suggests that gratitude may be in chimpanzees an attitude that disposes them to reciprocation rather than an emotion that excites displays of feeling. Or in Hume’s terms, it suggests that gratitude in chimpanzees is a calm passion rather than a violent one. If so, then the conclusion to draw from de Waal’s evidence is that chimpanzees can discern in the behavior of fellow chimpanzees the latter’s being well-​disposed towards them and retain this cognition in memory for later reciprocation. What his evidence does not support is the stronger conclusion that they are liable to an emotion that stirs displays of feeling as an immediate response to their receiving benefits for which they are grateful. At the same time, de Waal does relate an incident first described by the psychologist Wolfgang Köhler in which two captive chimpanzees spontaneously expressed gratitude as an immediate response to Köhler’s having rescued them from inclement weather.13 The chimpanzees had been left stranded in a rainstorm during cold weather as a result of their shelter’s having been inadvertently shut before they could re-​enter. Köhler, finding them outside the shelter soaked and shivering, opened it to let them in. Before they took refuge, however, they hugged him exuberantly. Enough evidence of this sort, if systematically gathered to rule out alternative interpretations, would support the stronger conclusion. Be this as it may, the weaker conclusion, which de Waal’s evidence supports, is sufficient to find in the social relations of chimpanzees reciprocity due to a combination of their discernment of the attitudes fellow chimpanzees have towards them and their responsiveness to what they discern. In an important sense gratitude is prior to resentment among the personal reactive attitudes. It emerges in social life from kindnesses that the members of a society do for one another. A precondition of its emergence is that the members can discern in kindnesses that others do for them the latter’s intentions and attitudes towards them. By contrast, resentment emerges in social life only after the members of a society have come to expect of at least some of the other members that the latter will act with goodwill towards them.

13.  Ibid., p. 215.

111

Reactive Attitudes Revisited  111

The preconditions of resentment, then, include more than that the members can discern in mean things that others do to them the latter’s intentions and attitudes towards them. They include, in addition, a prior acquisition of an expectation, with respect to those who act meanly towards them, that they will behave well towards them. To be sure, without such an expectation, one who is injured by another and who sees in the latter’s injurious behavior a mean or malicious intention towards one might still respond with anger. Such anger, however, would not amount to resentment, for it would lack a sense of one’s having been betrayed or wronged by the offender. To have such a sense, plainly, one must have expected better of him or her at the time he or she injured one. De Waal misses this distinction between resentment and other forms of anger. He works instead with a more general notion of a retributive emotion that he variously labels “resentment”, “indignation”, and “moralistic aggression”. He bases his attribution of this emotion to chimpanzees on his observations of hostile, retaliatory behavior by some of them towards others in a variety of contexts. Nonetheless, his firm evidence in support of attributing gratitude to chimpanzees supports in turn, in view of some of his observations of this retaliatory behavior, attributing resentment to them as well. Specifically, while one cannot attribute resentment on the basis of observations of hostility between male rivals for the favors of a female, observations of aggression against a fellow chimpanzee who had once been an ally of the aggressor but who later refuses to help her defend against an opponent in the colony supports attributing resentment to explain the aggression. For the anger the aggressor expresses towards her former ally in this context includes a sense of being betrayed at a time of need. Here is de Waal’s description of one such incident, which he observed while researching the colony of chimpanzees at the Arnham Zoo in the Netherlands: A high ranking female, Pruist, took the trouble and risk to help her male friend, Luit, chase off a rival, Nikkie. Nikkie, however, had a habit after major confrontations of singling out and cornering allies of his rivals, to punish them. This time Nikkie displayed at Pruist shortly after he had been attacked. Pruist turned to Luit stretching out her hand in search of support. But Luit did not lift a finger to protect her. Immediately after Nikkie left the scene, Pruist turned on Luit, barking furiously. She chased him across the enclosure and even pummeled him.14

14.  Frans B. M. de Waal, Good Natured: The Origins of Right and Wrong in Humans and Other Animals (Cambridge, MA: Harvard University Press, 1996), p. 97.

2 1

112  From Psychology to Morality

II Strawson is chiefly interested in resentment.15 Accordingly, he places it within the context of relationships of mutual affection or respect. In general, he thinks of these relationships as based on the parties to them expecting and demanding “some degree of goodwill or regard” from each other.16 This is another of the plain, natural facts about us as social beings he adduces. Resentment occurs, then, against the background of these expectations and demands. Thus, when one party in a relationship of mutual regard injures another and the injured party feels resentment in response, the offender may try to remove these feelings with expressions of regret or apologies and such explanations as “It was an accident,” “I didn’t mean to,” “I didn’t see you coming” or similarly “I couldn’t help it,” “I slipped,” or “I had no choice.” Each of these explanations is typically offered to assuage resentment by denying that the offending action manifested ill will. They are meant, that is, to restore the relationship to the prior state of mutual goodwill or regard by giving the injured person reason to abandon his belief that the offender acted with indifference or malice towards him. The offender in making such pleas is in effect saying that he did not, despite his having committed the offense, lack the goodwill towards the person he injured that is expected of him. He is saying, in other words, that while he is an appropriate object of reactive attitudes particularly at the time he committed the offense, the resentment the injured person feels is misplaced in this case. Strawson compares such pleas with pleas that are offered to assuage resentment by denying that one was even, at the time one committed the offense, a proper object of reactive attitudes. “I lost my head,” “I was only a child,” “It happened during a psychotic breakdown,” “I was in a complete trance” etc. Each of these explanations is meant to exempt the offender, at least at the time of the offense, from the class of people of whom one expects some degree of goodwill in their dealings with one. It is in effect a plea to look on the offender, not as a fellow participant in the kind of friendly or cooperative relations that presuppose mutual goodwill and regard, but rather objectively as someone to be managed, controlled, or handled. In taking this outlook towards someone, one ceases to see his actions as manifesting good or ill will towards one and consequently one loses one’s disposition to feel

15.  I am repeating at this point material that appears in Essay 4. The repetitious material extends to the end of this section. Readers already familiar with this material may skip ahead to section III. 16. Strawson, p. 76.

13

Reactive Attitudes Revisited  113

resentment, gratitude, or any of the other reactive attitudes towards him. One may of course still be liable to some feelings towards him. That one looks on the offender objectively does not mean that one becomes incapable of feeling sympathy, fear, or irritation towards him. What it means, though, is that one no longer engages with him with an attitude of goodwill or regard and an expectation of its being reciprocated. While his actions can, then, provoke in one emotional responses, these responses do not belong to the class Strawson defines as that of the reactive attitudes. Hence, the objective view of someone, which pleas of this second type invite, is sharply opposed to the view one takes of another with whom one participates in an interpersonal relationship. Strawson makes this comparison between the two types of plea to bring out the difference between our viewing someone as a participant in an interpersonal relationship with us and our viewing him as an object whose behav­ ior is the product of the external forces and conditions to which he is subject and who is to be managed or controlled accordingly. The latter is consistent with the view we would take of people and their actions if we were to accept the thesis of determinism and regard them in its light. We are disposed to take such a view of someone when we see him as immature or dysfunctional in a way that incapacitates him for normal interpersonal relationships. In that case, we abandon the view of him as a participant with us in such relationships and regard him, instead, with detachment of a sort that an objective view of him entails. Nor is this the only occasion on which we would view someone objectively. We also, typically, take such a view of people generally when we are making social policy or designing a program to provide certain services to a given population. And we sometimes take such a view when, because of the strain of our relations with someone, we need to put distance between him and us of the sort that a detached attitude towards him creates. In each of these cases, we maintain an objective view only temporarily or only towards certain individuals. We could not, Strawson argues, permanently hold such a view of all people, for doing so would require our doing something we cannot in fact do, namely, cease to have interpersonal relationships. Nor, to continue Strawson’s argument, would it be wise permanently to adopt the objective view even if we could, for to do so would mean giving up human life as we know it, and abandoning our humanity is an unreasonable sacrifice. Determinism, Strawson concludes, is therefore no threat to the personal reactive attitudes. We would not cease to be liable to them even if we did accept its truth, nor would we have good reason to suppress them. Whether Strawson’s argument, at this point, succeeds in vanquishing the bogey of determinism is not my concern, however. What interests me

4 1

114  From Psychology to Morality

is how he uses the plain, natural facts about human sociability he adduces to explain our attributing moral responsibility to people and the correlative practices of moral censure and punishment. The explanation will, Strawson believes, bring out the omission in the optimists’ explanation of what justifies those practices that provokes the pessimists to endorse the metaphysical excesses characteristic of their position. It will thus furnish the optimists with a different way to justify these practices and at the same time offer an account of them that the pessimists can accept without resorting to their “panicky metaphysics.”17 To make this explanation Strawson adds to the class of reactive attitudes vicarious analogues of the personal reactive attitudes. These are attitudes one takes towards the goodwill or ill will of another, not in reaction to the manifestation of that will in the person’s actions towards one, but in reaction to its manifestation in his actions towards others. Strawson, to mark this difference between these attitudes and the personal reactive attitudes, assigns them, somewhat arbitrarily, different names. Specifically, he calls the vicarious analogue of resentment “indignation” or “disapproval.” This use of the terms ‘indignation’ and ‘disapproval’ is stipulative: in common speech, for instance, we are as comfortable with saying that one feels indignation at one’s own mistreatment as that one feels it at the mistreatment of another. Strawson restricts the term’s use to the latter case solely to enforce his distinction between personal and vicarious reactive attitudes, and I will observe the same restriction. Strawson also remarks that the vicarious character of indignation, that one can conceive of it as “resentment [felt] on behalf of another, where one’s own interest and dignity are not involved,” makes the adjective ‘moral’ an apt modifier of the attitude.18 Strawson, in other words, takes indignation to be a moral attitude in virtue of its being impartial or disinterested. He does not, then, in conceiving of it as a third-​party analogue of resentment, suppose that to feel it in response to someone’s injuring another in a way that shows a lack of due regard for the victim one must have a special attachment to or specially identify with the victim. Rather he supposes that people’s expectations of and demands for a reasonable amount of goodwill towards each other that form the backdrop of interpersonal relationships generalize as they, as individuals, come to recognize that everyone reasonably expects and demands such goodwill and regard on the part of others towards himself or herself in their transactions with him or her. The recognition of this general fact, Strawson further supposes, naturally disposes a person to respond with 17. Ibid., p. 93. 18. Ibid., p. 84.

15

Reactive Attitudes Revisited  115

indignation or disapproval when, say, the person witnesses an injury done to someone to whom he has no special attachment by another, who manifests attitudes towards the injured that fall short of the goodwill and regard for others that is generally expected of people in their transactions with others. Strawson, apart from a quick reference to human connection, does not say why it is natural for people’s expectations of and demands for a reasonable amount of goodwill from others towards themselves to generalize in this way. Nonetheless, one could, to explain this development, appeal in the way Hume did to something like our general sympathy with humankind. Strawson, in fact, seems to be tacitly following Hume in this part of his essay. Hume too distinguished attitudes that reflect a disinterested view from those reflecting an interested one and took the former to be moral because they reflected this view and occurred in response to a perception of the quality of the will manifested in the action that is their object. Indeed, Hume took the distinction between the personal or interested view and the general or disinterested view to be crucial to the explanation of our praising some people for their moral virtue and censuring others for their moral viciousness. Such praise and censure, Hume held, express the sentiments of moral approbation and blame that we experience in response to our perceiving behind the actions of those whom we praise and censure such motives as kindness and compassion, on the one hand, envy and malice, on the other, and for our perception of such motives to give rise to these sentiments we must perceive them from a general or disinterested view. We could not otherwise, Hume argues, explain our praising and censuring the heroes and villains of the ancient world, like the Roman senators who defended the republic at the time of its crumbling and the Roman emperors who came later and whose cruelty and injustice are infamous, since their actions can hardly affect our interests.19 What Strawson then adds to Hume’s account is the idea of certain moral feelings and attitudes being analogues of feelings and attitudes that reflect a personal or interested view. In addition to the vicarious reactive attitudes, Strawson introduces reflexive ones into the class of reactive attitudes. He calls these “self-​reactive attitudes.” They naturally arise, he suggests, as a result of one’s generalizing the expectations of and demands for goodwill on the part of others towards one and thus coming to expect and demand such goodwill of others towards each other in their transactions with themselves. For once one has generalized these expectations and demands, one realizes that they apply to oneself as 19. David Hume, Enquiries concerning Human Understanding and concerning the Principles of Morals, P. H. Nidditch, ed., 3rd ed. (Oxford: Clarendon Press, 1975), pp. 215–​218.

6 1

116  From Psychology to Morality

well, that one is no different from others in being someone of whom such goodwill in interpersonal relations is expected and demanded. Accordingly, one becomes liable to feeling obliged or bound to act as the goodwill towards others that is expected and demanded of one requires, to feeling compunction, and to feeling guilt or remorse if one fails to meet this requirement. These are what Strawson refers to as the self-​reactive attitudes. He mentions them, he says, for the sake of completing the picture.20 But he also notes that an agent who was insusceptible to the self-​reactive attitudes would be abnormal in a way that supported, were he to injure another, a plea like the pleas of immaturity and derangement that invite taking an objective view of him and thus suppressing the resentment or indignation towards him that otherwise would be forthcoming.21 Strawson, having completed his picture of the reactive attitudes, returns to the question of the threat determinism allegedly poses to our practice of attributing moral responsibility, and to show that it doesn’t, he reprises the argument he gave to show that accepting the thesis of determinism would not affect our being liable to the personal reactive attitudes. The argument, he believes, applies mutatis mutandis to the vicarious reactive attitudes, since they are analogues of the former. At the same time, he suggests that we can directly see in the argument’s application to this subclass of the reactive attitudes that our accepting the thesis of determinism is no threat to our practice of attributing moral responsibility. The argument, as applied to the vicarious reactive attitudes, affords this observation, he argues, because our susceptibility to these attitudes, unlike our susceptibility to the personal reactive attitudes, is the same as our expecting and demanding of everyone in our community some degree of goodwill or regard towards others generally in his or her dealings with them, and this generalized, disinterested expectation and demand amounts to an understanding of everyone in the community as lying under moral obligations to act with some degree of goodwill towards others generally in his or her dealings with them and as therefore someone to whom moral responsibility for failures to meet these obligations and for harms that result some such failures is, in the absence of an exculpatory plea, properly attributed. Attributions of moral responsibility, according to Strawson, require an understanding of people as moral agents and as members of a moral community, and the interpersonal relationships in which one participates are seen as falling among the relationships that a moral community comprises

20. Strawson, p. 84. 21. Ibid., p. 86.

17

Reactive Attitudes Revisited  117

only when one regards people in one’s community from a general or disinterested standpoint. Strawson initially makes this point by supposing someone whose view of himself and others is entirely interested or personal.22 Such a person is liable to the personal reactive attitudes but not to the vicarious ones. He is someone, then, who expects and demands from others a degree of goodwill towards himself in their dealings with him but does not expect of or demand from others such goodwill or regard in their dealings with each other. Such a person, Strawson contends, would be a moral solipsist. He would not, that is, see the interpersonal relationships in which he participates as embedded within a moral community. He would not see the demands he makes on others that they act with a degree of goodwill or regard towards him as general requirements to act with that degree of goodwill towards others to which everyone is subject and that as such represent moral obligations under which everyone lies. The ideas of belonging to a moral community, of having in common with others reciprocal moral obligations to act with a degree of goodwill towards one another, and of being, as everyone else is, morally responsible for actions that affect others do not arise, then, until one takes up the standpoint of others in one’s community and their expectations and demands on those with whom they interact.23 These ideas do not arise, in other words, until one regards oneself and others from a general or disinterested standpoint. Here too Strawson may be meaning to follow Hume in supposing that perceptions and judgments of moral relations and moral conduct must come from a view we share with others. The intelligibility and practical force of the language of morals, Hume argued, presuppose a common point of view, and therefore we must vacate our respective interested or personal viewpoints, which are peculiar to each of us, and take up a general or disinterested one

22. Ibid., p. 85. 23.  Strawson then reaffirms the point in considering how pleading, in response to injuring someone, that one was at the time incapacitated for normal participation in interpersonal relationships (“I was suffering hallucinations, undergoing a psychotic breakdown, in a trance, etc.”) affects our susceptibility to moral indignation. Thus, Strawson, after introducing the case by writing of seeing the agent in a different light, “as one whose picture of the world is an insane delusion; or as one whose behavior . . . is unintelligible to us” etc., goes on, “[A]‌bstracting now from direct personal interest, we may express the facts with a new emphasis. We may say: to the extent to which the agent is seen in this light, he is not seen as one on whom demands and expectations lie in that particular way in which we think of them as lying when we speak of moral obligation; he is not, to that extent, seen as a morally responsible agent, as a term of moral relationships, as a member of the moral community.” Ibid., p. 89.

8 1

118  From Psychology to Morality

(or suppose that we have), which is common to all, to take in the moral character of people’s relationships and actions within them and to speak properly of it. Nonetheless, though Strawson may be meaning to follow Hume here, he departs, I think, from him in supposing that we get our ideas of moral obligation and moral responsibility directly from generalizing the expectations of and demands for goodwill towards ourselves that we make on others, which is to say, by vacating our personal view and taking up a general or disinterested one. For Hume does not think we come to have these ideas until we apply the generalized expectations and demands that we acquire in taking up the general view to ourselves in those circumstances in which we find ourselves lacking the natural motives by which we typically manifest goodwill or due regard in our dealings with others. Specifically, Hume thinks we acquire these ideas as a result of a kind of indignation with ourselves for lacking the motives of goodwill towards others that we come to demand of everyone including ourselves when we take up the general or disinterested view.24 That is, seeing that we lack the motive of benevolence, say, or fellow feeling that in our dealings with another would show us to bear goodwill towards him and at same time expecting and demanding on his behalf that we so act, we reproach ourselves for our lack of benevolence and fellow feeling and are thereby moved by a sense of duty to do the action that we would have done had we been naturally moved by either to do it. It is this experience of self-​reproach for lacking the proper natural motive of goodwill and the consequent sense of duty to do what we would naturally do if we had that motive that gives us the idea of moral obligation, the idea that acts displaying goodwill or due regard for others in our dealings with them are obligatory even when we lack in ourselves the motives of benevolence or fellow feeling that would naturally produce them. And presumably, though Hume does not expressly say this, the same experience that gives us the idea of moral obligation gives us as well the idea of being morally responsible for doing or failing to do what we now understand as obligatory. Be this as it may, the important point is that Strawson, like Hume, holds that the acquisition of these ideas occurs only after one becomes susceptible to the vicarious reactive attitudes. Being susceptible to the personal reactive attitudes is by itself insufficient for having them.

24.  See Hume, A Treatise of Human Nature, p. 479.

9 1

Reactive Attitudes Revisited  119

III How do the recent, influential accounts of our practice of attributing moral responsibility that purport to draw on Strawson’s view in their appeal to the reactive attitudes come to misconstrue his understanding of them and their relation to the practice? How do they come to misappropriate his ideas? Our chief examples are the accounts of R.  Jay Wallace and Stephen Darwall that I  mentioned at the outset. A  crucial factor in their misunderstanding of Strawson’s view is a certain assumption about the reactive attitudes that they make in giving their accounts. Both assume that the interpersonal relationships to which these attitudes are integral are relationships among rational agents. Both assume, first, that a reactive attitude is essentially an attitude one takes towards a rational agent, either oneself or another. That is, both take the actions to which the reactive attitudes are responses as actions seen as done for reasons. Second, both conceive of the attitudes as containing beliefs or judgments about those reasons. On their conception of them, the reactive attitudes contain cognitions that are significantly more sophisticated than the cognitions that Strawson takes to be implicit in the personal reactive attitudes if not also the vicarious and self-​reactive ones. For Wallace and Darwall, the reactive attitudes have cognitive content that takes the form of a proposition, which necessarily makes them attitudes only rational beings can have. For Strawson, the cognitive content of a reactive attitude need not be propositional. Indeed, nothing in Strawson’s account of the intentionality of the reactive attitudes excludes social animals other than humans from being susceptible of them. His account fits well within de Waal’s observations of chimpanzees and their liability to gratitude and resentment. They too can discern in the behavior of other chimpanzees the latter’s goodwill or ill will towards them, and such discernment does not require a capacity for propositional thought. “[E]‌ven a dog distinguishes between being stumbled over and being kicked,” to quote a famous remark of Justice Holmes.25 Generally speaking, then, Wallace and Darwall conceive of the reactive attitudes as attitudes of rational minds responding to the actions of rational agents. This conception of them, particularly in the case of resentment, is, as I shall argue, at the root of their misunderstanding and misappropriation of Strawson’s ideas. Let me, then, first describe Wallace’s and Darwall’s accounts of the reactive attitudes as attitudes of which rational beings alone are susceptible and as appropriate only when taken in response to the actions of rational 25. Oliver Wendell Holmes, Jr., The Common Law (Boston:  Little, Brown and Co., 1881), p. 3.

201

120  From Psychology to Morality

agents. For the sake of brevity, I will, to the extent that I can, treat Wallace and Darwall together. Consider first their understanding of resentment. Like Strawson, Wallace and Darwall take resentment as integral to interpersonal relationships and take such relationships as the primary context in which to conceive of resentment. They hold that for a person to be susceptible of resentment he must believe that others with whom he is engaged in interpersonal relationships are required to act in certain ways towards him. The belief is implicit either in the person’s readiness to blame others for failing to meet expectations of how they should treat him to which he holds them or in his regard for himself as having authority to demand of others that they act in certain ways towards him.26 What provokes one person P to feel resentment towards another Q, then, is Q’s violating a requirement imposed on him to act in certain ways towards P, which is to say, Q’s failing either to meet an expectation to which P holds him or to comply with the authoritative demands P addresses to him. Q’s failure, then, appears to P as an offense that Q commits against him. In other words, resentment implies a judgment that the person whom one resents has committed an offense against one. Consider next Wallace’s and Darwall’s conceptions of indignation and guilt, the other types of reactive attitude that are of particular interest to them. These conceptions are modeled on their understanding of resentment. Accordingly, susceptibility to indignation requires a belief that others are required to act in certain ways towards each other. This belief is implicit in one’s readiness to blame others for failing to meet expectations of how they should treat each other to which one holds them or in one’s regard for oneself as having, in virtue of belonging to the same moral community as others, the authority to address demands to them to act in certain ways towards each other. Thus, when one feels indignation at someone, one judges that he has committed an offense against someone else, distinct from oneself. Likewise, susceptibility to guilt requires a belief that one is required to act in certain ways towards others. This belief is implicit in one’s readiness to blame oneself for failing to meet expectations of how to act towards others to which one holds oneself or in one’s regard for oneself as subject to the authority of others to make demands on oneself to act in certain ways. Thus, when one feels guilt, one judges that one has committed an offense against another. Hence, resentment, indignation, and guilt are, on Wallace’s and Darwall’s account of them, direct correlates of each other. Each entails the same sort of judgment, namely, with regard to two people x and y, that x has committed an offense against y. What differentiates these attitudes is the standpoint from 26.  Wallace, pp. 21–​22; Darwall, p. 68.

2 11

Reactive Attitudes Revisited  121

which this judgment in each case is made. In feeling resentment, one makes the judgment from the standpoint of the victim; in feeling indignation, one makes the judgment from the standpoint of a third party; and in feeling guilt, one makes the judgment from the standpoint of the offender. Thirdly, on Wallace’s view, when one holds others to an expectation, one believes they have a reason to meet that expectation.27 Similarly, on Darwall’s view, when one takes oneself to have authority to address demands to others, one believes that those demands, being authoritative, give them reason to comply.28 Consequently, both Wallace and Darwall conceive of the agents towards whom or whose actions the reactive attitudes are taken as rational. That is, on their conceptions of the reactive attitudes, rational agents or their actions are the intentional objects of those attitudes. Moreover, each builds this feature into his conception of the reactive attitudes for the same reason. For both take as a condition of one’s being morally responsible for an action that one is capable of being moved by reasons that morality supplies. Hence, to explain, as they do, our practice of attributing moral responsibility to people for their actions as an expression of our susceptibility to reactive attitudes of resentment, indignation, and guilt, requires conceiving of those attitudes as essentially attitudes towards rational agents and their actions. At the same time, they build this feature into their conceptions in different ways, and this difference needs to be grasped in order to see why susceptibility to the reactive attitudes makes no real contribution to either of their explanations. To do this, it is sufficient to consider the susceptibility to resentment and how they conceive of its intentionality. Wallace takes from Strawson the idea that resentment is an emotional response to intentions and attitudes manifested in the actions that produce it. But instead of taking these intentions and attitudes to be states of ill will or indifference towards the person who feels resentment, as Strawson does, he takes them to be choices.29 They are the choices from which those actions issue, and being choices, they are necessarily made on the basis of what their agents see as reasons for those actions. In feeling resentment towards someone, one judges that the person’s reasons for acting as he did are inadequate to justify his action, and since his action fails to meet the expectations to which one holds him, one thereby judges his action to be an offense committed against one. Hence, Wallace, in shifting to choices from states of ill will and indifference, abandons Strawson’s conception of resentment as a distinctly personal 27.  Wallace, pp. 35–​36. 28.  Darwall, pp. 74–​79. 29.  Wallace, p. 128.

221

122  From Psychology to Morality

attitude. On Wallace’s conception, that is, resentment towards someone does not include any sense of that person’s being ill-​disposed towards or contemptuous of one. What it includes is a belief that the reasons on which the person based his choice were weaker than the reasons he had to meet the expectations to which one holds him. When those expectations are based on moral obligations that the person has to one, then what Wallace does is to replace Strawson’s conception of resentment as a distinctly personal attitude with a conception of it as a moral attitude. It is the attitude of someone whose rights have been violated. But because Wallace does not conceive of the reactive attitudes as essentially moral attitudes, there is a hole in his account of the intentionality of resentment. For nothing explains why, when one’s resentment of another is a nonmoral attitude, one should think the reasons on which the person based his choice are weaker than the reasons he has to meet the expectations to which one holds him, or even that he has any reasons to meet those expectations.30 Wallace’s general account of resentment is therefore problematic. In this respect, Darwall’s account is an advance over Wallace’s. For on Darwall’s account, to be susceptible to resentment is to regard oneself as having authority to address demands to others.31 And to regard oneself as having such authority is to assume that others recognize and will be guided in their actions by that authority. It is to assume, in other words, that the demands one addresses to others are authoritative and therefore give them reasons to do what one demands of them. One’s authority, moreover, on Darwall’s account, is the moral authority one has as a member of the moral community. Darwall, like Strawson, takes the interpersonal relationships to which the reactive attitudes are integral as relationships one can see as constituting a moral community, but unlike Strawson, he does not suppose that to see them as such one must regard them from the disinterested view or that one must generalize the expectations of and demands one has and makes for the goodwill of others towards one by virtue of which one is susceptible to resentment. Rather Darwall holds that susceptibility to any of the reactive attitudes, including resentment, implies regarding oneself and others as members of a moral community. He holds this because he conceives of each attitude as a form of addressing authoritative demands to others or oneself or making 30.  Wallace, pp.  36–​37, gives as an example resentment felt in response to a breach of etiquette. The example, however, is not intelligible without further explanation. A breach of etiquette might provoke contempt or disdain since it is normally taken by those to whom such things as where one places the dessert fork in setting the table matter as a sign of the offender’s inferiority or vulgarity. But it could hardly as such provoke resentment. 31.  Darwall, pp. 57–​61 and 67–​68.

213

Reactive Attitudes Revisited  123

valid claims on them or oneself and thinks that such demands and claims require, to be authoritative or valid, norms commonly recognized as justifying them and thus serving to bind people together in a community. These norms, then, and the justification of the demands and claims addressed through the reactive attitudes that they provide explain why the demands and claims give reasons for action to those to whom they are addressed. Thus, Darwall avoids the problem in Wallace’s conception of resentment by conceiving of it and indeed of every reactive attitude as a moral attitude. Yet on this point, the distance between Darwall’s understanding of the reactive attitudes and Strawson’s could not be greater. Nothing in Strawson’s account supports Darwall’s idea that susceptibility to the reactive attitudes implies a regard for oneself as having authority to address demands to others or oneself. To be sure, Strawson says that one’s being susceptible to resentment consists in one’s expecting and demanding of others a certain degree of goodwill towards oneself, but that one expects and demands something from others hardly implies that one regards oneself as having authority to make demands on them. Bright pupils who like to shine in the classroom or small children who like to perform and receive applause from adults often expect and demand attention, but this doesn’t mean they think of themselves as having authority to make demands that others pay attention to them. Nor is there anything in Strawson’s account to support Darwall’s idea that susceptibility to any of the reactive attitudes implies regarding oneself and others as members of a moral community. To the contrary, the point of Strawson’s reflections on the possibility of moral solipsism, the possibility of someone’s being susceptible to the personal reactive attitudes but not to the vicarious ones, is to show the necessity of taking the disinterested view for seeing the interpersonal relationships in which one participates as embedded in a moral community. Strawson’s strategy for finding a way to reconcile the optimists and the pessimists is, as I suggested, to show how certain natural facts about human sociability, facts that uncontroversially fit within a deterministic universe, support our practice of attributing moral responsibility. To understand the reactive attitudes as moral attitudes in the way Darwall does, that is, as attitudes to which we are susceptible only by virtue of our belonging to a moral community in which each of us has authority to address demands to others or himself and in which there are norms justifying that authority, ill-​ serves that strategy. On Strawson’s account of resentment, the attitude is a natural response to the manifestation, in an action that harms one, of the actor’s ill will or indifference towards one. It does not in its original form contain any thought of the actor’s being morally responsible for harming one, though after one

241

124  From Psychology to Morality

becomes susceptible to the vicarious and self-​reactive attitudes such thought may also characterize it. One’s focus is on the actor’s appearing to be ill-​ disposed or indifferent to one, and one is thus sensitive to pleas from him or others to the effect that the reality is different from the appearance. Such pleas take the form of excusing the person from responsibility for his actions only after he is understood to be subject to expectations and demands that apply to everyone engaged in the relevant network of interpersonal relationships. In other words, they take this form only when one regards him from the gen­ eral or disinterested view. On Strawson’s account, then, resentment, being a personal reactive attitude, does not originally include the thought of the person resented as morally responsible for his actions.32 Matters are decidedly different on Wallace’s and Darwall’s accounts. On Wallace’s account, resentment is a response to another’s choice to contravene the expectations to which one holds him that he act in certain ways towards one. That one should feel resentment at the contravention of such expectations makes sense when one regards those expectations as having the force of moral obligation. It does not otherwise, as I’ve argued. So let us limit Wallace’s account to the former. To take the expectations to which one holds another as having the force of moral obligation is to suppose that that person must answer to someone if he contravenes them. This point is no doubt what Darwall means to be capturing by attributing to the members of a moral community authority to address demands to one another. For to have authority to demand that another act in certain ways is for that person to be answerable to one for failing to act as one has demanded. Since answerability for violating a moral obligation is essentially the same thing as moral responsibility for that violation, resentment at someone’s having violated a moral obligation to one, on Wallace’s and Darwall’s accounts, contains the thought of the offender’s being morally responsible for his offense. The same holds for the other reactive attitudes that interest them, indignation and guilt. Consequently, neither of their accounts of the reactive attitudes can advance our understanding of our attributions of moral responsibility for actions. On their accounts, our being susceptible to these attitudes presupposes our attributing moral responsibility, so it cannot help to explain our practice of attributing moral responsibility. To the extent that Wallace and Darwall are

32.  Here the same observation applies as the one Strawson makes about the arbitrariness of his limiting his use of ‘indignation’ to its being a name for an attitude that arises towards someone only when one regards him from a disinterested view. Ordinarily, we use ‘resentment’ to name an attitude that arises from such a view, but Strawson has somewhat arbitrarily restricted its use in taking it as a name for a personal reactive attitude.

215

Reactive Attitudes Revisited  125

concerned with essaying that practice, the reactive attitudes, as they understand them, merely decorate their essays.

IV At the crux of my criticism of Wallace’s and Darwall’s appropriation of Strawson’s ideas is the observation that they, unlike Strawson, treat our susceptibility to resentment as an aspect of our rationality. Strawson, by contrast, takes it as an aspect of our sociability. The two, human rationality and human sociability, are not the same. Nor is there reason to think that our social nature springs from our rational nature. Children, after all, form interpersonal relationships before they show any signs of having rational powers. And though Strawson concentrates on interpersonal relationships between adults, his understanding of the personal reactive attitudes does not exclude them from the relationships young children have with their parents, other caregivers, siblings, and playmates. As I noted in Essay 4, his understanding is consistent with accounts of a child’s moral development that take as the catalyst for its acquiring a sense of justice the capacity to recognize in its parents and other caregivers’ nurturing behavior their concern for its well-​being.33 On these accounts gratitude, resentment, love, and hurt feelings are not alien to the experiences of young children. One might object to this reading of Strawson by citing Strawson’s account of the different pleas someone who has become the object of another’s resentment might make to persuade the latter to give up feeling resentment towards him. That is, one might object that this account shows that Strawson understands resentment as implying a belief that the person who is the object of one’s resentment is rational. For among the pleas of the second type that Strawson describes are those of immaturity and lunacy, and either, understood as offering reason to cease feeling resentment, suggests that only the actions of rational agents are properly resented. Further, the account, so the objection continues, shows as well that Strawson takes being rational to be a condition of being susceptible to resentment. For one could not have the belief that another was rational unless one was rational oneself. Hence, so the objection concludes, Strawson too took being rational as a condition of this susceptibility and took the belief that those towards whom resentment is felt are rational agents as essential to the attitude.

33.  See p. 91 above.

261

126  From Psychology to Morality

This objection, however, rests on a mistake about the import of the pleas of immaturity and lunacy. To see this mistake, it is necessary to attend to Strawson’s division of the various pleas he surveys into two types. For not all of these pleas tell us something about the character of resentment’s intentional object. In fact, only pleas of the first type do. Only pleas of this type indicate what someone must believe about another or discern in the other’s actions if he is to be truly described as feeling resentment towards that person. Pleas of the second type, which include pleas of immaturity and lunacy, on the other hand, tell us something about the features of people that incline us to view them objectively and so to suppress our tendency to view them as engaged with us in interpersonal relationships. But this is not to tell us anything about the character of resentment’s intentional object. Consequently, Strawson’s acknowledging pleas of this second type does not show that he takes resentment to be, in virtue of the character of its intentional object, a response to harmful actions whose agents are viewed as rational. Nor, then, does it show that he thinks being rational is a condition of being susceptible to resentment.34 Thus, consider pleas of the first type. They are offered to people who feel resentment at having been injured in the hope of disabusing them of their view of the person who injured them as having acted with ill will or indifference towards them. For ease of reference I’ll use ‘S’ to refer to anyone who feels such resentment. S expects and demands a certain degree of goodwill or regard towards himself by others in his dealings with them, and the resentment S feels at having been injured is a response to that action and to its manifesting in the offender a deficiency in the goodwill or regard towards him S expects and demands. If S accepts a plea of the first type (e.g., that the harm was accidental, unintentional, inadvertent etc.), he will cease to feel resentment towards the offender because he will cease to regard the latter as having acted with either ill will towards him or indifference to how he fares. A plea of the first type is therefore a representation of the action as not having the character that S sees it as having. Hence, a plea of this type tells us something about the character of resentment’s intentional object. It tells us, that is, how S views the action he resents since it is offered to persuade S that his view of the action is false. Pleas of the second type, by contrast, are offered to S in the hope of inducing him to cease viewing the offender in the way one views those with 34. My point here about the difference between the two types of plea corresponds to Bennett’s distinction between disqualifying and dispelling an attitude. See “Accountability,” pp. 28–​30.

217

Reactive Attitudes Revisited  127

whom one is interpersonally engaged and to view him objectively instead. If S accepts one of these pleas, he withdraws from being susceptible to feeling any of the personal reactive attitudes towards the offender. Accordingly, one could say that pleas of the second type tell us that S views the actions of those towards whom he feels resentment or any of the other personal reactive attitudes as actions of people with whom he is engaged in interpersonal relationships. But this is just to say that to be susceptible to the personal reactive attitudes is to view those towards whom one is liable to feel those attitudes as each someone with whom one is engaged in some form of interpersonal relationship, a form of relationship defined by one’s expectations of and demands for a certain degree of goodwill towards one on the other’s part. It tells us, in other words, nothing more about the character of resentment’s intentional object than that it is the action of someone with whom one is so engaged. Since being rational is not a condition of engagement in an interpersonal relationship, S’s ceasing to feel resentment towards the offender upon acceptance of a plea of the second type, a plea of immaturity or lunacy, in particular, does not show that S, in initially resenting being injured, viewed the offender as a rational agent. It shows only that S viewed him as someone from whom he expected and demanded a certain degree of goodwill. Plainly, then, even if S accepts a plea of the second type because it persuades him that viewing the offender objectively is warranted, S need not think his initial view of the latter as having shown ill will towards him was mistaken. As Strawson observes, S always has the option of viewing someone with whom he participates in an interpersonal relationship objectively. So S’s taking an objective view of someone does not necessarily mean that he regards the person as unsuitable for being viewed alternatively as someone from whom he expects and demands a certain degree of goodwill. S may so regard this person even when the reason he takes the objective view of him is some feature about him S thinks warrants taking that view of him. For instance, Strawson notes that seeing someone as “peculiarly unfortunate in his formative circumstances” disposes one to take the objective view of the person.35 Yet being unfortunate in one’s formative circumstances hardly implies that one cannot bear ill will towards others or be indifferent to how they fare. To the contrary, some of the most disturbing acts of malice or profound indifference to the humanity of others are the work of people brought up in horrific circumstances and by monstrous parents. In viewing such a person objectively, one does not deny the ill will or indifference behind his

35. Strawson, p. 79.

281

128  From Psychology to Morality

actions, ill will or indifference that would inspire one to feel resentment were one to view him as someone of whom one expects and demands a certain degree of goodwill or regard. Rather one judges him to be more suitably viewed as someone to be controlled, managed, or perhaps just avoided than as someone with whom to engage interpersonally. Of course, the younger or more deranged a person is the harder it is to view him as someone with whom to engage interpersonally. The less a person’s behavior is evidence of a stable set of attitudes and intentions, the less one deals with him with the expectation of his having goodwill or regard towards one and the more one’s dealings with him are attempts at managing, controlling, or engineering the events involved in those dealings. A tendency, in other words, to take the objective view of someone becomes stronger the more the person’s actions appear mindless, erratic, or detached from reality. This tendency no doubt gives rise to a consensus about not expecting or demanding of the actions of the young or the deranged that they manifest the degree of goodwill or regard that is expected and demanded of adults of sound mind. Hence, it gives rise to social norms proscribing feeling resentment at the actions of the young and the deranged. Or to put this development as Strawson would, the consensus first gives rise to social norms concerning the impropriety of feeling indignation at the actions of the young and the deranged since it would represent a generalized judgment, one made from the disinterested view, and sensitivity to this norm, when one has been injured by a child or a lunatic, would then yield the corresponding norm concerning resentment. Thus, we can understand how pleas of immaturity and lunacy gain the backing of social norms and so how small children and lunatics come to be considered to be outside the class of morally responsible agents. This exclusion of them from the class of people whose actions are the proper objects of the reactive attitudes is due, not to the character of those attitudes’ intentional objects or anything else essential to those attitudes, but rather to the emergence and entrenchment of social norms that proscribe feeling resentment, indignation, and the like towards them. One cannot, in other words, take our susceptibility to these attitudes as in itself a source of knowledge about the conditions of moral agency, specifically that it requires being rational or capable of acting for reasons.

291

6

Is Empathy Required for Making Moral Judgments?

1.  Hume’s Account of Our Moral Sensibilities The problem I’ll be discussing in this essay begins with Hume. At any rate, Hume’s ethics is as far back as I have traced it. Hume, in A Treatise of Human Nature, asks at the beginning of book III whether moral distinctions derive from reason or feeling. The chief distinction that interests him is that between virtue and vice, and the way he puts the question is to ask whether we distinguish between virtue and vice by means of our ideas or our impressions (1978, p. 456). Nonetheless, we can be certain, given the argument in part 1 of book III, that Hume in asking this question understands ideas and impressions as proxies for reason and feeling. His principal question, then, is whether we distinguish between virtue and vice by relying on the deliverances of reason or those of feeling. Asking it in terms of ideas and impressions suits his using, to answer it, apparatus and results that he had previously introduced and derived in books I and II. In asking the question, Hume presupposed that the human mind has certain sensitivities or powers that enable its possessor to distinguish between virtue and vice. For answering it requires determining which of the mind’s sensitivities and powers are stimulated or exercised when its possessor

129

301

130  From Psychology to Morality

concludes of an action that it is noble and exhibits virtue or, alternatively, that it is depraved and springs from vice. Hence, in asking the question, he fastened on us, if not for the first time, then at least much more securely than any predecessor who gave a similar answer, the idea that to understand the nature of moral judgment requires understanding the distinctive mechanisms of the mind that produce such judgments. On this idea, moral judgment is a type of judgment that we distinguish from other types, not by its subject, but rather by the peculiar operations of the mind that produce its tokens. That is, on this idea, we distinguish moral judgments from other types in the same way as we distinguish visual judgments from tactile ones and not in the way we distinguish arithmetic judgments from geometric ones. The idea fits Hume’s project of applying the methods of natural science, what he calls experimental philosophy, to the subject of morals since it focuses on the mechanisms of the mind that produce moral judgments. Such a focus lends itself to a scientific search for the natural causes and conditions that explain the phenomenon in question. This focus, I believe, is problematic, and in the discussion to follow I will lay out the trouble it causes in the study of moral judgment. Let us begin the discussion with an explanation of the answer Hume gave to the question with which he opens book III. His answer is that moral distinctions derive from feeling. And he argues for this answer by excluding ideas from being the means by which we distinguish between virtue and vice. Summarizing his argument, he writes, “Our decisions concerning moral rectitude and depravity are evidently perceptions; and as all perceptions are either impressions or ideas, the exclusion of the one is a convincing argument for the other. Morality, therefore, is more properly felt than judg’d of” (1978, p. 470). His project going forward, then, is to identify the impressions by which we distinguish between virtue and vice and to explain how they arise in the mind. Those impressions, he tells us, are certain qualitatively distinct pleasant and unpleasant feelings that follow ideas that have a specific type of content. The objects of those feelings are human beings, either oneself or others, and the ideas the feelings succeed represent features of their objects the thought of which we find satisfying or displeasing. Hume abstracts this schema from observing that we take pleasure in contemplating or witnessing generous, heroic, and patriotic actions and are distressed when we see or hear about cruel, duplicitous, and cowardly actions. The former and similar acts of moral rectitude or goodness gladden our hearts and lift our spirits; the latter and similar acts of moral degeneracy and weakness unsettle and displease us. In the former cases, the pleasure we take gives way to sentiments of moral

131

Is Empathy Required for Making Moral Judgments?  131

approbation or esteem; in the latter the displeasure gives way to sentiments of moral disapprobation or blame. These sentiments are a distinct or, as Hume says, “peculiar kind” of pleasure and pain, which “makes us praise or condemn” (1978, p. 472). Moral approbation and disapprobation, Hume goes on to observe, are not restricted to ourselves or those near to us. The heroism of our enemies as well as that of our friends can elicit our esteem. The cruelty of tyrants who ruled centuries ago and over people unrelated to us can arouse our disgust. And because they can, our feelings of approbation and disapprobation do not depend on any personal interest we take in how the objects of these sentiments, by their actions, benefit or harm us. Thus to explain how these sentiments arise in the mind, Hume concludes, one must look to mental operations that do not come from self-​love. To explain them, that is, one must look to mental operations that occur when one views things generally and disinterestedly. Hume’s explanation is complicated. It is based on two facts about human psychology. First, Hume writes, “The minds of men are similar in their feelings and operations” (1978, p. 575). Second, he observes immediately afterwards, “As in strings equally wound up, the motion of one communicates to the rest; so all the affections readily pass from one person to another, and beget correspondent movements in every human creature” (1978, p. 576). The communication of feeling to which Hume refers in stating this second fact takes place through the workings of sympathy. Sympathy, as Hume understands it, is not a distinct passion, like pity or benevolence. Rather it is a principle of the mind whose operation turns an idea of another’s feeling into that very feeling. Thus, when one sees in another’s face or posture or infers from his conduct that he is feeling pain, pleasure, anger, joy, sadness, or the like, one forms an idea of what he is feeling. And then, as a result of the influence of sympathy, this idea is enlivened into the feeling itself. Hume, accordingly, constructs from this principle of sympathy and his premiss that the minds of human beings are all similar to each other an explanation of how feelings that are concordant with the feelings of another and that result from the workings of sympathy give rise to esteem or blame for whoever brought pleasure or pain, as the case may be, to the object of one’s sympathy. The details of this explanation need not concern us. What matters is Hume’s making sympathy its pivotal factor. The conversion that sympathy brings about of an idea of another’s feeling into the feeling itself connects one with that person in a way that makes his feelings matter to one when, as feelings of pleasure or pain, they serve as tokens of his happiness or

321

132  From Psychology to Morality

unhappiness. Having those feelings, Hume holds, yields favorable or unfavorable sentiments towards whoever’s actions brought pleasure or pain to that person. In other words, the sentiments of approbation and disapprobation, esteem and blame, result from one’s first being so connected to another’s pleasure or pain and then taking the party or parties responsible for it as the sentiments’ objects. Hume thus, in attributing this mechanism of connection to the human mind, gave an account of the phenomena of having feelings in common with another that we now typically explain as resulting from empathy. His explanation of how we make moral distinctions, distinctions between virtue and vice, in particular, thus represents a view of moral judgment as requiring empathy.1 And what is more important is that it represents more generally an explanation on which the ability to make these distinctions is due to certain native sensitivities and tendencies of mind that are distinct from the mental powers by which we make other distinctions, like those between sweet and sour, rain and snow, night and day. At the same time, this feature does not alone distinguish Hume’s explanation of moral judgment from the traditional explanations against which he was arguing. The prevailing tradition at the time was that of Christian ethics. In this tradition, the seat of moral knowledge is located in conscience, and we determine whether an act is one of virtue or vice by exercising this knowl­ edge. Hence, given the Christian tradition’s understanding of conscience as a distinct power, native to the mind, its explanation is like Hume’s. It too attributes our ability to make distinctions between virtue and vice to a certain mental power that is distinct from mental powers by which we make other distinctions. Nonetheless, there is a significant difference between them. On the traditional explanation, because the judgments of virtue and vice are exercises of moral knowledge, the location of that knowledge in conscience, however important that may be to the broader theory of moral agency, is not essential to the ability of men and women to make these judgments. If moral knowledge were located in some other organ of thought than conscience, 1. Hume allows one small exception:  when one finds another’s witty conversation or charming demeanor, for instance, immediately agreeable, the pleasure of the other’s company can produce the sentiment of approbation directly. In this case, the sentiment and so one’s discernment of virtue arise without the intermediation of empathy since one does not need to enter empathically into anyone else’s mind to experience the pleasure that elicits the sentiment. Noting these exceptions, Hume then quickly moves on to observe that taking such conduct as virtuous in other cases “has a considerable dependence on the principle of sympathy so often insisted on” (1978, p. 590).

313

Is Empathy Required for Making Moral Judgments?  133

they could still exercise it. They would still, therefore, have the ability to determine whether an action was virtuous or vicious, and the power in virtue of which they would have this ability would then be the distinct mental power to which moral judgments were attributed. On Hume’s explanation, by contrast, such judgments could not be made if men and women lacked sympathy and were unable to view things disinterestedly, that is, from a view uninformed by self-​love or other kinds of partiality, since without sympathy or the ability to take this disinterested view, human beings would be incapable of the sentiments of approbation and disapprobation that are the vehicles of moral opinion. The difference, then, between the two explanations rests on a difference between the content of moral judgments, on the one hand, and their psychological profile, on the other. On the traditional explanation, one makes these judgments using knowledge of moral laws or precepts the typical exposition of which reduces them to one or a few basic precepts of right and wrong. These basic precepts supply, then, the content of morality and consequently the content of moral judgments. If, for instance, the sole basic precept were the Golden Rule, then the content of morality would be conduct among human beings that promoted reciprocity in their relations with each other. On Hume’s explanation, by contrast, one makes moral judgments by means of viewing things disinterestedly and under the influence of sympathy. Given his understanding of sympathy as a mechanism that works to convert an idea of another’s feeling into that very feeling and the subsequent favorable or unfavorable sentiment that results from that feeling as representing moral judgment, the determinant of moral judgment is its psychological profile and not its content. It will be useful, at this point, to distinguish moral judgment from other moral opinion. Hume suggests this distinction when he declares in the passage I  earlier quoted that “Morality  .  .  .  is more properly felt than judg’d of” (1978, p. 470). Plainly, this declaration implies that Hume takes judgments to be exercises of reason, for his point is to affirm that moral distinctions derive from feeling and not from reason. Accordingly, in deference to Hume, I propose to take moral judgments to be those moral opinions that one reaches by applying moral knowledge or, more exactly, moral beliefs that, if true and well-​founded, would be knowledge to the events or facts before one and to understand that in so applying such knowledge or beliefs one exercises reason. Thus, the explanandum of the traditional explanation is moral judgment. The explanandum of Hume’s explanation, by contrast, is not, or at least not in the first instance. Hume, indeed, denies that the

341

134  From Psychology to Morality

opinions one forms directly about people’s virtues and vices are judgments since he denies that they are exercises of reason.2 Hume, to be sure, allows that to be consistent in one’s moral opinions of people who engage in actions of a given type sometimes requires that one adjust one’s feelings to conform to the opinions one forms of them when one’s view of those actions is clear and undistorted. It is possible, then, to interpret him, as some commentators do, as having thereby introduced moral judgments into his theory and, as a result, gone against his thesis that we distinguish between virtue and vice by means of our feelings. These commentators’ argument is that such adjustments require one to judge what feelings one would have if one were viewing the actions from a clear and undistorted view and such judgments follow from applying what, on Hume’s theory, would amount to moral knowledge or beliefs (Sayre-​McCord, 1994). But I do not believe such an interpretation is necessary. The adjustments to feeling Hume allows are like the adjustments we make to our visual perceptions of objects of a given type to bring them in line with how such objects appear to us from a similarly clear and undistorted view. That we perceive an object as having, say, the same hue through various changes in light is due to the way the mind transforms the images of the object on the retina into perceptions of the object’s color, and it seems better to suppose that such transformations are built into the machinery of perception than that they consist in exercises of reason in which we apply beliefs about color and light to the objects before us. In general, therefore, Hume’s explanation represents the view that what makes one’s favorable or unfavorable response to someone a moral opinion is its consisting in qualitatively distinct pleasant and unpleasant feelings that accompany the idea of that person and that occur when one views things disinterestedly and under the influence of sympathy. It thus represents the view that the determinant of moral opinion is its psychological profile and not its content.

2.  Hume, despite implying that our discernment of moral distinctions consists in feeling rather than judgment, nonetheless sometimes uses ‘judgment’ to describe it. For example, immediately after introducing the subject of book III in its opening paragraphs, he refers to “those judgments by which we determine moral good and evil” (1978, p. 456). And later in his discussion of natural virtue, he writes of the two ways in which a person’s motives or character gives rise to sentiments of approbation or disapprobation, “My opinion is, that both these causes are intermix’d in our judgments of morals; after the same manner as they are in our decisions concerning most kinds of external beauty” (1978, pp. 589–​590). I take these occurrences of ‘judgment’ as showing a lack of consistency in Hume’s use of ‘judgment’ and not deviation from his thesis about the means by which we make moral distinctions. On his use of ‘judgment’ in such contexts see Stroud, 1977, p. 172n1.

315

Is Empathy Required for Making Moral Judgments?  135

2.  A Problem with Hume’s Account A view like Hume’s is obviously untenable without further explanation. One cannot, after all, deny that people make judgments about duties they have and do so by applying a set of beliefs they hold to the facts or events before them. G. E. M. Anscombe’s example of owing a grocer money as a result of his both supplying one with potatoes and sending one a bill nicely makes the point (1958a and 1958b). Defenders of the view must therefore give some explanation of these judgments as either not really moral or not really judgments. Hume, plainly, accepts them as judgments. He too sees that when we are presented with certain facts or events, we apply a set of beliefs to those facts or events in determining that someone has trespassed on another’s land or otherwise invaded his property or that someone has gone back on his word or otherwise reneged on a promise. What Hume denies is that these judgments are moral judgments. To be sure, he observes that they do produce in the mind of people who view these facts or events disinterestedly and under the influence of sympathy the sentiment of blame the object of which is the trespasser or promise-​breaker. But in a mind that views them self-​interestedly they do not. His famous worry about the sensible knave who acts justly contrary to his own selfish interests only when his omitting to do so would cause “a considerable breach in the social union” implies as much (1975, p. 282). In this way, Hume keeps separate judgments that can give rise to moral sentiments from the sentiments themselves. Of course, these judgments, being judgments about people’s rights and duties, are moral judgments in a perfectly ordinary sense of the word ‘moral’. They concern matters of justice, and judgments about matters of justice are moral judgments. So to defend a view like Hume’s one must justify taking the word ‘moral’ in a different sense. The justification comes from the argument in book III, part  1, of the Treatise whose summarization by Hume we earlier noted. Hume rests this argument on his thesis that morality is essentially practical (1978, p.  475). Let us call this thesis “Hume’s internalist thesis.” In saying that morality is essentially practical he means that it is in the nature of morality to have influence on our passions and actions. Accordingly, Hume takes the word ‘moral’, when used to qualify names of mental states, to restrict the mental states that those names denote to states that have an influence on our passions and actions. His internalist thesis is thus what supports his use of ‘moral’ in this way. It justifies, that is, his taking ‘moral’ in a different sense from the one on which judgments concerning matters of justice are ipso facto moral. To defend a view like Hume’s, then, one must offer a similar justification.

361

136  From Psychology to Morality

The dispute over Hume’s internalist thesis, as he understood it, has engaged the interest of moral philosophers for over a half century. It arose in metaethics when specialists in the field came to realize that the question the thesis raises about the motivational influence of moral judgments is independent of questions about the definability of ethical terms and the deducibility of ethical sentences from sentences that are not ethical (Frankena, 1958). As a result of this realization, studies in metaethics expanded into areas of psychology that had previously been at most tangential to the inquiries into the semantics of ethical language and the logic and epistemic status of ethical arguments that had been the primary preoccupations of moral philosophers in the analytic tradition. This expansion corresponded to a shift from treating questions about the analysis of moral judgments as questions about their content, that is, whether the basic terms expressing this content were definable in naturalistic terms, indefinable, or noncognitive, to treating them as questions about the psychological profile of moral judgments, whether motivation was built into them and, if so, by virtue of which operations of the mind. This shift helped to entrench Hume’s idea that to understand the nature of moral opinions requires determining the sensitivities and powers of the mind whose stimulation and workings produce them. At the same time, because many philosophers still held to the use of the word ‘moral’ in the ordinary sense in which the only criterion of application is the content of the cognition or affect one uses the word to describe, the dispute in metaethics over the question Hume’s internalist thesis raises reached, after several decades of fierce debate, an impasse. To date philosophers have not found a way out of this impasse (Enoch, 2011, pp. 249–​252).

3. A Proposal In a paper I  wrote some years ago, “Empathy and Universalizability,” I proposed, in view of this impasse, to study the question that Hume’s thesis raises as a question in psychology rather than as a question in metaethics (Deigh, 1995). My proposal was to accept the common description of psychopaths as people who know the difference between right and wrong but are nonetheless amoral in the sense of lacking such marks of a moral personality as compunctions of conscience and a liability to feelings of guilt, and to ask whether their amorality, while consistent with their being competent to make moral judgments of one type, is due at bottom to a cognitive deficit and therefore to an inability to make moral judgments of a different type. My thought was that an affirmative answer could vindicate Kant’s supposition of

317

Is Empathy Required for Making Moral Judgments?  137

the possibility of our making judgments of the type he characterized as categorical imperatives, that is, judgments of duty that, on their own, influence their makers to act in accordance with them. The cognitive deficit I attributed to psychopaths was their lacking the capacity for empathy, where I understood empathy to be a state of understanding what is going on in another’s mind that one attains by taking that other person’s perspective. Accordingly, my hypothesis was that having a capacity for empathy in addition to the cognitive capacities that people who bear the marks of a moral personality have in common with psychopaths gives a person the capacity to make judgments of the type Kant characterized as categorical imperatives. And I understood the examination of this hypothesis to lie outside of metaethics since one could undertake it, as I had proposed, independently of questions about the nature of moral judgment. Subsequent criticism of my hypothesis has convinced me that I did not go far enough in proposing to treat the question that Hume’s internalist thesis raises as a question in psychology rather than metaethics. One criticism, due to Jeanette Kennett, rests on a comparison of psychopathy with autism (Kennett, 2002). While both conditions, according to Kennett, include a lack of empathy, only psychopathy is a condition of amorality. Autism, by contrast, is a disorder whose severity varies from the extremes of mental retardation to more mild forms that are not so disabling as to make those subject to them incapable of functioning productively in society, and such people can have a conscience and be guided by it. Hence, Kennett concludes, having a conscience is consistent with being incapable of empathy. This conclusion, Kennett argues, supports Kant’s conception of moral judgment and motivation as products of reason. Taking empathy to be essential to Hume’s account of our moral sensibility and inessential to Kant’s, she then finds that Kant’s account, but not Hume’s, fits the moral judgments of high functioning autistics who have a conscience. In other words, vindication of Kant’s thesis about the possibility of judgments that qualify as categorical imperatives cannot come from our capacity for empathy. Oddly, Kennett takes the need that high functioning, conscience-​driven autistics have for order and their intolerance of inconsistency as evidence of this fit between Kant’s account and the moral judgments they make. Drawing on anecdotal and personal descriptions of them as both applying and following moral rules rigidly and expecting similar behavior from others, Kennett maintains that such behavior manifests a “reverence for reason” indicative of Kant’s conception of moral judgment (2002, p. 355). Typically, though, such rigid adherence to rules is seen as symptomatic of some emotional disturbance that the agent is keeping in check through such rigidity.

381

138  From Psychology to Morality

Someone who punctiliously arranges the books on his shelves alphabetically by author and vigilantly restores the arrangement whenever it is altered is not usually thought to be moved by reason alone but rather by anxieties about disruptions to an inviolate order. Yet Kennett offers no argument to distinguish such rigidity in the adherence to a common rule for organizing written materials from the similar rigidity in the adherence to moral rules on which she bases her argument for Kant’s conception of moral judgment (McGeer, 2008; Maibolm, 2008; de Vignemont and Frith, 2008). Be this as it may, the problem ultimately with Kennett’s argument is her using the comparison of psychopaths with high functioning autistics to make it. Her assumption is that the comparison will throw light on the determinants of moral judgment given the amorality of psychopaths and the possession of a conscience by high functioning autistics. Thus, her case against including empathy among the determinants rests on this assumption together with the observation that neither psychopaths nor autistics are capable of empathy.3 And her rationale for favoring Kant’s conception of moral

3.  The problem occurs as well in Sean Nichols’s argument for excluding empathy from the determinants of moral judgment (Nichols, 2004, pp. 8–​11). Nichols uses a different comparison, one between psychopaths and small children before they have developed a capacity for empathy, but his strategy is the same as Kennett’s. The small children Nichols compares with psychopaths have the same role in his argument as the high functioning autistics have in Kennett’s:  they exemplify people who are capable of making moral judgments despite lacking the capacity for empathy. Comparing them with psychopaths, who also lack the capacity for empathy, serves to fix the features of these small children’s capacity for making moral judgments as the determinants of moral judgment. Like Kennett, then, Nichols too assumes that the comparison with psychopaths throws light on the determinants of moral judgment. At the same time, since the children with whom Nichols compares psychopaths are too young to have developed the powers of reason that Kant’s conception of moral judgment implies, the conclusion he reaches about the nature of moral judgment is the opposite of Kennett’s conclusion about the nature of moral judgment. No matter, his argument suffers from problems that are as serious as the ones affecting hers. Specifically, Nichols bases his attribution of a capacity for moral judgment to small children who lack the capacity for empathy on studies that use a test that educational psychologists devised for determining whether a person recognizes a distinction between what these psychologists label “moral rules” and what they label “conventional rules”. He similarly bases his excluding psychopaths from the class of people who are capable of making moral judgments on studies using the same test. The latter studies use a personality inventory—​the Hare checklist—​to divide their subjects, incarcerated felons, into psychopaths and nonpsychopaths. Unless one defines a psychopath as someone whose score on the Hare checklist indicates the presence of psychopathy, the use of the checklist will produce some false positives, people who have a conscience but nonetheless count as psychopaths, and false negatives, people who do not count as psychopaths even though they lack a conscience. The first problem with Nichols’s argument, then, is that a statistically significant difference between the prisoners who were grouped by the Hare checklist as psychopaths—​the test group—​and those who were not—​the

391

Is Empathy Required for Making Moral Judgments?  139

judgment over Hume’s similarly presupposes that what determines whether someone’s judgments of duty are moral judgments is their psychological profile and not their content. This presupposition is at the root of her criticism. Consequently, the most effective answer to the criticism is to challenge its presupposition about the determinants of moral judgment. Such a challenge requires a more clear-​cut division between metaethics and psychology than I drew in proposing to study Hume’s thesis about internalism as a question within the latter and exclusive of the former. And drawing this more clear-​cut division means recovering the field of metaethics as it stood before its studies expanded into areas of psychology as a consequence of the special interest philosophers came to take in the question Hume’s thesis raises.

4. Metaethics Anglo-​American philosophers originally used the word ‘metaethics’, it is reasonable to suppose, to mean a kind of study in ethics that was roughly analogous to the studies in mathematics and logic of metamathematics and metalogic.4 control group—​in their sensitivity to the distinction between “moral” rules and “conventional” ones does not imply that the difference between the prisoners who have a conscience and the prisoners who lack one in their sensitivity to this distinction (given the possibility of correcting for the false positives and the false negatives) is statistically significant. Of course, if the validity of the Hare test—​the correlation between those it scores as psychopaths and true psychopaths—​ is high, then the seriousness of this problem diminishes as the difference between the test group and the control group widens. But without knowing either the validity of the Hare test or the statistical results of the studies, one can draw no conclusions from them. Even if one allows that the Hare checklist yields an insignificant number of false positives and false negatives, there is a second problem with Nichols’s argument. The studies he cites found a statistically significant difference between how psychopaths do on the recognition test for the “moral/​conventional” distinction and how nonpsychopaths do on it. Nichols, however, treats the test as if it were a litmus test for psychopathy, whereas at most these studies establish a significant association between being a psychopath and being unable to draw the distinction. It is statistically significant in the same sense in which the association of heart disease with high blood pressure is statistically significant. And just as it is possible for one to have heart disease without having high blood pressure, so too it is possible for a psychopath to draw the distinction. After all, nothing about lacking a conscience implies an inability to give the answers on a paper and pencil test or in an interview that show one recognizes what is “moral” from what is “conventional”. Hence, one cannot conclude from these studies that psychopathy excludes the capacity for making moral judgment. At most, one can conclude from them that it is statistically associated with the lack of that capacity. 4.  The OED cites a 1938 article in The Philosophical Review as the source of the earliest occurrence of ‘meta-​ethics’ or a cognate in English. This citation, however, is misleading. The article, “Philosophy in France 1936–​37” by André Lalande, is a translation from the French, and ‘meta-​ethics’ is the translator’s translation of ‘métamorales’, a French term from the writings of the anthropologist Lucien Lévy-​Bruhl. The earliest use of ‘meta-​ethics’ or a cognate by an

401

140  From Psychology to Morality

These studies emerged in the 1920s and 1930s. Metamathematics came first. David Hilbert invented it to advance his investigations into the consistency of arithmetic. His idea was to abstract from arithmetic a formal symbolic system that could, when the symbols were assigned meanings, express arithmetic’s axioms and theorems and be used to display the proofs of the latter. Hilbert then took this formal system as what metamathematics, or proof theory, as he also called it, studies. The study’s uppermost goal was to show that within the system no string of symbols of the form p. ~ p, a form that represents a contradiction, could be the last line of a proof. Subsequently, Alfred Tarski extended Hilbert’s program beyond arithmetic to an area of study he called semantics. Tarski, to capture Hilbert’s idea of taking a formal symbolic system as itself an object of study, called the system the object language and called the language in which the study was conducted the metalanguage. Where Hilbert’s program was narrowly focused on foundational questions of mathematics, specifically those of the consistency and completeness of arithmetic, Tarski’s was more broadly focused on modeling the notions of truth and reference in formalized languages generally. Hence, his program came to be known as “metalogic.” Metalogic was, in turn, adapted by Rudolph Carnap for the purpose of advancing logical positivism. Carnap, in his Logische Syntax der Sprache, expounded, with this purpose in mind, a program of study whose main proj­ ect was to construct formalized languages for expressing the arguments and conclusions of science (Carnap, 1937). His idea was that these languages, once constructed, could serve as object languages for the study of the logic of science. Such a study, Carnap held, would serve to substantiate the doctrines of logical positivism. Metaethics, conceived as a study analogous to metamathematics and metalogic, was a child of this program. Carnap argued, in the last chapter of his book, for replacing traditional philosophy with the sort of study of which he intended these formalized languages to be the object. He argued, that is, for the abandonment by philosophers of their traditional inquiries in metaphysics and for their taking up instead the logical analysis of the syntax of these languages when they were enriched with symbols whose interpretation made them suitable for expressing the knowledge found in a branch of science. Because Carnap expounded his program in the service of logical positivism, he treated ethics,

Anglo-​American philosopher that I have found occurs in Edel, 1942, p. 141. Edel’s program for expounding G. E. Moore’s ethics is roughly modeled on studies in metamathematics and metalogic, and his use of ‘meta-​ethical’ reflects that model.

4 11

Is Empathy Required for Making Moral Judgments?  141

when undertaken as a normative discipline, as continuous with metaphysics. He therefore denied that it was a branch of science. Or, what came to the same thing for him, he denied that it was a study that contributed to knowl­ edge. Consequently, the languages whose syntax he took to be the proper object of philosophical study had no symbols whose interpretation would yield ethical sentences, since in his view there were no conceptual or epistemological questions on which the logical analysis of a language enriched with such symbols would throw light. Carnap’s adherence to logical positivism, however, is inessential to the program he expounded in the service of it. One can therefore apply his program to ethics just as one applies it to any of the special sciences when one conceives of ethics as a branch of knowledge. Accordingly, one can engage in a study that consists in the logical analysis of the syntax of a language enriched with symbols whose interpretation makes it suitable for expressing ethical arguments and their conclusions. Such a study would go beyond the analysis of the syntax of the language only in its identifying certain symbols of the language as distinctly ethical and its focusing on the relation of formulae that contain those symbols to other formulae of the language. The study would therefore correspond to the understanding of metaethics as the study of the formal nature of ethical language and the logic and epistemic status of ethical arguments. Suppose, then, a formalized language, EL, suitable for expressing ethical arguments. This language can serve as an object language for studies in metaethics. The symbols whose interpretation makes EL suitable for this purpose are, in effect, ethical predicates, and one can accordingly take the sentences of EL in which they occur as ethical sentences.5 Using this notion of an ethical sentence we can then define a moral judgment as the mental act of affirming a proposition or proposition-​like thought that matches the speech act of asserting an ethical sentence. We need to include in this definition that the mental act it identifies as a moral judgment may be the affirmation of a proposition-​like thought rather than of a proposition so as to accommodate noncognitivist interpretations of the symbols of EL that correspond to ethical predicates To be sure, some noncognitivist theories of the meaning of ethical predicates imply that ethics is not a branch of knowledge, but this implication does not exclude them from being used 5.  For example, if the formalized language contains a first order, monadic functional constant E whose English translation is the predicate ‘is evil’ and an individual constant N whose English translation is the name ‘Nero’, then ‘E(N)’ is an ethical sentence of the language whose translation is ‘Nero is evil’.

421

142  From Psychology to Morality

to interpret symbols of EL in studies that consist in the logical analysis of the language’s syntax. While Carnap had no interest in the question of how to enrich a formalized language with symbols whose interpretation would make that language suitable for expressing ethical arguments, his younger contemporaries, A. J. Ayer and Charles Stevenson, who argued for taking ethical terms as having a distinctive noncognitive meaning, gave theories of this meaning that one could use to interpret the symbols of EL that correspond to ethical predicates. Consequently, we can understand their studies too as fitting the conception of metaethics that results from applying Carnap’s program to ethics. Their studies too are compatible with defining moral judgments as the mental acts of affirmation that match the speech acts of asserting ethical sentences. This definition represents an understanding of moral judgments on which the only mental powers whose exercise is necessary to making such judgments are those required for affirming propositions or proposition-​like thoughts. Since every judgment consists in the exercise of such powers, it follows that anyone who is capable of making judgments is capable of making moral judgments. Only unfamiliarity with the concepts or concept-​ like thoughts expressed by the symbols of EL that on a given interpretation of the language correspond to ethical predicates could render such a person incapable of making moral judgments. Hence, what determines, according to this definition, whether a judgment someone makes is a moral judgment is the concept or concept-​like thought she applies. In other words, what determines, on this definition, whether a judgment someone makes is a moral judgment is the content of that judgment and not its psychological profile. On the understanding of moral judgment that the definition represents, one can then study how different operations of the mind yield the same moral judgment. Such a study, being a study in psychology as distinct from metaethics, avoids the stalemated controversies that are born of Hume’s idea that to understand the nature of moral opinions requires understanding the distinctive mechanisms of the mind that produce them.

5.  Two Ways of Making Moral Judgments The upshot is that the study is open to the possibility that minds make moral judgments in different ways, that is, in ways that involve the stimulation of different mental sensitivities and the exercise of different mental powers. This possibility is well illustrated by different ways of making geometric judgments. These different methods of geometry were the inspiration for

413

Is Empathy Required for Making Moral Judgments?  143

Hilbert’s popular book, co-​authored with Stefan Cohn-​Vossen, Anschauliche Geometrie (Hilbert and Cohn-​Vossen, 1932). Hilbert, in the book’s preface, writes, In mathematics, as in any scientific research, we find two tendencies present. On the one hand, the tendency toward abstraction seeks to crystallize the logical relations inherent in the maze of material that is being studied and to correlate the material in a systematic and orderly manner. On the other hand, the tendency toward intuitive understanding fosters a more immediate grasp of the objects one studies, a live rapport with them, so to speak, which stresses the concrete meaning of their relations. As to geometry, in particular, the abstract tendency has here led to magnificent systematic theories of Algebraic Geometry, of Riemannian Geometry, and of Topology; these theories make extensive use of abstract reasoning and symbolic calculation in the sense of algebra. Notwithstanding this, it is still as true today as it ever was that intuitive understanding plays a major role in geometry. And such concrete intuition is of great value not only for the research worker, but also for anyone who wishes to study and appreciate the results of research in geometry. (1932, p. iii)6

Hilbert then goes on, in the chapters that follow, to present these results intuitively by means of visual constructions and commentary that enable the reader to see and understand propositions pictorially that one could also come to understand and affirm abstractly by means of the equations with which one works in algebraic geometry. In ethics too there is something like this difference between tendencies to abstract and intuitive thought. On the one hand, there is the tendency to take moral precepts abstractly as formulae that individually express inflexible principles of action and that jointly represent order in the world deviation from which is a serious offense against its authority. On the other, there is the tendency to take them pragmatically as rules that reflect mutual agreement among people about how to conduct their lives together and whose 6.  In the original, the passage reads, “In der Mathematik wie in aller wissenschaflichen Forschung treffen wir zweierlei Tendenzen an:  die Tendenz zur Abstraktion—​sie sucht die logischen Gesichtspunkte aus dem vielfältigen Material herauszuarbeiten und dieses in sstematischen Zusammenghang zu bringen—​und die andere Tendenz, die der Anschaulichkeit, die vielmehr auf ein lebendiges Erfassen der Gegenstande und ihre inhaltlichen Beziehungen ausgeht. Was insbesondere die Geometrie betrifft, so hat bei ihr die abstrackte Tendenz zu den grossartigen systematischen Lehrgebäuden der algebraschen Geometrie, der Riemannschen Geometrie und der Topologie geführt, in denen die Methoden der begreifflichen Überlegung, der Symbolik und des Kalküls in ausgiebeigem Masse zur Verwendung gelangen. Dennoch kommt auch heute dem anschaulichen Erfassen in der Geometrie eine hervorragende Rolle zu, und zwar nicht nur als einer überlegenen Kraft des Forschens, sondern auch für die Auffassung und Würdigung der Forschungsergebnisse.”

441

144  From Psychology to Morality

underlying rationale is the production of a common good in which all can share. The first tendency is manifested in stricter compliance with moral precepts and greater concern with their exact formulation. It fosters a more legalistic conception of morality. The second tendency is manifested in compliance with moral precepts from a sense of one’s having some latitude in determining what actions satisfy them and thus in a view of them as open to interpretation so as best to suit them to the ends they serve. It fosters a more instrumental conception of morality. Each yields moral judgments, of course, but the sensitivities and powers of the mind whose stimulation and exercise produce such judgments differ according as the first or second of these tendencies characterizes the thinking behind them. The difference between taking moral precepts as inflexible principles of action and taking them as rules reflecting mutual agreement among those whom they govern figures importantly in Jean Piaget’s study of the development of moral thought in children (Piaget, 1965). Piaget locates the first tendency in the thought of children at a younger age, whose way of relating to moral rules he refers to as “the morality of constraint” (1965, p.  197). Children, Piaget explains, come to the understanding of moral rules that the morality of constraint implies as a result of their being subject to their parents’ supervision and discipline. Accordingly, the child sees the rules its parents impose as fixed limits on its behavior, limits that require obedience by virtue of its parents’ authority. As children grow older and leave the shelter of their parents’ domain, they enter into social relations with peers and engage with them in activities that are free of close adult supervision and regulated by rules whose adjudication and enforcement they manage themselves. Piaget locates the second tendency in the thought of children at this later stage. He refers to the way they relate to moral rules as “the morality of cooperation” (1965, p. 197). The new understanding of moral rules children acquire at this stage, he suggests, comes from their taking responsibility for how the rules that regulate the activities in which they are cooperatively engaged are to be applied and for what to do when someone breaks or ignores these rules. It thus consists, on Piaget’s account of this later development in moral thought, in a child’s abandoning an egocentric understanding of the world on which rules are fixed, structural parts, put in place by adults, and its coming instead to see rules as more variable in view of their being based in mutual agreement among all who participate in the activities they regulate. This later development includes as well the child’s learning to take the perspectives of others who are co-​participants in activities that are regulated by rules. Presumably, the development of this skill is due at least in part to the child’s engaging in these activities independently of close adult supervision.

415

Is Empathy Required for Making Moral Judgments?  145

Presumably, that is, it is due at least in part to the child’s taking, in cooperation with others, responsibility for adjudicating and enforcing the rules that regulate their joint activities. To exercise this skill requires empathy with those whose perspective one is taking. To be sure, empathy emerges in children earlier in their lives. Piaget does not suggest otherwise. But children’s exercise of their capacity for empathy in judging the moral quality of their own or others’ rule-​following and rule-​breaking conduct is unnecessary at an earlier stage, given an understanding of moral rules as elements of a morality of constraint. When that understanding changes, however, and children come to see moral rules as elements of a morality of cooperation, then they acquire as well the habit of taking the perspective of the other participants in the activities in which they are engaged. Such perspective taking is essential to the success of these activities, for it is essential that each participant in them relate to every other as a cooperating partner in a joint enterprise, and they could not do so successfully if they were unable to comprehend either the views that each of their fellow partners had of his or her own situation within the enterprise or the beliefs, intentions, and motives with which their fellow partners acted as participants in it. So too the child’s exercise of its capacity for empathy in judging the moral quality of rule-​following and rule-​ breaking conduct is essential to the success of the enterprise. Empathy, then, is necessary for making moral judgments that reflect an understanding of moral rules as elements of a morality of cooperation. That is, while it is unnecessary for making moral judgments in which rules are taken abstractly as formulae to be applied to the facts at hand and followed to the letter as their authority requires, it is necessary for making moral judgments in which rules are taken pragmatically as instruments for achieving a common good made possible by cooperation with others in activities the rules regulate. In the former, one’s thought is concentrated on the rule’s formulation as an authoritative proscription or prescription and on the conduct required for conformity to it. In the latter, one’s thought is directed towards one’s relations with others as the object of the rule’s regulation, and through empathy with those with whom one shares these relations one comes to a more concrete understanding of the rule’s import and how it should be applied. Here Hilbert’s description of the tendency towards an intuitive understanding of the objects of judgment as yielding a live rapport (“ein lebendiges Erfassen”)7 with them seems especially apt as does his subsequent emphasis on the visual aspect of the intuitive understanding of geometry that he and his co-​author

7. See fn. 5.

461

146  From Psychology to Morality

present. Empathy too is a kind of visualization at least when it includes an understanding of its object as a whole person engaged in projects that give structure to his or her life and not simply as a site of emotions whose present feelings one shares (Deigh, 1995, pp.  759–​761). Such visualization alone fixes the difference in psychological profile of moral judgments that require empathy from those that do not.

5.  A Kantian Objection The hypothesis I advanced in my earlier paper, the hypothesis that through the exercise of the capacity for empathy one could make moral judgments of the type Kant characterized as categorical imperatives, is viable, I explained, only on the assumption that the empathy that informs the judgments includes an understanding of its object as a whole person, someone with projects and concerns that give structure to his or her life. The visualization that such empathy entails plausibly explains how a cognition alone could move one to action since it is plausible that one could be moved to further someone’s ends in life or alleviate his or her concerns just from seeing how those ends and concerns matter to him or her. On this supposition, one may be so moved independently of any emotional identification with the person who is the object of one’s empathy. The visualization of what he or she is thinking and feeling with regard to ends and concerns that matter to him or her would be sufficient. It is not yet, of course, a moral judgment since the cognition would have to comprise more than visualization for it to be an affirmation of a proposition corresponding to the assertion of an ethical sentence. Still, it offers the possibility of moral judgments of the type Kant characterized as categorical imperatives. Or so I argued. Students of Kant might object at this point. The hypothesis, after all, is part of an account of moral judgment on which some judgments are moral judgments despite their lacking the character of categorical imperatives, and in Kant’s ethics categorical imperatives are the bases of all moral judgments. Though Kant opposes Hume on the question of whether moral distinctions derive from reason or feeling, his opposition does not include rejection of Hume’s idea that to understand the nature of moral judgment requires understanding the distinctive mechanisms of the mind that produce such judgments. To the contrary, Kant’s argument for the possibility of categorical imperatives is an argument for morality’s having its grounding solely in the operations of reason. Specifically, it is an argument for identifying moral judgments with exercises of what Kant described as pure practical reason, and

417

Is Empathy Required for Making Moral Judgments?  147

his endorsement of Hume’s idea is unmistakable in this identification. Hume may have been the first to treat the nature of moral judgment as a question for psychology, but the treatment received a powerful second from Kant. Yet one should not so hastily regard Kant as a backer of Hume’s campaign to treat the nature of moral judgment as a question for psychology. Kant’s attribution of moral judgments to the operations of reason is not the result of a psychological study. Kant does not adduce observational evidence gathered through such a study for it. Rather he follows traditional Christian ethics in taking morality as an object of study that stands apart from the actual conduct and customs of human beings and that one applies in critically judging that conduct and those customs. It is, in this respect, an ideal. Kant expounds the ideal through analysis of the concepts of duty and practical law (or law that determines the will). He concludes from this analysis that the fundamental principle of morality is a precept requiring that lawfulness itself determine the will. One could not, he then argues, follow this precept in determining one’s will from natural desire or passion, for in that case the object of one’s desire or passion and not lawfulness itself would be what determines one’s will. Hence, morality—​the ideal that stands apart from human conduct and custom—​can be realized in human life only if the human mind is equipped with a power that can determine the will independently of natural desire and passion. If there is no such power, Kant says, morality is “a mere phantom of the brain” (Kant, 1785, p. 112). Such a power cannot, mutatis mutandis, be subject to natural causes. Else natural events and not lawfulness itself would determine the will. In other words, the mental power that Kant holds men and women must possess if morality is to be realized in human life necessarily operates outside of the natural world, even if the effects of its operation occur within that world. Its possession and exercise by human beings, if they do possess and exercise it, cannot, then, be known empirically. Consequently, Kant does not come to his understanding of moral judgment as by nature the product of such a power through empirical study of the sensitivities and powers of the mind that produce such judgment. Specifically, he does not come to understand the nature of moral judgment as a product of pure practical reason through observation of the workings of reason. Rather he posits, as a necessary condition of a judgment’s being a moral judgment, the power to determine the will independently of natural causes. Unless a judgment of duty issues from an exercise of that power, it is not a moral judgment. Thus, his procedure, which starts with the exposition of an ideal and then moves to explaining the conditions for its being realized, is the reverse of Hume’s. Hume assumes at the start that people actually discern moral distinctions, that they are in fact

48 1

148  From Psychology to Morality

capable of such discernment. He never questions the reality of this capacity. His interest, rather, is in whether its exercise is rooted in reason or feeling, and he concludes that it is rooted in the latter by observing its workings. The irony, then, of citing Kant as endorsing Hume’s idea that to understand the nature of moral judgment requires understanding the distinctive mechanisms of the mind that produce it is that the endorsement is inseparable from Kant’s denying the possibility of acquiring such understanding through scientific study.

References Anscombe, G. E. M. (1958a). “On Brute Facts,” Analysis 18, 69–​72. ——— (1958b). “Modern Moral Philosophy,” Philosophy 33, 1–​19. Carnap, Rudolph (1937). The Logical Syntax of Language, A. Smeaton, trans., Patterson, N.J., Littlefield, Adams & Co., 1959. Deigh, John (1995). “Empathy and Universalizability,” Ethics 105, 743–​63. Edel, Abraham (1942). “The Logical Structure of G. E. Moore’s Ethical Theory,” in The Philosophy of G. E. Moore, Paul Schilpp, ed., New York, Tudor Publishing Co., pp. 137–​76. Enoch, David (2011). Taking Morality Seriously, Oxford, Oxford University Press. Frankena, William (1958). “Obligation and Motivation in Recent Moral Philosophy,” in Essays in Moral Philosophy, A. I. Melden, ed. Seattle, WA, University of Washington Press, pp. 40–​81. Hilbert, David and Stefan Cohn-​ Vossen (1932). Anschauliche Geometrie, Berlin, J. Springer. The English translation is Geometry and the Imagination, P.  Nemenyi, trans. (New York: Chelsea Publishing Co., 1952). Hume, David (1978). A Treatise of Human Nature, 2nd ed., L. A. Selby-​Bigge, ed., Oxford, Oxford University Press. ——— (1975). Enquiries concerning Human Understanding and concerning the Principles of Morals, 3rd ed., L. A. Selby-​Bigge, ed., Oxford, Oxford University Press. Kant, Immanuel (1785). Groundwork of the Metaphysics of Morals, H. J. Paton, trans., New York, Harper & Row, Inc., 1964. Kennett, Jeanette (2002). “Autism, Empathy, and Moral Agency,” Philosophical Quarterly 52, 340–​57. Maibolm, Heidi (2008). “The Will to Conform,” in Moral Psychology: The Neuroscience of Morality: Emotion, Brain Disorders, and Development, W. Sinnott-​Armstrong, ed. Cambridge, Mass., MIT Press, pp. 266–​72. McGeer, Victoria (2008). “Varieties of Moral Agency:  Lessons from Autism (and Psychopathy),” in Moral Psychology: The Neuroscience of Morality: Emotion, Brain Disorders, and Development, W. Sinnott-​Armstrong, ed. Cambridge, Mass., MIT Press, pp. 227–​57. Nichols, Sean (2004). Sentimental Rules: On the Natural Foundations of Moral Judgment, New York, Oxford University Press. Piaget, Jean (1965). The Moral Judgment of the Child. Marjorie Gabain, trans., New York, Free Press.

491

Is Empathy Required for Making Moral Judgments?  149 Sayre-​McCord, Geoffrey (1994). “On Why Hume’s ‘General Point of View’ Isn’t Ideal—​ and Shouldn’t Be,” Social Philosophy and Policy 11, 202–​28. Stroud, Barry (1977). Hume, London, Routledge & Kegan Paul. Vignemont, Frederique de and Uta Frith (2008). “Autism, Morality, and Empathy,” in Moral Psychology:  The Neuroscience of Morality:  Emotion, Brain Disorders, and Development, W. Sinnott-​Armstrong, ed. Cambridge, Mass., MIT Press, pp. 273–​280.

501

7

Williams on Practical Reason

I An account of practical reason, if it is an account of the place of the power of reason in our ability to act rationally, is at the same time an account of our rational agency. As such, it falls within the scope of psychology. Specifically, it belongs to the study of the cognitive and conative machinery of human behavior. In contemporary Anglo-​American philosophy rational action is typically understood as action done for a reason, and accordingly the ability to act rationally is the ability to see the reasons one has for action and to act on them. Thus the interest in understanding the nature of practical reason comes to focus on understanding what it means for someone to have a reason to do a certain action. Bernard Williams gave a distinctive answer to this question.1 He held that a sentence of the form ‘A has a reason to φ’, where A names a rational agent and φ stands for a verb of action, is true only if there is among A’s desires, inclinations, appetites, commitments, dispositions of evaluation, and other 1.  “Internal and External Reasons” in Bernard Williams, Moral Luck (Cambridge: Cambridge University Press, 1981), pp. 101–​113.

150

151

Williams on Practical Reason  151

motivational states—​what Williams called the elements of A’s subjective motivational set S—​a motivational state D that would be furthered by A’s φing, provided that D’s existence does not depend on a false belief or A’s belief that D would be furthered by φing is not false. Later, Williams came to prefer a different formulation of the same answer: a sentence of the above form is true only if there is a sound deliberative route from some element D in S to A’s φing.2 On his preferred formulation, then, Williams’s answer is that the existence of a sound deliberative route from some element D in S to A’s φing is a necessary condition of A’s having a reason to φ.3 The answer is distinctive, as I said. Yet it also has a definite affinity with Donald Davidson’s account of rational action. This last statement may seem surprising, so let me explain. Davidson initially offered his account in opposition to those philosophers of action who maintained that rational action was categorically distinct from the movements of the human body through which it was realized.4 These philosophers held that one explained a rational action by citing the reasons for which it was done whereas one explained the movements through which the action was realized by citing the events and conditions that caused those movements, and they further held that these were two radically different types of explanation. Reasons, they thought, were different from causes. Davidson, to the contrary, held that to explain a rational action by citing the reason for which it was done just was to explain the action by citing events and conditions that caused it. The explanation of a rational action φ consisted, in the first instance, of a desire, passion, appetite, etc. for some object O—​what Davidson called a pro-​attitude towards O—​together

2.  “Internal Reasons and the Obscurity of Blame” in Bernard Williams, Making Sense of Humanity (Cambridge: Cambridge University Press, 1995), pp. 35–​45 and Bernard Williams, “Postscript:  Some Further Notes on Internal and External Reasons” in Varieties of Practical Reasoning, Elijah Milgram, ed. (Cambridge, Mass.:  MIT Press, 2001), pp. 91–​97. See also Bernard Williams, “Replies” in World, Mind, and Ethics, J. E. J. Altham and Ross Harrison, eds. (Cambridge: Cambridge University Press, 1995), pp. 185–​224, esp. 186–​194. 3.  While Williams did say that he thought the existence of a sound deliberative route from some element D in S to A’s φing was a sufficient condition of A’s having a reason to φ, he demurred from defending this further thesis. Accordingly, I  will take his internalist view of reasons to be limited to holding that the sound deliberative route from some element D in S to A’s φing is a necessary condition of A’s having a reason to φ. See “Internal Reasons and the Obscurity of Blame,” pp. 35–​36 and “Postscript,” p. 91: “The essence of the internalist position is that [the existence of a sound deliberative route] is a necessary condition.” 4.  “Actions, Reasons and Causes” in Essays on Actions and Events, pp.  3–​19. The local source of the view that Davidson opposed was Wittgenstein’s query, after observing that his arm goes up when he raises his arm, “And the problem arises: what is left over if I subtract the fact that my arm goes up from the fact that I raise my arm” (Philosophical Investigations, I, §621).

521

152  From Psychology to Morality

with an interlocking belief that φing will get one O, and this combination of a pro-​attitude and interlocking belief is what Davidson dubbed “the primary reason” for the agent’s φing. For example, suppose, having parked my car downhill, I turn the steering wheel so that the front wheels are angled towards the curb. The explanation for this action is that I want to keep my car from rolling down the hill in case the brake fails and I believe that turning the steering wheel so that the front wheels are angled towards the curb will prevent the car from rolling should the brake fail.5 Every element in the subjective motivational set that Williams attributes to an agent is a pro-​attitude in Davidson’s sense. Williams, however, unlike Davidson, does not take a reason to do some action to be a combination of a pro-​attitude in an agent’s motivational set and an interlocking belief that doing the action will further the attitude. Nor does Williams take either the attitude alone or the interlocking belief that doing the action will further the attitude alone to be a reason to do the action. Indeed, he avoids identifying reasons with anything. He concentrates instead on the truth conditions of sentences of the form ‘A has a reason to φ’, where A names an agent and φ stands for a verb of action.6 Let us, for convenience’ sake, call these “R-​sentences.” If one specified all the truth conditions of R-​sentences, one could, then, contextually define ‘having a reason to φ’. Accordingly, we would understand ‘reason to φ’ as a nonreferring expression in the same way as definite descriptions are nonreferring expressions on Russell’s theory of definite descriptions.7 But Williams never says that he means to be understanding ‘reason to φ’ in this way. What is more, he takes R-​sentences to imply in each instance the possibility of A’s φing for the reason the sentence attributes to A and that in such case, he says, the reason is A’s reason for φing. So, presumably, he does regard the expression as referring to something. The best fit for what he says is that ‘reason to φ’ refers to a feature of the agent’s circumstances in virtue of which those circumstances provide the agent with an opportunity to further some element in his or her motivational set.

5.  Independently, Carl Hempel gave essentially the same account of reasons for action in “Aspects of Scientific Explanation” in his Aspects of Scientific Explanation and Other Essays in the Philosophy of Science (New York: The Free Press, 1965), pp. 331–​496, esp. pp. 463–​487. 6.  Williams notes that a sentence of the form ‘There is a reason for A to φ’ can express the same proposition as the sentence of the form ‘A has a reason to φ’ if φ is the same verb of action in either sentence. For simplicity I will limit the discussion to sentences of the latter form. 7.  Bertrand Russell, “On Denoting,” Mind 14 (1905): 479–​493. See also Jeremy Bentham’s account of expounding a word by paraphrases in Bentham, A Fragment on Government, J. H. Burns and H. L. A. Hart, eds. (Cambridge: Cambridge University Press, 1988), p. 108n.

513

Williams on Practical Reason  153

Let us use the word ‘desire’ as an all-​purpose term for any pro-​attitude in Davidson’s sense. Both Williams and Davidson accept, as a subjective condition on the truth of any sentence that attributes a reason to do an action φ to an agent A, that A has a desire that bears a certain relation to φ. They differ, however, in how they understand that relation. Davidson takes any desire of A’s as generating a reason for A to φ if (but only if) A believes that φing will serve to satisfy that desire. Williams, by contrast, takes a desire of A’s as generating a reason to φ only if φing in fact serves to satisfy that desire and regardless of what A believes. The upshot is that Williams takes some R-​sentences to be false that are true on Davidson’s account. One can see this difference in Williams’s example of the man who wants a gin and tonic and falsely believes that the stuff in the bottle from which he is about to pour is gin.8 The belief is false because the stuff is gasoline. On Davidson’s account, the man has a reason to mix the stuff with tonic, for he wants a gin and tonic and believes that mixing the stuff with tonic will produce a gin and tonic. Williams, by contrast, denies that the man has a reason to mix the stuff with tonic.9 Rather, Williams says the man thinks he has a reason to mix it with tonic, but in fact he does not. On Williams’s preferred formulation, the man lacks a reason to do the action because there is, owing to his false belief, no sound deliberative route from his desire for a gin and tonic to his mixing the stuff in the bottle with tonic.10 Williams allows that the man’s action would be rational relative to his belief that the stuff in the bottle is gin.11 Hence, he does not disagree with Davidson’s account of rational action. The disagreement between them, rather, is over the interpretation of R-​sentences. Davidson’s account, as I mentioned, is essentially concerned with explanation. Specifically, Davidson would interpret the use of an R-​sentence to attribute a reason to do an action to an

8.  “Internal and External Reasons,” p. 102. 9.  To be sure, the sentence might be true owing to the man’s having a different desire e.g., a desire to increase the amount of liquid in a glass that contains tonic, for in that case there would be a sound deliberative route from this other desire to the man’s mixing the stuff in the bottle with tonic. Clearly, Williams, in taking the sentence to be false, is assuming that the man has no such desire in his motivational set. 10.  Williams likewise denies that the man’s desire to mix the stuff in the bottle with tonic, a desire he has as a result of his desiring a gin and tonic and his believing that the bottle contains gin, generates a reason for the man to mix the stuff in the bottle with tonic. The desire’s being based on a false belief disqualifies it, on Williams’s answer, from generating a reason to mix the two liquids. In this case, there is no sound deliberative route from the desire to his mixing the stuff in the bottle with tonic because the desire is an unsound starting point from which to deliberate. “Internal and External Reasons,” p. 103. 11. Ibid.

541

154  From Psychology to Morality

agent as an attribution of a combination of a belief and a desire that is such as to motivate the agent to do that action. On Davidson’s interpretation, the sentence describes a causal relation between the agent’s state of mind and this action. Williams’s interpretation, by contrast, includes an ideal element, that of a sound deliberative route, among the truth conditions of the sentence. This ideal element precludes the sentence, when it is used to attribute to an agent a reason to do an action, from describing a causal relation between the agent’s state of mind and the action. Because of this element, the sentence may be true even if there is no such causal relation, even, that is, if the agent is presently unmotivated to do the action. On Williams’s interpretation, one uses the sentence not to describe a relation between the agent and a certain action but to evaluate the relation between the two. One uses the sentence to evaluate the action as one that it is or would be rational for the agent to do. In short, while Williams agrees with Davidson that a rational action is one in which the agent acts from a desire he or she believes the action will serve to satisfy, he departs from Davidson in understanding the characterization of the agent as having a reason to do an action as evaluative rather than merely descriptive of the agent’s relation to the action. When an agent is unmotivated to do an action despite there being a sound deliberative route from some desire he has to that action, the agent, on Davidson’s account, lacks a reason to do the action because he lacks the requisite interlocking belief that doing the action will serve to satisfy that desire. Williams, as we have seen, disagrees with Davidson on this point. The agent, Williams holds, has a reason to do the action despite being unmotivated to do it. What is more, Williams argues, the agent can come to see that he has this reason, and the agent’s coming to see the reason coincides with his becoming motivated to do the action. Grasping the truth of the relevant R-​sentence opens the agent’s eyes, as it were, to an opportunity for satisfying one of his desires that he had not previously recognized. Similarly, others can get the agent to see this opportunity by advising him about the reason he has to do the action. In either case, the agent through evaluation of his relation to an action comes to see an opportunity for satisfying a desire. These cases show, Williams observes, that R-​sentences have normative force.12

12.  “Internalism and the Obscurity of Blame,” p. 36. Davidson also believes there is a normative element in his account of a reason for action, “since the action must be reasonable in light of the [agent’s] beliefs and desires.” See Davidson, “Intending,” in Essays on Actions and Events, p. 84. Whether or not he is correct, the element he specifies as normative is not what Williams means in attributing normative force to R-​sentences.

515

Williams on Practical Reason  155

Here is an example from an episode of the 1970s detective drama “Columbo.”13 In this episode a wine snob, played by Donald Pleasance, kills his brother by imprisoning him in an airtight wine cellar for a week. The murder is reminiscent of the one in Poe’s “The Cask of Amontillado” except that after the brother dies, the wine snob removes his body from the cellar, dresses it in a wet suit and scuba diving gear, and dumps it in the Pacific Ocean to make the death look like an accident. Columbo, who suspects foul play and believes the victim did not die in the water, visits the wine snob at his winery. The inquisitive detective learns of the wine snob’s collection of valuable wines and gets him to show them to him, at which point Pleasance’s character leads Columbo into the wine cellar where the murder occurred. Columbo immediately recognizes the cellar’s possibilities as a trap and tomb and with the aid of a few other clues solves the crime. Thus, the wine snob, who of course wanted to escape detection, had reason to steer Columbo clear of the cellar, though he did not appreciate this at the time he showed it to him. Wanting to escape detection, he could have come to the conclusion that he had reason to keep Columbo from seeing the place where his brother died and so been moved to invent some excuse for not showing Columbo his valuable wines. Misplaced confidence in the undetectability of his crime led him to think and do otherwise. Similarly, the wine snob’s secretary, given her marital designs on him, might have advised him that he has reason to keep Columbo from seeing the cellar. In this case, she would be using an R-​ sentence to express the same thought in the second person as the first-​person conclusion the wine snob might have reached on his own. As Williams observes, an account of R-​sentences must cover their use to give advice in the ‘if I were you mode’.14 Davidson’s account fails to do this. The reason is that to use an R-​ sentence to give advice to some agent is to bring to that agent’s attention something normative about her circumstances of which she is unaware. On Davidson’s account, however, telling some agent that she has a reason to φ would be telling her only that she had a certain combination of desire and belief and not that her circumstances presented her with an opportunity for satisfying a desire by φing. Indeed, it would hardly be advice to tell someone who had the requisite combination of desire and belief that she had a reason to φ when one knew that the belief was false and that her acting on it, given her desire, would lead to an outcome she would regret. One would not, for

13.  “Any Old Port in a Storm,” directed by Leo Penn; written by Stanley Ralph Ross. 14.  “Internalism and the Obscurity of Blame,” p. 36.

561

156  From Psychology to Morality

instance, be giving advice to someone who one knew wanted a gin and tonic and believed falsely that the stuff in the bottle from which she was about to pour was gin if one told her she had a reason to mix the stuff in the glass with tonic. Yet on Davidson’s account the person would have such a reason. On Davidson’s account, an R-​sentence, when true, is a basis for predicting what someone will do and not for advising someone about an opportunity she has to satisfy some desire. It therefore lacks normative force. Williams’s account of R-​sentences, then, appears to be an advance on Davidson’s. To be such an advance, however, Williams’s account must lend itself to explanations of rational actions that consist in citing the reasons for which the actions are done. Otherwise his account would not be a competitor of Davidson’s. It would merely exist on a different plane. Yet one may wonder whether Williams’s account can lend itself to such explanations. Specifically, one may wonder whether the interpretation of R-​sentences that Williams defends, because it includes an ideal element in the truth conditions of such sentences, can yield an understanding of an agent’s reason as a factor in such explanations. Williams clearly thinks that it can. He explicitly denies that his account yields an understanding of R-​sentences as having only normative force. “It must be a mistake,” he writes, “simply to separate explanatory and normative reasons. 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 But how does it figure? And how can Williams square its figuring with his taking R-​sentences to have normative force? The answers to these question are inobvious. The obscurity is due to Williams’s division of rational actions into two types, those done for reasons and those done for what their agents mistake for reasons. The second type are the actions Williams characterizes as rational relative to their agents’ beliefs. Presumably, then, Williams would characterize the first type as categorically rational. The same form of explanation, according to Williams, fits either type. That is, regarding an agent A and an action φ, and regardless of which type of rational action A’s φing belongs to, one can explain A’s φing by citing A’s having through deliberation accepted that he has a reason to φ, or more exactly by citing A’s having through deliberation accepted that he has more reason to φ than to do anything else.16 The agent, on this form of explanation,

15.  “Internal and External Reasons,” p. 106. 16.  “Internalism and the Obscurity of Blame,” p. 39. See also “Postscript,” p. 91. Williams does not explain how the agent comes to the stronger conclusion that he has more reason to

517

Williams on Practical Reason  157

deliberates; the deliberation leads to his accepting that he has more reason to do a certain action than to do anything else; and in coming to this conclusion, the agent is moved to do the action. Williams calls this form of explanation “the-​reason-​for-​action form.”17 Plainly, given Williams’s division of rational actions into two types, an agent’s reason to do a certain action can figure in an explanation of his doing that action, where the explanation has the reason-​for-​action form, only if the action is a rational action of the first type. If it is a rational action of the second type, then no such reason can figure in an explanation of this form since either there is no such reason, despite the agent’s thinking he has one, or though there is, it is not the one the agent thinks he has and therefore does not enter into the deliberation that leads to his doing the action. Let us, then, concentrate on rational actions of the first type. How does an agent’s reason for doing a certain action figure in explanations of actions of this type? When an agent, on Williams’s view, has a reason to do a certain action, there is a sound deliberative route the agent can take from one of his desires to that action. And when the agent acts for that reason, the deliberation leading to the agent’s doing the action proceeds from that desire and follows that route. Hence, the agent’s reason figures in the explanation of the action as the sound deliberative route the agent must take in deliberating if in acting as a result of his deliberation he acts for a reason. The deliberation itself has a causal role in the explanation, but its being sound, that is, the soundness of the route it follows, figures only as a logical consequence of the action’s being one that is done for a reason. In other words, the soundness of the deliberative route the deliberation follows makes a difference as to whether the agent acts for a reason or acts only for what the agent thinks is a reason. It does not make a difference as to whether the agent does the action. If his deliberation were the same but unsound, he would still do the action.18 φ than to do anything else, but filling in, we can suppose that the agent proceeds to deliberate from a desire on the assumption that satisfying this desire will not thwart any of his desires whose satisfaction is more important to him and is at the same time more important than his satisfying any of the remaining desires he could satisfy in the circumstances. If the agent, then, were to discover in the course of deliberation that he could not satisfy the desire from which he was deliberating by φing without thwarting a desire whose satisfaction was more important to him, he would conclude that he did not have more reason to φ than to do something else. 17.  “Internal and External Reasons,” p. 102. 18.  The reason figures in the explanation in the same way that a particular fact about a person’s circumstances figures in the explanation of the person’s emotional response to that fact. Explaining your anger at your boss for cutting your pay, you might say, “He cut my pay and that made me very angry.” In this case, you would be angry even if your boss hadn’t cut your pay as long as you believed that he did. The belief, together with a desire not to be treated unjustly

581

158  From Psychology to Morality

Williams’s departure from Davidson’s account on the question of the normative force of R-​sentences thus bleeds into how Williams understands the explanatory dimension of such sentences. It produces two further differences between the two accounts. First, Williams, unlike Davidson, limits the class of rational actions that are done for reasons to a proper subclass of rational actions generally. And second, he takes the reasons for which the actions in this subclass are done to have qua reasons a logical rather than causal role in the explanations of those actions. This latter result puts him at odds with Davidson’s primary thesis that such reasons are causes of the actions for which they are reasons. Here, then, is the answer to the question of how Williams squares the normative force of R-​sentences with their having an explanatory dimension. What explains the actions, on Williams’s account, are certain mental states and processes, which when sound imply the existence of a reason, and the implication that this complex of mental states and processes is sound is what gives the sentences’ attribution of a reason normative force. Davidson’s primary thesis, one might say, removes the ideal element in our understanding of having a reason to act and correspondingly enforces an understanding of the concept of being rational, as applied to actions, as a strictly psychological concept. Williams’s interpretation of R-​sentences then corrects for these departures from our ordinary understanding of having a reason and acting rationally, but it can do so only by adding complications to the account of the sentences’ explanatory dimension.

II A decade before Davidson first defended the view that to explain an action by citing the agent’s reason for doing it is to give a causal explanation of it, Kurt Baier advanced a very different account of reasons for action. Baier took such reasons to be considerations that an agent weighed if the agent properly deliberated about what to do. More exactly, he took them to be facts that constituted the agent’s circumstances and that bore on the question what should the agent, given these circumstances, do.19 Reasons, then, on Baier’s account, determine the answer to this question not only for the agent but for anyone else who asks it. This account had germinated in Stephen Toulmin’s

and a belief that the pay cut was unjust, causes the anger whether or not it is true, and when it is true, we commonly cite the fact that makes it true to explain the emotion. 19.  Kurt Baier, “Good Reasons,” Philosophical Studies 4 (1953): 1–​15.

591

Williams on Practical Reason  159

groundbreaking work An Examination of the Place of Reason in Ethics.20 And Baier himself subsequently, in his best-​known work, enlarged it into a full ethical theory.21 Plainly, on Baier’s account, attributing to an agent a reason to do a certain action has normative force. Such force derives from the reason’s being a factor in determining what the agent should do. At the same time, Baier’s understanding of the normative force of such attributions—​his understanding of the normative force of R-​sentences—​sharply differs from Williams’s. One can see this difference in the disagreement in their respective conceptions of deliberation. Williams’s conception is Aristotelian. Aristotle characterized deliberation as reasoning that proceeds from a moving principle and is shaped by the agent’s perception of his circumstances. That perception includes judgments about the effects of different actions available to him. It terminates in the agent’s choice of action.22 Similarly, Williams conceives of deliberation as reasoning that proceeds from a desire the agent has and is shaped by the agent’s perception of his circumstances. It terminates, on Williams’s view, in the agent’s accepting that he has more reason to do a certain action than any other, a judgment that motivates the agent to do that action. Baier, by contrast, conceives of deliberation as an exercise of reason that an agent undertakes to answer the question ‘what should I do?’ The exercise has two stages. At the first stage, the agent gathers, through a survey of her circumstances, the facts that bear on this question. At the second, she weighs those facts to determine, all things considered, what is the best thing to do in these circumstances. The answer the agent arrives at is the terminus of the deliberation. Action in accordance with the answer then follows, not because the conclusion of the exercise is a motivating judgment, but because the agent would not have deliberated in the first place if she did not already want to do whatever action she determines she should do.23 Williams’s conception disagrees with Baier’s, then, in the role that it assigns to deliberation in the explanation of the action to which deliberation leads. Again let φ refer to this action. On Williams’s conception, deliberation serves to transmit the motivating force of the desire from which it proceeds to the thought with which it concludes, the thought that one has more reason to

20. Stephen Toulmin, An Examination of the Place of Reason in Ethics (Cambridge, Cambridge University Press, 1950). 21. Kurt Baier, The Moral Point of View:  A Rational Basis of Ethics (Ithaca:  Cornell University Press, 1958). 22. Aristotle, Nicomachean Ethics, bk. III, ch. 3. 23. Baier, The Moral Point of View, p. 147.

601

160  From Psychology to Morality

φ than to do anything else.24 On Baier’s conception, deliberation presupposes a desire to act in accordance with the conclusion of the deliberation, and that desire then motivates the agent to φ when the agent concludes that she should φ, that φing is the best thing she can do. The desire, however, stands outside of the deliberation. Baier supposes that men and women are brought up to want to act as they should, and this standing desire explains their doing what they conclude from deliberating they should do. The thoughts with which their deliberations conclude, he holds, are themselves inert. As Baier goes on to say, the reasoning in which deliberation consists is identical to the reasoning one would engage in retrospectively to determine whether what one did was justified. Indeed, on Baier’s conception, one would engage in the same reasoning even if one were merely curious about the answer to the question what one should do in these circumstances. Such reasoning is essentially theoretical.25 It leads to the acceptance or rejection of a proposition or set of propositions but not to action. Action then follows only when the agent independently has a desire to act in accordance with the conclusion. The reasoning in which deliberation consists on Williams’s conception, by contrast, is practical. It issues in a conclusion that has in itself motivational force.26 When one determines whether an act one did was justified, one determines whether it was the or at least a right or proper thing to do in the circumstances. A fortiori, when one justifies what one did, one shows that it was the or at least a right or proper thing to do in the circumstances. In the former case, the reasons one considers are facts about one’s circumstances, and as such they constitute evidence for or against the proposition that one’s action was a right or proper thing to do. In the latter, the reasons one cites are facts about one’s circumstances that constitute evidence for the same proposition. Hence, the normative force of attributing such reasons to an agent derives from the normative force of the propositions about the rightness or propriety of the actions for which they are reasons. The normative force of

24.  See fn. 16 above. 25.  Baier, in effect, acknowledges this when he distinguishes between the theoretical and practical tasks the agent has in deliberation. The theoretical task is to answer the question ‘what should I do?’ The practical task is to act on the answer. The Moral Point of View, pp. 143–​144. 26.  This way of distinguishing between the two types of reasoning follows Aristotle’s notion of the practical syllogism. See Donald Davidson, “How is Weakness of Will Possible?” in Actions and Events (Oxford: Oxford University Press, 1990), p. 39, and G. E. M. Anscombe, Intention, 2nd ed. (Ithaca: Cornell University Press, 1963), §33: pp. 57–​62. Hume’s thesis that morality is essentially practical presupposes the same notion of practical. The thesis is pivotal in Hume’s argument for his subjectivist view of ethics. See A Treatise of Human Nature, p. 470.

6 11

Williams on Practical Reason  161

an R-​sentence that one interprets to be about such a reason derives from the normative force of a sentence of the form ‘It is right or proper for A to φ’. By contrast, the normative force of an R-​sentence on Williams’s interpretation of such sentences derives from the rationality of A’s φing in view of its relation to A’s desires. Williams, that is, takes attributions of reasons to φ to A to have normative force because A is a rational agent and as such acts according to the standard of rationality in action that one be guided by credible beliefs and coherent deliberation. Normative force that is due to being a rational agent is thus fundamentally different from normative force that derives from the concept of rightness or propriety.

III In justifying an action we sometimes speak of the grounds on which the action is justified. In this context, we use the words ‘grounds’ and ‘reasons’ interchangeably. At the same time, we also use ‘reasons’ to refer to considerations that oppose those that support doing an action even when the latter outweigh the former, and in this context we do not use the word interchangeably with ‘grounds’. Let us then say that the grounds of an action are the reasons that justify it when it is justified. The same holds as well when we speak of the grounds of a belief or an attitude. For beliefs we cite as grounds credible evidence of their truth. For attitudes we cite facts that show their objects to be fitting or appropriate. At the same time, we would not say that explaining a person’s belief or attitude as the belief or attitude of a rational mind consists in citing the grounds a person has or thinks he has for her belief or attitude. Having grounds or thinking one has them is not necessary for a belief or attitude to be the product of rational thought. Having grounds means that there is a basis in some system of norms or set of values that justifies the belief or attitude, and it is one thing for a belief or attitude to be the product of rational thought and another for one to be justified or think one is justified, relative to some system of norms or set of values, in having that belief or attitude. Other things than grounds or one’s thinking one has them can explain, consistently with one’s rationality, one’s having a belief or attitude. And the same is true of an action. Rationality in thought and action is not circumscribed by justification. The hardest case, one might think, is belief. Yet it is not a very hard case. Believing is often a function of memory, and memory, particularly our short-​term memory, is notoriously faulty. I  believe I  left my coffee cup on the kitchen counter because it’s the last place I remember putting it. But it’s

621

162  From Psychology to Morality

not there. Rather, it’s on the buffet in the living room where I put it down when looking for a book. The memory I draw on to form the belief is not grounds for my belief. I would not try to justify my belief by citing it. I know all too well that my memory in such circumstances is unreliable. Still my belief, notwithstanding the faultiness of my memory, is the product of rational thought. There is an evident connection between what I remember and what I believe, and I make it. Columbo’s belief that the wine snob murdered his brother may at first be the result of a hunch. If it were, then he would not, at that point, think he had credible evidence for his belief. Indeed, he might admit right away that he didn’t. A hunch, after all, though the source of a belief, is not grounds for it, and Columbo presumably understands this. Yet who would deny that such a belief would be the product of rational thought or hold that to have a belief as the result of what one knows to be just a hunch is irrational? Perhaps some philosophers would object. I can’t be said to believe that the cup is on the kitchen counter, they would argue, if I realize that I am relying on what may be a faulty memory in thinking so. I  can’t be said to have this belief because I must, in realizing that I am relying on what may be a faulty memory, realize too that I have grounds for being uncertain about the cup’s being on the kitchen counter, and I cannot believe that it is and at the same time realize that I have grounds for being uncertain about it. But this objection at most shows that belief is a special type of cognitive state and does not argue against the larger point that the state I am in, which I have described as my belief that the cup is on the kitchen counter, is the product of rational thought. If you object to calling this state a belief, then call it a supposition, an assumption, a conjecture, or even just an idea that I got into my head. Whatever we call it, the process of thinking by which I arrived at it could still be that of a rational mind, even though it does not supply me with grounds for being in that state. And similarly for Columbo’s cognitive state when he begins, as a result of a hunch, to suspect the wine snob of murdering his brother. In textbook accounts of scientific method, it is common to distinguish between the thinking that goes into the discovery of scientific hypotheses and theories and the thinking that goes into their confirmation and justification. The latter, if it is sound, conforms to the relevant science’s standards and protocols for testing hypotheses and proving theories, and it concentrates on determining whether there are grounds for accepting or rejecting them. The former, by contrast, is not subject to such constraints. While there may be techniques within a given science for generating hypotheses, there are no standards or protocols a scientist must follow to come up with them. The

613

Williams on Practical Reason  163

thinking is inventive and often calls for imagination. It does not, except perhaps incidentally, yield grounds for believing whatever hypothesis the scientist may come up with. The hypothesis is an idea that the scientist gets into his head, and the thinking by which it gets there is that of a rational mind, notwithstanding there being no grounds for the conjecture it represents. Take as an example Kekulé’s celebrated account of how he came to his conjecture about the molecular structure of benzene.27 One day, he wrote, he was sitting before a fire in his bachelor quarters in Ghent, working on his textbook when he began to doze. In the reverie that followed, he imagined rows of atoms “twining and twisting in snake-​like motion.” Suddenly one of the snakes seized its own tail and the ring it formed “whirled mockingly before [his] eyes.” He thus got the idea of the benzene molecule’s having the structure of a ring, and he “spent the rest of the night working out the consequences.”28 His conjecture is the product of rational thought, though the image of a snake seizing its tail and whirling before him mockingly is hardly grounds for making it. And similarly for the hunch that leads Columbo to suspect the wine snob of murder and the potentially faulty memory that leads me to believe that the cup is on the kitchen counter. The case of action is different. When the question of the grounds for a belief is raised, one looks for an answer relative to norms for the evidentiary value of the belief’s source. Using any other system of norms for judging the soundness of belief would be odd and would at least require its own justification. But when the question of the grounds for action is raised, one must first determine which system of norms or set of values is presupposed in raising the question.29 For unless one lives in some closed, tightly knit community, like that of a cult or the isolated villages of yore, in which a single religion or way of life regulates everyone’s conduct, there is no one system of norms or set of values by which the rightness or propriety of all actions is generally judged. Thus, it should be even plainer in the case of action than in that of belief, that actions may follow from rational thought applied to the object

27. August Kekulé, “August Kekulé and the Birth of the Structural Theory of Organic Chemistry in 1858,” O. T. Benfey, trans., Journal of Chemical Education 35 (1958): 21–​23. (A translation of Kekulé’s 1890 address at the twenty-​fifth anniversary celebration of the discovery of the structural formula for benzene held at the Berlin City Hall.) 28. Ibid., p. 22. 29.  Some contemporary philosophers hold that reasons one cites to justify doing an action are themselves fundamental and thus derive neither from norms nor values. For simplicity I  restrict the ultimate grounds that justify actions to norms and values. For one influential statement of the view that reasons are fundamental, see T. M. Scanlon, Being Realistic about Reasons (Oxford: Oxford University Press, 2015).

641

164  From Psychology to Morality

of some ungrounded desire, felt interest, natural appetite, or primitive emotion and independently of any thought of one’s having grounds relative to some system of norms or set of values for so acting. A person may have such grounds and know that he has, but that may not be why he does what he does. He may, after all, do the right thing for the wrong reason. Consider, for example, a man—​ let’s call him Rupert—​ who takes pleasure in making small children suffer.30 Rupert also wants to live comfortably and be respected by his colleagues and neighbors. He finds a way to satisfy both his sadistic and conventional desires by becoming a pediatric nurse. He goes to work at a clinic that specializes in health care for small children, where his responsibilities include using a syringe to administer vaccines to his young patients. The job thus gives him plenty of opportunity for gratifying his sadistic desires. In addition, Rupert takes a second job at a lab where he can again stick syringes into frightened children and cause them to suffer, for many come to the lab for tests that require his drawing their blood. Rupert is always careful to insert the needle into the child in accordance with accepted medical practice and to hide from onlookers the thrill he experiences at the child’s distress by keeping a blank face and a calm demeanor when administering the vaccine or drawing blood, lest his sadism be suspected. He is of course aware of the medical reasons he has for doing the things that cause his patients pain. These reasons are the grounds on which his actions are justified, and having received excellent training at nursing school, he understands this. But none of these reasons moves him. If he had gotten no pleasure from causing small children to suffer, he would have become an accountant who worked for a large firm that specialized in corporate taxes. They are not, then, the reasons for which he sticks needles into his patients.31 Williams begins “Internal and External Reasons” by distinguishing two opposing interpretations of R-​sentences, internal and external.32 The former is the one he espouses. An R-​sentence, on this interpretation, is true only if the agent has a desire that would be satisfied by her doing the action that

30.  The example is patterned after an example of Francis Kamm’s in her “Terrorism and Intending Evil,” Philosophy & Public Affairs 36 (2008): 157–​186, see esp. pp. 163–​166. 31.  Whether Baier would even count enjoyment of others’ suffering as a reason for causing it is unclear. He acknowledges that one’s enjoying an activity is a reason to engage in that activity, but it is possible that he meant only enjoyment of things that are not in themselves bad. In that case Baier would say that Rupert does not act for the wrong reason but rather acts for no reason. And in either case, Baier’s preferred explanation of Rupert’s action is that it springs from a bad motive. See The Moral Point of View, pp. 156–​160. 32.  “Internal and External Reasons,” p. 101.

615

Williams on Practical Reason  165

the sentence says she has a reason to do. On the external interpretation, an R-​sentence may be true even if the agent has no desire that would be satisfied by her doing the action that the sentence says she has a reason to do. The external interpretation, in short, contradicts the internal. Baier’s account of reasons for action clearly gives the external interpretation of R-​sentences. Such reasons, on Baier’s account, are equivalent to grounds that can justify an action, and there may be grounds justifying someone’s doing a certain action even if none of her desires would be satisfied by her doing it.33 Rupert’s sticking needles into his patients, for instance, is justified regardless of whether it satisfies any of his desires. The grounds on which it is justified are medical reasons for administering a vaccine or drawing blood using a syringe, even if none of them is Rupert’s reason for doing so. This seems uncontroversial, yet Williams, after distinguishing the two opposing interpretations and explaining the internal one, criticizes the external interpretation as either false or incoherent. How can this be? The answer lies in the issue over which Williams takes the two interpretations to be divided. The issue is that of the nature of rational agency. Williams assumes that exponents of either interpretation conceive of deliberation as an exercise of practical reason that leads to action. The exponents of the internal interpretation—​hereafter ‘internalists’—​hold that deliberation must proceed from some desire of the agent’s, some element in his or her motivational set, as Williams would say. Their opponents, the externalists, disagree. They hold that it is possible for an agent’s deliberation to proceed from a thought or thoughts that are distinct from any of the agent’s desires and that are not members of the agent’s motivational set. Baier’s conception of deliberation falls outside of this disagreement. As we saw, Baier denied that deliberation led to action. It leads, rather, on his conception, to the acceptance or rejection of a proposition that is an answer to the question ‘what should I  do?’ Hence, while Baier’s account of reasons for action gives the external interpretation of R-​sentences, its treatment of them as equivalent to grounds on which an action could be justified excludes it from being among the externalist accounts at which Williams directed his criticism. Does this save Williams’s criticism of the external interpretation from being defeated by Baier’s account?

33.  The element of an agent’s subjective motivational set that Baier believes moves the agent to do what he or she has reason to do is the desire to act justifiably, but Baier does not contend that every agent must have this desire. He holds only that it is common due to how people are traditionally brought up to deliberate before acting. Nothing in Baier’s account, then, precludes the possibility of Rupert’s lacking this element.

6 61

166  From Psychology to Morality

One could, of course, fault Williams for overlooking externalist interpretations like Baier’s in making his criticism. But one could also say, in Williams’s defense, that he took R-​sentences to be about reasons to do an action and not reasons to accept a proposition that doing an action would be pro tanto right or proper. That is, one could defend Williams as having understood attributions of reasons to act as having “a dimension of possible explanation” and, accordingly, taken R-​sentences to be about practical and not theoretical reasons.34 Yet even if one defended Williams in this way, one would still have to concede that there was nothing exceptional in how Baier used the term ‘reason for action’. One of its accepted uses, after all, is to refer to a consideration that would be among the grounds justifying an action if the action were justified. Accordingly, to avoid being defeated by interpretations like Baier’s, Williams’s criticism must be explicitly limited to externalist interpretations on which R-​sentences are about practical reasons. It needs, that is, to be explicitly limited to externalist interpretations that are consistent with an Aristotelian conception of deliberation. The importance of taking Williams as so limiting his criticism’s target is evident in what he regards as the criticism’s upshot, that it gives new support for Hume’s thesis that no motivation originates in reason.35 This upshot becomes clear, Williams thinks, when we take R-​sentences as representative of the conclusions of deliberation.36 On an Aristotelian conception, deliberation is an exercise of practical reason that leads to action. Accordingly, the rational process in which deliberation consists is understood to transmit the motivation that produces the action from its source in the agent’s psychology. Could this source be reason itself? This is what Hume denies. To show, contrary to Hume, that it could requires, then, showing how the motivation for the action could arise through a rational process independently of any already existing element in the agent’s motivational set. And Williams’s thought is that by letting R-​sentences stand in for the traditional conclusions of practical reason one can better see the difficulty in meeting this requirement. Accordingly, either interpretation of R-​sentences, external or internal, must render them as expressing thoughts that could be conclusions of deliberation on the Aristotelian conception. In short, Williams’s purpose in distinguishing the two opposing interpretations is to clarify our understanding of

34.  “Internal and External Reasons,” p. 106. 35.  Ibid., p. 108. 36.  More exactly, Williams holds that thoughts expressed by sentences of the form ‘A has more reason to φ than to do anything else’ are the conclusions of deliberations, but he uses R-​ sentences as placeholders for such thoughts. See p. 156 above.

617

Williams on Practical Reason  167

the place of practical reason in rational motivation. He does this by, first, distinguishing the two different meanings R-​sentences can have when they express thoughts that could be the conclusions of such deliberation and, then, considering how, under each interpretation, those thoughts could come to motivate actions in the agents who had them.

IV Standard philosophical accounts of reasons for action agree with Baier’s account in taking such reasons to be equivalent to grounds that one would cite to justify the action if it were justified. Many philosophers also agree with Baier in taking deliberation to be an exercise of theoretical reason that concludes with acceptance or rejection of a proposition that a given action is a right or proper thing to do (or that it should be done or would be best to do), where acceptance or rejection of such a proposition does not in itself motivate one to act accordingly. They therefore often see their accounts as opposed to Williams’s because they suppose that Williams’s criticism is directed at accounts like theirs. Once Williams’s criticism is understood to be limited as I have proposed, then the supposed opposition between their accounts and his dissolves. Like Baier, they have not given accounts of practical reason. Some philosophers, however, who expound one of these standard accounts disagree with Baier about the nature of deliberation. They conceive of deliberation as an exercise of reason that leads to action. Contrary to Baier, then, they treat such reasons as being potentially explanatory. That is, they agree with Williams that an attribution of a reason to do an action to an agent has a possible explanatory dimension. His criticism, therefore, extends to their accounts. The question, then, is whether such a combination of a standard philosophical account of reasons for action with an Aristotelian conception of deliberation can defeat Williams’s criticism. To do so, philosophers who advance this combination must show how an understanding of reasons to do an action as facts that one would cite to justify the action if it were justified coheres with an Aristotelian conception of deliberation. The problem is that the former concerns reasons to accept a proposition about the rightness or propriety of the action, and as we saw in the case of Rupert, one can know that one’s actions are justified and, indeed, know how to justify them and not be in the least moved by the reasons one would cite in justifying them. Rather what leads one to do the action is one’s seeing how by doing it one will attain the object of the desire that initiated one’s deliberation, and seeing and being moved by this feature of the action

681

168  From Psychology to Morality

has nothing to do with whether the action is justified. The problem, in short, is that reasons one cites to justify an action are theoretical reasons, and the reasons that enter into deliberation, on an Aristotelian conception, are practical. Williams’s criticism, in effect, is that externalism is the attempt to convert theoretical reasons into practical reasons, and all such attempts are about as successful as alchemy. One prominent attempt is T. M. Scanlon’s in his book What We Owe to Each Other.37 Scanlon expounds this account in the book’s first chapter. He expounds it as part of a general account of reasons, one that covers reasons for belief and other attitudes as well as reasons for action. He identifies the latter as “practical reasons” but implies that their logic is the same as the former. “The sense of ‘reason’ that I will be concerned with,” he declares, is the sense in which when one asks someone for reasons for believing P one asks for a justification of his belief that P.”38 Scanlon dubs this sense of the word “the standard normative sense.” Clarifying, he writes, “[A]‌sking what reason there is for believing that P [is to ask] for an assessment of the grounds for taking P to be the case.”39 Clearly, then, though Scanlon characterizes reasons for action as “practical reasons,” what he means by this characterization is theoretical reasons about practical matters. That this is what he means is evident in his taking them to be the same as grounds that one would cite to justify an action. His characterizing them as practical merely picks out their subject, just as we distinguish medical reasons from financial ones, moral reasons from prudential ones, or military reasons from diplomatic ones, according to their subject.40 How then does Scanlon move from theoretical reasons to practical reasons? How does he move from reasons that support the rightness or propriety of doing an action to reasons to do an action? Ironically, he does it by appealing to Williams’s interpretation of R-​sentences. Having observed

37. Scanlon, What We Owe to Each Other (Cambridge, Mass.:  Belknap Press, 1998), pp. 363–​373. 38. Ibid., p. 18. 39. Ibid., p. 19. 40. Anscombe made the same point about how philosophers of her day commonly characterized as practical reasoning ordinary reasoning that leads to a conclusion that such-​ and-​such an action ought to be done. Specifically, these philosophers typically took practical syllogisms to be syllogisms about practical matters. Anscombe acerbically remarked that their references to practical syllogisms had as much philosophical import as references to mince pie syllogisms that one might make to specify a type of syllogism that was about mince pies. Either was far from Aristotle’s understanding of practical syllogism as a distinctive type of reasoning that led to action as its conclusion. Anscombe, p. 58. See also fn. 26 above.

6 91

Williams on Practical Reason  169

that Williams, in expounding this interpretation, attributes normative force to such sentences, he wrongly assumes that Williams therefore takes ‘reason’ to occur in them in what he, Scanlon, had dubbed “the standard normative sense”. Since Scanlon takes himself to be using ‘reason’ in this sense, he concludes that reasons as he understands them can figure, just as they do on Williams’s account, in explanations of action that take the reason-​for-​action form. That is, since Williams takes ‘reason’ as it occurs in an R-​sentence to refer to a practical reason, Scanlon wrongly infers that ‘reason’, when it occurs in what he dubbed “the standard normative sense”, refers to a practical reason. He makes this inference as a result of having, when he takes reasons to do an action as grounds that justify doing the action, blinded himself to the fundamental distinction between theoretical and practical reasons. Scanlon’s error, in short, is to misread Williams as, when giving his internalist interpretation of R-​sentences, using ‘reason’ in the same sense as he uses it. The sense in which Scanlon uses ‘reason’, the so-​called “standard normative sense”, is the sense in which Baier understood and used the word in his account of reasons for action. Accordingly, the normative force of attributing a reason to do an action to an agent, on Baier’s account, reflects this understanding and usage. This force derives from propositions about the rightness or propriety of the agent’s doing the action. Such force is different from the normative force, on Williams’s account, of attributing reasons to do an action to an agent. The force Williams attributes to R-​sentences on his internalist interpretation, derives from the standard of rationality as applied to action. Clearly, given the difference between the normative force that Williams attributes to R-​sentences and the normative force that Baier takes attributions of reasons to have, Williams does not use ‘reason’ in what Scanlon dubbed “the standard normative sense”. Scanlon’s account of reasons then collapses into incoherence from this basic confusion. The muddle that he makes of a theory of reasons is complete within the first four pages of his exposition of the theory.41 Subsequently, in an appendix, Scanlon offers an extended critique of Williams’s view. This critique too is built on his confusion about practical reasons. That the critique depends on this confusion is evident in the formal

41. The same confusion infects Scanlon’s account of the explanatory dimension of R-​ sentences in his Being Realistic about Reasons. In this work, the confusion arises from his misconstruing Davidson’s account of reasons for action as an account of theoretical reasons. Davidson’s account, to the contrary, incorporates an Aristotelian conception of deliberation. See Being Realistic about Reasons, pp.  53–​54, for Scanlon’s misappropriation of Davidson’s account.

701

170  From Psychology to Morality

principle Scanlon makes its chief vehicle. He calls this principle “the universality of reason judgments.” Briefly, it is that anyone who judges he has a reason to φ must regard anyone else in relevantly similar circumstances as also having a reason to φ. This principle applies to reasons one would cite as grounds for a proposition that a certain action is a right or proper thing to do, for an action’s circumstances determine whether it is right or proper to do. Rupert, for instance, could cite medical reasons to justify his sticking needles into the small children who are his patients just as any other pediatric nurse could cite these reasons. And when he judges that he has medical reason to stick needles into his patients, he regards other nurses as having this same reason to do so. At the same time, the principle does not apply to his reasons for sticking needles into his patients, for Rupert is an anomaly among pediatric nurses in desiring to cause his patients to suffer and in being indifferent to promoting their health. Pediatric nurses, with few exceptions, do not desire to cause their patients suffering, and Rupert is well aware of this. Consequently, he could find himself in circumstances in which he judges he has a reason to cause a child pain—​say, a chance opportunity in a crowded bus to step on a six-​year-​old’s toes—​without ever supposing that other pediatric nurses also have a reason to do the same thing. Scanlon’s principle, in other words, does not apply to practical reasons. His critique of Williams comes to nothing more than an elaborate commission of the fallacy of equivocation.42

V While much of the opposition to Williams’s account of R-​sentences is similarly due to a misidentification of his account with one like Baier’s, not all of it is. Some of the opposition is due instead to disagreement over the nature of practical reason. Specifically, some of Williams’s opponents, notably, Christine Korsgaard, object directly to Williams’s suggestion that one can find in his criticism of the external interpretation new support for Hume’s thesis.

42. Scanlon’s critique is an elaborate ad hominem objection to Williams’s account. Williams, Scanlon argues, plants a seed that destroys his account when he includes in an agent’s subjective motivational set dispositions of evaluation. Scanlon then tries to show that these dispositions, owing to the universality of reason judgments, entail that the agent has reason to do a certain action despite lacking a sound deliberate route from any of his desires to his doing that action. What Scanlon fails to see, however, is that the dispositions of evaluation to which the universality of reason judgments applies are dispositions to justify one’s action, and dispositions of that sort need not belong to the agent’s subjective motivational set.

171

Williams on Practical Reason  171

The criticism could yield such support, it is contended, only if it showed that none of an agent’s desires originated in principles of practical reason, and nothing in the criticism excludes this possibility. The criticism, that is, could yield such support if the agent’s motivational set did not contain principles of practical reason, and nothing in the criticism excludes such principles from an agent’s motivational set.43 I will call this the rationalist objection to Williams’s view of practical reason. The objection seems well-​t aken. Indeed Williams has conceded its force.44 Nonetheless, it should not pass without examination. Williams, it is worth noting, does not go so far as to maintain that his criticism of the external interpretation entails Hume’s thesis. What he says is that it clarifies the issue between Hume and his critics and in doing so shifts the “onus of proof” onto them by relieving Hume’s side of the burden of specifying what counts as a “purely rational process.”45 The rationalist objection we are considering ostensibly meets this burden by appealing to principles of practical reason. “They determine what counts as a purely rational process in practical thought,” I imagine our rationalist objector would say. “So the burden shifts back to Hume’s side. It is up to the defenders of Hume,” the objector would go on, “to show that no desire of a rational agent originates in such a principle, that no such principle is contained in a rational agent’s motivational set.” The assumption on which the rationalist argues is that both sides agree on there being such principles and disagree over their number and character. In particular, they disagree over whether the defenders of Hume can make a solid case for limiting such principles to ones that are consistent with Hume’s view.46 But it might be granting the rationalist too much to accept this assumption. Perhaps rationalists are mistaken about the common ground between their side and Hume’s. Perhaps there are no principles of practical reason that specify what counts, in practical thought, as a purely rational process. Consider how Williams characterizes deliberation near the end of “Internal and External Reasons.” Responding to a criticism he supposes someone might make that his account of deliberation is unclear and even

43.  Christine Korsgaard, “Skepticism about Practical Reason,” Journal of Philosophy 82 (1986): 5–​25; see esp. §6. 44.  “Internal Reasons and the Obscurity of Blame,” p. 44n3. 45.  “Internal and External Reasons,” p. 108. 46.  Christine Korsgaard, “The Normativity of Instrumental Reason” in Ethics and Practical Reason, G. Cullity and B. Gaut, eds. (Oxford: Clarendon Press, 1997), p. 215.

721

172  From Psychology to Morality

allows for the use of imagination to alter the elements of an agent’s motivational set, he writes, I regard it as a basically desirable feature of a theory of practical reasoning that it should preserve and account for that unclarity. There is an essential indeterminacy in what can be counted as a rational deliberative process. Practical reasoning is a heuristic process, and an imaginative one, and there are no fixed boundaries on the continuum from rational thought to inspiration and conversion.47

The question is whether one can take such a view of deliberation and at the same time recognize principles of practical reason that specify what counts, in practical thought, as a purely rational process. Presumably, if there are such principles, then deliberation that conforms to them is sound and deliberation that does not is unsound. In this respect, they would be like the principles of theoretical reason to which the reasoning exemplified in a derivation of a proposition must conform if the derivation is sound. The reasoning consists in a series of inferences, each of which is sound if it conforms to a principle of theoretical reason that represents a purely rational process and unsound otherwise. But if we similarly determine the soundness of deliberation by applying principles of practical reason to the inferences that it consists in, then it is hard to see how Williams’s notion of sound deliberation, to which he appeals in his preferred account of the truth conditions of R-​ sentences, can work given his characterization of the deliberative process as heuristic. Sound use of imagination is not a matter of conforming one’s use of this power to certain principles of practical reason. To be sure, one could have a mixed view of deliberation such that it included both rational processes whose cogency consisted in conformity to principles of practical reason and rational processes whose cogency lay elsewhere. But the two seem to me to sit uneasily together when a general assessment of the soundness of deliberation that combines both is called for. Heuristics and demonstration are simply different models of rational thought. Since Williams’s preferred account of the truth conditions of R-​sentences calls for such an assessment in its appeal to a sound deliberative route, I am inclined to think that his view of deliberation is best regarded as unmixed. That is, I am inclined to think that there is an implicit rejection, in his view, of the rationalist idea of deliberation as regulated by principles of practical reason.48 Thus, if you are attracted to Williams’s view, then you should find this idea suspect. 47.  “Internal and External Reasons,” p. 110. 48.  Williams, in “Postscript” (pp. 93–​94), returns to Korsgaard’s objection. He characterizes it as proposing that “the structure and not simply the content of practical reason can ground

713

Williams on Practical Reason  173

Of course, you might be convinced that there are principles of practical reason that specify what counts, in practical thought, as a purely rational process. In that case, your thought might be, “So much the worse for Williams’s view.” In any event, I do not mean to be saying at this point that Williams’s view is superior to the rationalist’s. What I am saying is that the rationalist objector cannot shift the burden back to Hume’s side merely by appealing to the possibility of there being desires that originate in principles of practical reason. For if the possibility of there being such desires is put into doubt by questions concerning the nature of practical reason, then the rationalist must say something more to show that this is a real possibility. One rationalist line of argument should be familiar from Thomas Nagel’s The Possibility of Altruism.49 Nagel writes in the book’s opening sentence, “Just as there are rational requirements on thought, there are rational requirements on action . . .”50 So one might take Nagel to be saying that just as there are principles or laws to which thought must conform to be rational, so there are principles or laws to which action must conform to be rational. The latter would then be principles of practical reason. But to what are they being compared when Nagel speaks of rational requirements on thought? One might think initially that they are being compared to the laws of logic. And it is certainly easy to suppose that the Law of Contradiction, ~ (p.~ p), represents a rational requirement on thought. For it is easy to suppose that it represents a prohibition on believing p while at the same time believing ~ p. But unfortunately not every law of logic is as easily regarded as a rational requirement on thought. Take the Law of the Excluded Middle, (p v ~ p). If this law represented a rational requirement on thought, say, believe p or believe ~ p, then one could not rationally be agnostic about a given proposition. Indeed, it would be a failure of rationality to regard any question that admits of true and false answers as open.51 reasons” and allows that if it could be made good, which he doubts, what it would yield would be a “limiting version of internalism” in this sense: “If it were true that the structure of practical reason yielded reasons of a certain kind as binding on every rational agent, then it would be true of every rational agent that there was a sound deliberative route from his or her S to actions required by such reasons.” What I’m suggesting is that his own characterization of deliberation excludes the possibility of its have this sort of structure. 49.  Thomas Nagel, The Possibility of Altruism (Oxford: Clarendon Press, 1970). 50. Ibid., p. 3. 51.  For convenience’ sake I use standard notations of symbolic logic to represent the two laws. Note that on the standard interpretation of the notations given in classical propositional logic the two laws are equivalent. So the point about there being a difference between the laws requires a different interpretation. See Alonzo Church, Introduction to Mathematical Logic: Volume I (Princeton: Princeton University Press, 1956), pp. 141–​142.

741

174  From Psychology to Morality

Similarly, it would be a mistake to think that rules of inference in logic are rational requirements on thought. A rule of inference in logic validates or invalidates inferences one makes, say, in the context of constructing a derivation, but it never requires that an inference be made. This is one of the main points of Lewis Carroll’s parable about Achilles and the Tortoise.52 And it is perhaps best seen with the rule of inference addition:53 p p⋅ v ⋅q

Given this rule, you could never rationally stop making inferences, once you had affirmed some proposition, if the rule required, as a matter of rationality, that you make the inferences it validates. Once you had affirmed, for example, that Jupiter is the largest planet in the solar system, you would be rationally required to infer every disjunction in which this proposition was one of the terms, and there are infinitely many disjunctions that fit this description. Oddly, some moral philosophers think that the lesson of Carroll’s parable is the necessity of rules of inference to theoretical reasoning. A good example is Stephen Darwall’s remark, in his book The Second-​Person Standpoint, that no one can have rational beliefs unless norms of rationality regulate his or her thinking. Citing Carroll as having “famously” made this point, Darwall writes, “Without guidance by some norm of inference, a set of beliefs or believed premises is impotent to yield conclusions.”54 And using this remark for support, Darwall goes on to say that the same thing is true of practical reasoning. Nothing, however, in Carroll’s parable supports the point Darwall attributes to it. Achilles, to be sure, repeatedly replaces modus ponens with conditional propositions. These, you may recall, he adds ad infinitum to an original list of three propositions from Euclid, the last of which, Z, he is trying to compel the Tortoise to accept, as Z follows logically from the first two and the Tortoise has accepted them. We can, therefore, understand the

52.  Lewis Carroll, “What the Tortoise Said to Achilles,” Mind 4 (1895): 278–​280. 53.  See Donald Kalish, Richard Montague, and Gary Mar, Logic:  Techniques of Formal Reasoning (New York: Harcourt, Brace, and Jovanovich, 1980), p. 60. 54. Stephen Darwall, The Second-​ Person Standpoint (Cambridge, Mass.:  Harvard University Press, 2007), pp. 154–​155. Similarly, Darwall writes, “The rational formation of new beliefs through inference cannot be fully explained by belief . . . ; it also requires regulation by a norm of inference” (ibid., p. 155). Darwall’s appeal to Carroll’s parable suggests the error I explain below. Alternatively, however, the error may be the one I pointed out in section III of thinking that theoretical reasoning only takes place in the context of justification.

715

Williams on Practical Reason  175

regress into which Achilles falls as resulting from his having the same confused thought at each stage of his efforts. And the confusion in this thought consists not only in his conflating conditional propositions with modus ponens but also in his misunderstanding the different roles that premisses and rules of inference have in a derivation of some proposition. Premisses are accepted and argued from; rules of inference are applied in so arguing. Hence, even if in a moment of partial clarity Achilles formulated modus ponens and elicited the Tortoise’s acknowledgment of it as a sound rule, the Tortoise would still not be forced to accept Z. This is because acknowledging a rule is not the same thing as applying it, and the Tortoise would be forced to apply the rule only if it required him to make inferences whenever he could. As we have seen, rules of inference in logic do not require that one make inferences. So there would be no inconsistency in the Tortoise’s acknowledging the rule and declining to make some inference it validates. “Logic,” Achilles says towards the end of the parable, frustrated by the Tortoise’s having refused yet again to accept Z, “will take you by the throat and force you to [accept] it.”55 Carroll’s point, of course, is that logic has no such power. This is a point about the character and role of rules of inference in logic, and from it nothing follows about the nature of rational thought. Indeed, Carroll’s larger point is that the rules of inference in logic are misunderstood when they are taken to be laws of thought or principles of reason. The latter, if they existed, would be psychological laws and psychological principles, and logic is not a branch of psychology. Hence, while some types of inference may, as a matter of human psychology, be compelling once one recognizes their soundness, it does not follow that the rule in a system of formal logic that validates inferences of that type also requires that one make the inference. Modus ponens, for instance, may be just such a rule. Once one affirms some simple proposition p and a conditional proposition p → q, then one may find that one automatically draws the conclusion q and that one cannot resist doing so. But these psychological phenomena have no bearing on the modal character of rules of inference in a system of formal logic. That one cannot resist drawing the conclusion does not mean that the rule should be formulated as a requirement. Some philosophers may find these points about rules of inference in logic unsatisfying. Surely, they think, rules of inference constrain our thinking in some way and are thus requirements on thought in the sense that we are required to conform our thinking to them when making inferences. But this

55.  Carroll, p. 280.

761

176  From Psychology to Morality

understanding of them is too strong. A cautious thinker, of course, might never add to his or her store of beliefs unless the addition followed in accordance with some rule of inference from a set of beliefs he or she already held. But a bolder thinker might add to the store of his or her beliefs by drawing analogies, say, for which there is no rule of inference, without putting his or her rationality into question. Darwall’s assertion that to reason is to infer a conclusion from premisses by following a rule—​that it is essentially norm-​regulated thinking—​ is nothing more than rationalist dogma, an echo, perhaps, of Kant’s doctrine that reason operates according to its own laws. When one reasons by analogy, the cogency of one’s reasoning depends on the goodness of the analogy one uses, and no rule determines the goodness of an analogy.56 In general, then, I think any attempt to model principles of practical reason on the laws of logic or its rules of inference will end up either misrepresenting logic or setting too pinched a standard for what counts as rational thought. At the same time, we have, in looking to logic for rational requirements on thought, uncovered one such rational requirement that might serve as a model for a rational requirement on action. This is the requirement of consistency in the sense of noncontradiction. Kant, of course, is famous for having proposed that there are requirements of consistency in the will that are in practical reason counterparts to this requirement, and Korsgaard plainly has Kant in mind when she appeals to principles of practical reason in objecting to Williams’s view. In particular, she thinks that there is one principle that even the defenders of Hume recognize and use to found their accounts of

56.  In particular, it would be a mistake to suppose that the rule of inference one follows in reasoning by analogy is: y is similar to x F (x ) F (y )

If this were a rule of inference, one could then cogently conclude that yellow jackets make honey from the similarity of yellow jackets to bees and the fact that bees make honey. The evident fallaciousness of the argument shows that similarity is not what makes an analogy a good one. Moreover, the same example shows that the goodness of an analogy between x and y is not a function of the degree of similarity between x and y. Rather, only certain kinds of similarity make good analogies, and for the similarity between x and y to be a good analogy and thus support inferring F(y) from F(x), it must not only be a similarity of the right kind but the property that F represents must belong to the same domain of properties as the properties the sharing of which by x and y is the basis of the analogy. See Dedre Gentner, “Structure-​ Mapping: A Theoretical Framework for Analogy,” Cognitive Science 7 (1983): 155–​170. The indeterminacy of the kinds of similarity that make good analogies alone entails the absence of a rule of inference governing cogent analogical reasoning.

717

Williams on Practical Reason  177

practical reason. This is the so-​called instrumental principle. This principle rationally requires taking the necessary means to our ends. The idea is that one’s deciding to pursue some end is an act of will with which one would be inconsistent if one resisted doing what was necessary to achieve that end, and such an inconsistency in the will is a mark of irrationality. I do not doubt that such an inconsistency is a mark of irrationality. But what I question is whether we need to introduce a principle of reason in order to explain such irrationality. Isn’t it sufficient to say that a person who resists taking the necessary means to ends she has decided to pursue willfully frustrates her own purposes, and to frustrate willfully your own purposes is a kind of irrationality? Surely, we don’t need the principle fulfill your purposes in order to understand this. What would be the point? After all, it is not as though you could have purposes but not know what to do with them without some principle that tells you. And as far as I can see, the so-​called instrumental principle is similarly idle. You do not need it to know that having an end entails taking the means necessary to achieving it, for means just are what a person takes to achieve his ends. The rationalist objection to Williams’s view rests, as I  said, on the assumption that the existence of some principles of practical reason is common ground between both parties to the dispute. Specifically, both parties, it is assumed, take the instrumental principle to govern deliberation about means, and what they dispute is whether there are any other principles of practical reason. Hume, it is said, is an instrumentalist. He and his followers think all practical reasoning consists in deliberation about means, and therefore they recognize only one principle of practical reason. Rationalists, by contrast, recognize additional principles of practical reason by virtue of which some practical reasoning consists in deliberation about ends. But the assumption on which this way of characterizing the dispute rests is unwarranted. Defenders of Hume’s side may challenge directly the existence of any principles that govern the operations of practical reason. The dispute, in that case, will be over the question of whether deliberation, when sound, is a rational process that proceeds according to principles of practical reason. Is it a process like the thinking that goes into the justification of scientific hypotheses and theories, which, to be sound, must conform to certain standards and protocols of the relevant science? Or is it more like the thinking that goes into the discovery of scientific hypotheses and theories, which need not conform to any standards or protocols to be sound?57 The 57.  Thus Aristotle takes deliberation to be a form of inquiry as distinct from demonstration and the forming of opinion. See Nichomachean Ethics, bk. III, ch. 3 (1112b12–​24) and bk. VI, ch. 9 (1142a33–​b15).

781

178  From Psychology to Morality

rationalist side of the dispute thus has the burden of showing that it is more like the former, that it proceeds according to principles of practical reason. And until they do, the onus of proving what counts in the area of practical reason as a pure rational process remains with them. Williams’s view of deliberation is more like the latter, and in articulating it, he has, as I hope to have shown, shifted the burden onto the rationalist side.

791

8

Sidgwick’s Conception of Ethics

J. B.  Schneewind’s Sidgwick’s Ethics and Victorian Moral Philosophy surpassed all previous treatments of Sidgwick’s The Methods of Ethics by showing how Sidgwick’s work follows a coherent plan of argument for a conception of ethics as grounded in practical reason. Schneewind offered his interpretation as the product of a historical rather than a critical study. This article undertakes a critical study of Sidgwick’s work based on Schneewind’s interpretation. Its thesis is that the conception of ethics for which Sidgwick argued is incoherent. As a result, it is argued, the coherent plan of argument in the Methods that Schneewind disclosed masks a deep incoherence in the argument itself. Sidgwick’s Methods of Ethics places great demands on its reader.1 It is a comprehensive study of ethics, and to follow and understand it more than superficially requires tremendous concentration and retention. C. D. Broad memorably captured the frustrating experience of reading the work when he wrote in his chapter on Sidgwick in Five Types of Ethical Theory:

1.  Henry Sidgwick, The Methods of Ethics, 7th ed. (1907; reprint ed., Indianapolis, 1981).

179

801

180  From Psychology to Morality [Sidgwick’s] style is heavy and involved.  .  .  . He incessantly refines, qualifies, raises objections, answers them and then finds further objections to the answers. Each of these objections, rebuttals, rejoinders and surrejoinders is in itself admirable, and does infinite credit to the acuteness and candour of the author. But the reader is apt to become impatient; to lose the thread of the argument; and to rise from his desk finding that he has read a great deal with constant admiration and now remembers little or nothing.2

Broad’s summary keenly displays the problem that confronts many readers of Sidgwick’s masterpiece. They cannot see in its wealth of analyses, criticisms, arguments, and constructions any overarching argument that would organize these different exercises into a coherent whole. For many readers, then, the book has seemed more like a reference work to be consulted on specific topics in ethics than a philosophical essay that defends a general account of the field. This, at any rate, was a common problem before the publication of Jerome Schneewind’s marvelous treatise on the Methods and its place in nineteenth-​century British moral philosophy.3 It was Professor Schneewind’s express aim to correct this problem, and his success in doing so has not only added immeasurably to our understanding of Sidgwick’s work but changed radically our view of its chief purpose. Schneewind succeeded by drawing on what, even to readers who get lost in all of the Methods’ twists and turns, is evident in Sidgwick’s treatments of specific ethical topics. For even to such readers it is evident that Sidgwick, in treating a specific ethical topic, seeks to advance our understanding of it by resolving differences among the main contending philosophical views on it, that his efforts are directed at finding a happy synthesis of those views, and that his genius lies in his ability to find some way of framing the topic’s central question and some underlying concept, principle, or procedure on which the disputants on that question can agree and on the basis of which an answer seemingly satisfactory to all can be worked out. Accordingly, Schneewind proceeded on the interpretative hypothesis that the Methods in toto is a study of this kind: that it seeks to advance our understanding of ethics by resolving differences among the main competing philosophical systems in the field; that its development is directed at finding a happy synthesis of those systems; and that to the extent that it achieves this synthesis, it does so through an argument that exhibits Sidgwick’s distinctive genius for fashioning and finding common ground among competing views. By following this hypothesis, then, Schneewind

2.  Five Types of Ethical Theory (London, 1930), pp. 143–​144. 3.  J. B. Schneewind, Sidgwick’s Ethics and Victorian Moral Philosophy (Oxford, 1977).

8 11

Sidgwick’s Conception of Ethics  181

produced an original and compelling interpretation of the Methods on which the work advances, from its introductory chapter on the nature of ethics to its famous concluding chapter on the problem of the dualism of practical reason, an integrated and cumulative argument for seeing ethics as primarily the systematic elaboration of the fundamental demands of practical reason as applied to the circumstances of individual human beings.4 Where previously the Methods had been read as a disparate collection of profoundly intelligent discussions whose most important ambition was to vindicate utilitarianism, it is now evident, thanks to Schneewind’s interpretation, that one can, to the contrary, comprehend the work as a coherent whole whose primary aim is to defend a conception of ethics as grounded in practical reason. Schneewind, in the preface of his book, describes his project as ‘a historical rather than a critical study’5 and explains that his reason for pursuing a study of this kind and not the other is that ‘it seemed necessary, before criticizing Sidgwick, to have a sound historical grasp of the problems he was trying to solve as well as a clear understanding of the solutions he offered’.6 This explanation makes good sense in view of the character of Sidgwick’s solutions, for they are, as I  noted, typically syntheses of the views of his predecessors and to criticize them without regard to these origins would invite misunderstanding, while criticizing them in tandem with an attempt to understand them through regard to their origins would be unmanageably complicated at best. At the same time, the decision to pursue a historical study over a critical one meant giving an interpretation that was intentionally uncritical, and such an interpretation does have a definite drawback. Being uncritical, it risks leaving the impression that Sidgwick succeeded in giving a coherent argument for his conception of ethics as grounded in practical reason when whether he did succeed is an open question. For although Schneewind’s interpretation, by laying out the Methods’ argumentative structure, shows the plan of argument that Sidgwick followed, it does not establish the coherence of its execution. The worry, then, is that since the interpretation is itself a complex argument for reading the Methods as having this argumentative structure, its cogency and success can easily lead one, in the absence of criticism of Sidgwick’s argument, unjustifiably to believe that

4.  “[T]‌he central thought of the Methods of Ethics is that morality is the embodiment of the demands reason makes on practice under the condition of human life and that the problems of philosophical ethics are the problems of showing how practical reason is articulated into these demands” (ibid., pp. 303–​304). 5.  Ibid., p. vii. 6. Ibid.

821

182  From Psychology to Morality

this structure is sound, that Sidgwick’s argument is fully coherent. Indeed, the very brilliance of Schneewind’s interpretation may tend to obscure the defects in Sidgwick’s argument that render it incoherent. The irony in that case of a decision to pursue a historical study over a critical one would be that the study ended up obscuring factors that would help explain why readers of the Methods have trouble seeing how the work’s various discussions fit together into a coherent whole at the same time as it provided relief from that trouble. None of these observations, I hasten to point out, is meant to be a criticism of Schneewind’s interpretation. The interpretation, in my view, is unsurpassed in its excellence, and I do not, in making these observations, mean to suggest that it could be improved or that a better one is available. Rather, in making them I mean to introduce an issue that the cogency and success of the interpretation raises but that lies beyond the scope of Schneewind’s study. It is an issue that belongs instead to the sort of critical study that Schneewind decided to forgo. The issue is whether Sidgwick, in giving the argument that according to Schneewind’s interpretation structures the Methods from beginning to end, succeeded in giving a coherent argument for his conception of ethics as grounded in practical reason. Sidgwick, on Schneewind’s interpretation, appears to have constructed a powerful argument for this conception, one that philosophers with a similar bent towards understanding ethics as the product of practical reason should find attractive. It is my contention, though, that the argument fails to establish this conception and that, after examining its construction and despite the impetus Schneewind’s interpretation gives us to see it as coherent, we should conclude that the argument is not coherent. Indeed, we should conclude that the argument is deeply incoherent. Of course, Sidgwick himself admitted that an incoherence existed in his final account of ethics. What he called “the dualism of practical reason” is, after all, an admission of the possibility of an irresolvable conflict between deliverances of practical reason each of which follows from the application of one or another of the bona fide methods of ethics to which he had reduced the field. It is an admission, in other words, that ethics, regarded as realizable through each of the rational procedures for “determin[ing] what individual human beings ought . . . to do”7 that Sidgwick identified as methods of ethics, is not necessarily free of encompassing contradictory judgments. And plainly, if an account of ethics leaves open this possibility, then it has failed to explain the discipline as a fully coherent study. But the account’s failure in this regard

7.  Methods, p. 1.

813

Sidgwick’s Conception of Ethics  183

does not mean that the conception of ethics behind the account is flawed. Sidgwick’s conception of ethics as grounded in practical reason, which is to say, his conception of ethics as realizable through rational procedures for determining what an individual ought to do, may be perfectly sound notwithstanding the failure of the account he developed from it. For under a different specification of what rational procedures counted as methods of ethics it might support an account of ethics as a fully coherent study. That is, under a different specification, the sets of procedures that count as methods of ethics might be ones whose deliverances were always and necessarily consistent with each other, and if they were, one could then give an account of ethics based on Sidgwick’s conception that did not entail the (or any) dualism of practical reason. The incoherence in his final account of ethics does not, therefore, entail an incoherence in either the conception of ethics behind that account or the overarching argument of the Methods that gives that conception its force. Nonetheless, this incoherence in the account may have its source in an incoherence in either the conception or the overarching argument that gives it its force. The dualism of practical reason may be a symptom of a knot Sidgwick got himself into in formulating the conception or developing the argument. And one thing suggesting this possibility is Sidgwick’s having formulated the conception so as to include egoism as well as intuitionism and utilitarianism among the methods of ethics. For egoism, unlike intuitionism or utilitarianism, is sometimes rejected as an ethical theory, and plainly any account of ethics that counts it along with the other two as a method will necessarily yield the dualism that renders the account a failure at representing ethics as a fully coherent study. Perhaps, then, the incoherence the dualism creates is due to some incoherence or confusion in Sidgwick’s having treated egoism as a rational procedure for determining what an individual ought to do that is comparable to intuitionism and utilitarianism. I will call this “the deep incoherence hypothesis”. Of course, egoism, as a matter of the history of the discipline, is an ethical theory and, indeed, an important one. Hence, it must correspond in some sense to a method of ethics. But the question that the deep incoherence hypothesis raises is whether egoism, on Sidgwick’s conception of ethics, qualifies as a method of ethics by the same criterion by which utilitarianism and intuitionism qualify as methods of ethics. And its being historically an important theory of ethics does not show that it qualifies by the same criterion. Rather, to qualify, egoism must be a rational procedure for determining what an individual ought to do, where ‘ought’ is understood in the stricter ethical sense of the word that Sidgwick isolated. In this sense, ‘ought’ is used

841

184  From Psychology to Morality

in judgments of the form ‘φ ought to be done’, where φ is an action whose performance is within the power of the individual to whom the judgment is addressed, and any such judgment is essentially a cognition of the rightness or reasonableness of φing. What is more, Sidgwick maintained, the cognition is such that “in rational beings as such [it] gives an impulse or motive to action”.8 Accordingly, he characterized it as a dictate or imperative of reason and noted that such cognitions appear most clearly as dictates of reason when some inclination or passion opposes them. Utilitarianism and intuitionism qualify as methods of ethics, therefore, because one can understand either as a rational procedure that produces this sort of cognition. One can understand either, that is, as producing, when applied to practical circumstances, cognitions of rightness that count as dictates of reason in Sidgwick’s sense. The question, then, is whether one can understand egoism as a rational procedure that produces cognitions of the same sort. Each method of ethics, on Sidgwick’s account, applies an ultimate criterion of rightness or reasonableness to an agent’s practical circumstances. The criterion that utilitarianism applies is the conduciveness of an action to advancing the general happiness of humankind. The criterion that intuitionism applies is the conformance of an action to duty as defined by fundamental and self-​evident principles of duty. Since many and perhaps most people have neither a desire for the general happiness of humankind nor a desire for doing their duty for duty’s sake, it is common for either of these two methods, being a rational procedure for determining what one ought to do, to yield dictates prescribing actions that would not satisfy any of the person’s desires. Consequently, if Sidgwick were right in maintaining that the cognitions the method produced gave in rational beings an impulse or motive to action, the impulse or motive they gave would not arise, at least not wholly, from any desire. It would arise, instead, at least in part, from the cognitions themselves. For this reason, these cognitions count as categorical dictates or imperatives. They count as categorical in the sense that, if they do give an impulse or motive to action in the mind of the person who has them, as Sidgwick maintained, then they have this motivating force independently of that person’s desires. The same point, however, cannot be made about the cognitions that egoism produces. Egoism too, as a method of ethics, is a procedure for determining what one ought to do. The criterion of rightness it applies is the conduciveness of the action to advancing one’s own happiness. But since almost everyone desires

8. Ibid., p. 34.

815

Sidgwick’s Conception of Ethics  185

his own happiness, it seldom, if ever, happens that a cognition produced by the method prescribes an action that would not satisfy a desire. Moreover, one could not have evidence of the method’s ever having produced such a cognition if the cognition must give an impulse or motive to action, for it would have to be an impulse or motive to promote one’s own happiness that was distinct from a desire for one’s own happiness, and it is hard to see what evidence one could give that identified the impulse or motive as distinct from that desire. Hence, Sidgwick’s thesis that the impulse or motive to action that the cognition gives in a rational being arises from the cognition itself is problematic in the case of egoism. For unlike the cognitions that utilitarianism and intuitionism yield in those who apply these methods, one cannot adduce conflicts between the cognitions that egoism produces in those who apply it and their desires to show that the impulse or motive to action the cognitions give arises independently of those desires, since one cannot identify the impulse or motive as being distinct from the desire for happiness. Consequently, one cannot use such conflicts to support taking these cognitions as categorical dictates or imperatives. And unless there is other evidence for taking them to be cognitions of this sort, taking them as such is problematic. Thus, in the absence of such evidence, it is problematic whether egoism is a rational procedure that yields cognitions of what one ought to do in the strict ethical sense of ‘ought’ that Sidgwick isolated. It is problematic, that is, whether egoism qualifies as a method of ethics in Sidgwick’s sense. Sidgwick, to be sure, did not ignore the question of whether there were adequate grounds for taking egoism to be a method of ethics in his sense. On the contrary, he gave reasons to justify his doing so. His main reason comes at the very beginning of his exposition, when he first identifies the different methods of ethics. His argument, in identifying them, consists in an appeal to what ‘the common sense of mankind’, to use his phrase, regards as the ultimate reasons for action.9 “This is”, as he later says, “. . . the starting-​point for the discussions of [his] treatise”.10 Each of these ultimate reasons corresponds to a method of ethics, and one of them is conduciveness to one’s own happiness. This is, as a matter of common sense, an ultimate reason for action, for—​as Sidgwick observed—​we commonly take an action’s being conducive to the actor’s own happiness as alone sufficient for its being rational. “We do not all look”, he wrote, with simple indifference on a man who declines to take the right means to attain his own happiness, on no other ground than that he does not care about

9. Ibid., p. 6. 10. Ibid., p. 78.

861

186  From Psychology to Morality happiness. Most men would regard such a refusal as irrational with a certain disapprobation; they would thus assent to Butler’s statement ‘interest, one’s own happiness, is a manifest obligation’. In other words, they would think that a man ought to care for his own happiness. The word ‘ought’ thus used is no longer relative:  happiness now appears as an ultimate end, the pursuit of which—​ at least within the limits imposed by other duties—​appears to be prescribed categorically.11

Accordingly, conduciveness to one’s own happiness is an ultimate reason for action and therefore an ultimate criterion of rightness and as such can be used as the basis for a method of ethics. This is not, however, a compelling argument for holding that reason prescribes one’s own happiness as an ultimate end. It may well be true that a person can never be criticized as irrational for pursuing his own happiness and would sometimes be criticized as irrational if he did not. But it does not follow from this that reason prescribes one’s own happiness as an end, much less an ultimate one. People who pursue the happiness of another out of love and affection for that person and at a sacrifice of their own happiness do not necessarily act contrary to reason even when that sacrifice is deliberate and even when they say that their own happiness does not matter. Nor are they necessarily pursuing an end that reason has prescribed. They are pursuing an end that love has elevated above their self-​interest, and it would be ludicrous to suppose either that this elevation must be irrational or that reason had prescribed it. Yet if reason categorically prescribed the pursuit of one’s own happiness, then one or the other of these suppositions would have to be true. That we never fault a person as irrational for pursuing his own happiness, assuming this is true, shows only that we think self-​love is too strong a motive in most people’s lives to be subject to abandonment through reasoning and argument, and nothing in this thought implies that reason lies behind self-​ love or endorses it. Indeed, we take the same view of love of another that is so deep as to lead a person to sacrifice his own happiness for the sake of his beloved. We think that love in such cases is too strong a motive to be subject to abandonment through reasoning and argument, and accordingly we do not fault as irrational the lover’s ignoring his own happiness in acting for the sake of his beloved’s. Nor do we think that reason lies behind such love or endorses it. The point is the same as a famous observation Pascal made centuries ago and much more elegantly.12 Sidgwick’s grounds for thinking

11.  Ibid., p. 7. See also, p. 119. 12.  “Le cœur a ses raisons, que la raison ne connaît point.” [“The heart has its reasons, which reason does not know”], Blaise Pascal, Pensées, §277.

817

Sidgwick’s Conception of Ethics  187

reason prescribes one’s own happiness as an ultimate end would seem, then, to have no more to recommend it than the common prejudice of identifying self-​love with reason rather than as a species of love. To Sidgwick’s credit, though, he did not insist on the thesis that reason prescribes one’s own happiness as an ultimate end. He did not think acceptance of this thesis was necessary for taking egoism as a method of ethics. It was enough that self-​love was a predominant motive in human life and that people adopted their own happiness as an “ultimate and paramount” end. Thus, shortly after first laying out the grounds on which he held that reason prescribes one’s own happiness as an ultimate end, he wrote: At the same time it is not necessary, in the methodical investigation of right conduct, considered relatively to the end either of private or general happiness, to assume that the end itself is determined or prescribed by reason: we only require to assume, in reasoning to cogent practical conclusions, that it is adopted as ultimate and paramount. For if a man accepts any end as ultimate and paramount, he accepts implicitly as his ‘method of ethics’ whatever process of reasoning enables him to determine the actions most conducive to this end.13

And two chapters later he observed: Here, however, it may be held that Egoism does not properly regard the agent’s own greatest happiness as what he ought to aim at; but only as the ultimate end for the realization of which he has, on the whole, a predominant desire.14

At bottom, then, Sidgwick thought he had adequate grounds for taking egoism as a method of ethics, whether or not reason prescribed one’s own happiness as an ultimate end, for he believed egoism would still qualify as a rational procedure for determining what ought to be done even if the ultimate end relative to which one made these determinations was fixed by desire rather than reason. As long as the agent had adopted the end as ultimate and paramount, Sidgwick held, his determination of the right means to that end would still yield cognitions of the requisite sort. It would still yield cognitions that gave an impulse or motive to action that did not arise, at least not wholly, from any of the agent’s desires and so a fortiori did not arise, at least not wholly, from the desire that fixed his own happiness as his ultimate and paramount end. These cognitions must therefore, in Sidgwick’s view, be judgments about what one ought to do in the strict ethical sense of ‘ought’ he had isolated. They must, that is, be judgments that one makes by applying to the agent’s circumstances the fundamental notion of ethics, the notion that ‘ought’ in 13.  Methods, p. 8. See also p. 77 n. 1. 14. Ibid., p. 36.

881

188  From Psychology to Morality

this sense expresses. Sidgwick says as much in a remark that follows the one about egoism I just quoted. “But even if we discard the belief, that any end of action is unconditionally or ‘categorically’ prescribed by reason”, he wrote, “the notion ‘ought’ as above explained is not thereby eliminated from our practical reasonings:  it still remains in the ‘hypothetical imperative’ which prescribes the fittest means to any end we may have determined to aim at.”15 This view, however, leads Sidgwick into trouble. Consider a sentence of the form ‘A ought to φ’. When the occurrence of ‘ought’ in such a sentence is hypothetical, which is to say, relative to an end whose adoption is contingent on A’s desires, then the sentence expresses the judgment that A’s φing is advisable in the circumstances in view of A’s desires. By contrast, when the occurrence of ‘ought’ is categorical, which is to say, absolute and so without regard to any end fixed by A’s desires, the sentence expresses the judgment that φing is required of A  in these circumstances regardless of any of A’s desires. This contrast suggests that the sense ‘ought’ has when it occurs in such a sentence depends on whether its occurrence is hypothetical or categorical. That is, the sense it has when its occurrence is hypothetical is different from the one it has when its occurrence is categorical. And the idea that ‘ought’ has these two different senses goes directly against Sidgwick’s view. This idea should be familiar. It is at the nub of Prichard’s powerful critique of the central problem of Plato’s Republic.16 If I  promise to repay a loan by a certain date, Prichard observed, then I am bound by the obligation I have undertaken to repay it. In other words, though I may not want to repay it, though repaying it may be inconvenient or have unpleasant consequences, I ought to repay it nonetheless, and to say this is to say that I am required to repay it regardless of any desire that I  have to do otherwise. Of course, seeing the disadvantages and unpleasantness I would suffer by repaying the loan, I might ask myself whether I really ought to repay it. Doesn’t reneging make more sense? But to understand this question as one that I could genuinely ask even while knowing that I am bound by an obligation to repay the loan requires our taking the ‘ought’ in a different sense from the sense with which I would use it to express the judgment that I am bound to repay it. So too, then, when Glaucon and Adeimantus raise the question for Socrates of whether one ought always to act justly, their question, if we are to understand it as a genuine one, must be understood as arising when one considers 15. Ibid., p. 37. 16. H. A. Prichard, “Does Moral Philosophy Rest on a Mistake?,” Moral Obligation (Oxford, 1949), pp. 1–​17.

891

Sidgwick’s Conception of Ethics  189

the advantages of occasional injustice, or indeed even wholesale injustice, if one can get away with it. Thus we must take ‘ought’ as it occurs in our formulation of their question as expressing a different sense from the sense it expresses when we use it to declare that an act of justice is obligatory. We must take it, that is, to express the sense of its being advisable that I do a certain act rather than of my being bound to do it. Plato’s error, then, according to Prichard, consisted in his thinking that a question about what it is advisable to do in view of one’s desires could undermine the obligatoriness of justice, and it cannot. Whether or not one thinks Prichard’s criticism of Plato is sound, one can still see the merit of the distinction between the two senses of ‘ought’ on which he based the criticism. And it is the distinction that puts Sidgwick’s view into doubt. How could he hold, for instance, that ‘ought’, when used to formulate the question Glaucon and Adeimantus raised, expresses the same sense whether one takes its occurrence in that formulation as hypothetical or categorical? The answer follows from the character of the cognition that the question solicits on Sidgwick’s view. For the cognition the question solicits is one that is supposed to give an impulse or motive to action independently of any desire, so the impulse or motive it gives must be due to a recognition of something about the action in the circumstances under consideration such that the latter call for or require the former. Consequently, the sense ‘ought’ expresses in the question, whether one takes it hypothetically or categorically, is that of a requirement on action of this sort. On Sidgwick’s view, then, ‘ought’ expresses the sense of a requirement even when it is used merely instrumentally, that is, to raise a question about what it is advisable to do in view of one’s desires. Sidgwick, moreover, explained the requirement the sense of which ‘ought’, when used instrumentally, expresses by noting that the unreasonableness of adopting an end and refusing to adopt the necessary means to its attainment implies that reason demands that one adopt the necessary means to the attainment of an end one has already adopted. “Practical reason,” he declared, “condemns as irrational” the inconsistency of the will that refusing to adopt such means entails, and this condemnation implies that practical reason requires that one adopt such means.17 In other words, anyone who adopts an end but refuses to adopt the necessary means to its attainment acts unreasonably and, if unreasonably, then contrary to what reason requires.

17.  Methods, pp. 37–​38.

901

190  From Psychology to Morality

This argument, however, is unsound. It rests on the faulty assumption that every unreasonable act is one that violates a requirement of reason. Thus Sidgwick, seeing that someone would be acting unreasonably if she adopted some end but refused to adopt the necessary means to its attainment, introduces a requirement of reason to explain the person’s irrationality. But there is no need to introduce such a requirement to explain her irrationality. It is enough just to point out that the person is frustrating her own purposes. For to refuse to take the necessary means to the attainment of one’s ends is plainly to frustrate one’s purposes, and frustrating one’s purposes is a form of irrationality and not a kind of action whose irrationality one must explain by citing a requirement of reason that the agent is violating. “Don’t frustrate your own purposes” is not a requirement of reason. It is not something reason prescribes. Someone who has to be told not to frustrate her own purposes is someone who does not know what it means to have purposes and not someone who knows she has purposes but does not know what she is supposed to do with them. And similarly reason does not require that one take the necessary means to the attainment of one’s ends. There is no such requirement of reason. It is not something that reason prescribes. Perhaps, though, I have under-​described Sidgwick’s argument. It is possible that Sidgwick thought the demands for consistency that reason places on practical thinking supplied the requirement on action by virtue of which the instrumental use of ‘ought’ qualifies as an expression of the fundamental notion of ethics and accordingly meant to base his argument on this thought. This possibility is suggested at the conclusion of Sidgwick’s argument, when he writes: There is a[n]‌ . . . inconsistency between the adoption of an end and a general refusal to take whatever means we may see to be indispensable to its attainment; and if, when the time comes we do not take such means while yet we do not consciously retract our adoption of the end, it can hardly be denied that we ‘ought’ in consistency to act otherwise than we do.18

But even if this recourse to consistency is Sidgwick’s ultimate appeal, it still does not keep his argument from collapsing. For the demands of consistency that reason places on practical thinking cannot supply the requirement on action Sidgwick needs to show that the instrumental use of ‘ought’ expresses the fundamental notion of ethics. If they did, then the impulse or motive to action to which judgments in which ‘ought’ is used instrumentally give rise would have to come at least in part just from the thought that one

18. Ibid.

9 11

Sidgwick’s Conception of Ethics  191

would be doing something inconsistent in neither taking the indispensable means to an end one had adopted nor abandoning that end. And while there are no doubt some people in whom such thoughts are motivating, the motivation in their case is usually seen as evidence not of reason’s conative power but rather of an unfortunate compulsiveness about consistency. Most people, I venture to say, are not typically motivated by such thoughts. The inconsistency of failing to take the necessary means to their ends is not typically something that worries them. What worries them, if anything, is failing to get what they want. Sidgwick, it would seem, in his effort to establish the univocality of ‘ought’ as it occurs in the primary judgments of egoism as well as intuitionism and utilitarianism, lost sight of what actually moves people to do what they judge they ought to do when such judgments reflect their determination of the best means to attaining the object of their desires.19 We now have some confirmation of the deep incoherence hypothesis that I proposed earlier. Sidgwick’s argument for his conception of ethics as grounded in practical reason depends, for its coherence, on his being justified in assuming that every method of ethics applies the same notion of right or reasonable conduct when it is used to determine what an individual ought to do. This notion, if it exists, is what Sidgwick called ‘the fundamental notion’ of ethics,20 and accordingly, on Sidgwick’s view, when one uses any method of ethics to determine what an individual ought to do, ‘ought’ as it occurs in the judgments one makes from using that method expresses this notion. It has, in other words, the same sense no matter what method of ethics one uses. Sidgwick’s argument for his conception of ethics as grounded in practical reason therefore depends, for its coherence, on there being one and only one sense of ‘ought’, the strict ethical sense, in which the word occurs when it occurs in the judgments one makes from using some method of ethics to determine what an individual ought to do. Yet this condition, as we have seen, fails in the case of egoism once Sidgwick allows that egoism would qualify as a method of ethics even on the supposition that the adoption of one’s own happiness as one’s ultimate and paramount end was due to a predominant desire for happiness rather than a prescription of reason. For on this supposition the judgments of what an individual ought to do that one makes from using the method are judgments in which ‘ought’ occurs with a different sense from the sense it has in the judgments one makes from using a method 19.  For further discussion of the difficulties in Sidgwick’s account of the instrumental use of ‘ought’ see my “Sidgwick on Ethical Judgement,” Essays on Henry Sidgwick, ed. Bart Schultz (Cambridge, 1992), pp. 241–​258. 20.  For example, at Methods, p. 25.

921

192  From Psychology to Morality

in which the agent’s ultimate end is prescribed by reason or the ultimate criterion of rightness is conformance to a principle of duty. In either of the latter methods, the sense ‘ought’ has in the relevant judgments, judgments of the form ‘A ought to φ’, is the sense of A’s being required or bound to φ, whereas in the former the sense it has in the relevant judgments is that of its being advisable for A to φ in view of A’s desires. Sidgwick’s argument for his conception of ethics as grounded in practical reason thus contains at its base a confusion about the notion of right or reasonable conduct that egoism applies when it is understood as a method whose determinations of what ought to be done are judgments about the means to attaining an ultimate end whose adoption is due to a predominant desire rather than a prescription of reason. Of course, you may well ask, “What is the connection between this confusion and the incoherence in Sidgwick’s final account of ethics that the dualism of practical reason creates?” After all, the incoherence arises when egoism is understood as a method of ethics whose determinations of what ought to be done are judgments about the means to attaining an ultimate end that reason prescribes and therefore it arises when egoism applies the same notion of right or reasonable conduct as utilitarianism and intuitionism. Indeed, this understanding of egoism is plainly the one Sidgwick favored. So whatever confusion he fell into as a result of holding that egoism is still a method of ethics even when its determinations of what ought to be done are judgments about the means to attaining an ultimate end whose adoption is due to a desire, it does not explain the dualism of practical reason. Hence, it cannot be a deep source of the incoherence in Sidgwick’s final account of ethics that the dualism creates. What bearing, then, can it possibly have on the explanation of that incoherence? The bearing it has comes, I believe, not from any logical connection between it and the incoherence but rather from Sidgwick’s reasons for taking egoism as a method of ethics. Thus, while it is true that the notion of right action egoism applies, on Sidgwick’s favored understanding of the method, is the same as the notion of right action that intuitionism and utilitarianism apply, it is also true that Sidgwick thought his argument for taking egoism, on that understanding of it, as a method of ethics, which is to say, his argument for holding that reason prescribes one’s own happiness as an ultimate end, was less than compelling. Consequently, to justify his taking egoism as a method of ethics, he either had to give a stronger argument or show that egoism would still qualify as a method even if reason did not prescribe one’s own happiness as an ultimate end. And lacking a stronger argument, he took this second option. That is, he expounded egoism according to his favored understanding of it, not because he thought there was a compelling

913

Sidgwick’s Conception of Ethics  193

argument for doing so, but because he believed that even if the argument he gave was rejected, he could still justify his taking egoism as a method of ethics by falling back to the alternative understanding of it on which adoption of one’s own happiness as an ultimate end is due to a predominant desire rather than a prescription of reason. To fall back to this alternative, however, means giving an account of ethics on which the dualism of practical reason and the incoherence it creates not only remain in the study but are also the result of there being two distinct and possibly conflicting notions of right conduct that the study must recognize as fundamental. In other words, it means giving an account of ethics on which Sidgwick’s confusion of distinct notions of right conduct is the deep source of the incoherence the dualism creates. Sidgwick’s final account of ethics avoids this result, then, because his fallback position gives him latitude to present egoism as more closely aligned with intuitionism and utilitarianism than he has sound reason to hold. But because of this, one must see his presentation of egoism according to his favored understanding of it as developing in the shadow of the alternative understanding, and thus the reason we see no connection on his final account of ethics between his confusion and the incoherence the dualism creates is that we view the account as if it were standing in sunlight. At this point, my reading of Sidgwick’s Methods appears to have significantly departed from Schneewind’s. For an important part of Schneewind’s interpretation is that Sidgwick saw egoism as, like utilitarianism, a method of ethics that could complete and systematize commonsense morality. As such, the method must have, as its fundamental notion of right conduct, the same notion as the method whose product it completes and systematizes, and that, of course, is intuitionism. In addition, though, it must also be a method corresponding to which is a principle that can serve as a first principle from which the principles of duty according to commonsense morality follow. And since those principles, being the first principles of intuitionism, are intuitively self-​evident, any principle from which they follow must be, as their first principle, “supreme in validity as well as in certainty” and therefore intuitively self-​evident as well.21 Clearly, egoism, on Sidgwick’s favored understanding of it, satisfies these conditions, for if reason prescribes one’s own happiness as an ultimate and paramount end, then the prescription could serve as such a first principle. But on Sidgwick’s alternative understanding of egoism reason does not prescribe one’s own happiness as an ultimate end. As a result, it appears that on this understanding of the method there is no principle corresponding to it whose validity is intuitively self-​evident and that can therefore serve as a 21.  Sidgwick’s Ethics, p. 264.

941

194  From Psychology to Morality

first principle from which the principles of duty that commonsense morality contains follow. And if, as I read Sidgwick’s Methods, he expounded egoism according to his favored understanding of it in the shadow of this alternative understanding, then it is questionable whether he regarded it as a method that could serve to complete and systematize commonsense morality. To be sure, Sidgwick allows that one could also understand utilitarianism as a method of ethics whose ultimate end is fixed by a predominant desire rather than a prescription of reason. So one might think that his alternative understanding of egoism is no more of a difficulty for Schneewind’s reading of the Methods than the analogous understanding of utilitarianism that Sidgwick also acknowledged. And perhaps this is right. At the same time, one could read Sidgwick, in light of the place that his demonstration of how utilitarianism completes and systematizes commonsense morality has in the Methods, as regarding this analogous understanding of utilitarianism as a mere abstract possibility and not a genuine alternative. By contrast, he appears, in view of his argument for taking the merely instrumental ‘ought’ as also expressing the fundamental notion of ethics, to regard his alternative understanding of egoism as a genuine alternative and not a mere abstract possibility. Consequently, he could not argue for egoism as the method that completes and systematizes commonsense morality with the same certainty as he expresses in making that argument for utilitarianism. And indeed one could say, to bring out my departure from Schneewind’s reading of the Methods, that what indicates a deep incoherence in the work’s argument is its exposition of egoism as companionate with intuitionism. Be this as it may, even to see the possibility of a deep incoherence in the argument of the Methods that Sidgwick’s alternative understanding of egoism implies requires seeing, in the first place, how that argument brings significant coherence to the work. And that we can now see this argument and how it unifies the different parts of the work is due to Jerome Schneewind’s extraordinarily clarifying and pathbreaking book.

915

9

Moral Ideals

Philosophers now commonly write about morality where two centuries ago they wrote about morals. Consider these great works of eighteenth-​ century ethics:  Hume’s Enquiry Concerning the Principles of Morals, Kant’s Groundwork of the Metaphysics of Morals, Bentham’s Introduction to the Principles of Morals and Legislation, and Price’s Review of the Chief Questions and Difficulties of Morals. By comparison a list of significant works in ethics produced over the last thirty years includes Gewirth’s Reason and Morality, Harman’s The Nature of Morality, Geoffrey Warnock’s The Object of Morality, Donagan’s The Theory of Mora1ity, Kagan’s The Limits of Morality, and Bernard Gert’s book, now titled Morality: Its Nature and Justification.1 Whether this comparison shows some real change in the subject of moral philosophy over the course of two centuries or just a shift in terminological preferences is hard to say. While I am confident that David Gauthier’s Morals by Agreement deals with the very same subject as Alan Gewirth’s Reason and Morality, I am less confident about there being a similar identity of subject between either of these works and book III of Hume’s Treatise. Hume meant 1.  Bernard Gert, Morality: Its Nature and Justification (New York: Oxford University Press, 1998). All references will be to this work.

195

9 61

196  From Psychology to Morality

the study he pursued in the third book of his Treatise to be continuous with the studies he pursued in the first two books, and this is to say that he understood the study of morals to be continuous with the studies of the understanding and the passions. Whether one could understand the study of morality to be similarly continuous with those of the understanding and the passions seems to me a real question, and consequently that the subject of ethics may have changed from Hume’s time to ours a real possibility. Be this as it may, in our time morality is what moral philosophers chiefly study. The prevailing philosophical view of morality goes something like this. First, morality consists of rules and principles. These rules, these principles, govern the behavior of human beings in their relations with each other as members of a community. They specify both the actions no one may do to another and the actions everyone must do for others. They define, in other words, a set of prohibitions and requirements on conduct towards others. Second, morality is universal. Its rules, its principles, apply to everyone who has the powers of reason and self-​control necessary to understand and follow them. No one who has these powers is exempt from their governance. Consequently, either the rules of morality do not vary significantly from one human community to another, or human beings as such form a global community and morality governs their behavior in the relations they have with each other as members of this global community. Finally, the rules and principles of morality are distinct from the moral customs of any particular society. Indeed, they represent the standards moralists and social critics use to evaluate and criticize such customs. Moral customs, being among the settled habits and practices of the members of a society, are concrete phenomena properly studied by ethnologists and sociologists. Morality, by contrast, being an abstract system of rules that men and women can comprehend by reflecting on and reasoning about their relations with and conduct towards each other, is a paradigm object of philosophical study. The chief aims of this study are to formulate, systematize, and justify morality’s rules and principles. Many of the most important works in contemporary moral philosophy have been organized around these aims. Bernard Gert’s book is exemplary. Gert’s project in the book is to provide rational foundations for morality, and he mainly pursues it by formulating a list of ten rules, working out a systematic account of how they provide the basis for all moral prohibitions and requirements, and arguing for their justification as what reason endorses as the fundamental rules of morality. At the same time, Gert acknowledges that morality comprises more than the ten rules on his list and the procedures for determining when their violation is justified. Recognizing that some moral judgments about the goodness or badness of an

917

Moral Ideals 197

action are different from moral judgments about an action’s being required or prohibited, he has incorporated into his theory a set of moral ideals through recourse to which the theory accounts for these judgments.2 Accordingly, he takes moral ideals as forming a part of morality that is distinct from the part the moral rules form. They are similar to the moral rules in being precepts, but they differ from them in that they do not set requirements or prohibitions on conduct. In sum, on Gert’s theory, morality is a system of precepts that is divided into two parts. One part consists of the ten moral rules and is the basis for determining whether an action is morally required or prohibited. The other consists of moral ideals and accounts for morality’s concern with an action’s being good or bad apart from its being morally required or prohibited. The dominant focus of Gert’s book in its earliest editions was the ten moral rules. As a result, the book inspired criticism that it misrepresented a part of morality for the whole, that it shortchanged the subject by concentrating on the minimal demands of morality, its requirements and prohibitions, to the exclusion of morality’s loftier concerns with conduct of merit and worth.3 In later editions, to correct this misperception, Gert has paid increasingly greater attention to moral ideals. Though the major focus of his project continues to be the ten moral rules and their justification, a full understanding of his theory, he now makes clear, requires reflection on his account of moral ideals. This account is the subject of my paper. It represents an attractive view of how moral ideals are related to the requirements and prohibitions of morality. Nonetheless, I believe that it is seriously flawed. In the first and longest section of the paper, I critically examine Gert’s account with the aim of bringing out the flaws in its view of how moral ideals are related to moral prohibitions and requirements. In the paper’s second and much shorter section, I briefly offer a different account of moral ideals and a correspondingly different of view of this relation.

I Gert’s account of moral ideals proceeds from the distinctions he draws between them and moral rules. Both are moral precepts. But—​and this is the first distinction Gert draws—​ only moral rules set requirements and prohibitions on conduct.4 Moral ideals, by contrast, merely encourage certain 2.  See Gert, Morality, 318–​319, 322. Note that Gert does not take every judgment that an action is morally bad to be one made apart from the act’s being morally prohibited, but he does hold that some are. 3.  See Gert, Morality, preface, ix–​x. 4. Gert, Morality, 247. See also 109.

981

198  From Psychology to Morality

conduct. Accordingly, they support judgments applauding such conduct as morally good or deploring its omission as morally deficient, whereas moral rules, together with the procedures for determining when their violation is justified, support judgments condoning or sanctioning certain conduct as at least morally permissible and condemning its opposite as morally wrong. This first distinction between ideals and rules turns, then, on a difference in their force. I  will refer to it as Gert’s functional distinction. The second distinction he draws, which I will refer to as his substantive distinction, turns on a difference in their content. Moral ideals, Gert holds, call for the prevention of harm or evil. Moral rules, by contrast, call for the avoidance of causing harm or evil.5 What Gert has in mind in drawing this distinction is not clear on its face, however, especially since some acts of preventing harm are also acts of avoiding causing harm. A parent, for instance, who sterilizes a needle before using it to remove a splinter lodged in his child’s finger is preventing the finger from being infected and also avoiding causing an infection. Similarly, the signs “Prevent Forest Fires” that the government posts along country roads and mountain highways are meant to be a call to avoid causing forest fires.6 So one has to look further into Gert’s text to figure out what he intends in drawing his substantive distinction. The search will take us to the heart of his account. One indication of what Gert intends by this distinction comes from comparing examples of each kind of precept with examples of the other. Consider the following rules from Gert’s list. “Do not kill,” “Do not cause pain,” “Do not disable,” “Do not deprive of freedom,” “Do not deceive,” and “Do not cheat.” Each speaks against doing a certain action that either causes harm directly to someone or leaves someone vulnerable to suffering harm, and conforming to any of these rules means not doing that action. These points then apply to the other rules on the list, though their application to three of them, “Keep your promises,” “Obey the law,” and “Do your duty,” needs some explanation because the negative element is not evident in their formulation. The points do not apply, however, to moral ideals. Gert offers several examples of these as representative of the entire class.7 His examples

5. Gert, Morality, 126. 6.  Of course, one might think, in view of these and similar examples, that preventing harm is a kind of act one of whose species is avoiding causing harm. But Gert, however he understands the distinction between preventing harm and avoiding causing harm, does not think that the latter is a species of the former. For if it were, then violations of moral rules would frequently count as failures to act on a moral ideal, and it is clear that Gert sees violations of moral rules as importantly different from failures to act on moral ideals. Thus he writes (157), “Failing to act on a moral ideal never warrants punishment, whereas violating a moral rule often does.” 7. Gert, Morality, 248.

9 91

Moral Ideals 199

include “Preserve life,” “Relieve pain,” “Help the needy,” and “Prevent immoral behavior.” These ideals do not speak against doing a certain action. On the contrary, they direct one to engage in certain actions. This contrast suggests, then, that what Gert has in mind when he distinguishes moral rules from moral ideals according to whether their object is the avoidance of causing harm or the prevention of harm is a difference between a call for negative action and a call for positive action. The text, moreover, supports this suggestion. In one place Gert glosses this distinction by characterizing it as one between precepts conformity to which “[does] not generally require positive action but only the avoidance of certain kinds of action” and precepts conformity to which generally requires positive action.8 And similar remarks occur elsewhere.9 At the same time, Gert never explains how he understands the distinction between positive and negative action or what it is about the distinction he thinks supports his account. This is unfortunate, for the distinction is a well-​known snare.10 Plainly, he cannot understand it simply as a distinction between acts described positively and acts described negatively. Too many acts can be described either way. Thus, one can either say that S ignored G or that S did not pay attention to G; or again one can either say that S forgot G’s birthday or that S did not remember G’s birthday; or that S revealed his sources or that S did not keep his sources confidential, and so forth. Gert himself makes such an observation about promise keeping when he points out that keeping a promise and not breaking a promise are equivalent descriptions.11 So there is good reason to assume that he does not mean to be using the grammar of an act’s description as the criterion of whether the act is positive or negative. Perhaps the safest interpretation is that Gert intends the distinction commonly made in the law between acts that proceed from the initiation of bodily movement and acts to which no initiation of bodily movement is essential.12 Accordingly, yelling counts as a positive act since one cannot yell without moving one’s vocal chords, but keeping quiet counts as a negative act since one need not move any part of one’s body to keep quiet. Yet 8. Gert, Morality, 123 and 126. 9.  See, e.g., Gert, Morality, 127, 189, and 249. 10.  I  follow N.  Ann Davis’s discussion in her very perceptive and instructive essay “The Priority of Avoiding Harm,” in Killing and Letting Die, ed. Bonnie Steinbock (Englewood Cliffs, N.J.: Prentice Hall, 1980), 172–​214. 11. Gert, Morality, 188–​189. 12.  See Bentham, Introduction to the Principles of Morals and Legislation, ed. H. L.  A. Hart and J. H. Burns (Oxford, Eng.: Oxford University Press, 1979), 75–​76 (ch. 7, pars. 8 & 9).

02

200  From Psychology to Morality

Gert’s substantive distinction between moral ideals and moral rules is just as untenable on this interpretation as on the last. For there are plenty of circumstances in which conformity to a moral rule requires moving one’s body. For instance, if some buck private were to go several weeks without showering or bathing, he would cause a good deal of visceral distress to his mates in the barracks in which he bunks. Hence, in these circumstances, his conforming to the moral rule “Do not cause pain” requires his showering or bathing. Similarly, suppose the same circumstances and that the private has contracted some highly contagious disease. Because he would threaten the health of his mates if he remained in the barracks, conformity to the rule “Do not disable” requires that he get himself to the infirmary. Other examples of this sort are easily generated. Indeed, the literature on killing and letting die is now replete with examples of homicide by omission, which is to say, of circumstances in which conformity to the rule “Do not kill” requires positive action. Nor are these examples mere philosophical fictions. The New York Times recently reported on a tragic case of a young mother whom the Bronx District Attorney is prosecuting for manslaughter on the grounds that she starved her infant son.13 Clearly, then, we cannot safely assume that Gert understands the distinction between positive and negative action in the way it is commonly made in the law. One might hope to mine Gert’s text for evidence of how he understands the distinction. But unfortunately this evidence is mixed. His several references to positive action are not consistent in their meaning. For example, in a passage reaffirming that moral rules, given their negative formulation, do not demand positive action, Gert illustrates this proposition by pointing out that “The rule prohibiting depriving of freedom can be obeyed by doing nothing.”14 This point suggests that he takes the initiation of bodily movement to be the criterion of positive action. But elsewhere it appears he does not. Thus, in the passage quoted earlier in which he opposes positive action to the avoidance of certain kinds of action, it would seem he has a different criterion in mind.15 For it is obvious that sometimes one must initiate some movement of one’s body to avoid doing certain actions, as when I must swerve to avoid hitting a pedestrian or change the subject to avoid causing embarrassment. And in a later passage one finds Gert using the grammar of an action’s description as the criterion of positive action. Thus, in discussing whether the 13.  Nina Bernstein, “Placing the Blame in an Infant’s Death:  Mother Faces Trial After Baby Dies from Lack of Breast Milk,” New York Times, 15 March 1999, l(B). 14. Gert, Morality, 189. 15. Gert, Morality, 123.

201

Moral Ideals 201

moral rules admit of positive formulations, he writes, “[The moral rules] do not require positive action, except in those cases where there is no difference between requiring action and prohibiting action. There is no difference, except in style, between saying ‘Keep your promises’ and ‘Don’t break your promises.’ ”16 Here Gert must be using the grammatical criterion, for otherwise one could not make sense of his conceding that a moral rule can be said to require positive action when it admits of a positive formulation. In all of these passages the uncertainty about what Gert intends by the distinction arises from difficulties in understanding his view that moral rules do not require positive action. It does not, however, arise only from these difficulties. Similar difficulties also attend his view that moral ideals require positive action. Take his first example of a moral ideal, “Preserve life.” To preserve something is to keep it safe from deterioration and destruction. To preserve life then is to keep it safe from these effects. “Do not let life deteriorate or be destroyed” is therefore an equivalent formulation of this ideal. Or in other words, one cannot say that the ideal requires positive as opposed to negative action if the criterion of whether an action is positive or negative is the grammar of its description. Nor would switching to the criterion common in the law improve the prospects for Gert’s view. Preserving something, after all, does not always require action that proceeds from initiating a bodily movement. As my wife and I have learned over the years, the best way to preserve our good china is not to use it. In sum, unless Gert has some other criterion in mind for distinguishing positive from negative action, we would seem forced to conclude that the distinction between these kinds of action is useless to understanding what he means by his substantive distinction between moral ideals and moral rules. It is possible, though, that Gert does have some other criterion in mind. He sometimes observes in support of his substantive distinction that because conformity to a moral ideal requires positive action, one cannot follow that ideal all of the time.17 He takes this observation to support his substantive distinction because, as he also observes, one can follow any moral rule all of the time since none requires positive action.18 It is a cardinal point of Gert’s account of moral ideals that, unlike moral rules, one can sensibly ask how much time one should devote to following them, and it is evident that he sees this point being reaffirmed in these observations.19 So, evidently, he sees 16. Gert, Morality, 189. 17.  See Gert, Morality, e.g., 126 and 249. 18. Gert, Morality, 122 and 126. 19. Gert, Morality, 122–​123.

0 2

202  From Psychology to Morality

some connection between an action’s being positive and its being such that one cannot always be following a precept conformity to which requires it. Of course, he may make this connection because he is applying the criterion for distinguishing positive from negative action common in the law. But his point, in that case, would be mistaken. The necessity of initiating movements of one’s body to conform to a moral rule could, if one’s circumstances were sufficiently strange, be as great as the necessity of initiating such movements to conform to a moral ideal. There is, however, a different explanation of why he makes the connection, and from this explanation one can extract a criterion for distinguishing positive from negative action that renders Gert’s substantive distinction meaningful. The explanation is this. The reason why it makes sense to ask how much time one should devote to following moral ideals is that the ideals prescribe ends, and people cannot be expected to live their lives exclusively and always in pursuit of these ends. By the same token, the reason why it shows confusion to ask how much time one should devote to following the moral rules, the rules on Gert’s list, is that these rules do not prescribe ends. Rather they set constraints on the pursuit of whatever ends people adopt, and everyone can be expected always to live within these constraints. Thus Gert makes the connection between an action’s being positive and its being such that one cannot always follow a precept conformity to which requires it, because he sees that no one can be expected always to act in pursuit of the same end. On this explanation, then, what Gert means by a positive action is an action with a certain end.20 Accordingly, on Gert’s account, moral ideals require positive action because each calls for the promotion of a certain end, whereas moral rules do not require such action because none of them calls for the promotion of some end. Gert, in other words, intends his substantive distinction to capture the difference between precepts that call for pursuing a certain end—​the lessening of evil—​and precepts that call for avoiding, in whatever pursuits one undertakes, certain consequences—​the instantiations of harm and evil. Hence, on this latest criterion of positive action, his thesis that the object of moral ideals is the prevention of evil, the object of moral rules the avoidance of causing evil is explicable. 20.  That this explanation yields a separate criterion of positive action from the other two should be clear. For example, on the criterion it yields, my growing a beard would be a positive action; my letting my beard grow would not. The two actions would thus be classified differently on this criterion. By contrast, on both the grammatical criterion and the criterion common in the law, they would not be. On the former, both my growing a beard and my letting my beard grow would count as positive actions, whereas on the latter neither would count as a positive action.

0 23

Moral Ideals 203

To be sure, Gert’s equivocal use of the term ‘positive action’ makes any interpretation of what he means by it less than certain. The passages cited above are evidence that sometimes he uses it with the grammatical criterion in mind and sometimes with the criterion common in the law in mind. Nonetheless, there is decisive reason to interpret him as having this latest criterion in mind when he draws his substantive distinction, for unlike the other two criteria, this criterion renders the distinction unmysterious. Accordingly, let us assume that Gert intends the distinction to be one between precepts that call for the promotion of certain ends and precepts that do not. The assumption then allows us to move forward. Gert’s account of moral ideals results from his combining his functional distinction with his substantive distinction, and we can now examine whether this combination is sound. The combination yields a thesis about the place of the promotion of ends in morality. The thesis is that morality merely encourages the promotion of some ends. This thesis is in fact a conjunction of two propositions, first that morality encourages the promotion of some ends, and second that there is no end the promotion of which morality requires or indifference to which morality prohibits. The first proposition, which follows from Gert’s functional characterization of ideals as encouragements and his substantive characterization of them as precepts calling for the promotion of some end, is uncontroversial. The second, however, which follows from Gert’s functional characterization of moral rules as setting requirements and prohibitions on conduct, his substantive characterization of them as precepts that do not call for the promotion of any end, and his assumption that each of his distinctions divides the universe of moral precepts into the same disjoint classes, is not uncontroversial. Many philosophers would reject it. For many philosophers believe that there is a general duty of benevolence towards humankind and that this duty amounts to a moral requirement that one take the good of human beings as an end and act accordingly. In other words, they reject Gert’s assumption that no moral precepts setting requirements and prohibitions call for the promotion of some end. They reject his assumption that his functional distinction divides the universe of moral precepts into the same disjoint classes as does his substantive distinction, that the two distinctions are isomorphic to each other. Gert, it is worth noting, rejects the possibility of a general duty of benevolence towards humankind when he rejects Kant’s distinction between perfect and imperfect duties.21 The general duty of benevolence, in Kant’s ethics, is one of two main imperfect duties, and what distinguishes them from perfect 21. Gert, Morality, 16.

0 42

204  From Psychology to Morality

duties is that they require the adoption of an end, whereas perfect duties require the conformity of conduct to an external standard. Gert regards his distinction between moral rules and moral ideals as an improvement over Kant’s distinction, and the reason he does is now clear: he believes there are no moral requirements to adopt certain ends. Call this the Gert-​contra-​Kant belief, or the GCK for short. What needs to be examined then are Gert’s arguments for it. These are the arguments on which he bases his assumption that his two distinctions, the functional and the substantive, are isomorphic to each other. He gives three arguments for the GCK. Two are implicit in the arguments he makes for his substantive distinction. We will examine them shortly. A third appeals to the necessity of understanding moral rules as precepts enforced by punishment. We can deal with it immediately, for it cannot be sound if neither of the other two is sound. Examining them, in other words, is sufficient for our purposes. So we can set his third argument aside once we see how its fate is tied to that of the other two. As a convenience, I will continue to refer to it as Gert’s third argument for the GCK and will refer to other two as Gert’s first and second arguments for this belief. The major premiss of his third argument is that moral rules are those precepts and only those precepts “toward which all impartial rational persons adopt the appropriate moral attitude,” where the appropriate moral attitude towards a precept includes favoring making unjustified violations of it punishable.22 From this premiss Gert argues that, since impartial rational persons would not favor making any violation of a precept that called for the promotion of some end punishable, such a precept cannot be one towards which all impartial rational persons would adopt the appropriate moral attitude. Hence, it cannot be a moral rule. And the reason Gert gives to explain why impartial rational persons would not favor making violations of such a precept punishable is that any distinction they tried to draw between justified and unjustified violations of it would be arbitrary, owing both to the impossibility of complying with the precept all of the time and to the impossibility of complying with it impartially with respect to everyone.23 These last two 22. Gert, Morality, 180. See also 126–​127. 23. Gert, Morality, 181. Using “Promote pleasure” as an example of a rule towards which impartial rational persons could not take the appropriate moral attitude, Gert’s argument, verbatim, is Unlike the . . . moral rules, this rule cannot possibly be obeyed all of the time. Nor is it likely that it can ever be impartially obeyed with regard to everyone. Thus an impartial rational person would either have to adopt the attitude that everyone be publicly

0 25

Moral Ideals 205

points, that neither constant nor universally impartial compliance with such a precept is possible, are also premisses of Gert’s first and second arguments for the GCK. Consequently, by virtue of these shared premisses, the fate of his third argument for the GCK is tied to the fate of the other two. Let us then turn directly to examining them. We are already acquainted with one of them. It occurs when Gert argues for his substantive distinction on the basis of the two observations, (a) that no precept calling for positive action can be followed all of the time and (b) that every moral rule can be followed all of the time. For, given our assumption that in this argument Gert takes ‘positive action’ to mean action with a certain end, (a) means that no precept calling for the promotion of a certain end can be followed all of the time, and given Gert’s functional distinction, (b) implies that one can always act such as to be in compliance with every moral requirement. Hence, the thrust of this argument, Gert’s first argument for the GCK, is that no moral precept can set a requirement to promote some end. Or, put simply, there are no moral requirements to promote certain ends. This is a puzzling argument. The conclusion appears to follow from the premisses, yet there are evident counterinstances. After all, no one expects the parents of small children to be advancing their children’s interests all of the time. Even the parents of small children are allowed some time to themselves and allowed on some occasions to put their interests ahead of their children’s. Yet it doesn’t follow that they are not morally required to promote their children’s welfare. Similarly, if I promise to take care of your cat while you are away for a week, you certainly wouldn’t expect me to spend every waking minute of that week attending to your cat’s welfare. Yet it doesn’t follow that I am not morally required to take care of your cat. Indeed, even allowed to violate the promote-​pleasure rule whenever and with regard to whomever he feels like doing so, or else to hold that everyone is liable to punishment all of the time. To do the former would make it pointless to adopt the moral attitude toward the rule; to do the latter would be to increase everyone’s chances of suffering evil. I conclude that no impartial rational person would adopt the moral attitude toward the rule ‘Promote pleasure.’ I have taken the liberty of interpreting this argument as one in which Gert is appealing to the arbitrariness of distinguishing between justified and unjustified violations of the promote-​ pleasure rule to explain why its being impossible to obey the rule all of the time and impartially with respect to everyone are grounds on which impartial rational persons would not favor making violations of the rule punishable. Nothing in what follows, however, depends on the soundness of this way of interpreting Gert’s argument. The important thing is that Gert takes the impossibility of constant and universally impartial compliance with a rule as the grounds on which impartial rational persons wou1d not favor making violations of the rule punishable, and that he does so is plain from the passage quoted in this footnote.

6 0 2

206  From Psychology to Morality

on Gert’s theory, parents are morally required to promote their children’s welfare, given the usual custodial duties parents have towards their children and the moral rule “Do your duty,” and likewise I would be morally required to take care of your cat, given the promise I made and the moral rule “Keep your promises.” Why Gert should deny that morality ever directly requires the promotion of some end when at the same time he allows that it can indirectly require the promotion of ends in virtue of these moral rules and the possibility of duties and promises to promote certain ends is a puzzle in itself.24 But the more pressing puzzle is why this first argument for the GCK appears cogent when applied to general precepts calling for the promotion of some end yet clearly fails in cases of specific duties and promises like the duties of parents to look after their children and the promises of friends to take care of their friends’ pets. To resolve this puzzle we need to look more closely at such duties and promises. A  duty to promote someone’s good or a promise to look after someone requires one to engage in action with a certain end, but it does not require that one constantly engage in that action. Rather what is required to fulfill the duty or promise is that one take as one’s end the good of the person who is its beneficiary and act in furtherance of that person’s good in all circumstances in which failure to do so would convict one of not having taken it as one’s end. The point is that to take something as an end is an inner state or condition whose existence does not depend on constant outward expression or manifestation. To take someone’s good as one’s end is to be disposed to be guided by that person’s good in one’s deliberations and decisions, and though there will certainly be circumstances in which one could not omit exercising this disposition without having lost it, one need not be constantly exercising it to continue to have it. Consequently, even though there are times when one does not promote the person’s good, one may still have his good as one’s end, just as I may do things unintended to promote

24.  Gert could undo this inconsistency by reversing himself on the question of whether keeping a promise and not breaking a promise, fulfilling a duty and not violating a duty, are equivalent descriptions and by reformulating the rules about promises and duties to read “Do not violate your duties” and “Do not break your promises.” For he could then say, using “positive action” in the sense we are assuming he means, “One can follow these rules all of the time, even at those times when one is not engaged in the positive actions necessary to fulfilling some duty one has undertaken or keeping some promise one has made.” But this inconsistency in his views is not the problem. It is just a symptom. The problem is that parents are morally required to look after their children, a promise to promote an end does create a moral requirement to promote the end, and Gert cannot affirm either of these propositions without abandoning one or the other premiss of the argument we are considering.

0 2 7

Moral Ideals 207

my health, such as driving to work or even indulging my taste for fried food, without having abandoned my health as an end. Hence, Gert’s point that one can always act such as to be in compliance with every moral requirement, whatever its merits, does not imply that there are no moral requirements to promote some end. For one does not need to be engaged in an action promotive of some end to be in compliance with a requirement to promote it. On many and perhaps most occasions, one’s having adopted the end and thereby become disposed to be guided by it suffices for meeting the requirement. Gert’s first argument for the GCK therefore appears cogent only when one loses sight of this fact about such requirements. Gert’s second argument for the GCK is similar to the first. It too is implicit in an argument he makes for his substantive distinction. Moral ideals, Gert argues, are distinct from moral rules because, as precepts calling for positive action, one cannot follow them impartially. That is, because a moral ideal is a precept calling for positive action, one cannot, in following it, act equally towards everyone whose interests the ideal protects. By contrast, moral rules, being precepts that do not call for positive action, can be followed impartially. When one follows a moral rule, one always acts equally towards everyone whose interests the rule protects.25 From these observations we can then obtain an argument for the GCK in the same way as we obtained Gert’s first argument for it from parallel observations about the impossibility of following a moral ideal all of the time and the possibility of always following a moral rule. The actual construction of the argument I leave as an exercise for the reader. This second argument, it should be clear, is no more cogent than the first. Its conclusion too has evident counterinstances. The executive director of UNICEF,26 for instance, who no doubt has a duty to promote the welfare of the world’s children, cannot possibly advance the interests of every child in the world equally. Yet it does not follow that she is not morally required to promote the welfare of the world’s children. Similarly, if I promise to take care of your ten children for a day, I cannot possibly look after all of them equally. Some are bound to get more of my attention than others. But it hardly follows that I  am not morally required to look after your children. The second argument thus goes wrong in the same way as the first. It too misrepresents what counts as meeting a requirement to promote some end. For what the duty the executive director of UNICEF has and the promise I make require is the promotion of the good of all who belong to a specific group, and one meets such a requirement if one adopts their good as one’s 25. Gert, Morality, 123–​124. 26.  Carol Bellamy.

0 82

208  From Psychology to Morality

end and acts in furtherance of it in every circumstance in which failure to do so would convict one of not having taken it as one’s end. Hence, one can meet the requirement even though one does not act equally towards every member of the group in advancing their good, since a disparity in the way one treats the members does not in itself convict one of not having adopted their good as one’s end. This point parallels the point on which the first argument founders. The latter, recall, is that one need not constantly be engaged in action that contributes to someone’s good to have adopted the person’s good as one’s end. One can, in other words, omit contributing to someone’s good on Tuesday, say, having contributed to it on Monday, without thereby showing indifference to the person’s good. Such disparity over a period of time in one’s conduct towards someone is consistent with constancy throughout the same period in one’s attachment to his good. Similarly, then, one need not contribute to the good of each member of some group, or contribute as much to each member’s good as to every other’s, to have adopted the good of all as one’s end and be equally well disposed towards each. One can, in other words, omit contributing to the good of some members, having contributed to the good of others, or contribute to the good of some less than one has contributed to the good of others, without thereby showing that one is indifferent to the good of some members or not as well disposed towards them as one is towards others. Such disparity across the group in one’s conduct towards its members is consistent with one’s being uniformly attached to the good of all. Thus Gert’s claim that every moral requirement can be complied with impartially is no reason to exclude requirements to promote some end from the class of moral requirements. For one need not contribute equally to the good of all who benefit from a requirement to promote some end in order to have complied impartially with that requirement. It is sufficient to have adopted their good as one’s end and to be equally well disposed towards each as a result. The temptation to think otherwise comes from thinking that any disparity in one’s treatment of the members of a group whose good one is required to promote must be due to some bias or prejudice favoring some members and disfavoring others. But this need not be so. One may simply have greater opportunities to contribute to the good of some than of others, and taking advantage of these opportunities is consistent with being attached to the good of all who belong to the group and equally well disposed towards each. A teacher who has several students all of whom are working on separate projects is required to help them advance their projects. At the same time, though, she may have more opportunities to help some than others just

9 0 2

Moral Ideals 209

because some have remained nearby and some have moved away. Yet one shouldn’t suppose that the greater time she therefore spends helping those who live nearby betrays a bias in their favor or a weaker commitment to the success of the projects of those who have moved away. As long as the teacher continues to take as an end her students’ success in their projects and education and to be equally well disposed in this regard to each, she will have complied impartially with the requirement to advance her students’ projects. To be sure, the question of when one’s actions convict one of not being attached to someone’s good or the good of all who belong to some group is not always cut-​and-​dried. A good example of this uncertainty can be seen in the controversy, currently popular in moral philosophy, over whether those of us who can save starving children in faraway lands by making donations to international relief organizations are morally required to do so, given that we certainly would be morally required to save a drowning child in a shallow pool whom we could save by wading into the pool and lifting him out of danger. Some philosophers argue that we are morally required to donate to international relief organizations, since we would be morally required to save the drowning child and there is no morally relevant difference between the circumstances in the two cases.27 Others, however, resist this conclusion, for they regard it as contrary to common sense about what morality requires. Like Gert, they hold that sending donations to international relief organizations is something morality encourages but does not require. While the issue in dispute here is complex and no doubt open to several interpretations, one of them is whether meeting the general duty of benevolence towards humankind is consistent with one’s not sending money overseas to save starving children, given that it is inconsistent with one’s doing nothing to save a drowning child in a shallow pool should one happen upon a child in such trouble. And on this interpretation one can see sufficient reason to resist the conclusion of the philosophers who hold that the moral necessity of saving the drowning child implies the moral necessity of sending money overseas to save starving children. The reason lies in the difference in how certain we can be in each case that the omission of life-​saving action is a sign of the actor’s indifference to the good of human beings. In the case of the drowning child, declining to wade into the pool and lift the child out of danger is a sure sign of such indifference, whereas in the case of the starving children, declining to send money overseas for their relief need not be. Its significance in this regard is much 27.  Peter Singer, Practical Ethics (Cambridge, Eng.: Cambridge University Press, 1979), 168–​171.

021

210  From Psychology to Morality

more open to doubt. For one thing, a person who had already sent money overseas to save starving children, say within the last month, might then decline to send money for this purpose without thereby showing indifference to the good of human beings, whereas letting a child drown whom one could save would be a sign of such indifference whether or not one had saved a child in similar danger within the last month. For another, a person who donates to local efforts to help the poor or the sufferers of some dread disease like AIDS might decline to send money overseas without thereby showing indifference to the good of human beings, whereas one could not, without showing such indifference, ignore a drowning child because one was too busy building houses for the homeless. Generally, then, we cannot be as certain about someone’s declining to send money overseas to save starving children, that it is a sign of deficient attachment to the good of human beings, as we can about someone’s declining to lift a drowning child out of a pool, for whether the latter is such a sign depends much less than whether the former is on the context of the omission and on the actor’s prior and ongoing efforts at promoting human welfare. It admits of determinate and firm answers, whereas the question of the former’s significance, in all but extreme cases of selfishness, clannishness, and chauvinism, does not. These last points reinforce our conclusion that Gert’s first and second arguments for the GCK fail. What is more, they remove the worry lying behind these arguments that any general moral requirement to promote the good of human beings places impossible demands on the time and resources of moral agents and entails the impossible task of involving them impartially in the lives of billions of people. They make clear that meeting such a requirement, at least if it is of the sort that Kant and others identify as the duty of benevolence towards humankind, need not entail an unreasonable expense of time and resources or excessive involvement in the lives of strangers. Hence, while there may still be difficulties in taking as a general requirement of morality a requirement of this sort, while there may, in particular, be difficulties in Kant’s understanding of it as an imperfect duty, neither of Gert’s first two arguments for the GCK has uncovered any. A duty of benevolence like the one Kant affirms would therefore appear to be one plausible form of the moral requirement to promote the good of human beings that escapes these arguments. And in the absence of some new and cogent objection to taking it as a general requirement of morality, Gert would appear to have no solid grounds for his denying that morality includes any such requirement. What of Gert’s third argument for the GCK? Could it be the source of some new and cogent objection? To be sure, the argument is unsound just as his first and second arguments for the GCK are. It too includes among its

21

Moral Ideals 211

premisses Gert’s faulty claims about the impossibility of constant and universally impartial compliance with precepts requiring the promotion of some end. But though the argument is unsound just as they are, the worry lying behind it is different from the worry that lies behind them. This worry, the one lying behind the third argument, is that such precepts cannot be fairly or usefully enforced by punishment, and unlike the other worry, it is completely separable from the faulty claims on which the argument it lies behind are based. Consequently, even though these claims make that argument unsound, the worry that lies behind it might still be a source of a cogent objection to taking a duty of benevolence like the one Kant affirms as a general requirement of morality. So we ought to consider whether Gert could give a different argument that substantiates that worry. Specifically, since the argument Gert gives to substantiate it consists in his presenting reasons why impartial rational persons would not favor making unjustified violations of precepts that require the promotion of some end punishable, we ought to consider whether Gert could present different reasons explaining why impartial rational persons would not favor making unjustified violations of such precepts punishable. If he could, then these reasons would constitute a new and different argument for the GCK. He would then have produced a new objection to taking as a general moral requirement the duty of benevolence that Kant and others affirm. This possibility may seem promising as a way of restoring cogency to Gert’s view. Gert does, after all, treat the attitude impartial rational persons would take towards unjustified violations of moral requirements as independently important to understanding the nature of those requirements. And a new argument for the GCK that drew on that attitude without relying on the faulty premisses of the third argument would represent this independent importance Gert assigns to his account of the attitude.28 But though this possibility may seem promising, it cannot succeed. It certainly could not succeed if Gert held that the attitude impartial rational persons took towards unjustified violations of moral requirements was one of favoring punishment for every such violation. To favor punishment for every unjustified lie and broken promise and for every petty utterance of cruel words would be to favor a cure for these wrongs that was far worse

28.  See Gert, Morality, 126–​127. Cf. 180–​181, where Gert does not as clearly treat his account of the attitude as being important to understanding the nature of moral requirements independently of his functional and substantive distinctions but does suggest its independent importance in citing the attitude to explain why such contrived precepts as “Offer to prevent pain for the first person you see each day” are not moral rules.

21

212  From Psychology to Morality

than the disease. What Gert must and does hold is that the attitude is one of favoring punishment of unjustified violations where the cost of imposing the punishment is not greater than the benefit of its deterrent effect.29 This by contrast is a sensible attitude to attribute to impartial rational persons, so sensible, in fact, that it is hard to fathom what reasons Gert could give to show that such persons would decline to take it towards unjustified violations of requirements to promote a morally significant end such as the good of human beings. The chief difficulty, in any case, is that whatever new argument for the GCK Gert constructed from such reasons would be subject to the same counterinstances that proved fatal to Gert’s first and second arguments. We undoubtedly think, for instance, that parents of small children, given the usual custodial duties such parents have, should be subject to some threats of punishment for neglecting their children’s welfare, and likewise think that people who promise to look after their friends’ children, given the obligations such promises create, should be subject to threats of punishment were they to neglect the welfare of these children once they were put in their care. We think, in other words, that such duties and promises entail requirements to promote their beneficiaries’ welfare and that unjustified violations of them should not be immune to punishment. No reasons Gert could give purporting to show how impartial rational persons could think otherwise would even be credible. And the same holds for reasons purporting to show how such persons could favor making immune to punishment unjustified violations of a requirement to promote the good of human beings like the one Kant and others identify as the duty of benevolence. To be sure, we can assume that impartial rational persons would, as a result of balancing the costs of punishment against the benefits of its deterrent effect in each case, favor stricter enforcement of the former requirements than the latter. But whatever the results of these calculations, the point is that there is nothing in the latter to distinguish it from the former as a requirement whose enforcement by threats of punishment impartial rational persons would, as a matter of principle, not favor. The duty of benevolence that Kant and others affirm represents, then, a general moral requirement calling for positive action that confounds Gert’s arguments against the possibility of such requirements. His actual arguments fail as a result of his misconstruing what compliance with such requirements involves when they are requirements to promote some end.

29.  See Gert, Morality, 182.

231

Moral Ideals 213

And even possible arguments suggested but not developed in the text that avoid this misconstruction fall short. The impact of this result on his account of moral ideals is immediate. The account follows directly from his combining his functional and substantive distinctions between moral rules and moral ideals, and his combination of the two distinctions depends on his assumption of an isomorphism between them. Since the possibility of a general moral requirement to promote some end is incompatible with the assumption, the failure of Gert’s arguments against this possibility means that the assumption on which his account depends is unwarranted. And insofar as the criticisms showing that failure make a strong case for including among the general requirements of morality a requirement to promote the good of human beings like the one Kant and others identify as the duty of benevolence, there is little hope for supporting the assumption by some other argument. The end result of these criticisms, therefore, is that Gert’s account of moral ideals cannot be sustained. For without the assumption on which it depends, Gert must either abandon his substantive distinction or abandon his functional one. That is, he must either allow that some moral rules call for positive action or allow that some moral ideals require action and do not merely encourage it. And neither is an option Gert can embrace. Plainly, he cannot embrace the second option, for if he did, he could no longer oppose moral ideals to moral requirements and prohibitions. Specifically, he could no longer point to moral ideals as accounting for moral judgments of good and bad action that are distinct from moral judgments of required and prohibited action. But equally clearly he cannot embrace the first option, for if he did, he would be hard pressed to find a precept that counted as a moral ideal rather than a moral rule. After all, nothing in his theory, apart from the substantive distinction, explains why “Preserve life,” “Relieve suffering,” “Help the needy,” and so on are moral ideals and not moral rules. Hence, once the theory abandoned this distinction and allowed that some moral rules call for positive action, it would need some basis for explaining why not all precepts calling for action that prevents evil are moral rules, that is, why some are moral ideals, and it is hard to see what that basis could be. Indeed, as Gert himself recognizes, once he admitted as a moral rule a precept that called for the promotion of the good of human beings, all of the precepts he counts as moral ideals would be consequences of this rule. Thus, short of discovering some wholly new and cogent objection to the possibility of moral requirements that call for positive action, there is no saving Gert’s account of moral ideals and their relation to moral requirements and prohibitions.

421

214  From Psychology to Morality

II Gert takes moral ideals to be precepts. On his account, they belong to the same category of standards as moral rules. They are standards one complies with or obeys. There is a different way to understand moral ideals, however. One can take them, instead, as models of conduct. Accordingly, they belong to a different category of standards from rules. Models of conduct are standards one lives up to or realizes. The difference between the two categories is evident both in their modes of expression and in their modes of guidance. Where precepts take the form of an imperative, models are presented in images, pictures, and portraits. Where precepts guide by explicit direction, models guide by inviting emulation or imitation. As children, for instance, we learn that standing during the playing of the national anthem at community events is the proper thing to do. Typically, we learn this both by being directed to stand by our parents or other authority figures when the national anthem is about to be played and by seeing our parents and these other figures stand during its playing and following their lead. In the one case we learn by prescription; in the other we learn by example. Having found serious flaws in the initially attractive account of moral ideals that Gert expounds, the account on which they are understood as precepts, let us briefly consider an alternative account on which they are understood as models of conduct.30 As my observation about learning suggests, we get our ideals from examples that people whom we look up to or admire set. These will not only be people, such as our parents and teachers, whom we know directly but also people whom we see from afar and learn about through the news or whom we meet in literature, history, theater, and film. The latter point is well illustrated in philosophy by the portrayal of Socrates in Plato’s early dialogues. Countless readers of these dialogues have taken Socrates’s conversations and activities, as portrayed by Plato, both as a model of how one conducts a philosophical inquiry and a model of intellectual and moral integrity. Admiring Socrates’s refusal to profess knowledge he does not have and his unwavering commitments to truth and right action, these readers have found in the early dialogues an ideal of truth seeking and justice that has captured their imagination and favor. Plato’s portrayal is perhaps a sufficient illustration of my point, but for the sake of variety let me offer two others. One is Thoreau’s description of his two-​year sojourn at Walden Pond. His experiment there

30.  On the notion of ideals as models of conduct presented in images and pictures, see P. F. Strawson, “Social Morality and Individual Ideal,” Philosophy 36 (1961): 1–​17.

251

Moral Ideals 215

to show that “a man is rich in proportion to the number of things he can afford to let alone”31 is our foremost ideal of a life of solitude lived close to nature and without dependence on material wealth. The other is the figure of Mohandas Gandhi, whose leadership in the campaigns against the injustice of British colonial rule, first in South Africa and then in India, inspired many to adopt his methods of nonviolent struggle against tyranny and to model their conduct in such struggles on his. Gandhi, both through his actions and his writings, thus conveyed to many disciples and admirers powerful ideals of a life dedicated to advancing justice through radical social reform and without violence or malice towards one’s opponents and of the courage, patience, and self-​discipline necessary to persevering in such a program. There are of course a great many other ideals that guide people in the conduct of their lives:  ideals of athletic prowess, artistic achievement, romantic love, family life, triumph over the elements of nature, and so on. Plainly, it would be a mistake to construe all of them as moral ideals. So the question is how one should distinguish these from all the others. One possibility is to suppose that they are ideals that have a certain content. This way of distinguishing them will recommend itself to anyone who, like Gert, holds that morality has a definite point. Thus, if like Gert you believe that the point of morality is to lessen harm and suffering, then it would be reasonable for you to define moral ideals as the ideals whose realization tends to have this result. Another possibility, though, is to suppose that moral ideals are ideals that bear a certain formal or functional relation to moral requirements and prohibitions. This way of distinguishing them will recommend itself to anyone who either holds that morality has no definite point or does not have a view on the matter. At the same time, it would not preclude anyone from distinguishing moral ideals by their content since they would, by virtue of being formally or functionally related to moral requirements and prohibitions, inherit whatever content was distinctive of those requirements and prohibitions. Hence, this second way of distinguishing moral ideals has the distinct advantage of avoiding controversies over whether morality has a definite point and what that point is if it does. It is, for this reason, a better initial way of distinguishing them than the first. The question therefore becomes one of finding a formal or functional relation that ideals can bear to moral requirements and prohibitions and that can serve to distinguish those that do as moral. Ideals, I suggest, give meaning

31. Henry David Thoreau, Walden, ch. 2, par. 1.  Walden, and Civil Disobedience (New York: W. W. Norton, 1966).

621

216  From Psychology to Morality

or purpose to a life. Certainly, people who devote their lives to the pursuit of some ideal, as Mother Teresa, for instance, devoted her life to ministering to the poorest of the poor, give to their lives in this way a distinct and compelling purpose. But even for those of us who have not devoted our lives wholly to the pursuit of a single ideal, whose lives consist of some of this and some of that, the ideals that we do hold to and that do guide our lives help to give meaning and purpose to our various activities and conduct. Suppose, then, that moral ideals are ideals allegiance to which gives meaning or purpose to moral conduct, specifically, to one’s compliance with moral prohibitions and requirements. They are, in other words, ideals that people can take themselves to be realizing through such compliance. On this proposal, moral ideals function in people’s lives by enlarging their understanding of the significance and value of compliance with moral requirements and prohibitions and thus strengthening their disposition to comply. Accordingly, any ideal that so functions in a person’s life is for that person a moral ideal. Distinguishing moral ideals in this way is consistent, needless to say, with recognizing other ways in which people find meaning or purpose in their complying with moral requirements and prohibitions. Nothing in this proposal, in other words, implies that holding to some moral ideal is a necessary condition of finding such meaning or purpose. Indeed, for the most part people may simply see the purpose of their complying with moral requirements and prohibitions in the necessity for social life that general compliance with moral requirements and prohibitions has or in the advantages of an untarnished reputation that personal compliance with them brings. That is, for the most part people may see a purpose to moral conduct without recourse to any moral ideal. Still, there are times when one’s circumstances invite acts of dishonesty or infidelity that would neither disrupt social life nor sully one’s reputation. One finds a lost wallet, say, containing a wad of cash, or one becomes privy to insider information from which one can gain significantly though illicitly. At these times, allegiance to a moral ideal will give meaning or purpose to compliance with the applicable moral requirement or prohibition that would otherwise escape one. At these times, allegiance to a moral ideal keeps one from falling prey to the perspective of Hume’s sensible knave. From that perspective, the perspective of Hume’s knave, moral requirements and prohibitions appear as externally imposed restraints on one’s conduct. One sees a general purpose as well as personal advantage to observing those restraints and, like Hume’s knave, is willing to observe them in circumstances in which their violation jeopardizes social harmony or risks damage to one’s reputation. But from the knave’s perspective one

2 71

Moral Ideals 217

finds nothing self-​affirming or self-realizing in observing them. By contrast, then, from the perspective of someone who holds a moral ideal, moral requirements and prohibitions are felt as self-imposed restraints. The ideal represents a way of acting or living to which one aspires, which is to say, with which one identifies, and in virtue of this identification, the restraints cease to appear external. I  think something like this point is expressed by Harry Frankfurt when he writes, “Especially with respect to those we love and with respect to our ideals, we are liable to be bound by necessities which have less to do with our adherence to the principles of morality than with integrity and consistency of a more personal kind. These necessities constrain us from betraying the things we care about most and with which, accordingly, we are most closely identified. In a sense which a strictly ethical analysis cannot make clear, what they keep us from violating are not our duties or our obligations but ourselves.”32 Moral ideals, therefore, on this understanding of them, do not belong to the system of rules and principles of which morality, on the prevailing philosophical view of it, consists. They are not, contrary to Gert’s understanding of them, a type of precept whose function in morality complements the function of those moral precepts that set requirements and prohibitions on conduct. Rather, moral ideals exist within the large class of images, pictures, and portraits of ways of acting and living that appear worthy of realization and inspire pursuit. These are images, pictures, and portraits on which we draw to give meaning or purpose to the various activities and practices that make up our lives. Compliance with the demands of morality, its requirements and prohibitions, is commonly a source in people of uncertainty about whether it serves any of their own purposes, and thus moral ideals answer directly a common uncertainty about the meaningfulness of one type of conduct. They specifically work to alleviate this uncertainty, as the development and strengthening of allegiance to moral ideals transforms their possessor’s sense of morality’s demands from that of being externally imposed constraints into that of being self-​imposed ones. Moral ideals, then, do not come into the study of morality as an integral part of the system of rules and principles that, in contemporary philosophy, is seen as the proper object of that study. Perhaps they don’t come into the study at all. Alternatively, though, if the study of morality were understood not so much, let  alone exclusively, as an examination of a set of social norms regulating conduct and were understood, instead, as more of an examination 32.  Harry Frankfurt, The Importance of What We Care About (Cambridge, Eng.: Cambridge University Press, 1988), 91.

821

218  From Psychology to Morality

of those habits and springs of action distinctive of a moral personality, then there might be a place for them in the study. In other words, if there is any merit to the thought at the outset of this paper that the subject of moral philosophy may have changed over the past two centuries, then one might find a place for them in a return to the older understanding of the subject. Accordingly, introducing moral ideals into the study might help to reverse this change. It might, that is, serve to bring the study of morality back to something closer to a study of morals.

921

10

The Emotional Significance of Punishment

To appreciate the emotional significance of punishment one must see its special place in the law. To see this requires distinguishing punishment from other sanctions that modern states use to enforce the law. The modern state makes use of a variety of such sanctions. All consist in losses and hardships that the state imposes on lawbreakers and thereby makes the laws to which it attaches them binding on everyone who is subject to its authority. Punishment is one type of sanction the state imposes, but it is hardly the only one. It has, that is, a distinctive character. It is not simply a tool of law enforcement. Unlike mere penalties, for instance, it has a larger purpose than deterring past and potential lawbreakers from breaking the law. Thus mere penalties, like the fees exacted from taxpayers who file late returns and the civil fines companies incur for discharging pollutants into the environment, are not punishments. Nor is forfeiture of assets for failure to pay one’s taxes or disqualification from voting for having been convicted of a felony punishment. And while these latter sanctions too serve purposes beyond that of inducing those subject to the state’s authority to comply with the law, their purposes are not punitive. In the one case, the state acts with the

219

0 2

220  From Psychology to Morality

purpose of recovering what it is owed from someone delinquent in paying his or her taxes. In the other, it treats voting as a privilege and acts to remove the privilege from someone it deems unfit for the civic responsibility whose performance the privilege facilitates (Deigh, 1988). For a sanction to have a punitive purpose the state must impose it on someone in retaliation for his or her having broken the law. Only such sanctions are punishments. The point is well illustrated in legal cases in which the courts must determine whether an action the government has taken against some individual or individuals is punishment. Consider, for example, the case of Flemming v.  Nestor.1 Ephram Nestor immigrated to the United States from Bulgaria in 1919 and was deported in 1956. The grounds for his deportation was his having been a member of the Communist Party. His membership lasted for six years, from 1933, when he joined, to 1939, when he quit. His deportation on these grounds triggered the denial of his old age insurance benefits under the Social Security Act as amended in 1954. He filed suit in federal court for reinstatement of these benefits, and after the district court ruled in his favor, the case came before the U.S. Supreme Court on appeal. One of the issues with which the Court dealt was whether the Constitution’s protections from punishment without trial and for acts that were legal at the time they were done applied to Nestor’s loss of his old age benefits. In short, the Court had to decide whether Nestor’s loss of these benefits was punishment that the government inflicted on him for having been a member of the Communist Party. If it were, then the law under which the government deprived him of them would violate his Constitutional rights. A deeply divided Court ruled in favor of the government. It held, in a five to four decision, that it must defer to Congress when the question of whether Congress had enacted a law repugnant to the Constitution is open and that it was open in the present case because one could read the amendments Congress made to the Social Security Act whose constitutionality Nestor was challenging as consistent with a valid exercise by Congress of its plenary powers. Specifically, the majority wrote, one could read them as falling within Congress’s power to regulate the administration of the old age and disability insurance programs it had established under the Social Security Act. The dissent disagreed. Justice William Brennan, who wrote for two other justices, put the disagreement most cogently. Only one thing, Brennan wrote, can plausibly explain Congress’s mandating that people who were deported on grounds of their past membership in the Communist Party be deprived of their social security benefits, and that is to “strike at” those who fall into 1.  Flemming v. Nestor, 363 U.S. 603 (1960).

221

The Emotional Significance of Punishment  221

this class of persons. Congress’s intent was “to inflict hurt upon those who by their conduct have incurred the displeasure of Congress.”2 In short, Brennan argued, Congress’s purpose was unquestionably punitive, and hence its enacting, for such a purpose, a law that inflicts losses on people who belong to a specified class is to impose punishment on them. Such an imposition, when carried out without a trial and for acts that were legal at the time they were done, is a clear violation of the Constitution’s bans on bills of attainder and ex post facto legislation. A second case, Gompers v.  Bucks Stove & Range Co.,3 concerns the difference between the sanctions for civil and criminal contempt. Samuel Gompers and two associates had been found in contempt for violating a court order not to publish statements either alleging that Bucks Stove & Range Company had engaged in unfair labor practices or blacklisting the company. Each defendant was sentenced to a fixed period of imprisonment for violating the order. Their sentences varied from six months to a year. They appealed, and the Supreme Court took the appeal. The Court, in a decision that overturned these sentences, explained that a sanction, if imposed for criminal contempt, is necessarily punitive but one that is imposed for civil contempt must be remedial. The sanction’s purpose in the latter case is to benefit the complainant. When the sanction is imprisonment and imposed for civil contempt, the imprisonment is a means of coercing the prisoner to comply with the order in disobedience to which his contempt consists. Accordingly, his imprisonment lasts only for as long as he refuses to comply. As the Court put it, “ ‘[H]‌e carries the keys to his prison in his own pocket.’ He can end the sentence and discharge himself at any moment by doing what he has previously refused to do.”4 It follows that a court cannot properly imprison someone for a fixed period as a sanction for civil contempt. Such imprisonment would not work to coerce the person to comply since he would remain imprisoned regardless of any change in his readiness to comply. It would not benefit the complainant. Rather, it would be punitive. The Court, observing that the complaint against Gompers and his associates had originated with Bucks Stove & Range Company, a private party, that it was the company who had instituted the contempt proceedings leading to the imprisonment of Gompers and his associates, and that the proceedings themselves were conducted as a civil trial, concluded that Gompers and his associates could not be made liable to punishment as a result of these

2.  Ibid., p. 640. 3.  Gompers v. Bucks Stove & Range Co., 221 U.S. 418 (1911). 4.  Ibid., p. 442.

2

222  From Psychology to Morality

proceedings. To be liable to punishment, they would have had to have been found guilty in a criminal trial, and such a trial would have afforded them certain Constitutional protections that were not available to them in the proceedings that led to their imprisonment. One would have been the protection against their having to testify against themselves—​the plaintiff had called each of them as witnesses in the proceedings. Another would have been the higher standard of proof—​guilt beyond reasonable doubt—​that would have been necessary for conviction on the evidence presented. Consequently, the imposition on them of fixed periods of imprisonment, being punitive rather than coercive, violated their Constitutional rights. The Court, recognizing these violations, reversed the lower court’s ruling. Punishment, as these cases show, is a sanction the state imposes with punitive intent or for a punitive purpose. What then is it to act with punitive intent or a punitive purpose? Plainly, to answer that it is to act with the intent to punish or in order to punish would be circular and so unilluminating. To answer in a way that avoids circularity requires, therefore, a notion that is distinct from that of punishment. Let us, as I  suggested earlier, use the notion of retaliation. This notion identifies the direct point or aim of punishment. Accordingly, the state acts with punitive intent if it imposes pain or loss on an offender, real or supposed, in retaliation for his or her having actually or supposedly committed some offense. Similarly, a legal sanction has a punitive purpose if the state imposes it in retaliation for some actual or supposed offense that the person on whom it is imposed did or is presumed to have done. The answer matches the understanding of punishment we express in ordinary talk about punishment in a wide array of areas. For example, when people in romantic relationships retaliate against their lovers by withdrawing affection from or ceasing to communicate with them, we commonly describe them as punishing their lovers for having angered or displeased them. Or to draw on a very different context, international relations, when one country sponsors unprovoked attacks on the citizens or property of another and the latter retaliates by wreaking destruction on the former, we describe the destruction the latter wreaks as punishment inflicted on the former for its belligerence. Thus when the United States, having determined that Libyan operatives carried out the 1986 bombing of a Berlin night club in which two American soldiers died and many others were injured, launched air strikes on several military targets in Libya in retaliation, it was natural to describe the United States as inflicting punishment on Libya for its lethal attack on American soldiers. For a third example consider an election in a representative democracy in which angry voters turn an incumbent out of office. In

23

The Emotional Significance of Punishment  223

such an election, especially when the incumbent’s defeat is decisive, it is common to describe the voters as having punished him or her for conduct in office that displeased them or raised their ire. Similar examples can be found in such disparate areas of life as business, competitive sports, and relations among neighbors. Once we identify acts of punishment as belonging to the class of retaliatory acts and observe that punishment is as much at home in personal and social relations as it is in the law, its emotional significance becomes evident. Retaliation in personal relations is naturally an act of hostility. Its target is someone who has injured one or caused one pain and as a result provoked in one displeasure, anger, resentment, enmity, or the like. In retaliating one then expresses this hostility towards him. Punishment inflicted in such institutional contexts as the law, being a retaliatory action, retains this significance.5 At the same time, its emotional significance transcends the emotions that any of the public officials responsible for inflicting it may feel towards the offenders upon whom they inflict it. In this respect, punishment is like other institutional actions that signify emotion. A country expresses gratitude to its heroic soldiers, for instance, by awarding them medals of valor. The emotional significance of such acts transcends any emotion that a public official or military officer involved in awarding these medals may have. Indeed, such officials and officers may take a detached view of their actions in keeping with their being merely the administrators of the practice of honoring soldiers who performed heroic actions in the service of their country. Elsewhere I  have proposed, in view of examples of punishment that occurs outside of any institutional context, a general definition of the phenomenon according to which its genus is pain or loss inflicted on someone in retaliation for something he or she actually or supposedly did and its differentiae are its being inflicted openly and by someone or some agency who is at least as powerful as its target (see Essay 11). These differentiae are necessary, I argued, to distinguish punishment from pain or loss that is secretly inflicted on someone in retaliation for something he did and also to distinguish it from pain or loss that is inflicted as an act of resistance to or rebellion against another who exercises power over one. Specifically, infliction 5.  James Fitzjames Stephens writes, “My other observation is that, in my opinion, the importance of the moral side of punishment, the importance that is of the expression which it gives to a proper hostility to criminals, has of late years been much underestimated” (Stephens, 1883, vol. 2, p. 91). While Stephens correctly identifies hostility as what punishment expresses, he mischaracterizes punishment as necessarily an institutional expression of hostility. The characterization is implicit in his famous remark that punishment stands to hostility as marriage stands to lust (Stephens, 1883, p. 82).

4 2

224  From Psychology to Morality

of pain or loss on someone in retaliation for something he did is not punishment, I argued, if the person on whom it is inflicted has no knowledge of its source or reason, and it is not punishment if it is inflicted in an attempt to oppose or overthrow someone or some organization who exercises power over one. To these differentiae I would now add that the pain or loss must be deliberately inflicted.6 A punch-​in-​the-​nose or drink-​in-​the-​face thrown spontaneously from sudden anger provoked by an insult, say, or rude treatment, while retaliatory, is not punishment. That this general definition fits legal punishment is evident in light of the distinctions we observed earlier between it and other legal sanctions. In addition, I believe the definition provides a deeper understanding of punishment than the consensus definition that has emerged in the philosophical literature, especially as it concerns the question of legal punishment’s justification. The consensus definition reflects an assumption the theorists who have contributed to this literature implicitly make about legal punishment’s being the touchstone of the phenomenon. As a result, they treat punishment in other contexts like romantic relationships as only weakly analogous to it. The upshot is that the aims of social policy or morality to which these theorists look for the justification of legal punishment tend to obscure the more primitive function punishment has in stable social groups of all kinds. A stable social group may be a couple in a long-​term relationship, a family, a clan, a tribe, or a people who make up a nation or polity. In such groups punishment, taken in the broad sense that I proposed, serves to maintain the group’s stability by working to preserve order within the group. That it does so is obvious in groups in which preserving order means nothing more than preserving the relations of power among the group’s members. It is obvious because to say that punishment works to preserve order in such groups is just to say that one thing more powerful members of the group do to maintain their dominance over less powerful members is to retaliate openly against the latter when the latter harm or displease them. In so retaliating, they demonstrate their power over the recipient of their inflictions, a demonstration that may be necessary to keep the latter or other would-​be challengers to their position from seeking to displace them. Order in many social groups, however, consists in more than the relations of power among the group’s members. It consists in relations among them that are defined and regulated by rules the members generally accept and follow. In social groups whose order is so determined, whose order is civil

6.  For the distinction between acting intentionally and acting deliberately see Austin (1966).

25

The Emotional Significance of Punishment  225

rather than natural, punishment works to preserve that order when either more powerful members retaliate openly against less powerful ones who break the rules or an agency that is invested with power greater than that of any of the group’s members and that acts in their name retaliates against members who break the rules. Such retaliation works to preserve civil order both by demonstrating the greater power that the members or the agency that inflicts punishment has over those on whom it inflicts it and by communicating to them, to others who may be disposed to engage in similar action, and even to members who are not so disposed that the rules are to be respected and that violations of them are not tolerated. This message, conveyed as it is through the pain or loss those on whom it is inflicted suffer, deters them and others who are disposed to act as they did from breaking the rules and at the same time strengthens the disposition to comply with the rules of those who are not so disposed. Punishment thus works to preserve civil order both by deterring those who are disposed to act in ways that would disrupt it and by assuring those who are not that they have not been gulled into complying with the rules.7 Let us call these ways in which it works to preserve civil order the deterrent and assurance functions of punishment. To say that they are functions of punishment is to say no more than that punishment has a causal role in preserving a group’s civil order. Specifically, it is not to say that the members of the group who inflict punishment or the agency empowered to do so inflicts it with the aim or purpose of deterring potential rule-​breakers or reassuring those members who have complied with the rules. Punishment, after all, can deter the former and reassure the latter even when its infliction is undertaken with the sole aim of retaliating against those who disrupt the civil order. It can also, of course, be inflicted with the further aim of producing either effect, though whether or not it is so inflicted is incidental to its emotional significance, however important it may be to justifying its infliction.8 How, then, does punishment’s deterring those who are disposed to disrupt the civil order and assuring those who are not work to preserve that order? The deterrence function’s contribution to preserving it needs little explanation. Open retaliation by someone or some agency at least as powerful as oneself inspires caution and calculation if not outright fear should one again contemplate crossing that person or breaking the rules the person or agency enforces, and it similarly influences others who may also be contemplating 7.  For the identification of assurance as a separate function of punishment, see Morris (1968) and Rawls (1971, p. 315). 8.  For discussion of justifying aims, see Hart (2008).

6 2

226  From Psychology to Morality

crossing the person or breaking those rules. If the pain or loss that is inflicted is great enough, calculation will lead to forbearance by the less powerful, and fear will likewise move them to forbear. Thus punishment’s deterrent effect helps to preserve the civil order by reducing the incidence of disruptions to it. The contribution of the assurance function, by contrast, needs a more developed explanation. The reason is that its contribution is due at least as much to a second mechanism by which a social group’s order is preserved. This second mechanism is the members’ allegiance to the group, and in the case of a group whose order is civil the mechanism includes the members’ commitment to upholding the rules that determine that order. The members’ allegiance and commitment are the basis of trust they place in each other and in the institutions that govern their conduct. This trust cements the social relations among them and consequently strengthens social stability. It follows, then, that the more widespread and stronger allegiance is among the group’s members, the less important the deterrent function of punishment is to the work of preserving civil order. For the broader the allegiance is across the membership and the stronger it is among the members, the smaller will be the class of members who are positively disposed to break the rules and the weaker, on average, will be their disposition to do so. Thus punishment assures those members who comply with the rules out of allegiance to the group and correlative commitment to uphold those rules that their trust in others and the institutions that govern their conduct is not misplaced. In sum, it both reinforces calculations they might make to the effect that breaking the rules does not pay and vindicates their regarding the rules that determine the group’s civil order as authoritative strictures on their conduct. Allegiance to the group and commitment to upholding the rules that determine its civil order come about through internalization. That is, a person, as a result of this process, comes to have a conscience about supporting the group’s good and respecting its rules. The first thinker to identify this process and treat it as formative of the moral personality men and women develop that marks them as respecters of law and duty was Friedrich Nietzsche. His account of the origin of bad conscience in Europeans traces the phenomenon to the beginnings of communities rooted in one place and the self-​stifling of aggressive drives that was necessary for such communal life to go forward. The process, Nietzsche wrote, consisted in turning these drives against oneself and thereby transforming the disposition to act aggressively towards others, which threatened the stability of the community, into susceptibility to compunctions of conscience and feelings of guilt when one contemplated engaging in such aggressive conduct (Nietzsche, 1969). Sigmund Freud, in writings on moral development in civilization, gave a similar account of how

2 7

The Emotional Significance of Punishment  227

through internalization humans acquired a conscience and a sense of guilt (Freud, 1969; see also Deigh, 1984). He too explained the process as one of hostility turned inward, for like Nietzsche he regarded the harshness of conscience, its bite and reproaches, as its most telling characteristic. Freud, however, saw the process as taking place in the inner life of small children in their relations to their parents who, as both providers of protection and nourishment, inspired love, but who, as supreme authorities demanding and exacting obedience, excited fear and anger. Conscience, Freud argued, was the product of the child’s resolution of these powerful ambivalent feelings towards its parents, a resolution it achieved through internalization of parental authority, which it invested with the anger towards its parents that had previously developed. Later thinkers have followed Freud in locating the process in the inner life of small children. At the same time, they have departed from him and Nietzsche in explaining it not as taking place through the turning around of aggression or hostility that was originally directed outward, but rather as the work of other emotions and capacities natural to human beings. Thus John Rawls, in his account of how people acquire a sense of justice, posits a primitive desire to reciprocate, when one sees that one is the recipient of another’s benevolence or enmity, to explain how a child of loving parents comes, in response to their protection and care, to respect their authority and accept their rules, and how the same process replays itself as the child ventures into different social surroundings and enjoys the benefits of participation with others in the organized activities distinctive of these environments (Rawls, 1971, pp. 453–​496). Martin Hoffman, by contrast, explains internalization as a process that combines a child’s being subject to parental discipline with its capacity for empathic understanding of the distress that those whom it harms experience (Hoffman, 2001, pp. 113–​171). However the process is explained, its result is a development in emotional capacities that implies significant change in the child’s attitudes and feelings towards its parents, which is later extended to other powerful figures and institutions in whose charge it is placed and who then exercise authority over it. Among other changes, the child looks to parents and other authorities for approval of its conduct, is concerned not to disappoint them, feels guilt if it either transgresses a limit that they have placed on its conduct or omits doing something they have mandated be done, and in that case seeks restoration of good relations with them. The guilt the child feels and the consequent orientation towards restoring good relations with the authorities it has disobeyed corresponds to anger or stern displeasure that its disobedience provoked in them. And while parents and other authorities may not inflict

8 2

228  From Psychology to Morality

punishment in anger or as an expression of their displeasure but rather do so strictly for the child’s own good—​to teach it about limits to individual lib­ erty in a world one shares with many others and to prepare it to take on the responsibilities of adult life—​the child will nonetheless experience punishment as an expression of anger or displeasure at its misbehavior and take it as a sign of a rupture to relations it values dearly. Consequently, the punishment the authorities inflict typically facilitates the restoration of these relations to good order as it relieves the child of the guilt it feels and at the same time, in the child’s imagination if not in fact, assuages the anger or displeasure that it provoked. One can see, in this familiar dynamic in relations between children and parents, a third way in which punishment works to preserve civil order. Let us call it the reparative function of punishment. When one disrupts the civil order by breaking the rules that determine it, others by virtue of their allegiance to the group and commitment to upholding those rules will regard one’s action as an instance of wrongdoing and, indeed, a crime if the rules in question make up the society’s criminal law. When so regarded, the act represents a rupture to the social relations that those rules define and regulate. Being the agent responsible for this rupture, one is seen as fittingly made to repair it, and the reparations, which are exacted through punishment, consist in one’s undergoing pain or loss consonant with the seriousness of the type of wrong one did. One’s suffering this pain or loss makes the relations once again whole as it both expiates one’s guilt and meets the demand for satisfaction that more powerful members or the agency responsible for enforcing the rules makes as requital for wrongdoing. Or rather it makes the relations once again whole provided that, on the one hand, one accepts the pain or loss as fairly inflicted in virtue of one’s guilt and, on the other, the members of society generally (or the public officials who represent them) are satisfied with the sufficiency of its infliction as requital for wrongdoing and are prepared to cease inflicting further pain or loss on one. These conditions mirror those that obtain in personal relations, such as relations between friends, a rupture to which is healed when the friend responsible for the rupture makes amends and these dissolve whatever resentment his friend, whom he injured, feels towards him. In making amends, the former acknowledges being at fault for injuring the latter, and if he then succeeds in gaining the latter’s forgiveness, his guilt should dissipate as well. For forgiveness entails cessation of resentment and signals that further amends are unnecessary. Of course, as the social distance between wrongdoer and the person he injures increases, it becomes at once easier for the

9 2

The Emotional Significance of Punishment  229

wrongdoer to avoid acknowledging his guilt and harder for him to find ways to dissolve the injured person’s resentment. And his avoiding acknowledging guilt is easier still when his wrongdoing consists in breaking a rule and there is no distinct individual whom he has injured or none at least with whom he can connect. For this reason, in large groups or whole societies whose social relations are ruptured when a member breaks the rules that determine its civil order, allegiance to the group and the correlative commitment to uphold its rules become increasingly important factors in the repair of these relations when they are ruptured. They replace the love and respect between friends that work to repair their relations when they are ruptured. Thus, when someone breaks the laws of her society and thereby ruptures the social relations between her and the other members that those laws define and regulate, punishment performs the reparative function amends serve in personal relations. Its infliction, being a form of retaliation for lawbreaking, signifies anger towards the offender, and the offender’s acceptance of the pain or loss inflicted on her as punishment signifies acknowledgment of guilt. Relations between her and the members of society generally are then made whole if the punishment is seen as cleansing her of guilt. The burden of reparation she has is lifted as the cessation of retaliation signifies the dissolution of society’s anger. Is it realistic to think that the punishment legal systems in modern societies inflict can perform this reparative function? Doesn’t the history of penal practices in such systems show the folly of thinking it can? Certainly, if one fixes on the most violent and incorrigible criminals and takes them as the prime object of such punishment, skepticism about the possibility will seem irrefutable. Nietzsche, for one, expressed such skepticism when he wrote (Nietzsche, 1969, p. 81), It is precisely among criminals and convicts that the sting of conscience is rare. Prisons and penitentiaries are not the kind of hotbed in which this species of gnawing worm is likely to flourish:  all conscientious observers are agreed on that . . . Generally, punishment makes men hard and cold; it concentrates; it sharpens the feeling of alienation; it strengthens the power of resistance. If it happens that punishment destroys the vital energy and brings about a miserable prostration and self-​abasement, such a result is certainly even less pleasant than the usual effects of punishment.

And Nietzsche surely speaks for many. Yet this view distracts from a larger picture. To be sure, the population of a modern society includes some who have no conscience about their dealings with people beyond a small circle of friends and family, if that, and whose estrangement from the larger community in which they live

0 32

230  From Psychology to Morality

makes them a danger to others. And for such people punishment the law imposes can have no more significance than that of being harm that society forces them to suffer when they ignore its rules. Further, they are doubtlessly overrepresented in the class of criminal offenders as compared to their numbers in the general population. Nonetheless, they do not exhaust that class and, notwithstanding the stereotypes fostered in popular culture, may not even be typical of its members. Other offenders, who have some allegiance to the larger community and a conscience about how they deal with people, could see punishment differently. For them acceptance of punishment could signify acknowledgment of guilt and also a means to its expiation and to their reintegration in society. This possibility, however, depends on penal practices that create favorable conditions for its realization. The irony of skepticism like Nietzsche’s is that the more it is taken as the prevailing wisdom, the more self-​fulfilling it is. As long as societies embrace policies like those characteristic of the penal practices in United States of the last thirty years, policies that have led to severe penalties and inhumane prison conditions, any hope of punishment’s serving more than its deterrent and assurance functions is idle.

References Austin, J. L. (1966). Three ways of spilling ink. Philosophical Review, 75, 427–​440. Deigh, J. (1984). Remarks on some difficulties in Freud’s theory of moral development. International Review of Psycho-​Analysis, 11, 207–​225. Deigh, J. (1988). On rights and responsibilities. Law and Philosophy, 7, 147–​178. Freud, S. (1969). Civilization and its discontents. In Standard Edition of the Complete Psychological Works of Sigmund Freud. J. Strachey, Gen. Ed., vol. 21. London, UK: Hogarth Press. Hart, H. L.  A. (2008). Prolegomenon to the principles of punishment. In Punishment and responsibility:  Essays in the philosophy of law (2nd ed.). Oxford, UK:  Oxford University Press. Hoffman, M. (2000). Empathy and moral development: Implications for caring and justice. Cambridge, UK: Cambridge University Press. Morris, H. (1968). Persons and punishment. Monist, 52, 475–​501. Nietzsche, F. (1969). Guilt, bad conscience and the like. In On the genealogy of morals and Ecce Homo (W. Kaufmann, Ed.) New York, NY: Vintage Books. Rawls, J. (1971). A theory of justice. Cambridge, MA: Belknap Press. Stephens, J. F. (1883). A history of the criminal law of England (3 vols.). London, UK: Macmillan.

231

11

Punishment and Proportionality: Part 1

The modern study of punishment in legal and political philosophy concerns the penal sanctions imposed by the state on people who commit crimes. The main question, the question with which writers on the topic have long been preoccupied, is the justification of these sanctions. Before answering it, however, many of these writers take up the question of how to define punishment, and they use the main question as their guide in answering it.1 The answer that has emerged in the literature as the consensus definition is that punishment is pain or loss inflicted on an actual or supposed offender O for some offense that O committed or is supposed to have committed. Further, to be punishment, the pain or loss must be intentionally inflicted by someone or some agency who is not O and who has authority to enforce the rules in the violation of which O’s offense, either real or supposed, consists.2 Finally, to distinguish punishment from mere penalty for violating a rule, like a late fee or disqualification, it is now common for writers on the topic to add as a defining

1.  See Duff, Trials and Punishment, 151–​152. 2. See Flew, “Justification of Punishment”; Benn, “Problems of Punishment”; Hart, “Prolegomenon”; Fletcher, Rethinking Criminal Law, 408–​412.

231

23

232  From Psychology to Morality

condition that the infliction of pain or loss conventionally expresses censure of O for the offense that O committed or is supposed to have committed.3 These writers readily concede that the word ‘punishment’ is used more broadly in everyday speech than adherence to this consensus definition would allow. Accordingly, they take the definition to apply to the central cases of punishment and at the same time acknowledge other cases in which one or more of the features specified in the definition is absent.4 These outlier cases, as it were, include the infliction of pain or loss for the purpose of discipline by parents on their children or schoolteachers on their pupils, the wreaking of destruction by more powerful countries on weaker ones in response to the latter’s having sponsored unprovoked attacks on the former’s citizens or property, the hazing of a member of a clique or other close-​knit group by the other members for deviation from the group’s norms, the turning-​out-​of-​ office by the voters of a legislator whose votes in the legislature angered them, and the delivery of heavy and repeated blows by one boxer to another during a prizefight. To explain this distinction between central and outlier cases, the definition’s exponents typically remark that the word ‘punishment’, when used to characterize the pain or loss in any of the outlier cases, is used in an extended sense or metaphorically.5 The remark reinforces their focusing on the question of justifying the punishment that the state inflicts on people who commit crimes by implying that the central cases are the ones that not only raise the question of justification perspicuously but also exemplify the very concept of punishment more truly than any of the outlier cases. And given that most writers’ interest in the topic is primarily on the question of justification, this thesis about the concept of punishment has gone largely unchallenged.6 The question of justification, however, is not the only normative question about punishment that calls for investigation. Questions about the justice of certain forms of punishment and about how to achieve a just fit between the severity of the punishment and the seriousness of the offense for which it is inflicted may arise after the question of justification has been settled and even independently of it. In the latter case, one assumes that the practice of punishing those convicted of criminal offenses is a fixture of the law, whether justified or not, and asks what justice requires in the carrying out of this practice.

3.  See Feinberg, “Expressive Function of Punishment.” 4.  See Flew, “Justification of Punishment,” 292–​294; Benn, “Problems of Punishment,” 326; Hart, “Prolegomenon,” 5. 5.  See Flew, “Justification of Punishment,” 292; Benn, “Problems of Punishment,” 326. 6.  One exception is Leo Zaibert. See Punishment and Retribution, 24–​28.

23

Punishment and Proportionality: Part 1  233

And while the consensus definition among the legal theorists and political philosophers who write about punishment well serves their interest in the question of justification, it may not be as serviceable for investigating these other questions. This article’s point of departure is the thesis that a different definition, one that covers more of the outlier cases, better serves the investigation of these other normative questions. Of course, if the consensus definition captured the very concept of punishment in virtue of its fitting the central cases, then it would be hard to motivate replacing it with a definition that included some of the outlier cases. So before proposing a different definition, it will be useful to point out the consensus definition’s shortcomings. Specifically, it will be useful to show how our common understanding of punishment is broader than the one the consensus definition implies, that oftentimes when we use ‘punishment’ to characterize the pain or loss exemplified in an outlier case, we use it neither in an extended sense nor metaphorically. The outlier cases so far mentioned, leaving aside those of parental and schoolroom discipline and that of voter anger, exemplify punishment inflicted by people who do not necessarily have the authority to do so. Thus the boxer who delivers heavy and repeated blows to his opponent, the country that wreaks destruction on another country, and the clique or other close-​knit group which subjects one of its member to hazing lack authority to impose pain or loss on those on whom they inflict it. Perhaps, then, contrary to what the consensus definition implies, someone or some agency can inflict punishment on another or others despite having no authority to do so. When war criminals, for instance, are convicted and sentenced to death or imprisonment by foreign tribunals, the harsh treatment to which they are sentenced is their punishment for the crimes of which they are convicted, even if the tribunals that convict them have no authority over the land or people in which these crimes were committed. A good example is Israel’s hanging of Adolph Eichmann after he was convicted in an Israeli court of being one of the chief administrators of the Nazi genocide of European Jewry. The court, upon his conviction, sentenced him to death by hanging for these horrific crimes, and this was the punishment Israel subsequently inflicted on him. These facts are uncontestable regardless of whether Israel had authority to try German citizens for crimes they committed on European soil.7 So, too, the vigilante lynchings of cattle rustlers that took place in the wild West of nineteenth-​century America are examples of punishment inflicted by people,

7.  Indeed, Eichmann’s crimes occurred before the state of Israel came into existence.

4 2 3

234  From Psychology to Morality

typically ranchers and their hired hands, who lacked authority to enforce the law in their territory. And similarly, the tarring and feathering of scoundrels by angry townsfolk, like the tarring and feathering of the duke and dauphin in The Adventures of Huckleberry Finn, two grifters last seen by Huck “astraddle of a rail . . . just look [ing] like a couple of monstrous big soldier-​plumes,”8 is the punishment the victims of their scams inflict on them despite these victims’ having no authority to do so. These and many other examples of punishment inflicted on criminals by the people who suffered at their hands or others closely associated with them argue against conceiving the pain or loss inflicted on someone as punishment only if the person or agency inflicting it has authority to do so. After all, being the victim of a crime or a close associate of victims of crime does not give one such authority.9 Surprisingly, perhaps, one can find further support in political theory for this conclusion. Specifically, one can find such support in the early modern theory of the social contract as the foundation of government. John Locke’s version of the theory, for instance, in his Second Treatise of Government, explains the government’s authority to punish lawbreakers as originating in a right that every man or woman of sound mind has in the state of nature—​the condition of men and women before they form a government—​to punish anyone who violates the law of nature. At the same time, no man or woman has, in the state of nature, authority over any other man or woman. All are political equals. The right to punish that each has is therefore a right in the sense of its holder’s having no duty to forbear punishing those who violate the law of nature. It is not a right in the sense of others’ having a duty to submit to punishment. No one in the state of nature has such a duty. The right, in other words, does not confer any authority to which others have a duty to obey. In the state of nature, punishment is inflicted on lawbreakers, principally, though not exclusively, by those whom the lawbreakers harm, and independently of anyone’s having authority to enforce the natural law. On Locke’s understanding of punishment it can be inflicted by anyone capable of retaliating against a lawbreaker.10 8. Twain, Tom Sawyer and Huckleberry Finn, 403. 9. Flew, citing Mabbott, “Punishment,” declares that “direct action by an aggrieved person with no pretensions to special authority is not properly called punishment, but revenge.” “Justification of Punishment,” 294. But Flew’s point would hold only if revenge and punishment were mutually exclusive, and I see no reason to think they are. A Nazi-​hunter aggrieved at the loss of family and friends at Auschwitz or other Nazi death camps can seek both revenge and justice through his or her efforts to track down former SS officers responsible for running those camps. 10. Locke, Second Treatise of Government, §§ 7 and 8. Hugo Grotius defines punishment “in its most general meaning” even more capaciously as “pain of suffering, which is inflicted

253

Punishment and Proportionality: Part 1  235

Locke, to be sure, introduces punishment and the right to inflict it into his theory as part of the hypothetical condition, the state of nature, whose inconveniences explain the need and character of the social contract on which government is founded. And one might wonder whether he stretches the concept of punishment to provide a theoretical basis in the state of nature for government’s having a right to punish those subject to its authority as a consequence of the social contract on which it is founded. The state of nature, however, though a hypothetical condition, matches the actual condition of the international community before the institution of quasi-​ governmental organizations and agencies, and it would be hard to maintain that someone distorts our everyday understanding of international relations in describing greater powers as punishing weaker states when they retaliate against the latter for breaking a trade agreement, say, or even just affronting the former’s power. Indeed, such considerations no doubt influenced how Locke and other seventeenth-​century social contract theorists understood the implications of their supposition of a state of nature.11 The outlier cases suggest as well a second respect in which our common understanding of punishment is broader than the concept the consensus definition captures. On the consensus definition, punishment is pain or loss inflicted on someone for an offense he or she actually or supposedly committed, where an offense is generally understood as a breach of a rule that is enforced by the infliction of pain or loss on those who do not comply with it. Cases of voter anger and parental discipline, however, need not be ones in which the elected official or the child who is punished has breached some rule. An elected official, to be punished at the polls, need only have exercised his or her office contrary to the voters’ wishes. And a child can be punished for doing things of which its parents disapprove even when

for evil actions.” On the Laws of War and Peace, ch. 20, par. 1. Thomas Hobbes, by contrast, defines punishment as “an evil inflicted by public authority, on him that has done, or omitted that which is judged by the same authority to be a transgression of the law.” Leviathan, ch. 28. par. 1. Hobbes’s definition reflects the difference in his account of the relations of men and women in the condition of mere nature from Locke’s account. Hobbes holds that such a condition is ineluctably one of war in which everyone is the enemy of everyone else. Locke opposes Hobbes on this point. 11.  Telford Taylor reflected this view in writing about the relations of modern countries at war with each other: “[A]‌s long as enforcement of the laws of war is left to the belligerents themselves, whether during the course of hostilities or by the victors at the conclusion, the scope of their application must be limited by the extent to which they have been observed by the enforcing party. To punish the foe—​especially the vanquished foe—​for conduct in which the enforcing nation has engaged, would be so grossly inequitable as to discredit the laws themselves.” Nuremburg and Vietnam, xx.

6 32

236  From Psychology to Morality

the parents have said nothing to the child in advance about not to do those things. To say, then, that the voters punish the official for an offense could mean only that what he or she did provoked their wrath and that the act went beyond what they would tolerate. And similarly for saying that the parents punish the child for an offense. Hence, to maintain that punishment, as commonly understood, is pain or loss inflicted on someone for an offense he or she actually or supposedly committed, one must understand by offense an act that gives offense to or displeases whoever is responsible for the infliction of the pain or loss. The shortcomings in the consensus definition, when one takes it to capture our common understanding of punishment, show the need for a broader definition to accomplish that purpose. Indeed, several of the outlier cases suggest that the broader definition that is needed express a different theme from that of a sanction imposed by an authority for disobedient conduct. They suggest instead the theme of retaliation for harmful or offensive conduct. This theme is present in the cases of parental and schoolroom discipline, vigilante justice, voter anger, the wreaking of destruction by one country on another, and hazing by a clique or other close-​knit group. Thus, in giving this broader definition, rather than modify the consensus definition by removing conditions it places on a pain’s or a loss’s being punishment, let us start anew by making the theme of retaliation the genus of punishment and then specifying its differentiae. The latter is necessary since not all acts of retaliation are acts of punishment. Accordingly, punishment is pain or loss that an individual or group R intentionally inflicts on someone S in retaliation for something S did that offended or displeased R, given that the pain or loss is openly inflicted and that R is at least as powerful as S. The latter two conditions, that the retaliation is open and that the retaliator is at least as powerful as his or her target, serve to distinguish punishment from other types of retaliatory pain or loss. One does not, first of all, inflict punishment on someone if one conceals what one is doing from him or her. To be punished, a person must know that the pain or loss he or she suffers is no accident but is rather intentionally inflicted as payback for something he or she did. Thus it is not punishment if the office manager of a small firm retaliates against an employee who had complained about his sexual overtures to the firm’s owner by secretly sabotaging her work or damaging her car. Similarly, for the financial loss suffered by the victim of the elaborate swindle at the center of the movie The Sting. The victim, a powerful mobster, is completely taken in by the fraudulent representations of the swindlers and as a result never realizes that the loss he suffers is payback for his having ordered the murder of a much-​beloved friend of theirs. Swindling

2 73

Punishment and Proportionality: Part 1  237

him out of half-​a-​million dollars is retaliation for what he did, but because its retaliatory character is concealed from him, it is not punishment. A second reason it is not punishment is the asymmetry in power between the mobster and the people who swindle him. He is the head of a powerful criminal organization. They are small-​time con artists who have come together to avenge their friend’s murder. The mobster’s power is so much greater than theirs that they do not dare try openly to harm him. To do so would make their lives forfeit. If they had tried openly to harm him, he would have seen their attempt as impudent. As such, it would have been rebellious rather than punitive within the order of common thieves, career crooks, and mobsters who practice in the same “business arena.” That is, the impudence of their openly trying to harm him would mark their efforts as an insurgency against this order and against him and the other criminal bosses who maintain it. Accordingly, retaliation against someone more powerful than one qualifies as an act of resistance, rebellion, insurgency, or simply striking back, and not as an act of punishment. This implication of the concept is evident too in the case of the boxer who delivers heavy and repeated blows to his opponent. Characterizing these blows as punishing is apt if they are delivered when the boxer delivering them is overpowering his opponent. If the opponent manages to retaliate by landing a few punches of his own, then we describe him as fighting back and cease to describe the first boxer as delivering punishing blows. Taking these two conditions, then, as the differentiae of punishment, let us define punishment as pain or loss that is intentionally and openly inflicted on someone S in retaliation for something S did by an individual or group who is at least as powerful as S. This definition avoids the shortcomings of the consensus definition and in doing so offers a different lens through which to investigate the normative questions that the topic raises. To investigate these questions using this broader definition, however, requires situating punishment within a stable social group. Such a group may be a couple in a long-​ term relationship, a family, a clan, a tribe, a nation, or the population of a polity. So situated, punishment serves to maintain the group’s stability by working to preserve order within the group. That it does so is obvious in groups in which preserving order means nothing more than preserving the relations of power among the group’s members. It is obvious because to say that punishment works to preserve order in such groups is just to say that one thing more powerful members of the group do to maintain their dominance over less powerful members is to retaliate openly against the latter when the latter harm or displease them. In so retaliating, they demonstrate their power over the recipient of their

8 2 3

238  From Psychology to Morality

inflictions, a demonstration that may be necessary to keep the recipient or other would-​be challengers to their position from seeking to displace them. Order in many social groups, however, consists in more than the relations of power among the group’s members. It consists in relations among them that are defined and regulated by rules the members generally accept and follow. In social groups whose order is so determined by rules, whose order is civil rather than natural, punishment works to preserve that order when either more powerful members retaliate openly against less powerful ones who break the rules, or an agency that is invested with power greater than that of any of the group’s members and that acts in their name retaliates against members who break the rules. Such retaliation works to preserve civil order both by demonstrating the greater power that the members or the agency who inflicts punishment has over rule-​breakers and by communicating to them, to others who may be tempted to engage in similar action, and even to members who are not so tempted that the rules are to be respected and that violations of the them are not tolerated. This message, conveyed as it is through the pain or loss the rule-​breakers suffer, at once discourages those tempted to break the rules from doing so and strengthens the disposition to follow the rules of those who are not so tempted. Legal punishment is punishment that the state, through the operations of law, inflicts on its subjects when they commit crimes. Plainly, then, when one uses the broader definition of punishment to investigate normative questions about legal punishment, the social order one takes such punishment as working to preserve is civil and not natural. Law, after all, is the paradigm of a body of rules that defines and regulates relations among the members of a social group, and legal punishment is therefore to be seen as working to preserve the order that this body of rules constitutes. Seen in this way, its function has bearing on the normative questions under investigation. In particular, it bears on the question of how the severity of legal punishment is to be matched to the seriousness of the crimes for which it is inflicted. For it yields a measure of the seriousness of different types of crime. Specifically, it implies that the more unsettling of the civil order the commission of a crime of a given type is, the more serious is that type of crime and correspondingly, then, the severer the punishment for such crimes. If, for instance, murder is more unsettling of the civil order than larceny, then the punishment for murder will be severer. While some may criticize this measure as insufficiently sensitive to the moral dimension of crime, their criticism, I suspect, merely reflects a tendency to replace the civil order that law, as a body of rules, constitutes with the order that morality, conceived as a set of universally valid precepts of right and wrong, constitutes and to regard legal

9 32

Punishment and Proportionality: Part 1  239

punishment as fulfilling the responsibility that every civilized society has to uphold this universal moral order. One consequence, then, of using the broader definition of punishment to investigate normative questions about legal punishment is that one must reject conceiving of such punishment as an essentially moral response to crime. Indeed, it is a consequence of using this definition that one must reject conceiving of crime as necessarily having a moral dimension. Some crimes in a civilized society whose civil order law constitutes may be merely mala prohibita.12 Call this question the ordinal question of proportionality. A  second question on which punishment’s function of working to preserve civil order bears is then aptly called the cardinal question. It is the question of how severe punishment for a given criminal act ought to be in view of the seriousness of the type of crime it is. What amount of pain or loss, in particular, would make the punishment of this act proportionately severe? The answer to the cardinal question is that the punishment ought to be severe enough to preserve the civil order but not more severe than is necessary for it to preserve that order. The principle of proportionality behind this answer is the same as the one in international law regarding how much military force a belligerent may use to achieve its goals, given that pursuit of these goals is itself permissible under the laws of war. The principle requires that a belligerent use no greater military force against its enemy than is necessary to meet its military objective. Aerial bombardment of a munitions factory is a common example. A belligerent’s use of bombs to destroy such a factory would be a violation of this principle if it used bombs that devastated the surrounding community when it could have used much less powerful bombs to destroy the factory. The additional destruction in this case represents wanton disregard of the lives that were needlessly lost or ruined, and such disregard is a crime of war. Accordingly, to inflict punishment of greater severity than is necessary to preserve civil order is a similarly excessive use of force against human beings if the function of punishment to preserve civil order is the function one presupposes in answering the cardinal question of proportionality. It, too, is unjust on this presupposition. Admittedly, this standard of proportionality is vague. It is doubtless impossible in practice to determine whether the law’s making ten years of incarceration the standard sentence for armed robbery is as conducive to preserving civil order as its making fifteen years the standard sentence. Still, the answer

12.  See Dolinko, “Some Thoughts about Retributivism,” 556–​557.

0 42

240  From Psychology to Morality

to the cardinal question when one uses the function of punishment to preserve civil order to answer it is different from other answers that theorists of punishment have given, and the differences among the answers correspond to differences in the severity of punishment that would result from the law’s following one answer rather than another in its sentencing policies. Consider, in particular, how the answer differs from the answer that classical deterrence theories of punishment give.13 The latter is that the severity of punishment for a given type of crime ought to be the amount of pain or loss whose infliction would have the optimal deterrent effect on all who are disposed to commit crimes of that type, where a punishment has optimal deterrent effect for a crime of a given type when its infliction reduces the incidence of crimes of that type to the point at which any increase in the severity of the punishment inflicted would either fail to reduce further the incidence of such crimes or, though it succeeded in reducing further the incidence of such crimes, would do so at the cost of diminishing the society’s general welfare. And while sentencing policies deriving from the answer to the cardinal question that classical deterrence theories give would, no doubt, work to preserve civil order, the severity of the punishments that these policies generate are nonetheless liable to be greater than what is necessary to preserve that order. Indeed, they will be greater whenever preservation of the civil order is consistent with greater incidence of crimes of the type in question than would occur under these policies. This result follows not only from there being a difference between something’s serving to preserve the civil order and its having the optimal deterrent effect, but also from the reasonable supposition that the smallest incidence of crimes of a given type that is necessary to unsettle the civil order is greater than the incidence of crimes of that type at which no increase in the severity of punishment would further reduce that incidence without diminishing the society’s general welfare. In short, sentencing policies that derive from the answer that classical deterrence theories give are liable to generate severer punishments than policies that derive from the answer one would give if one took the function of punishment to be the preservation of the social order. A second reason why classical deterrence theories are liable to yield sentencing policies that generate severer punishments than those generated by policies that result from taking preservation of the social order as the function of punishment is that legal punishment is not the only mechanism 13.  See Bentham, Principles of Morals and Legislation, ch. 14, “Of the Proportion between Punishments and Offences”; Bentham, Rationale of Punishment, book 1; Posner, “Economic Theory of the Criminal Law.”

241

Punishment and Proportionality: Part 1  241

in a society that works to preserve the social order. In addition, the members’ allegiance to principles of right action that call upon them to support the society’s civic institutions works in this way.14 Hence, the more widespread and stronger that allegiance, the less important is legal punishment to the work of preserving social order. For the broader the allegiance is across the membership and the stronger it is among the members, the smaller the class of members who are disposed to break the laws. Presumably, then, the incidence of crimes decreases as the size of this class decreases, and accordingly each criminal act is less unsettling of the social order than it would be if the class were larger. Punishment that works to achieve the smallest number of such acts consistent with maximizing the society’s general welfare is likely, therefore, to be more severe than what is necessary to preserve the social order, since the order may be well-​enough preserved even when the number of criminal acts is not reduced to its smallest size consistent with maximizing the general welfare. In short, society may well be able to tolerate a greater incidence of criminal acts without the social order’s being threatened than the incidence that would occur if the punishments that the state inflicted had optimal deterrent effect. Part of the allegiance people have to those principles of right action that call upon them to support the society’s civic institutions, when they have such allegiance, is an attitude of respect for law. Their allegiance, in other words, makes them not just law-​abiding but also law-​respecting. That is, they recognize the law and the state that stands behind it as having the authority and not merely the power to govern their lives. They thus see it as something worth having and keeping. Such recognition represents their having internalized the core of the social order that the law determines, and, consequently, if they violate the law, they will tend to be less resistant to its treatment of them first as defendants whom the state tries and, ultimately, should they be found guilty, as convicted offenders. Their acceptance of this treatment depends, of course, on their regarding the procedures of their trial as fair and the sentence they receive as just. If either is in their eyes unjust, if the prosecution appears to them overzealous or the judge or the jury biased against them, or if the sentence seems to them excessive, then they are not open to accepting their punishment as a just desert. Such treatment will, instead, embitter them and, as a result, weaken considerably their allegiance. These considerations further highlight the space between the two answers.

14.  See Rawls, Theory of Justice, 496–​504.

2 4

242  From Psychology to Morality

With regard to the answer that classical deterrence theories give, they complicate how these theories determine the severity of the sentences for each type of crime. They argue, on the one hand, for making the punishment for a first criminal offence less severe whenever doing so decreases the incidence of repeat offenses more than it increases the incidence of first offenses. They argue, on the other, for increasing the severity of the punishment for repeat offenders whenever increasing the severity of the punishment for the nth offense, for n > 1, decreases the incidence of offenses each of which is the offender’s mth offense, m > n. By contrast, with regard to the answer that one would give if one took the function of punishment to be the preservation of the social order, these considerations do not affect the determination of the severity of sentences for first offenses but may affect the determination of the severity of sentences for repeat offenders. They do not affect the determination of the severity of sentences for first offenses since these sentences must already be the minimum necessary for preserving social order and therefore cannot be made less severe without making them inadequate for that purpose. But they may affect the determination of the severity of sentences for repeat offenders since any weakening of these offenders’ allegiance that results from the punishment they received for their first offence would mean that their subsequent offenses might be more unsettling than their first offense and consequently call for more severe punishment than they received for the first offense. This contrast between the two answers thus shows that the sentencing policies on the second answer, but not on the first, match the severity of the punishment to the seriousness of the crime for which it is imposed. The first instead adjusts the severity of the punishment to produce the lowest incidence of the type of crime for which it is imposed consistent with maximizing the society’s general welfare. These observations of the space between the two answers do not by themselves argue for the superiority of one of the answers to the other. They do, however, raise the question whether there is reason, in light of the space between them, to take one answer as superior to the other. Of course, if the standard by which the answers should be measured is the efficiency of sentencing policies in stopping crime that the answers yield, then the answer that classical deterrence theories give is superior. But the question of proportionality is not a question of efficiency but rather one of justice. The standard, in other words, by which the answers should be measured is the justice of the sentencing policies they yield in view of the severity of the punishments those policies generate. Now either answer, as I  noted earlier, yields sentencing policies that would work to preserve the social order. But the answer classical deterrence

2 4 3

Punishment and Proportionality: Part 1  243

theories give tends to yield policies that generate severer punishments than the policies the answer that is based on taking the function of punishment to be the preservation of social order yields. Hence, its answer would be superior only if taking optimal deterrent effect as the goal of the state’s sentencing policies better served the interests of justice than taking preservation of the social order as that goal. The main objection to classical deterrence theories, however, suggests otherwise. For on the main objection, the state’s taking the optimal deterrent effect as the goal of its sentencing policies is morally problematic. Indeed, the tendency of sentencing policies with this goal to generate more severe punishments than policies whose goal is the preservation of the social order highlights this problem. One way to put the problem is this: for the state to take the optimal deterrent effect as the goal of its sentencing policies is for it to treat the criminal offenders whom it punishes merely as instruments of its efforts to reduce the incidence of crime as much as possible without decreasing the general welfare more through punishment than through the additional criminal mischief that punishment would deter. The criminal offenders whom the state punishes are treated as instruments in these efforts in that the state intentionally harms them as a means to promoting the society’s general welfare. Specifically, it publicly inflicts this harm on them so as to make credible the threats to do such harm that it issues when it attaches sanctions to the acts it prohibits as crimes, for only by issuing credible threats can it create in those who are disposed to commit crimes enough fear of suffering such harms to inhibit many of them from doing so. Punishing criminal offenders thus serves to decrease the amount of mischief that crime causes in the lives of people, and so long as the unhappiness that punishment causes is less than the unhappiness that the criminal mischief it deters would cause, a decrease in the amount of such mischief that the state brings about through punishment means an increase in society’s general welfare. Of course, to treat a person as an instrument is not always morally objectionable. It is not morally objectionable, for instance, if the person consents to being so treated. Nor is it morally objectionable if he or she voluntarily participates in a practice or institution whose stability requires that its participants sometimes serve as instruments of other participants’ interests, provided that the person shares in the benefits of participation and that these benefits and their correlative burdens are fairly distributed among the participants. And finally, it is not morally objectionable if the person, having been found to be responsible for damage to or destruction of something of value that he or she does not own, is made to repair or replace it, or to compensate its owner if neither repair nor replacement is feasible. In this case,

42

244  From Psychology to Morality

the person is treated not merely as an instrument, but also as a member of the group who, as a participant in its life, is expected to answer for harm he or she does to other members or the group as a whole.15 Using someone as an instrument becomes morally objectionable, however, when it is done against his or her will and for the sake of advancing the welfare of others or the welfare of some group. It becomes morally objectionable, that is, when a person is forced to make sacrifices solely for the benefit of others or some group. Thus, what makes sentencing policies whose goal is to inflict punishments that have optimal deterrent effect morally problematic is their requiring that the state force the people whom it punishes to suffer some harm for the sake of advancing the society’s general welfare. The problem is especially evident when one considers the greater severity of the punishments that such policies generate in comparison with the punishments that policies whose goal is to preserve the social order generate. Accordingly, the offenders who are subject to the former suffer greater pain or loss as punishment for their offenses than they would suffer if they were subject to the latter. And given that the severer the punishment the greater its deterrent effect, the additional pain or loss they suffer brings about a decrease in the incidence of the type of crime they commit from the incidence that would result from their being subject to the latter. This decrease then translates into an increase in the society’s general welfare over what the general welfare would be if they were subject to the latter. It follows that offenders who are subject to the punishments generated by the sentencing policies classical deterrence theories yield are forced to suffer this additional pain or loss as a means to achieving the corresponding increase in the society’s general welfare. The state, in other words, achieves this incremental rise in the society’s general welfare by treating these offenders merely as instruments of its efforts to increase that welfare. Admittedly, those defenders of classical deterrence theories who are untroubled by the main objection to these theories will be no less sanguine about this instantiation of it. But even they should find it harder to maintain that the optimal deterrence effect of punishment is a superior goal to preservation of the social order in light of the collateral consequences of making the former the goal. When punishment, for instance, consists in incarcerating the offender for months or years, then the offender’s spouse and children suffer significant losses too. Such losses, in some circumstances, extend to the offender’s parents, siblings, and others who may be dependent on him or her

15.  See Strawson, “Freedom and Resentment.”

2 4 5

Punishment and Proportionality: Part 1  245

for their well-​being. These bystanders are themselves innocent of the offense for which the offender is punished, yet they too bear a portion of the burdens that the state imposes as a condition of its upholding the law through its penal system. While their having to bear a portion of these burdens is unavoidable, the question is whether their having to suffer the additional hardships that result from the state’s having made optimal deterrent effect the goal of its sentencing policies is just. Do the additional hardships they suffer due to the greater severity of punishments offenders receive when the state makes optimal deterrent effect rather than preservation of the social order the goal of its sentencing policies imply that the policies are unjust? It is hard to deny that it does, for it is hard to deny the injustice of making innocent people suffer additional hardships against their will so as to bring about an increase in the general welfare of the society. Indeed, it is harder still, the more likely those who are made to suffer these additional hardships are not only innocent bystanders but also among the poor and socially disadvantaged in the society. Such a consequence would be a clear violation of the principle of justice that John Rawls famously dubbed the difference principle.16 This last consideration would become even more important to the case for taking the preservation of the social order as superior to the optimal deterrent effect as the goal of the state’s sentencing policies if the main objection to the latter’s being the goal applied equally well to the former. And that one might think the former’s being the goal was as open to the objection as the latter is clear, given, as we noted earlier, that preservation of the social order requires deterring not only the offender on whom punishment is inflicted from committing future offenses but also others who too might be disposed to commit offenses. But unlike bringing about the optimal deterrent effect, there is more to preserving the social order than preventing crime. There is more to it than preventing crime, just as there is more to preserving one’s health than preventing disease. And what more there is explains why the main objection to making optimal deterrent effect the goal of the state’s sentencing policies does not apply to making preserving the social order the goal of those policies. Thus one preserves one’s health both by strengthening the vital organs on which it depends and seeing that they are functioning well and by taking measures, such as getting vaccinated, that help prevent disease. To strengthen one’s bones, for instance, one includes in one’s diet food rich in calcium or vitamin D, and one does the same to prevent osteoporosis or rickets. And if

16.  See Rawls, Theory of Justice, 75–​80.

6 42

246  From Psychology to Morality

one’s heart develops an arrhythmia, then one may correct it with medication, and the same medication may help to prevent stroke as well. In these examples of preserving one’s health, the same expedient works both to correct a condition of weakness or irregularity and to prevent disease. Analogously, the state preserves social order both by correcting the disruptions to the social order that are due to criminal actions and by preventing future crimes, and it uses the same expedient, punishment, to achieve both ends. For punishment not only works, as I pointed out earlier, to prevent future crimes by making credible the threats the state issues when it attaches sanctions to the acts it prohibits as crimes, but also works to correct present disruptions to the social order that are due to criminal acts. And while the former consists in the state’s treating the offender merely as an instrument in its efforts to prevent crime, the latter does not. Rather, it consists in the state’s treating the offender as the agent of a crime, someone who is responsible for disruption in the social order. To disrupt the social order by means of criminal action is, in other words, to breach or rupture the relations between one and the other members of society generally (or between one and the society’s supreme ruler) that the laws one broke define and regulate. Being the agent responsible for this rupture, one is seen as fittingly made to repair it. And the reparations, which are exacted through punishment, consist in one’s undergoing pain or loss consonant with the seriousness of the type of crime one committed. This undergoing of pain or loss makes the relations once again whole, provided of course that the members of society generally (or its supreme ruler) are by and large satisfied with it and prepared to cease inflicting further hardships on and to renew relations with one. When punishment works in this way to preserve social order, it retains the recognition of the offender as a participant in the social relations he or she ruptured, as still a member of society, and therefore does not reduce him or her to a mere instrument of the state’s efforts to prevent crime. Hence, the main objection to the state’s taking optimal deterrent effect as the goal of its sentencing policies does not apply to its taking preservation of the social order as the goal of those policies. I will close with an observation. It may seem as though the argument of this article is too abstract to make any concrete difference to how social policy as regards the severity of sentences for criminal offenses is made. To think so, however, would be a mistake. Among other things, one can understand the argument for the superiority of the preservation of the social order as the goal of the state’s sentencing policies to the optimal deterrent effect as a caution about popular campaigns against certain types of crime like those that have in the last forty years in America too frequently placed strong pressure on, if

4 2 7

Punishment and Proportionality: Part 1  247

not fed the zeal of, public officials to stop these crimes. Such efforts easily fit an understanding of punishment as primarily inflicted to prevent crime, and taking the optimal deterrent effect as the goal of the state’s sentencing policies dresses them in the guise of rationality and gives them the imprimatur of science. At the same time, they have led to the incarceration of record-​high numbers of Americans for periods of time far beyond what was once the norm for crimes of these types. The Rockefeller drug laws enacted in the 1970s and similar laws that were later enacted in other states, with the heavy penalties they imposed for mere possession of contraband drugs and other nonviolent crimes, painfully illustrate how ruinous of people’s lives such campaigns can be. If policymakers, in their devising of sentencing policies, could be brought to see the function of punishment as preserving the social order rather than optimally preventing crime, then they would be better able to understand the injustice of imposing such harsh sentences on people whose crimes do not greatly threaten the social order and better prepared as well to resist the pressures to sacrifice those people that arise with popular movements for highly punitive laws whose aim is to stamp out these crimes.

Bibliogra p hy Benn, S. I. “An Approach to the Problems of Punishment.” Philosophy 33 (1958): 325–​341. Bentham, Jeremy. An Introduction to the Principles of Morals and Legislation. Oxford: Clarendon, 1907. http://​www.econlib.org/​ library/​Bentham/​bnthPMLCover.html. Bentham, Jeremy. The Rationale of Punishment. London: Robert Heward, 1830. http://​ www. laits.utexas.edu/​poltheory/​bentham/​rp/​. Dolinko, David. “Some Thoughts about Retributivism.” Ethics 101 (1991): 537–​559. Duff, R. A. Trials and Punishment. Cambridge: Cambridge University Press, 1986. Feinberg, Joel. “The Expressive Function of Punishment.” In Doing and Deserving:  Essays in the Theory of Responsibility, 95–​118. Princeton:  Princeton University Press, 1970. Fletcher, George. Rethinking Criminal Law. Boston: Little, Brown, 1978. Flew, Anthony. “The Justification of Punishment.” Philosophy 29 (1954): 291–​307. Grotius, Hugo. On the Laws of War and Peace. Translated by A. C. Campbell. London, 1814. http://​www.constitution.org/​gro/​djbp.htm. Hart, H. L.  A. “Prolegomenon to the Principles of Punishment.” In Punishment and Responsibility: Essays in the Philosophy of Law, 1–​28. Oxford:  Oxford University Press, 1968. Hobbes, Thomas. Leviathan, London: Andrew Crooke, 1651. http://​www.gutenberg.org/​ files/​3207/​3207-​h/​3207-​h.htm. Locke, John. Second Treatise of Government. Edited by C. B. McPherson. Indianapolis: Hackett, 1980. http://​www.gutenberg.org/​files/​7370/​ 7370-​h/​7370-​h.htm. Mabbott, J. D. “Punishment.” Mind 48 (1939): 152–​167.

8 4 2

248  From Psychology to Morality Posner, Richard. “An Economic Theory of the Criminal Law.” Columbia Law Review 85 (1985): 1193–​1231. Rawls, John. A Theory of Justice. Cambridge, MA: Belknap Press, 1971. The Sting [film]. Directed by George Roy Hill. Screenplay by David S. Ward. 1973. Strawson, P. F. “Freedom and Resentment.” Proceedings of the British Academy 48 (1962): 1–​25. Taylor, Telford. Nuremburg and Vietnam: An American Tragedy. New York: Quadrangle, 1970. Twain, Mark. The Complete Adventures of Tom Sawyer and Huckleberry Finn. New York: Harper & Row, 1978. Zaibert, Leo. Punishment and Retribution. Burlington: Ashgate, 2006.

9 42

12

Punishment and Proportionality: Part 2

1. Retributive justice requires that the punishment for a crime match in its severity the offense’s gravity. This requirement’s earliest variant is lex talionis. The Hebrew Bible contains its best-​known formulations: “If any harm follows [the striking of a woman with child], then you shall give life for life, eye for eye, tooth for tooth, hand for hand, foot for foot, burn for burn, wound for wound, stripe for stripe.”1 The measure of justice in the biblical formulations is that of equality. Inflict on the injurer the same injury as he inflicted on his victim. This measure survives in American law, if at all, as a retributivist rationale for capital punishment. It has otherwise been abandoned throughout Western legal systems as both barbaric and unworkable. Imprisonment and fines have replaced the kinds of corporal punishment that lex talionis calls for, and because imprisonment and fines admit of quantitative determinations, proportionality has replaced sameness of injury in the 1.  Exodus 21:23–​25; see also Leviticus 24:21, and Deuteronomy 19:21. There is an even earlier formulation in the Hammurabi code of ancient Babylon. The principle, in both the Hammurabi code and Hebrew law, functioned as a principle of recompense to victims. See Daube, Studies in Biblical Law, 102–​153.

249

0 52

250  From Psychology to Morality

specification of the measure of equality that the requirement of retributive justice applies. Accordingly, to achieve the match that such justice requires between the crimes people commit and the punishments they receive in return, a legal system must impose punishment in keeping with the rule that the graver the crime, the severer the punishment. To implement this rule, lawmakers need a way to assess the gravity of the crimes they affirm or create. Perhaps the most natural way is to apply a notion of moral desert. Naive lawmakers, let us suppose, naturally focus on giving offenders what, owing to their crimes, they deserve, and the basis of an offender’s desert is the evil of his or her criminal act. Thus, the more evil a criminal act is, the more severe is the punishment its agent deserves. Of course, these lawmakers need a way to assess the degree of an action’s evil that is consonant with the broad range of acts that are crimes in a modern legal system. Typically, though not uncontroversially, theorists sympathetic to the naive lawmakers’ view identify two components of a criminal act that combine to determine how evil it is and therefore how much punishment the offender deserves.2 These components are the harm the act causes and the blameworthiness of its agent. Provided, then, that one can determine both the amount of harm that an act causes and the degree of its agent’s blameworthiness, this analysis shows how one can use moral desert as a way to assess a crime’s gravity. It can, therefore, serve lawmakers as a guide to enacting a schedule of punishments that meets the rule that the graver the crime, the severer the punishment. Such a schedule, however, is not alone sufficient to meet the requirement that punishments match in their severity the gravity of the crimes for which they are imposed. After all, if the punishments in a given schedule are systematically too severe or too lenient, then the requirement will not have been met. Lawmakers, in other words, need not only a way to determine the relative gravity of the crimes they affirm or create, but also a way to determine for each type of crime the severity of its punishment as determined, say, by the amount of time an offender spends in prison or the amount of money he or she pays as a fine. More exactly, they need a way to determine a range in the severity of punishment for each type of crime, so that judges and juries can make the punishments they impose on particular offenders who have been convicted of crimes of that type more or less severe within that range according to whether the circumstances of their crimes either aggravate or mitigate their gravity. How lawmakers do this, however, depends on

2.  See McMahan, “Proportionate Defense,” 23, and Moore, Placing Blame, 191–​193.

251

Punishment and Proportionality: Part 2  251

their understanding of the end or ends of punishment. Our naive lawmakers understand the sole aim of punishment to be giving criminals what they deserve. More sophisticated lawmakers see punishment as serving larger social ends than retribution, ends such as promoting public safety or preserving civil order. Accordingly, they use their interest in promoting these ends to guide them in determining for each type of crime the severity of the punishment to be imposed. In doing so, they, too, can remain faithful to the rule that the graver the crime, the severer the punishment, though their fidelity to the rule requires their understanding the gravity of a crime, not as a matter of how evil it is, but rather as a matter of how much it retards the social ends they seek to advance through punishment. How, then, do our naive lawmakers determine for each type of crime the severity of the punishment to be imposed? How does their interest in exacting retribution guide them in making these determinations? Because they have no further aim than retribution, their intention is strictly punitive. A  punitive intention is an intention to inflict pain or loss on someone solely because of something bad that he or she did to one or to someone on whose behalf one acts. It is, in other words, an intention to retaliate against that person. Accordingly, let us understand punishment generally as pain or loss inflicted on someone in retaliation for something he or she did.3 To be sure, not every act of retaliation is an act of punishment. Rather, retaliation is the genus of such acts. What differentiates them within this genus is their being done deliberately, openly, and by someone or some group that is at least as powerful as the person on whom the pain or loss is inflicted. These are the chief differentiae. Perhaps there are others. In any case, our naive lawmakers, in determining for each type of crime the severity of the punishment to be imposed for committing it are guided by an interest in retaliation for its own sake, whereas our more sophisticated lawmakers are guided by an interest in retaliation for the sake of promoting public safety or preserving civil order. Legal punishment is an institutionalized form of retaliation. This feature, however, is not one of the differentiae of acts of punishment. Many such acts occur outside of institutional contexts. When a gang of thieves, for instance, inflicts pain or injury on one of its members in retaliation for his having disclosed their hideout to outsiders, the pain or injury the gang inflicts on the errant member is his punishment for exposing them to danger. Such 3.  See Essay 10 and Essay 11 for more extensive discussion of taking pain or loss inflicted on someone in retaliation for something he did as the genus of punishment in a definition of punishment by genus and differentiae.

25

252  From Psychology to Morality

inflictions of punishment are a common way in which a group maintains the allegiance of its members and keeps them subservient to its leaders. No institutional context is necessary to seeing them as inflictions of punishment. When they occur outside of an institutional context, they naturally express hostility towards whoever is their target, for retaliation, as a natural response to someone who has done something bad to one, expresses one’s anger, enmity, resentment, irritation, or the like towards him or her. When punishment is inflicted within an institutional context, it, too, signifies hostility towards the person on whom it is inflicted, notwithstanding the possibility that none of the officials or administrators responsible for imposing it feels any anger, enmity, resentment or the like towards that person.4 The hostility it signifies, moreover, guides our naive lawmakers in their determining for each type of crime the severity of the punishment to be imposed. On the one hand, they make the punishment for crimes of a given type as severe as is necessary to express adequately the hostility that members of the society are presumed to have towards people who commit such crimes. But on the other, they make it no more severe than what would express the hostility that such members would be warranted in having. You might wonder, of course, whether recourse to the hostility that the punishment for crimes of a given type expresses need have any role, let alone the guiding role, in how our naive lawmakers determine the severity of that punishment. After all, what warrants such hostility as well as shows how severe the punishment for such crimes must be to express it adequately is the evil of those crimes. So why not simply suppose that these lawmakers assess the magnitude of this evil directly when they determine the severity of the punishment? The answer is found in the way modern theorists of retributivism construe the notion of evil so as to render it applicable to the broad range of types of crime that a modern legal system comprehends. As I  observed earlier, these theorists take the evil of a criminal act to be a combination of the harm the act causes and the blameworthiness of its agent.5 And while we might grant that our naive lawmakers can directly assess the contribution that the first component makes to the magnitude of the act’s evil, the same cannot be said of their assessing the contribution that the second component makes. That assessment requires recourse to the hostility members of the society would be warranted in having towards agents of crimes of the type in question. 4.  See Essay 10, p. 223 above. 5.  See note 2 above.

235

Punishment and Proportionality: Part 2  253

To see why, let us first note that the blameworthiness of any such agent depends partially if not wholly on the character of his will. The contribution that the character of the will makes to the agent’s blameworthiness is a synthesis of two factors. One is the amount of proscribed harm, if any, that the agent foresaw or should have foreseen he would cause or risk if he succeeded in achieving the end he willed. Thus one would be more blameworthy for assaulting a coworker with the intention of killing him than one would be for assaulting the same coworker with the intention of causing him intense but bearable pain, even if the bodily movements the assault consisted in and the injuries that resulted from them were the same in either case. For death is a greater harm than intense but bearable pain. The other factor is the agent’s disposition towards the harm he caused or risked in acting as he did. Thus one would be more blameworthy for intentionally breaking someone’s ribs by throwing an elbow than one would be for negligently breaking that person’s ribs by throwing an elbow, even if the bodily movements in which the thrown elbow consisted were the same in either case. For if one’s disposition towards an injury one caused is blameworthy, then one deserves more blame for intending the injury than for being insufficiently concerned about its resulting from one’s action to heed whether it is likely to result. And even if we again grant that one can assess the amount of harm involved, both actual and potential, our naive lawmakers must still, to determine the severity of punishment, take recourse to the hostility the members of society would be warranted in having towards the agent. The reason is that knowing the character of an agent’s will and so his disposition towards any harm he caused or risked does not by itself tell one whether the agent is blameworthy or blameless. For instance, an intention to harm is blameworthy if its execution yields a wrongful act like battery, but blameless if its execution yields a permissible act like that of bloodying the nose of one’s opponent in a boxing match. Similarly, being indifferent to the potential lethal effects of one’s action is blameworthy if the action puts innocent bystanders at risk, but blameless if the people whose lives it puts at risk are enemy soldiers in a war in which one is a combatant serving one’s country and its enemy is the aggressor. Or again, an intention to injure another is blameworthy in a sane, adult assailant, but blameless in a psychotic one whose violence is the result of delusions, or an immature one who is not yet old enough to be held responsible for his or her actions. Plainly, what explains the difference between the two cases in each of these contrasts is not some discernible property of the will that the blameworthy agent has

4 2 5

254  From Psychology to Morality

and the blameless one lacks. Rather, to take a leaf from Hume’s Treatise,6 the difference arises from the difference in the feelings the cases provoke in an unbiased observer who takes in all the circumstances of the actions he observes. The blameworthy dispositions provoke hostility that implies blame; the blameless ones provoke at most dismay that carries no implication of blame. Hence, there being no discernible property or properties of a will that produce acts causing or risking harm by which taken alone one can tell whether its possessor’s disposition towards that harm is blameworthy, our naive lawmakers must, to determine the severity of punishment for crimes of a given type, take recourse in the hostility the members of society would be warranted in having towards agents of such crimes.

2. We have now come to the crux of the traditional problem of punishment in Western ethics. Because punishment is a form of retaliation, society, by making it a practice, sanctions general hostility towards those whom it determines should be, by virtue of their having acted wrongly, liable to punishment. When the practice is then designed with no further aim than to give these wrongdoers what they deserve, society does not merely sanction the hostility punishment signifies, but more grandly, under some such rubric as righteous anger or moral indignation, exalts it as the touchstone of justice. As a result, the pain or loss in which punishment consists is inflicted for its own sake. It is meant, that is, to do no more good than satisfy the desire for retribution that is implicit in such hostility. This view of punishment, notwithstanding its claim to be a vindication of justice, is hard to square with basic principles of humanity in Western ethics. Fidelity to these principles does not sit comfortably with the exaltation of such hostility. The problem goes back to Plato. It first appears as part of the great speech that Protagoras gives near the beginning of the dialogue that bears his name.7 Perhaps, then, the problem originated with the sophists since scholars generally place the dialogue among the ones they take to be 6.  See Hume, A Treatise of Human Nature, 468. “Take any action allow’d to be vicious. Willful murder, for instance. Examine it in all lights, and see if you can find the matter of fact, or real existence, which you call vice. In which-​ever way you take it, you find only certain passions, motives, volitions and thoughts. There is no matter of fact in the case. The vice entirely escapes you, as long as you consider the object. You can never find it, till you turn your reflexion into your own breast, and find a sentiment of disapprobation, which arises in you, towards this action.” 7. Plato, Complete Works, 320d–​328d.

25

Punishment and Proportionality: Part 2  255

imaginative reconstructions of conversations Plato heard as a youthful follower of Socrates. In any case, the problem is one that Plato cognizes in this early dialogue and revisits in later ones. The dialogue unfolds through a recounting by Socrates of a conversation he had with Protagoras at the home of a wealthy Athenian. Socrates tells of his challenging Protagoras’s claim to teach virtue and of Protagoras’s answer to this challenge. Protagoras’s great speech contains his answer. The speech is a lengthy account of how the citizens of a city shape each other’s conduct and character through both civic and private practices. The punishment of wrongdoers is one such practice. Punishment, Protagoras declares, when inflicted rationally, is done for the sake of straightening out the wrongdoer on whom it is inflicted and deterring others who witness this punishment from acting as the wrongdoer did. When citizens punish a wrongdoer solely on account of his wrongdoing and without thought to the future, Protagoras observes, they do not act in a rational way. Punishment, after all, cannot undo what was already done. Plato returns to this theme, at least implicitly, in the Gorgias and then again much later and explicitly in the Laws. In the Gorgias, it occurs in the exchange that Plato reconstructs between Socrates and Polus.8 Examining Polus’s thesis that it is better for a man to escape punishment if he has acted unjustly than to be punished, Socrates leads Polus to agree that the unjust man who is justly punished is thereby benefited in being made more just and so less bad, from which, as he gets Polus to concede, the very opposite of Polus’s thesis follows. And in the Laws, Plato’s last dialogue, Plato offers a theory of punishment on which the point and justification of the practice lies in the good that punishment does to those on whom it is justly inflicted.9 Only by benefiting the wrongdoer, Plato argues, can the seeming ugliness of being punished be removed. Plato’s theory of punishment follows from general principles about how one ought to treat one’s fellows that he attributed to Socrates and presumably held himself. The principles serve as premisses of the argument Socrates makes in the Crito for remaining in prison and accepting the death penalty to which the Athenian jury had sentenced him.10 The dialogue is in its entirety a conversation between Socrates and his friend Crito, and the argument begins at the point at which Socrates secures from his friend assent to the proposition that the worst sort of harm a person can suffer comes from his corrupting his soul by doing wrong. The first principle Socrates asserts is that 8.  Ibid., 468e–​481b. 9.  Ibid., 859d–​864c 10.  Ibid., 49b–​49e.

6 52

256  From Psychology to Morality

one ought never to do wrong, from which he immediately draws a corollary that one ought not to do wrong in return for being wronged. Here perhaps is the earliest articulation in writing of the common maxim that two wrongs do not make a right. The second principle Socrates asserts is that harming a person is a kind of wrongdoing, and, having asserted this principle, Socrates then draws from it that one ought never harm a person whatever that person may have done to one. Because punishment entails inflicting pain or imposing some loss on a person, it appears to violate this second principle and to violate as well the corollary that forbids harming a person whatever he may have done to one. Hence, to justify punishment, given these principles and their corollaries, Plato must show that, despite appearances to the contrary, punishment does not harm the person on whom it is inflicted. His account in the Laws of how punishment, if just, benefits the wrongdoer on whom it is inflicted meets this requirement. Plainly, the principle that makes punishment problematic for Plato is the second one that Socrates affirms in the Crito, the principle that one ought never to harm another. Plato’s use of this principle and presumably Socrates’s before him are early instances in Western philosophy of the recognition of a principle of humanity as a core principle of ethics. A later, notable instance in ancient philosophy, notable because of the singular importance of the work in which it occurs as a source of knowledge of ancient ethics for medieval and modern Europe, is Cicero’s affirming, in De Officiis, duties of beneficence along with duties of justice as the chief components of social virtue.11 Nor obviously is recognition of a principle of humanity as an axiom of Western ethics solely a legacy of ancient Greek and Roman thought. Its place in Christian ethics in the form of the command to love one’s neighbor as oneself makes it, if anything, an even more profound progenitor as well as one that is independent of Greco-​Roman thought. The same command is found in Hebraic ethics among the laws of Leviticus.12 Unsurprisingly, then, modern secular ethics, given the joint influence of Greco-​Roman thought and Judeo-​Christian ethics on its development, includes principles of humanity at its core. Their being among the core principles of modern ethics ensured that what troubled Plato about punishment and led Jesus to reject lex

11.  I crib from Henry Sidgwick: “there is probably no ancient treatise that has done more than [Cicero’s] De Officiis to communicate a knowledge of ancient morality to medieval and modern Europe.” Sidgwick, Outlines of the History of Ethics, 95. 12. Leviticus 19:11.

2 75

Punishment and Proportionality: Part 2  257

talionis13 would continue to sustain well into the modern period the problem of punishment in moral philosophy. Typically, principles of humanity appear in modern secular ethics as principles prescribing aid to the needy and benevolence towards all. Secular philosophers who have expounded systems of commonsense ethics often identify these principles with the duties of benevolence, though sometimes they formulate them as wholly action-​oriented principles. W.  D. Ross, for instance, who sharply distinguished right action from praiseworthy motive, replaced the duty of benevolence in the system he expounded with two distinct duties, one of non-​ maleficence and the other of beneficence.14 Similarly, secular philosophers who have gone beyond systematizing the several duties that commonsense ethics traditionally recognizes and given a theory that unifies these duties by deriving them from some foundational principle include in their derivations the duties of benevolence or their action-​oriented analogs. Kant, for instance, derives such duties from the Categorical Imperative.15 Mill and Sidgwick ground them in the Greatest Happiness Principle, what Sidgwick sometimes calls the principle of Rational Benevolence.16 Punishment, then, because of the principles of humanity, is a general problem in Western ethics. Sidgwick puts the problem succinctly when he writes, “[I]‌f the retributive view of Punishment be strictly taken—​ abstracting completely from the preventive view—​it brings our conception of Justice into conflict with Benevolence, as punishment presents itself as a purely useless evil.”17 The problem, as Sidgwick notes, is sharpest when punishment is inflicted solely with the aim of giving wrongdoers what they deserve. At the same time, it is avoided when larger social ends, such as promoting public safety and preserving civil order, inform the infliction of punishment. The reason is

13.  See Mathew 5:38: “You have heard that it was said, ‘An eye for an eye and a tooth for a tooth.’ But I say to you, Do not resist one who is evil. But if any one strikes you on the right cheek, turn to him the other also.” 14.  See Ross, The Right and the Good, 21. 15.  See Kant, Groundwork for the Metaphysics of Morals, 40. 16.  See Mill, Utilitarianism, 49–​50; Sidgwick, The Methods of Ethics, 238–​241. 17. Sidgwick, Methods, 72. Sidgwick’s observation echoes Joseph Butler’s including, in his sermon on resentment, among the abuses of deliberate resentment, “when pain or harm is inflicted merely in consequence of, and to gratify, that resentment, though naturally raised.” Fifteen Sermons Preached at the Rolls Chapel, 129. See Sidgwick, Outlines, 193–​194. Butler’s version of the problem of punishment arises from the presence of resentment in human nature, given the Deity’s “perfect goodness . . . [and that] general benevolence is the great law of the whole of moral creation” (Fifteen Sermons, 121).

8 2 5

258  From Psychology to Morality

that when these larger ends inform punishment’s infliction, the act, though it still signifies hostility towards the wrongdoer by virtue of being an act of retaliation, also signifies a desire to promote the general happiness or common good. To act on such a desire is to act benevolently towards others generally. It is, in other words, to act in accordance with principles of humanity. Put differently, the duty of general benevolence is the duty to take the good of others both individually and collectively as one’s end.18 Accordingly, when punishment is inflicted with the aim of promoting public safety or preserving civil order, it satisfies this duty. When it is inflicted solely with the aim of giving wrongdoers what they deserve, its infliction clashes with the duty. Its only end, in this latter case, is to cause the recipient of punishment pain or loss in retaliation for something he or she did, and to inflict pain or loss on someone solely for the sake of retaliation violates the principles of humanity. Let us, following Sidgwick’s lead, call the view that the sole aim of legal punishment is to give wrongdoers what they deserve the strict retributive view. Defenders of this view would strenuously object to the charge that punishment, on their view, is morally problematic. Any strict retributive view, they would point out, is based on principles of justice, specifically, the principle that wrongdoers receive their just deserts. And justice, they would further observe, is the first virtue of social institutions. This means that its principles take precedence in the regulation of such institutions over all other principles. That is, when other principles, particularly principles of humanity, collide with principles of justice, these other principles must yield to the principles of justice. Hence, duties of benevolence cannot pose a problem for the strict retributive view. Inflicting punishment with the sole aim of giving wrongdoers what they deserve, being what justice requires, is morally justified notwithstanding that it may be incompatible with doing what duties of benevolence require. This defense of the strict retributive view echoes the standard defense of retributivism in the modern dispute between its defenders and defenders of the deterrence theory of punishment. In the modern dispute, the charge against deterrence theory is that it supports the punishment of innocent people in circumstances in which the deterrent effect of the punishment is sufficiently beneficial to the community to outweigh both the damage such injustice does to the institution and the harms it brings to its victims. Retributivism is not open to this charge since it is based on principles of

18.  I follow Kant’s account of the duty. See Kant, The Metaphysics of Morals, 200–​201, and Essay 9 above.

9 52

Punishment and Proportionality: Part 2  259

justice including, particularly, that only people guilty of wrongdoing should be liable to punishment. Hence, defenders of retributivism in the modern dispute have the advantage over defenders of deterrence theory of being uncompromisingly on the side of justice. The dispute with which we are concerned, however, is different from this modern dispute. It is over the question of proportionality, specifically, how to determine the severity of punishment so that it matches the gravity of the crime for which it is inflicted. This dispute arises within the theory of retributive justice. Once punishment is understood as pain or loss inflicted on someone in retaliation for something he or she did, then any practice of punishment that hasn’t been thoroughly corrupted or perverted is necessarily retributive and the kind of justice it delivers, when it is just, is retributive justice. Consequently, defenders of the strict retributive view have no advantage over the other disputants in being uncompromisingly on the side of justice. In this dispute, all the parties are uncompromisingly on justice’s side. They all agree with the requirement of retributive justice that the punishment for a crime match in its severity the crime’s gravity. Where they disagree is over how to determine that match. The defenders of the strict retributive view maintain that justice requires giving wrongdoers what they deserve. Their conception of justice is desert-​based. Their opponents, however, hold different conceptions of justice. Some maintain that justice as regards conduct regulated by the rules of an institution obtains when those rules represent fair terms of cooperation among the people whose conduct they regulate, where fair terms of cooperation are the terms these people would agree to under conditions in which they were free and equal. This conception of justice is democratically based. Legal punishment, being the primary sanction imposed within the criminal law, clearly falls within the scope of this democratically based conception. And in view of such an alternative to the desert-​based conception, one can see that the defenders of the strict retributive view beg the question when they maintain that justice requires giving wrongdoers what they deserve. This principle is too strong. At most one can say that justice requires giving wrongdoers their due. This latter principle is weaker than the principle the defenders of the strict retributive view invoke, since the notion of one’s due is broader than that of one’s desert. The former encompasses not only what one deserves, but also what one is owed, entitled to, or has coming to one according to the rules of an institution. Hence, inflicting punishment on wrongdoers according to rules that represent fair terms of cooperation among people whose conduct those rules regulate is giving them their due. It is doing what justice requires.

0 6 2

260  From Psychology to Morality

The upshot, then, is that legal punishment remains morally problematic as long as it is inflicted on wrongdoers solely for the sake of inflicting pain or loss on them in retaliation for their having acted wrongly. The strict retributive view takes retaliation for wrongdoing as the sole end of inflicting punishment. Hence, legal punishment, when informed by this view, is morally problematic. Inflicting pain or loss on people solely for the sake of retaliating against them violates principles of humanity, for one who retaliates against others just for the sake of retaliating has either not taken their good as an end or abandoned their good as an end. In either case, he or she acts contrary to the duty of benevolence towards humankind. Nor can the act be defended morally by maintaining that it is an act of justice and as such fulfills a moral requirement to which the duty of benevolence must yield when the two conflict. Such a defense fails because the conflict between a requirement of justice and the duty of benevolence that the defense presupposes is avoidable. Retaliating against wrongdoers for the sake of promoting public safety or preserving civil order can be an act of retributive justice and at the same time one that reflects one’s having taken the good of others collectively as one’s end. Legal punishment undertaken in pursuit of such larger social ends can therefore fulfill both the requirement of justice that punishment of an offense match in its severity the gravity of that offense and the duty of benevolence.

3. Sometime in the last century a subtly different way of explaining why punishment raises a moral problem emerged in the philosophical literature on the topic.19 Legal punishment, the contributors to this literature ever more frequently observed, entails doing things to people, like killing them, imprisoning them, and forcing them to part with portions of their property, that would ordinarily be violations of their rights. Hence, to keep the law from being morally problematic, these writers concluded, one must show that its inflicting punishment on wrongdoers does not violate their rights. The problem, on this explanation, is subtly different from the traditional problem because the worry it raises, which is about whether inflictions of punishment violate the rights of those who receive it, differs from the one the traditional problem raises, which is about whether such inflictions are incompatible with principles of humanity or are breaches of the general duty of benevolence. To ask whether an action violates someone’s rights is to ask 19.  The earliest suggestion of this alternative explanation of the problem can be found in Ross, The Right and the Good, 60.

261

Punishment and Proportionality: Part 2  261

whether it is unjust and not whether it is malicious or mean-​spirited (that is, whether it was done to cause the other pain or loss for its own sake or with complete indifference to its causing pain or loss). The shift in explanation, in other words, identifies the problem as one that falls entirely within the sphere of justice. The traditional problem, by contrast, arises when the two main spheres of the part of morality that concerns other-​regarding action, justice and benevolence, appear to collide. As a result of this shift, considerations of justice have eclipsed those of benevolence in many contemporary discussions of the morality of punishment. These discussions tend now to focus exclusively on the question of whether legal punishment is compatible with principles of justice and so to lose sight of the question of whether it is compatible with principles of humanity. Plainly, when discussions of the morality of punishment focus exclusively on the former question, the strict retributive view is no more subject to criticism for rendering punishment morally problematic than the competing views on which punishment serves larger social ends than retribution. If the apparent incompatibility is due to the possession by wrongdoers of rights against the state not to be subjected to hard treatment of the kind in which punishment consists, then punishment’s serving the state’s interest in upholding justice is as strong a ground for justifying its abridging those rights as punishment’s serving the state’s interest in promoting public safety or preserving civil order. Alternatively, if the apparent incompatibility is due to people’s generally having rights not to be subject to hard treatment of the kind in which punishment consists, then defenders of the strict retributive view are in as good a position as defenders of competing views to dissolve this apparent incompatibility by holding that wrongdoers, by virtue of their wrongdoing, forfeit their rights not to be subject to hard treatment of the kind in which punishment consists.20 Such forfeiture, after all, leaves the state free to inflict punishment on wrongdoers when the only moral barriers to inflicting it are rights people generally have not to be subject to hard treatment of the kind in which punishment consists. 20. What distinguishes the first alternative from the second is that it proceeds on the assumption that wrongdoers do not forfeit their rights not to be subjected to hard treatment of the kind in which punishment consists. In place of forfeiture, it takes the rights in question to be ones that may yield to more important public interests, but only on condition that their so yielding is in accordance with due processes of law. Such an understanding of moral rights is a common feature of the major declarations and bills of rights from the eighteenth century forward. See, for example, article 4 of the Declaration of the Rights of Man, article 29 of the Universal Declaration of Human Rights, and arguably article 5 of the United States Bill of Rights before Marbury v. Madison established that the U.S. Constitution was to be treated as written law that the American people enacted through ratification.

6 2

262  From Psychology to Morality

While either alternative represents blindness to principles of humanity in traditional discussions of whether punishment is morally problematic, the second alternative is especially liable to induce such blindness in those who advance it. The reason is that forfeiture of rights implies loss of the protection against harm the rights provide, and one is liable, in viewing wrongdoers as having lost such protection, to see them as no longer having any moral protection against being so harmed. One is liable, that is, to see a wrongdoer’s loss of general rights as relieving others of all duties not to harm him or her in the ways against which those rights protect. In effect, then, the emphasis that the second alternative places on rights and its account of what triggers their loss makes it all too easy for one to focus exclusively on the duties people have correlative to those rights and to lose sight of other duties they have that also protect individuals from the harms against which those rights protect. A recent article by Kit Wellman crisply illustrates this point.21 Wellman, at the beginning of his article, identifies the permissibility of punishment as the question that has received the most attention from writers on the topic. Subsequent remarks show that this question, on Wellman’s understanding of it, represents a different problem from the traditional one. Thus he writes, “Given that punishment, by definition, involves visiting hard treatment upon those who are punished, and persons typically have a right not to be subjected to this hard treatment, it would seem natural to conclude that the permissibility of punishment is centrally a question of rights.”22 The statement signals Wellman’s understanding of the question of punishment’s permissibility as turning solely on whether or not those who are liable to punishment have rights that the infliction of punishment would violate. Wellman, in other words, ignores the possibility of moral barriers to punishment other than those grounded in individual rights. This implication is then confirmed when he declares that forfeiture of the right not to be subjected to hard treatment of the kind in which punishment consists is “both necessary and sufficient to explain the permissibility of punishment.”23 As he sees it, the theory he is presenting, the rights forfeiture theory of punishment, is complete in itself and stands as a competitor to the traditional theories of punishment that justify the practice by appeal to its serving such ends as the promotion of justice or the prevention of crime. Indeed, Wellman argues, the traditional theories, while they explain the good the practice brings about, do not explain how

21.  See Wellman, “The Rights Forfeiture Theory of Punishment.” 22.  Ibid., 371–​372. 23. Ibid., 375.

6 23

Punishment and Proportionality: Part 2  263

bringing about that good through punishment is compatible with people’s having rights not to be subjected to hard treatment of the kind in which punishment consists.24 The rights forfeiture theory, by contrast, he maintains, explains how the practice is compatible with people’s having these rights and also makes clear that nothing besides the forfeiture of these rights need be cited to explain punishment’s permissibility. The error in Wellman’s argument is easily identified. The argument, as stated, is a non sequitur. Forfeiture of a right to others’ forbearance from treating one harshly does not imply the permissibility of their so treating one. Though one has forfeited the right, they may still have a duty to forbear from treating one harshly. Accordingly, it may still be impermissible for them so to treat one. Of course, Wellman may have rested his argument on an unstated assumption that removes this non sequitur. Specifically, he may have rested it on the assumption that there is a correlative right to every duty, which is to say that every beneficiary of a duty has a right to its performance. On this assumption, forfeiture of a right necessarily extinguishes any duty of which the person who holds the right is the beneficiary, and therefore it becomes permissible to treat him or her in ways that were not permissible when he or she had that right. Yet even if we take Wellman as resting his argument on this assumption, the argument is still faulty. While the attribution of rights to people in view of their being the beneficiaries of duties is common, it is not always sound. Sometimes, it does not follow from one’s being the beneficiary of a duty that one has a right to the performance of that duty. Not every duty, after all, correlates to a right that its beneficiary has to the good or service its performance provides. Hence, the error in Wellman’s argument remains. A wrongdoer’s forfeiture of his or her right not to be subjected to hard treatment of the kind in which punishment consists does not imply that it is permissible to impose such treatment on him or her since there may still be a duty not to subject him or her to such treatment that is not grounded in that right. A clear example of a duty whose beneficiary does not have rights to the good or service the performance of the duty provides is the duty of gratitude.25 24.  Wellman, in implicitly criticizing traditional theories for failing to explain the compatibility of punishment with people’s having rights not to be subject to hard treatment of the kind in which punishment consists, misses the possibility that the exponents of those theories understand the ascription of moral rights to people as presupposing a legal system in which, if it is just, these rights are subject to further specification in accordance with due processes of law. See note 20 above. 25.  See Kant, The Metaphysics of Morals, 203–​204; Sidgwick, Methods, 259–​261; Berger, “Gratitude,” 299–​301.

6 42

264  From Psychology to Morality

A neighbor sees you in need of help—​your car, say, is stuck in the snow—​and on his own volition he comes to your aid. With you at the wheel, he gives your car the needed shove from behind that gets it off the ice and rolling. You now have acquired a duty to return the favor. He, however, has not thereby acquired a right to your doing so. Similarly, a young couple announces their upcoming marriage and later receives gifts from friends of their parents, gifts that their parents’ friends intend as boosts to the couple’s setting up a household. Accepting the gifts, the couple then has a duty formally to acknowledge their having received these gifts and to thank those who gave them, but the gift-​givers do not thereby have a right to these formalities. Accordingly, you would be doing something wrong if you did not return the favor your neighbor did you when the opportunity for doing so arose, and the couple, too, would be doing something wrong if they failed to thank formally those from whom they received these gifts. But neither you nor this couple would be violating any rights that your neighbor or their parents’ friends had. Neither your duty nor the couple’s correlates to a right that the duty’s beneficiary has to its fulfillment. What accounts for the difference between these duties and duties correlative to which are rights that the beneficiaries of the duties have to their fulfillment? To understand Wellman’s error, given our supposition that his argument rests on the assumption that every duty correlates to a right its beneficiary has to its performance, we need to account for the difference between duties to which no rights are correlative and duties correlative to which are forfeitable rights. The difference is that the beneficiaries of the latter, unlike those of the former, are in a position to demand of the bearers of these duties that they perform them and to complain and press for redress from them if they fail to perform them. The rights in these cases are claims that their possessors have against anyone who owes them either forbearance from interference with that to which they have rights or positive service in the provision of that to which they have rights. As such, they entail ancillary rights to enforce these claims either directly or indirectly by enlisting people who are authorized to use force to compel those who have the duties to perform them. Such rights are thus claims that their possessors assert, exercise, insist on, stand upon, or invoke when such actions are necessary for securing from those who have the correlative duties the good or service their performance provides. When, by contrast, one is merely the beneficiary of a duty, when one, though the duty’s beneficiary, has no right to its performance, then one is in no position to compel its performance or to enlist others as agents authorized to use force to compel it. One who has the duty is of course subject to criticism and even blame for his or her

6 25

Punishment and Proportionality: Part 2  265

failure, but no action by the beneficiary to compel performance or secure redress would be justified.26 Forfeiture of a right to some good or service thus entails that one is no longer in a position to demand of those who have duties to provide that good or service that they provide it. Likewise, it entails the loss of the ancillary rights to enforce the claim in which that former right consisted. It does not, however, entail that one is no longer the beneficiary of a duty the performance of which would secure for one that good or service or a similar good or service. One is, to be sure, no longer the beneficiary of duties that are correlative to the right if the right is itself the ground of those duties. But one may still be the beneficiary of duties to which there is no correlative right and whose performance provides one with the same or a similar good or service as duties that are grounded in the right. Wellman’s error, in short, is that he ignores this possibility. One can thus agree with him that wrongdoers forfeit their rights not to be subjected to the kind of hard treatment in which punishment consists without also agreeing with him that nothing need be added to this observation to answer successfully the question of the permissibility of punishment. That we have a duty of benevolence towards all human beings if not all sentient ones, a duty to which there is no correlative right to our benevolence, raises by itself the question of the permissibility of punishment even if those on whom punishment is inflicted have forfeited their rights not be subjected to the hard treatment in which punishment consists. The question it raises is the traditional problem of punishment in Western ethics. Wrongdoers, having forfeited their right not to be subjected to such hard treatment, are in a position similar to that of the religiously intolerant in a liberal democracy. The religiously intolerant, by virtue of their attempts to rid society of the religions they hate, have lost the right to complain about attacks on their religion by those who wish to rid society of it. But though the religiously intolerant have no right to complain, the question remains whether their religion should nevertheless be tolerated in a liberal democracy.27 Similarly, though a wrongdoer has no right to complain about being punished for his or her wrongdoing—​indeed, even if punishing him or her violates none of his or her rights—​the duty we have to promote the good of human beings generally nevertheless puts the permissibility of our practice of punishing wrongdoers into question. Unfortunately, this question gets lost 26.  See Feinberg “On the Nature and Value of Rights,” 249–​250, and Wasserstrom, “Rights, Human Rights, and Racial Discrimination,” 630–​631. 27.  See Rawls, A Theory of Justice, 216–​221.

62

266  From Psychology to Morality

when the moral problem that punishment raises is explained as the problem of punishment’s apparent violation of the rights people generally have not to be subjected to hard treatment of the kind in which punishment consists.

4. I have argued for the thesis that legal punishment, when informed by the strict retributive view, is morally problematic in a way that such punishment, when undertaken in pursuit of larger social ends, such as public safety and civil order, is not. At the same time, I grant that the former does not necessarily bring a schedule of more severe penalties than the latter. One cannot, after all, infer from the thesis for which I have argued that naive lawmakers, who embrace the strict retributive view and are guided by an interest in inflicting pain or loss on wrongdoers solely for the sake of retaliating against them, enact laws that systematically impose more severe punishments than do the laws more sophisticated lawmakers, who reject the strict retributive view and are instead guided by an interest in promoting public safety or preserving civil order, enact, for there is no a priori way of knowing which criterion of severity will yield the severest punishments. Still, if we consider the large role that the public’s hostility towards crime has had in the last forty years in creating in America laws that impose exceptionally harsh punishments, we have reason to think that at least in practice more sophisticated lawmakers, given their attitude to such hostility, would enact laws that impose more lenient punishments. Being focused on the larger social ends of public safety and civil order, they would, as a result, resist, when determining for each type of crime the severity of the punishments to be imposed for crimes of that type, indulging the powerful retributive impulses that contemplation of such crimes excites in them or in the citizens they represent. Indeed, lawmakers who are conscious of being restrained by principles of humanity should recognize that whatever hostility they or the citizens they represent have towards people who commit crimes of that type does not provide a good reason to increase the severity of the punishment beyond what promotes public safety or preserves civil order. Certainly, then, the schedule of punishments such lawmakers enact will be less severe than if they indulged these retributive impulses. And in view of how severe the schedule of punishments is that lawmakers who in recent years did indulge such impulses or caved in to the public’s punitive demands that expressed them, the schedule of punishment they enact will almost certainly be less severe as well.

6 2 7

Punishment and Proportionality: Part 2  267

Take, for example, the harshest of these recent laws, the laws popularly known as “three strikes and you’re out.” Their main rationale, as implied by their moniker, is to incapacitate habitual offenders, and the specific targets of the laws are offenders with a history of engaging in violent crime. The laws typically require that such offenders, upon being convicted of a third serious felony, be punished by being imprisoned for decades if not life. Ostensibly, these laws, in view of their main rationale, were enacted to prevent violent crime, since imprisoning a habitual violent offender for decades means removing from the general population someone who, whenever he does live as a member of that population, is much more likely than the average member to engage repeatedly in violent crime. At the same time, incapacitation through imprisonment is also a way of expelling someone from society for the duration of his time in prison, and such expulsion for decades is clearly an expression of great hostility towards the expellee. The true motives behind the three strikes laws could therefore be hostility and resentment towards habitual violent offenders, notwithstanding to the contrary that the laws’ supporters commonly champion them as tools of crime prevention. And one clear indication that hostility towards violent offenders, and not an interest in promoting public safety, is the true motive behind them is that the laws themselves, in the severity of the punishment they impose on repeat offenders, go far beyond what incapacitation, as a means of preventing violent crime, can achieve. It has long been known that most violent crimes are committed by adolescent boys and young men, and the likelihood of a man’s committing such a crime after he has reached, say, the age of fifty, whether or not he committed violent crimes in the past, is vanishingly small. Hence, imprisoning violent offenders for terms that expire only when these offenders have long since ceased to pose a serious risk of repeating their violence does not significantly advance the interest in public safety through incapacitation. Shorter terms would work just as well. To be sure, supporters of these laws also declare that the laws are effective tools of crime prevention by virtue of their deterring people, and especially convicted felons, from committing violent crimes. Yet this claim too can be challenged. It is entirely speculative. There is no solid evidence that threatening violent and potentially violent offenders with imprisonment for decades will significantly lower the incidence of those crimes for which such severe punishment is not already what the law imposes from what it would be with a smaller enhancement of the penalty imposed for such crimes on offenders whose third conviction triggers a three strikes law. Indeed, conventional wisdom tells us that at some point along the curve that charts increases in the severity of punishment for a given type of crime with their correlative

6 82

268  From Psychology to Morality

deterrent effects, the value of the latter steeply declines relative to that of the former, which implies that the laws’ threats will not significantly lower the incidence of these crimes. Nor is there solid evidence that the threats will have a significantly greater deterrent effect on offenders who have already, once or twice, been convicted of a violent crime than a smaller enhancement of the penalty imposed on thrice-​convicted offenders whose third conviction triggers the law. Indeed, a study of the deterrent effect of California’s three strikes law on first-​and second-​time offenders who had been convicted of violent crimes showed no such effect on the former and only minimal effect on the latter.28 The belief that the laws through their deterrent effect will significantly lower the incidence of violent crimes generally or that first-​and second-​time violent offenders will be significantly less likely to commit a second or third serious offense appears, then, to be largely wishful thinking that merely rationalizes the hostility behind these laws. Sophisticated lawmakers, we may conclude, being conscious of how such hostility can be the source of morally problematic sentencing laws and thus immune to its influence, would not author or back laws as harsh as these. Three strikes laws are, no doubt, extreme examples of laws that, owing to the influence of the public’s hostility towards criminals in their enactment, impose morally problematic punishments. Nonetheless, the lesson they teach is clear. To render legal punishment morally unproblematic, lawmakers must avoid being guided by their retributive impulses or the public’s hostility when determining for any type of crime the severity of the punishment to be imposed for crimes of that type. They must be guided instead by larger social ends such as public safety and civil order. In being so guided, they would remain faithful to principles of humanity as well as to those of justice and, in practice at least, enact laws imposing more lenient punishments than those that have characterized American penal law for the past forty years.

Bibliogra p h y Berger, Fred. “Gratitude.” Ethics 85, no. 4 (1975): 298–​309. Butler, Joseph. Fifteen Sermons Preached at the Rolls Chapel. London: G. Bell, 1914. Daube, David. Studies in Biblical Law. Cambridge: Cambridge University Press, 1947. Feinberg, Joel. “On the Nature and Value of Rights.” Journal of Value Inquiry 4, no. 4 (1970): 243–​257. Hume, David. A Treatise of Human Nature. Edited by L. A. Selby-​Bigge. Oxford: Oxford University Press, 1978.

28.  See Zimring, Hawkins, and Kamin, Punishment and Democracy.

9 6 2

Punishment and Proportionality: Part 2  269 Kant, Immanuel. Groundwork for the Metaphysics of Morals. Edited and translated by Allen Wood. New Haven: Yale University Press, 2002. Kant, Immanuel. The Metaphysics of Morals. Translated by Mary Gregor. Cambridge: Cambridge University Press, 1996. McMahan, Jeff. “Proportionate Defense.” Journal of Transnational Law and Policy 23 (2014): 1–​36. Mill, John Stuart. Utilitarianism. 2nd ed. Edited by George Sher. Indianapolis: Hackett, 2001. Moore, Michael. Placing Blame. Oxford: Oxford University Press, 1997. Plato. Complete Works. Edited by John Cooper. Indianapolis: Hackett, 1997. Rawls, John. A Theory of Justice. Cambridge, MA: Belknap, 1971. Ross, W. D. The Right and the Good. Oxford: Oxford University Press, 1928. Sidgwick, Henry. The Methods of Ethics. 7th ed. Indianapolis: Hackett, 1981. Sidgwick, Henry. Outlines of the History of Ethics. 5th ed. Indianapolis: Hackett, 1988. Wasserstrom, Richard. “Rights, Human Rights, and Racial Discrimination.” Journal of Philosophy 61, no. 20 (1964): 628–​641. Wellman, Christopher H. “The Rights Forfeiture Theory of Punishment.” Ethics 122, no. 2 (2012): 371–​393. Zimring, Franklin E., Gordon Hawkins, and Sam Kamin. Punishment and Democracy: Three Strikes and You’re Out in California. New York: Oxford University Press, 2001.

0 72

271

Index

aggressivity, 67, 70, 74–​75, 78–​81, 226 allegiance, 216–​217, 226, 228–​230, 241–​242 altruism, reciprocal, 108–​109 ambivalence, 80, 227 amorality, 136–​138 analogical reasoning, 176 anger, 9, 18, 20, 23, 36, 40, 47, 54, 57, 84, 87–​89, 111, 157–​158n, 222–​ 224, 227–​229, 252, 254. See also resentment of voters, 232–​233, 235–​236 Anscombe, G. E. M., 135, 160n, 168n anxiety, 67–​70, 74–​79 appraisal, 40–​42 (see also judgment: evaluative) Aquinas, Thomas, 2–​3, 72n Aristotle, 2, 7, 11, 22, 70, 72n, 107–​108, 159, 160n, 168n, 177n Augustine, 2, 4, 72 Austin, J. L., 224n

authority, 9, 66, 74–​78, 80, 85, 102, 145, 219, 227, 231, 233, 241 internal, 74, 80–​81 moral, 76–​77, 120–​124, 143 parental, 66, 74–​77, 78, 80, 144, 214, 227 autism, 137–​138 Ayer, A. J., 142 Baier, Kurt, 158–​160, 164n, 165–​167, 169–​170 Bain, Alexander, 49 Bellamy, Carol, 207n benevolence, 3, 8, 18, 44, 47, 98, 118, 131, 227 duty of, 8, 10, 203, 209–​213, 257–​258, 260–​261, 265 Benn, S. I., 231n, 232n Bennett, Jonathan, 101n, 126n Bentham, Jeremy, 152, 195, 199n, 240n Berger, Fred, 263n

271

27

272 Index Bernstein, Nina, 200 Brentano, Franz, 60 Breuer, Josef, 61 Brown, Thomas, 49 Brücke, Ernest, 60 Butler, Joseph, 186, 257n Carnap, Rudolph, 140–​142 Carroll, Lewis, 174–​175 Categorical Imperative, 75, 257 categorical imperatives, 137, 146 Charcot, Jean, 60–​61 Church, Alonzo, 173n Cicero, 256 Cohn-​Vossen, Stephan, 143 conscience, 1–​2, 5, 33, 66, 69–​70, 74–​77, 80–​81, 83–​86, 99, 107, 132, 136–​139, 226–​227, 229–​230 Damasio, Antonio, 29n, 31–​32, 34, 38–​40, 41n Darwall, Stephen, 84–​87, 89, 102–​103, 119–​124, 174, 176 Darwin, Charles, 5, 53n Daube, David, 249n Davidson, Donald, 151–​156, 158, 160n, 169n Davis, N. Ann, 199 deliberation, 39, 66, 156–​161, 165, 167–​168, 171–​173, 177–​178, 206 Aristoelian conception of, 7, 159, 166–​168, 169n, 177n Descartes, René, 24, 47, 51, 56, 72 determinism, 89, 91, 93, 96, 103–​105, 113, 116 Dixon, Thomas, 43 Dolinko, David, 239n Donagan, Alan, 195 Duff, R. A., 231n Edel, Abraham, 140n ego, 62, 66–​70, 74, 78, 80 egocentricity, 84, 86, 89, 99, 144 egoism, 3, 7–​8, 183–​185, 187–​188, 191–​194 psychological, 3

emotions, 5–​7, 11–​12, 14–​15, 50–​52, 62, 146, 157–​158n. See also passion cognitive theory of, 13, 21–​29, 38 identified with feelings, 13, 16–​18, 29–​34, 40–​42, 52–​57 identified with evaluative judgment, 21–​24 identified with perception, 24–​29, 32–​34, 40–​42 and punishment, 223, 225–​230 retributive (see reactive attitude) unconscious, 19–​21, 30, 78n empathy, 7, 99, 132, 137–​138, 145–​146, 227. See also sympathy: Hume’s account of empiricism in psychology, 5, 14, 17–​18, 50–​51, 53, 55 Enoch, David, 136 feeling, 11, 13–​16, 18–​22, 30–​31, 35–​37, 57, 61, 67, 71, 110, 129–​134, 146, 148, 254 of bodily changes, 16–​17, 21–​22, 29–​35, 38–​44, 49n, 52–​54, 56 of guilt, 33–​34, 78n, 85, 95, 116, 120–​121, 136, 226–​228 (see also guilt: sense of) necessarily conscious, 19–​20, 78 Feinberg, Joel, 232n, 265n Fellow feeling, 98, 118. See also empathy Fletcher, George, 231n Flew, Anthony, 231n, 232n, 234n Forgiveness, 76, 85, 90, 101, 106, 228 Frankena, William, 136 Frankfurt, Harry, 217 freedom of the will, 2–​4, 6–​7, 73, 81, 103–​104 Freud, Sigmund, 5–​6, 19–​21, 29–​30, 32, 35–​36, 59–​72, 74–​82, 226–​227 Gandhi, Mohandas, 215 Gauthier, David, 195 Gert, Bernard, 8, 195–​215, 217 Gewirth, Alan, 195 Goldie, Peter, 35–​37

273

Index 273 gratitude, 85, 90, 92, 101, 106, 109–​111, 113, 119, 125, 223 duty of, 263 in chimpanzees, 109–110, 119 Grotius, Hugo, 234n guilt, 78–​79, 85–​87, 89, 94, 120–​121, 124, 222, 241, 258–​259 expiation of, 228–​230 sense of, 20, 65, 69–​70, 74–​81, 227. See also feelings: of guilt Hamilton, Patrick, 84 Hare, R. D., 85, 138–​139n Harman, Gilbert, 195 Hart, H. L. A., 225n, 231n, 232n Hempel, Carl, 152n Highsmith, Patricia, 84 Hilbert, David, 140, 143, 145 Hills, David, 42 Hitchcock, Alfred, 83–​85, 88n, 100 Hobbes, Thomas, 3, 8, 71, 72n, 108, 235n Hoffman, Martin, 100n, 227 Holmes Jr., Oliver Wendell, 119 Hume, David, 3–​7, 14–​19, 31, 40, 43–​55, 72, 95, 97–​98, 108, 115, 117–​118, 142, 146–​148, 195–​196, 216 on calm passions, 6, 18–​19, 31, 44–​49, 55, 71, 110 on inertness of reason, 46–​48, 166, 170–​171 on morality’s essential practicality, 135–​137, 139, 160n sentimentalism, 10–​12, 129–​134, 254 id, 66, 68 ideals, 8–​9, 65–​66, 76–​77, 196–​204, 207, 213–​218 indignation, 85, 87–​89, 94, 96–​98, 101, 111, 114–​118, 120–​121, 124, 128, 254 instinct, 4, 5, 17n, 39, 44, 46, 53–​55, 64–​70, 74–​75, 78–​82, 91 death, 67, 70, 75, 78–​82 intentionality, 13, 21–​22, 29–​32, 34–​37, 90, 119, 121–​122, 126–​128 internalization, 69, 74, 80, 226–​227 introspection, 57

James, William, 6, 14–​17, 20–​21, 29–​33, 35, 38–​42, 50–​57 Jesus, 256–​257 Jones, David, 75n judgment, 44, 133–​135 evaluative, 22–​25, 29, 38–​39, 41–​42. See also moral judgment justice, 1–​4, 84–​85, 89, 107, 108n, 135, 188–​189 retributive, 10, 232, 234n, 236, 238–​240, 242–​247, 249–​255, 257–​260, 266–​268 sense of, 91, 125, 227 Kagan, Shelly, 195 Kamm, Francis, 164n Kant, Immanuel, 8, 72–​75, 77–​78, 81, 136–​139, 146–​148, 176, 195, 203–​204, 210–​213, 257, 258n, 263n Kekulé, August, 163 Kennett, Jennette, 137–​138 Köhler, Wolfgang, 110 Korsgaard, Christine, 7, 170–​171, 172n, 176 Lalande, Andre, 139 Lange, C. G., 50n Lear, Jonathan, 19 Levy-​Bruhl, Lucien, 139 libido, 64–​65, 67–​69, 79–​81 Locke, John, 14–​16, 50, 53, 234–​235 Lyons, William, 38 Mabbott, J. D., 234n Maibolm, Heidi, 138 McGeer, Victoria, 138 McMahan, Jeff, 250n metaethics, 136–​137, 139–​142 Mill, John Stuart, 257 Moore, G. E., 140n Moore, Michael, 250n moral judgment, 7, 84–​89, 99–​100, 120–​121, 130, 132–​139, 141–​142, 144–​146 Gert’s typology of, 196–​197, 213 Kant’s account of, 136–​139, 146–​148 Sidgwick’s account of, 183–​184, 187–​188, 191–​192

4 72

274 Index moral responsibility, 2, 6–​7, 84–​85, 89, 96–​99, 101–​104, 114, 116–​119, 121, 123–​124, 128, 228, 243, 246, 253 morality, 4–​5, 8, 10–​12, 63, 75, 81, 133, 135, 144–​147, 181n, 195–​197, 203, 206, 209, 215, 217–​218, 224, 238, 261 commonsense, 193–​194 Morris, Herbert, 225n

infliction by superego, 69–​70 need for, 76–​80 parental, 76–​77, 227–​228, 232–​233 proportionality as a condition of justice in, 10, 238–​244, 246–​247 retributive theory of, 252, 257–​261, 266 rights forfeiture theory of, 262–​265 Putnam, Hilary, 27n Quine, W. V. O, 43n

Nagel, Thomas, 173 naturalism, 1–​12, 70–​75, 136 neurosis, 61, 63, 68–​69 Nichols, Sean, 138–​139n Nietzsche, Friedrich, 5, 75, 81, 226–​227, 229–​230 Nussbaum, Martha, 22, 24 Oedipus complex, 64–​65, 69, 74, 79–​80 ‘ought’. See moral judgment: Sidgwick’s account of Pascal, Blaise, 186 passions, 6, 11, 15–​19, 31, 40, 42–​50, 52–​56, 71–​73, 135, 147, 184, 196 Paul, 72 Piaget, Jean, 7, 144–​145 Pitcher, George, 38 Plato, 1n, 2–​3, 26n, 54, 71, 73, 107, 188–​189, 214, 254–​256 Posner, Richard, 240n practical reason, 7–​8, 10, 146–​147, 150, 160, 165–​167, 170–​178 Sidgwick’s treatment of, 179–​195 practical syllogism, 160n, 168n Price, Richard, 195 Prichard, H. A., 188–​189 Prinz, Jesse, 29n, 32, 38, 40–​42 Protagoras, 1–​2, 107, 254–​255 psychoanalysis, 20, 61–​66, 70–​71 psychopathy, 83–​89, 99, 136–​139 punishment, 9–​10, 103–​105, 114, 219–​226, 228–​238 deterrence theory of, 240, 242–​247, 258–​259 in Gert’s theory, 8, 198n, 204–​205, 211–​212

rational agency, 119–​121, 125–​127, 150, 161, 165, 171, 172–​173n Rawls, John, 91, 225n, 227, 241n, 245, 265n reactive attitude, 6, 9, 85, 89–​96, 98–​103, 105–​128 reason, 3, 5–​6, 10–​12, 44, 46, 48–​49, 55, 71–​72, 129, 133–​134, 138, 166–​167, 175–​176, 196 Kant’s account of, 72–​73, 75, 137–​138, 146–​147, 176 practical (see practical reason) theoretical, 160n, 167, 172 reasons for action, 7, 119, 121–​122, 128, 150–​151, 152n, 158–​161, 164–​165, 170, 185–​186 Davidson’s account of, 151–​156, 158, 169n as grounds, 161, 163n, 164 moral, 86–​87, 121–​123, 168 normative force of, 154–​156, 158–​161, 168–​169 theoretical v. practical, 166–​169 Williams’s account of, 150–​161, 164–​172 Reid, Thomas, 47–​49 reparations, 9, 76, 106, 228–​229, 243, 246 repression, 19–​20, 62–​69 resentment, 6, 44, 84–​90, 92, 94, 96, 98–​ 99, 101, 106, 112–​114, 116, 119–​128, 223, 228–​229, 252, 257n, 267 in chimpanzees, 109–​111, 119 retaliation, 9–​10, 220, 222–​225, 229, 236–​238, 251–​252, 254, 258–​260 retributivism. See punishment: retributive theory of

275

Index 275 revenge, 84, 234n rights, 122, 135, 260–​261, 265–​266 Constitutional, 220, 222, 261n and correlative duties, 85, 262–​265 forfeiture of, 261–​263, 265 Robinson, Jenefer, 32, 38, 41n Ross, W. D., 257, 260n Rousseau, Jean-​Jacques, 4–​5, 91–​92, 108 Rozin, Paul, 23n Sayre-​McCord, Geoffrey, 134 Scanlon, T. M., 163n, 168–​170 Schacter, Stanley, 21n Schneewind, J. B., 179–​182, 193–​194 Scholastics, 56 seeing as, 25–​28 Sidgwick, Henry, 7–​8, 179–​194, 256n, 257–​258, 263n Singer, Jerome, 21n Singer, Peter, 209n Socrates, 1–​2, 188, 214, 255–​256 Solomon, Robert, 21n, 22, 38 Spinoza, Baruch, 3 Stephens, James Fitzjames, 223n Stevenson, Charles, 142 Stoics, 22, 24 Strawson, P. F., 6–​7, 85, 89–​107, 112–​128, 214n, 244n Stroud, Barry, 134n superego, 67–​70, 74–​81 sympathy, 17n, 54, 93, 95, 99, 113, 115 Hume’s account of, 131–​135

Tarski, Alfred, 140 Taylor, Telford, 235n Thoreau, Henry David, 214–​215 Toulmin, Stephen, 158–​159 Trivers, Robert, 108–​109 trust, 107, 109, 226 Twain, Mark, 234n unconscious, 5, 19–​21, 30, 35–​36, 61–​67, 71–​72, 78–​79 dynamics of, 5, 62, 66, 72 emotions (see emotions: unconscious) feelings (see feeling: unconscious) thought, 19, 62, 64–​65, 67, 71–​72 Velleman, J. David, 75n Vignemont, Frederique de, 138 Waal, Frans de, 108–​111 Wallace, R. J., 84n, 102–​103, 119–​125 Warnock, Geoffrey, 195 Wasserstron, Richard, 265n Watson, Gary, 84n Wellman, Christopher, 262–​265 Williams, Bernard, 7, 10, 150–​161, 164–​173, 176–​178 Wittgenstein, Ludwig, 12, 101, 151n Zaibert, Leo, 232n Zimring, Franklin, 268n