Morality and the Emotions

Table of contents :
List of Figures/Tables
List of Contributors
Part I
1. Craving the Right: Emotions and Moral Reasons
2. Emotions and the Categorical Authority of Moral Reason
3. Self-Love and Practical Rationality
4. The Nature and Morality of Romantic Compromises
Part II
5. Values and Emotions: Neo-Sentimentalism’s Prospects
6. Emotions, Perceptions, and Reasons
7. Conscience: What is Moral Intuition?
8. Empathy and Empirical Psychology: A Critique of Shaun Nichols’s Neo-Sentimentalism
Part III
9. Reactive Attitudes Revisited
10. Responsibility and Dignity: Strawsonian Themes
11. Guilty Thoughts
12. Moral Sentiment and the Sources of Moral Identity
13. On Alienated Emotions

Citation preview

Morality and the Emotions

This page intentionally left blank

Morality and the Emotions edited by

Carla Bagnoli



Great Clarendon Street, Oxford ox2 6dp 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 in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York # the several contributors 2011 The moral rights of the authors have been asserted Database right Oxford University Press (maker) First published 2011 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, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose this same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data Data available Typeset by SPI Publisher Services, Pondicherry, India Printed in Great Britain on acid-free paper by MPG Books Group, Bodmin and King’s Lynn ISBN 978-0-19-957750-7 10 9 8 7 6 5 4 3 2 1

Contents List of Figures/Tables List of Contributors Introduction Carla Bagnoli

vii viii 1

Part I 1. Craving the Right: Emotions and Moral Reasons Patricia Greenspan


2. Emotions and the Categorical Authority of Moral Reason Carla Bagnoli


3. Self-Love and Practical Rationality Edward Harcourt


4. The Nature and Morality of Romantic Compromises Aaron Ben-Ze’ev


Part II 5. Values and Emotions: Neo-Sentimentalism’s Prospects Christine Tappolet


6. Emotions, Perceptions, and Reasons Michael S. Brady


7. Conscience: What is Moral Intuition? Paul Thagard and Tracy Finn


8. Empathy and Empirical Psychology: A Critique of Shaun Nichols’s Neo-Sentimentalism Lawrence Blum


Part III 9. Reactive Attitudes Revisited John Deigh 10. Responsibility and Dignity: Strawsonian Themes Bennett W. Helm

197 217



11. Guilty Thoughts Angela M. Smith


12. Moral Sentiment and the Sources of Moral Identity Jacqueline Taylor


13. On Alienated Emotions Talbot Brewer




List of Figures/Tables Figures 1.1. Discounting a reason against postponement


1.2. An emotional reason against discounting


7.1. The EMOCON model of emotional consciousness


10.1. Pattern of emotions focused on tomatoes, constituting import


10.2. Pattern of reactive attitudes focused on being a person (solid arrows) and subfocused on Alice (broken arrows), constituting respect for her as a part of a more general respect for persons generally


Table 7.1. Predicted appraisal patterns for some major emotions


List of Contributors Carla Bagnoli is Professor of Philosophy at the University of Wisconsin, Milwaukee, and Professore Straordinario di Filosofia Teoretica University of Modena. Aaron Ben-ze’ev is Professor of Philosophy at the University of Haifa. Lawrence Blum is Distinguished Professor of Liberal Arts and Education at Boston University. Michael S. Brady is Professor of Philosophy at the University of Glasgow. Talbot Brewer is Associate Professor of Philosophy at the University of Virginia. John Deigh is Professor of Philosophy and Law at the University of Texas. Tracy Finn is a PhD candidate at the University of Waterloo, Ontario. Patricia Greenspan is Professor of Philosophy at the University of Maryland. Edward Harcourt is Lecturer in Philosophy at Keble College, Oxford. Bennett W. Helm is Professor of Philosophy at Franklin & Marshall College. Angela M. Smith is Associate Professor of Philosophy at Washington and Lee University, Virginia. Christine Tappolet is Professor of Philosophy at the University of Montreal. Jacqueline Taylor is Associate Professor of Philosophy at the University of San Francisco. Paul Thagard is Professor of Philosophy at the University of Waterloo, Ontario.

Introduction Carla Bagnoli

The place of emotions in morality is the subject of widespread and divisive philosophical controversies. This is hardly a peculiarity of present debates; as the history of moral philosophy shows, the relation between morality and the emotions has always been problematic. On the one hand, emotions are more often names of vices rather than virtues, as in the case of envy, or jealousy. When they are not regarded as vices in themselves, emotions are taken to represent a pervasive and persistent source of obstacles to morality, as in the case of self-love. Some virtues, such as prudence, temperance, and fortitude require or simply consist in the capacity to counteract the disruptive effect of emotions. The basic worry is that emotions interfere with the deliverances of reason, and often provide motives that are in competition with morality. A more radical worry is that emotions undermine our status of rational agents insofar as we are not in control of them, but we are possessed by them. Emotions undermine autonomy, which is a requirement for rational agency. On the other hand, venerable traditions of thought place emotions such as respect, love, and compassion at the very heart of morality. Many ethical theories take emotions to ground general duties toward humanity as well as special obligations that arise out of personal relations such as friendship, cultural kinship, or family ties. Emotions such as love and compassion are perceptions of the value of others. More generally emotions seem to play a distinctive role in practical reasoning, by supplying motives and reasons for action. Emotions such as blame, guilt, and shame speak the voice of moral conscience. Such emotions are recognizably central to the functioning of morality as a general normative practice. They serve as sanctions against moral transgression and provide incentives to abide by moral norms. These emotions constitute the very ‘stuff of moral life’, and are intrinsically connected to our identity and character. To discount them would be to ignore important aspects of how morality relates to our humanity. It seems that a plausible account of the place of emotions in morality should start by recognizing their diversity. Arguably, this is to be expected, since the term ‘emotions’



may not indicate a homogeneous category of states but group disparate phenomena.1 To be fruitful, the exploration of the diversity of emotions should proceed from two complementary perspectives. First, it should be considered from the point of view of theories about what emotions are and what place they hold in the topography of the mind. Second, it should be investigated from the point of view of ethical theories, which propose the normative standards for moral assessment and specify the conditions under which emotions contribute to morality. This volume is born out of the conviction that philosophy provides a distinctive approach to the cluster of problems about the emotions and their relation to morality. The task of this Introduction is to motivate this conviction by highlighting some aspects of the novelties and peculiarities that characterize current debates.

1. The Retrieval of Emotions in Moral Philosophy 1.1 Bernard Williams’ diagnosis of the neglect of emotions In his seminal essay ‘Morality and the Emotions’, Bernard Williams complains that contemporary moral philosophy does not make room for emotions ‘except perhaps for recognizing them in one of their traditional roles as possible motives to backsliding, and thus as potentially destructive of moral rationality and consistency’ (Williams (1966/ 1973): 207). At the time Williams writes, there are already some important albeit isolated attempts to investigate ‘emotions’ as an autonomous category, distinct from ‘desires’ and ‘beliefs’. For instance, Edward Bedford (1957) and Anthony Kenny (1963) recognize the role of emotions in the explanation of action, and consider their relation to the will and to reason. But Williams is right that, by and large, post-war analytic moral philosophers discount the role of emotions in moral life and deny that they could be sources of moral knowledge. According to Williams, the neglect of emotions depends on three main factors: a simplistic view of emotions as blind causes for action, a legalistic conception of moral rationality associated with Kantian ethics, and a correspondingly simplified view of moral language as prescriptive or expressive. In the last three decades, the emotions have become one of the main foci of philosophical attention, and it is useful to reconsider Williams’ assessment. Is this resurgence of interest in the emotions a sign that the causes that Williams identified as the reasons for neglect have been removed? Or was Williams’ diagnosis partial, or mistaken? An interesting aspect of Williams’ diagnosis is that it considers the relation of the emotions to morality from a double perspective, which combines a view about the ontology of emotions and their place and role in the topography of the mind, with a

1 The claim that emotions are a homogeneous category may seem to commit to the stronger claim that they represent a natural kind, cf. Griffith (1997), Panksepp (2000). However, supporters of the unity of emotions do not generally defend emotions as a natural kind.



view about morality, its language, scope, and function. Williams’ own work has contributed in a large measure to making room for emotions in practical reasoning and restoring them to their central place in morality. However, several other factors lie behind the dramatic changes in the philosophical attitude toward emotions that we are witnessing. The purpose of section 1 is to account for the circumstances of such changes and highlight their philosophical implications. 1.2 One step back: Iris Murdoch and the demand for moral psychology Williams’ dissatisfaction with a style of ethical theorizing that ignores moral psychology becomes widely shared in the 1980s. His words echo those of isolated but prominent figures of post-war analytic philosophy. The first astounding thesis of G.E.M. Anscombe’s article ‘Modern Moral Philosophy’ is that moral philosophy should be set aside, because it is not profitable to do it without an adequate philosophy of psychology (Anscombe (1958); cf. Thompson (2008): 5–7). Starting from a similar analysis, Iris Murdoch makes the ‘inner moral life’, and particularly the emotions, the center of her moral philosophy (Murdoch (1956, 1957, 1970)). It is mainly thanks to her efforts that moral psychology becomes a firm point in the agenda of analytic philosophy. In her early essays, Murdoch argues that this narrow view of the scope of ethics has very specific roots. She identifies logical positivism as the main cause of the disinterest in the ‘life of the mind’ that characterizes post-war analytic ethical theory. Logical positivism considered the mind elusive and favored an ethical theory that dispenses with any hypotheses about human nature and the mind (Murdoch (1956): 34; cf. also (1957, 1970)). For fear of metaphysics, analytic moral philosophy discarded both metaphysical concepts and psychological concepts alike, so that the idea of ‘action’ was reduced to outward performance, ‘without any transcendent background’ (Murdoch (1957): 105). These critical remarks about analytic ethics are condensed in her more famous The Sovereignty of Good (1970), which argues in favor of moral psychology and attempts to refocus the philosophical debates on the exploration of moral life. The negative part of her argument targets the existentialist and behaviorist approaches to ethics, which account for action as an outward performance. Following this model, moral philosophers set out to treat morality independently of any moral psychology or philosophy of mind, and they see their philosophical mission as that of providing norms for the performance of rational actions. Kantian ethics is elected as the best candidate for carrying out this alleged task of moral philosophy. Murdoch’s critique of Kantian ethics must sound uncharitable to the reader familiar with existentialism, and with current Kantian action theory, for reasons that I will present in section 1.5 (Moran (2002, 2011)). However, Murdoch’s actual polemical targets are Richard M. Hare and Stuart Hampshire as the representatives of Kantian ethics. She pairs their (allegedly Kantian) conception of rational agency as performance with a view of moral language as emotive, prescriptive, or persuasive (Murdoch (1957): 102–5). This view of moral language is appealing because of its simplicity. Murdoch’s objection is that simplicity is



the trade-off for a depleted moral vocabulary that lacks crucial moral-psychological concepts such as virtue, will, and the emotions. Without these concepts moral philosophy misses the philosophical resources for exploring our moral experience. It becomes incapable of recognizing the nuances and varieties of moral life, and thus it becomes impractical (Antonaccio (2001), Lear (2004), Moran (2011), Bagnoli (2011)). Murdoch’s polemic against the limitations and ‘philosophical blindness’ of this account of concepts has inspired more sophisticated views of moral language and concepts (Putnam (2004): 144 n. 6, Williams (1985, 1995), McDowell (1985), Diamond (1988)).2 But for present purposes we should focus on her positive argument for moral psychology. The call for moral psychology is presented as a demand on moral philosophers. Moral philosophy should account for how moral consciousness can be ‘oriented toward the good’, and it is in this context that emotions become a matter of interest for moral philosophy. Emotions are recognized as modes of moral awareness (Antonaccio (2001), Bagnoli (2011)). The point I am pressing is that the call for a new moral psychology is what explains the resurgence of interests in the emotions within moral theory. When we consider the poverty of the analytic models of morality, the contrast with ancient philosophy is obviously striking. Not surprisingly, then, philosophers turn to Plato and Aristotle to overcome the problem that Murdoch identifies. The development of moral philosophy in the 1980s is marked by the attempt to respond to this quest for moral psychology by revisiting traditions of thought that have a place for emotions. 1.3 Aristotelian themes: emotions and practical reason Murdoch’s view inspires the pioneering works in moral psychology of the 1980s and 1990s, such as those by Ame´lie O. Rorty (1980a), Martha C. Nussbaum (1986, 1990, 2001), Lawrence Blum (1980, 1986, 1991), and Owen Flanagan (1991). Thanks to the seminal work of Rorty (Rorty 1980b) and of Nussbaum (1986), the retrieval of the emotions has partly coincided with a revival of Aristotelian ethics, which has had the extraordinary effect of refocusing debates in moral philosophy on the role of emotions in rational deliberation, character, and moral consciousness. The merits and promise of Aristotelian ethics should be apparent from two perspectives concerning practical reason and the nature of emotions respectively. First, Aristotle’s ethics centers on practical reason and conceives of the excellences of character as chiefly contributing to the flourishing life.3 The Aristotelian approach to emotions supports 2 Murdoch’s legacy expands through feminism, communitarianism, and particularism, as critiques of rationalist and liberal ethical theory; see Antonaccio (2001), Blum (1991): 3; cf. Ruddick (1980). 3 It is often argued that the concerns about emotional responsiveness and moral consciousness are best vindicated by ‘virtue ethics’. The category itself is dubious, since it incorrectly suggests that the concept of virtue does not have a place in Utilitarian and Kantian ethics; see Nussbaum (1999). However, its presence in the contemporary taxonomy of ethical theory shows that moral philosophy has made at least some progress toward vindicating Murdoch’s main concerns for moral life. I would venture to suggest that the category of virtue ethics has little unity, and perhaps it has lost taxonomical utility, exactly because the concept of virtue has become so widely recognized as vital to any account of morality.



the claim that moral cognitions result from the appropriate habituation of the emotions, which is the primary task of moral education. Moral education consists in shaping and orienting the emotions toward goals and ends that are choice-worthy. This coincides with the development of a second nature (McDowell (1995), Stark (2001, 2004), Korsgaard (2009): 19).4 Second, and as a consequence, the Aristotelian approach made the controversy about the criteria for assessing the rationality and appropriateness of emotions central to the debates in moral psychology. Since it aims at building a character in which emotions are aligned with reason, this ethical view takes it that the emotions can be habituated because they are responsive to judgment. The distinctive appeal of the Aristotelian approach to the emotions best emerges in the contrast with the ‘simple view of emotions’, which takes them as ‘blind causes’, belonging to the category of feelings or sensations, explained by physiological conditions and subjectively experienced as having distinctive qualities (de Sousa (2010b): }2). This rival view of emotions is a major obstacle to recognizing a place for emotions in morality, since it denies that they are educable according to the standards of reason and apt for rational assessment. In its crudest form, which reduces them to sensations, this view treats emotions as involuntary states, before which we are helpless, and hence as a threat to moral and rational agency. It is no surprise that in order to restore emotions to a central position in morality, philosophers have typically focused on their cognitive cores. Thus the revival of emotions as a topic of philosophical inquiry in the 1980s took the form of a defense of cognitivism. The thesis that emotions are intentional, and importantly linked to the will and practical reason, is already present, although not prominent, in post-war analytic philosophy (Bedford (1957), Kenny (1963)). The canonical objection to the simple view is that it does not account for the differences among emotions, since not all of them can be reduced to sensations or identified by their associated bodily changes.5 Moreover, the phenomenology of deliberation shows that we are not completely helpless in the face of emotions. Some arguments appeal to traditions of thought that judge agents by the sort of emotional responses that they exhibit. If emotions were mere sensations, we would not be able to explain the widespread moral practices of praising and blaming people for their emotions. The renewed interest in Aristotle’s ethics generates a radical form of cognitivism, which takes emotions as equivalent to judgments. On this view, emotions have cognitive cores and intentional contents, and therefore function as affective appraisals (Solomon (1980, 1984, 2003), Nussbaum (1990, 2001)). For some philosophers, this 4 On this view, moral education amounts to the development of a ‘second nature’. This is to say that the sort of ‘naturalism’ that is peculiar of Aristotelian ethics importantly differs from empiricist sorts of naturalism; see McDowell (1995), Foot (2001), Annas (2005), and Thompson (2008). 5 These obvious defects are mitigated in more sophisticated versions of this view, which take emotions to be somatic markers; see Damasio (1994). However, critics object that even in the most sophisticated versions, the view is still too simple to account for the complexity and subtlety of our emotional moral life.



cognitive core is the basis for assessing their relation to the will and to reason (Wallace (1994)). A more moderate cognitivist thesis proposes that emotions are non-cognitive appraisals—rough-and-ready automatic assessments of salient features of the environment (Robinson (2005): chs 1–3).6 The preoccupation with making the emotional life intelligible has led philosophers such as Ame´lie Rorty to identify the target, focus, or propositional objects of emotions as the key way to recognize their varieties (Rorty (1980a, 1987, 1998a, 1998b)). Rorty’s efforts to rehabilitate emotions are in many important respects connected to Murdoch’s and Williams’ polemics about the pretenses of ethical theory (Rorty (1988)). Her view is that emotions play an important epistemic role: they allow agents to frame and understand the deliberative situation by highlighting the aspects of salience. They are, in this sense, modes of valuing and of perceiving something as relevant by directing the agent’s attention toward specific aspects of the situation. According to Rorty, emotions have a narrative structure and are crucial for the development and exercise of critical thinking. The narrative structure of our moral sensibility is socially formed. This means that the categories pertaining to practical rationality are psychological as well as moral and political (Rorty (1988): 295–7, Rorty (1998a, 1998b)). As a consequence, the intelligibility of emotions has both political and moral dimensions (Rorty (1998a, 1998b), Solomon (2003): ch. 9).7 In more recent debates, cognitivism has been criticized as implying an over-intellectualized view of emotions (Greenspan (1981), Deigh (1994), Pugmire (1998), Pugmire (2005): 14, Goldie (2001): 3). Some express reservations about the cognitivist view that emotions are fully or necessarily intentional; others doubt that the cognitivist view conforms to common sense (Carr (2009)), or they simply deny that emotions involve judgments (Peacocke (2004): 252–65). The critique of cognitivism is not always directed at objecting that emotions deserve a place within morality. Rather, the point of critics is often that the evaluative and moral importance of emotions depends on their affectivity, that is, on their being feelings (Greenspan (1981, 1988, 1995), Deigh (1994), Stocker (1996): 54–5). Thus, the standards of emotional rationality should not be reduced to the standards of rationality for beliefs and desires (de Sousa (1987), Ben-Ze’v (2000), Goldie (2001), Helm (2001)). Finally, a distinct and promising approach, which originates in these debates, denies that we can distinguish between conative and cognitive cores. It proposes that we treat emotions as states that affectively perceive their intentional objects as falling under ‘thick affective concepts’ which cannot be analyzed into separate affective and cognitive independent components (Zagzebski (2003); cf. Goldie (2001, 2004)). This view importantly resonates with Murdoch’s and Williams’ arguments against the noncognitivist semantic analysis of moral concepts, which mechanically separates descrip-

6 Some use the term ‘cognitive core’ in a broader sense, to refer to any state that has representational content. 7 Similarly, de Sousa construes the intelligibility of emotions in terms of paradigm scenarios, where biological and cultural criteria dictate their criteria of rationality and appropriateness; see de Sousa (1987).



tive and expressive semantic components (Murdoch (1970), Williams (1966), Wiggins (1987a, 1987b), McDowell (1985)). In order to accommodate the claim that moral emotions are perceptions of value or modes of moral discernment, philosophers face two tasks. On the one hand, there is pressure to offer an account of the ontological status of moral properties that emotions are supposed to discern. On the other hand, we need to make sense of the claim that emotions are conative and drive agents to action. Sensibility theories are designed to respond to each of these concerns. They suggest that moral properties are perceptions of properties whose ontological status depends on our own emotional response. The recognition of the evaluative role of the emotions is traditionally cast in terms of the perceptual model. On this model, emotional experiences are taken to represent evaluative properties, in much the same way as perceptual experiences represent non-evaluative properties. For instance, in analogy with colors, John McDowell argues that values are response-dependent concepts and insists that our sensibility responds to genuine properties in the world (McDowell (1985); cf. Helm (2001)). While it remains highly controversial whether the criteria of appropriateness of moral emotions should focus on the cognitive or conative and affective cores, the Aristotelian approach represents a prominent option in this debate. It centers on the interplay between reason and the emotions, and thus proposes a distinctive conception of ‘practical reason’. To this extent, the Aristotelian view of emotions is also opposed to some other sentimentalist views that take reason to be inert and separate from our affective and emotional life. I now turn to such views. 1.4 Humean themes: emotions and motivation Williams’ critique of Kantian ethics has brought back to life another tradition of thought that makes morality the province of the emotions, which dates back to eighteenth-century moral sentimentalists. Broadly speaking, sentimentalism is the view that moral evaluation should be understood in terms of human emotional response. Lord Shaftesbury, Francis Hutcheson, David Hume, and Adam Smith all make the claim that sentiments are the source of moral judgment and the drive for moral conduct. Some admit a special ‘moral sense’, in analogy with the five senses (Slote (2006): 219–26). Contemporary sentimentalism appeals to this venerable tradition and often deploys the analogies between values and colors, even though it makes no use of the obscure claim about a special moral sense (Baier (1985), Mackie (1977), Blackburn (1985); cf. Skorupski (2010a)). The retrieval of Hume’s theory points toward a way of investigating the role of emotions in morality, which turns out to be a significant alternative to the Aristotelian approach (Baier (1985, 1991, 2010b)). The interest of Humean theorists in the emotions lies especially in the investigation of moral motivation. Humean theories deny that reason directly motivates us, and thus deny that there is anything like ‘practical reason’. In respect to the capacity to motivate, emotions are more similar to desires than to beliefs. Humean accounts of emotions thus tend to insist on the conative cores of



emotions, directed to objects or states of affairs (Zemach (2001)). This feature puts the Humean view in radical contrast to the Aristotelian view of emotions as modes of moral discernment, and to the cognitivist view of emotions, which insists on their cognitive cores. Unlike rationalist value theories and straightforward cognitivist accounts of emotions, neo-sentimentalists argue that ‘something akin to the aesthetic is a central part of the ethical life of a rational animal. The affective disclosure of value is the beginning of our ethical life’ ( Johnston (2001): 183). But neo-sentimentalism comes in different versions. Some appeal to the Humean picture of the emotions as elaborations of affective desire in the light of the subjects’ beliefs about their relation to the appealing and the repellent (Wollheim (1999), Johnston (2001), Abramson (2010)).8 Critics of post-war analytic moral philosophy pointed out that a simplistic view of emotions is paired with an impoverished account of moral language. We should then expect that the changes in the account of emotions documented in sections 1.3–1.4 should have a correlate in meta-ethics. Simon Blackburn provides the paradigmatic example. He recognizes the challenge of adequately dealing with moral phenomenology (1985, 1988, 1998). In contrast to McDowell, he deploys the analogy with perception to argue that values are projected onto the world, rather than being original parts of its fabric. In contrast to Mackie, Blackburn insists that projectivism does not entail that we are systematically mistaken about the nature of value, because ‘the way in which we gild or stain the world with the colors borrowed from internal sentiment gives our creation its own life, and its own dependence on facts’ (Blackburn (1984): 219). Projectivism supports a quasi-realist view of moral judgments: it shows the mechanism by which moral discourse acquires its right to truth without admitting a moral ontology. The surface grammar of our ordinary moral discourse appears to be realist, but moral properties are mere projections onto the fabric of the world, like colors. Along Blackburn’s line, Allan Gibbard develops a sentimentalist theory where moral judgments are explicated as expressions of norms for governing the appropriateness of guilt and anger, which are taken to be natural emotions (Gibbard (1990)). Gibbard elaborates a systematic account of the logical relations among norms, and explains morality as a cooperative enterprise grounded on emotions. It might be objected that expressivist sentimentalism lacks the sort of depth and richness that Murdoch would hope and expect from an adequate meta-ethics, since it does not deal with the moral life as it is felt. But it is beyond question that its supporters made significant progress in


By contrast, other neo-sentimentalists invoke hybrid theories that are designed to account for moral judgments as having both cognitive and conative cores; see Wiggins (1987a, 1987b), McDowell (1985): section 1.3. A variant of this position, dispositionalism, holds that the judgment that something is good is true if and only if subjects in the relevant conditions would approve of it. A third form of sentimentalism develops the model of ‘tertiary properties’, which are genuine properties but dependent on the observer. However, several other distinct traditions of thought deny that moral distinctions are grounded in reason, and propose that emotions be considered sources of moral judgments and modes of moral appraisal; cf. Brentano (1889/ 1969), Husserl (1988), Scheler (1913–1916), Meinong (1917).



this direction since Williams’ complaint about the poverty of the meta-ethics. In particular, Gibbard produces a detailed account of the phenomena of normative governance, such as the attitudes and motives associated with being guided by a norm. This research is firmly grounded on evolutionary psychology and is mainly interested in explaining the emotional mechanisms that allow cooperative interactions to prove successful. Gibbard offers the most sustained defense of normative psychology to date. 1.5 The Kantian approach to moral psychology To be sure, the least controversial claim in Williams’ diagnosis of the neglect of the emotions in analytic moral philosophy is that Kantian rationalism is importantly responsible for it. There is a large agreement that the narrow conception of action as outward performance has roots in Kant’s ‘legalism’. The debate about the moral significance of emotions often takes for granted that an adequate moral psychology starts with a farewell to Kant (Rorty (1988), Oakley (1992)). As for the first claim about the neglect of emotions in the machinery of action, Kant is certainly an easy and appropriate target, because of his infamous remarks about the ‘deadly sensibility’ from which the moral agent ought to tear herself away (Kant (1785): 398). However, in the light of recent developments in Kantian scholarship, it is worth reconsidering the grounds of this accusation. What matters, Kant writes, is the ‘inner principles of actions that one does not see’ (Kant (1785): 407/61). It is, indeed, curious that an ethics centered on the concept of intention, such as Kant’s, is taken to support an ethics that reduces action to outward performance, divorced from the moral life. Murdoch might have a more plausible case against post-war Kantian ethics, but her attack on Kant seems to rest on a misunderstanding of Kant’s account of rational choice and agency (Reath (1980), Bagnoli (2003, 2009), Wuerth (2011)). For Kant, the outward performance of an act does not even qualify as the object of judgments of attributability, responsibility, or moral assessment. What is indicative of the moral worth of an action is the intention that motivates the agent in choosing it. The agent takes responsibility for her action by considering whether its subjective principle conforms to the requirements of practical reason. The primary (and perhaps sole) purpose of Kant’s method of the categorical imperative is to test the inner moral worth of subjective maxims, not to determine dutiful actions (O’Neill (1985): 164, 182). In this respect, then, Murdoch’s complaints against Kant’s action theory seem misplaced. Secondly, Kant’s conception of the emotions and moral sensibility is more complex than the references to the deadly influence of inclinations suggest (Kant (1785): }1, p. 398). On the basis of some exegetical evidence (Kant (1785): }1; cf. Kant (1797b): 211–12), scholars typically attribute to Kant the simple view of emotions as ‘mere feelings’, which do not perceive anything external but express a relation to the subject (Sabini & Silver (1987), Korsgaard (2009): 18–19). Insofar as they are pre-cognitive, physiological reactions, emotions are involuntary phenomena that cannot play any role in Kant’s account of moral value. However, Kant’s later writings show a complex



taxonomy of the categories of sensibility and moral sensibility, which does not fit with this simple view of emotions (Wood (1999); Louden (2000); Borges (2004)). For instance, emotions such as sympathy and gratitude cannot be explained according to this model, because they are not completely involuntary and passive. Kant holds that they can and should be cultivated, even if there cannot be any duty to have them (Kant (1797a): 401–2, (1797b): 236). Some scholars take this to prove that for Kant we are not passive regarding such emotions, and that they play a crucial role in moral agency as well as in building moral character (Sherman (1990), Baron (1995): 197, Cagle (2005), Anderson (2007)). Kant recognizes four fundamental ethical concepts that are grounded in our affective sensibility, namely, respect, conscience, moral feeling, and love of humanity. He places such moral feelings ‘at the ground of morality’, insofar as they make us responsive and receptive to moral duty (Kant (1797a): 399). These feelings make us feel pleasure or displeasure from the consciousness that our actions conform to moral duty (ibid.). Without these sensibility-based concepts, we would be ‘morally dead’, as Kant famously remarks. In particular, the moral feeling of respect is identified as the mark of rational agency, as well as the normative ground of duties to ourselves and to others (Reath (1980), Bagnoli (2003)). Scholars are divided about the status and role of these moral feelings. For some, Kant would agree with those contemporary theorists who think of emotions as subjective, non-cognitive, perceptions of unspecified bodily states, or states of the sympathetic nervous system, hence disconnected from value, character, and reason (Sabini & Silver (1987)). For others, Kant would instead accept a broader view of emotions, some of which are intentional states with propositional contents as well as feelings (Borges (2004)). In either case, it is an open question whether the Kantian account of moral sensibility has distinctive merits vis-a`-vis other theories of emotions. But the point here is that there is a role for emotions to play in Kant’s ethics and action theory. In fact, Kantian scholars identify several roles for emotions to play in Kantian ethics. Character in all its aspects, including dispositions and emotions, is inseparable from the very idea of practical reason (Sherman (1990)). Moral emotions, such as love and compassion, enable us to fulfill our moral duties (Baron (1995), Cagle (2005)). Perhaps more importantly, emotions themselves are ‘moral responses’ that determine what is morally relevant and, in some cases, what is required (Sherman (1990): 2). Emotions also play a significant role in the practice of moral judgment. To acknowledge this role, it is important to revisit the fundamental Kantian requirement that moral agents act on principles. To figure out what to do, we must adopt a method of rational deliberation that requires us to form an intention that all rational beings could endorse. An important problem for this model of rational deliberation is that moral principles do not always determine exactly and precisely what to do. One of our deliberative tasks consists in trying to figure out which principles are applicable and how. Emotions help us achieve this task by perceiving salient traits and specifying the domain of application and relevance of moral principles (Herman (1993)).



Current debates show that Kantian philosophers address and sympathetically engage with concerns about the emotions similar to those raised by Murdoch and Williams (Velleman (1999, 2006), Darwall (2006), Herman (1993, 2007), Korsgaard (2009): 16–20, 112–14). Recent reassessment of Kant’s moral psychology and action theory requires a corresponding reassessment of Kant’s legacy in current debates on emotions. What do we gain in introducing Kantian moral psychology? To begin with, attention to new work in Kantian scholarship might cast some doubt on the widely shared assumption that sentimentalism is the only or the most hospitable theory for emotions. But the most important advantage in introducing Kantian moral psychology in the debate about emotions is that we acquire a distinctive theoretical apparatus for revisiting the dichotomy between reason and sensibility. In this respect, the Kantian conception of moral sensibility is best appreciated as a mode for specifying the domain and function of practical reason, rather than in contrast to sentimentalism. Both Aristotelians and Kantians propose that we see reason as an activity or process, and as efficacious rather than inert. Moral sensibility is directly relevant to characterizing the practical use of reason. The basic claim of these theories is that a proper understanding of moral evaluation and practical reason should include the emotions. Different conceptions of practical reason provide competing accounts of which emotions should be included and of their role (Helm (2001, 2002, 2009), Jones (2004), Velleman (2006), Korsgaard (2009): 174–230). But these conceptions share the view that questions about motivation and questions about reasons for action are inextricably woven together. Theories that accept a sharp division between reason and sensibility, such as Humean theories, regard reason as inert and represent emotions as drives to action. But if reason is taken to have a direct, practical function, emotions should be acknowledged as playing a different and more pervasive role. In theories of practical reason, by contrast, emotions intervene in the explanation of rational action, as they supply moral motives. Moral sensibility represents the subjective conditions of receptiveness to moral duties. The contribution of emotions must be sought at the normative level of the ‘incentives’ of practical reason, the grounds for action, rather than at the motivational level of immediate drives. This suggests that we take the task of characterizing moral motivation as part of the larger project of elucidating the subjective and psychological grounds of our responsiveness to normative requirements. To the extent that Kant’s theory of practical reason requires a specific form of moral sensibility, it responds to Murdoch’s call for moral psychology and her demand that moral theory must have a philosophical vocabulary to address the role and import of emotions. Kantian and Aristotelian traditions share not only a marked concern for practical reason, but also the conviction that an adequate treatment of moral motivation does not belong to empirical psychology because it involves a priori concepts, which belong to a ‘pure psychology’ (Nagel (1970), Thompson (2008)). Thus, in reviving those traditions of practical thought, we face the question of whether the call for moral psychology leads to the retrieval of metaphysics. This claim raises large methodological issues, which are at the heart of the controversy over the status of moral psychology.



2. The Emergence of the Cognitive Sciences: Some Methodological Issues Many philosophers of the emotions share Williams’ view that the little consideration ethics accords to the emotions partly depends on a simplified and unrealistic view of our psychology (Williams (1973): 235, 222–3). Such advocates of moral psychology may have thought that moral philosophy had only to gain from the encounter with an adequate empirical psychology. The invitation to leave the philosopher’s armchair and meet the psychologist’s team in the lab has become particularly pressing with the development of the cognitive sciences. The rise of the cognitive sciences is probably the most significant factor in the philosophy of emotions, and we should expect this factor to have an immediate impact on moral philosophy as well. However, it is not at all obvious that the cognitive sciences have helped moral philosophy to advance in the direction that Anscombe, Murdoch, or Williams had hoped. These philosophers, call for moral psychology was a demand addressed primarily to moral philosophers. A striking fact is that the cognitive sciences have largely changed the general character of the debates about the relation of emotions to morality. As noted in the previous sections, moral philosophers have engaged in animated controversies about how to redesign the agenda of moral philosophy. By contrast, the debates that arise with the emergence of the cognitive sciences are neither confined to moral theory, nor engaged only or primarily with moral philosophy. At the center of these new debates lies a question about whether philosophy can legitimately claim the study of the mind with distinctive methods of inquiry. 2.1 The status of moral psychology: some methodological issues How to assess the philosophical relevance of the results of the cognitive sciences for the study of the emotions is a large and difficult question, which raises a host of important methodological issues. Ultimately, these methodological issues depend on how philosophy understands its own activity. Is philosophy an empirical or humanistic discipline? What are its driving concerns, criteria, and methods? With some simplification, we may identify two views on the matter. One the one hand, there are empirically minded scholars who understand their philosophical activity as mimicking science, and treat moral psychology as part of the empirical sciences (Doris & Stich (2005, 2006), Levy (2006)). This view bears important methodological implications, since its core claim is that philosophy needs to borrow empirical methods from science, if it is to stand a chance to be fruitful as a form of inquiry. The cognitive sciences do not simply provide new grounds for philosophical argumentations; nor do they simply represent a model to which philosophy should aspire. More radically, their task is to replace philosophy as a humanistic discipline with a thoroughly empirical approach. Understood as a humanistic endeavor, philosophy is thus an obsolete enterprise, whose only achievement is to ‘anticipate what the presumed scientific solutions to all metaphysical problems will eventually



look like’, as Hilary Putnam puts it in the course of his polemics against scientism (Putnam (1992): X). In contrast to this empirical approach to philosophy and psychology, others argue that both disciplines are autonomous with respect to the cognitive sciences (Williams (2000), Wallace (2005)). Of course, empirical findings may be indirectly relevant to philosophical arguments. Some kinds of philosophical argument use empirical evidence, and in such cases philosophy makes progress by taking into account evidence that was previously unavailable. For instance, empirical research may provide new evidence that undermines some philosophical arguments about the moral significance of emotions, and bolsters others. Even if they recognize that a dialogue with the empirical sciences is inevitable and rewarding, these philosophers argue that it is misleading to think of the activity of philosophy as modeled on the empirical sciences. At issue, then, is not science, but ‘scientism’, or the philosophical view that assimilates philosophy to science and borrows its methods (Williams (2000): 182). Those who defend the first view, and consider the emotions and moral behavior as the objects of empirical research, typically aim at discovering the influences to which we are subject, and at identifying patterns into which our behavior falls. On this approach, there is hardly any way to distinguish moral motivation from non-moral motivation. However, some prominent traditions of practical thought, such as the Aristotelian, the Scholastic, and the Kantian traditions, define ‘moral motivation’ in contrast to natural influences and patterns. The contrast between ‘moral’ and ‘natural’ should not be overstated, though. Aristotelians do not deny that there is some interesting continuity between natural emotional dispositions and the excellences of character; but they also argue that in order to become good we have to shape and develop our emotional dispositions according to reason. This is the task of moral education, which produces a ‘second nature’. Likewise, Kantians focus on the distinctiveness of moral motivation as opposed to non-moral motivation; but they capture such distinctiveness in terms of the requirements of practical reason. In both cases, however, morality is rooted in the distinctively human capacity for self-reflection, which introduces a significant difference in motivation, and sets us apart from other animals. As Thomas Nagel observes, ‘the suggestion that there must be motivational requirements on which to base ethical requirements seems to demand a priori reasoning in motivation theory’ (Nagel (1970): 5). This appeal to a priori concepts is what makes moral psychology problematic, and raises issues about the appropriate methodology for it. More recently, Michael Thompson has pointed out that any action theory and moral psychology that take seriously Anscombe’s (and Murdoch’s) critique of analytic ethics (section 1.1 above) have to come to terms with the Kantian critique of empiricism (Thompson (2008): 5–9). To be sure, there are significant differences between the Kantian and the neo-Aristotelian approaches to moral psychology (Korsgaard (2009): 174–206). Aristotelianism attributes special importance to the concrete concept of human (Foot (2001), Hursthouse (1999)), as opposed to the Kantian preference for the more abstract concept of person as rational being. However,



both the Aristotelian and the Kantian approaches consider ethical concepts, such as ‘action’, ‘virtue’, and ‘social practice’, significantly different from concepts such as ‘sea’, ‘cypress’, or ‘horse’, whose content we learn by a sort of direct association with experience (Thompson (2008): 6). Ethical concepts have a status akin to a priori concepts, and this indicates that they cannot be appropriately treated and articulated from within an empiricist framework. Kant did not acknowledge the relevant similarity of his program to the Aristotelian approach, because he mistakenly thought that the concept of human life, which is central to the Aristotelian ethics, was a biological or empirical concept. But neo-Kantians do not have to agree with Kant on this point, and many have insisted on a profound similarity between the two philosophical agendas (Sherman (1990), Korsgaard (1996, 2008, 2009), Engstrom & Whiting (1997), Engstrom (2009)). In fact, some scholars have also argued that Kant’s own ethics and moral psychology show a profound appropriation of Aristotelian themes, despite his rejection of Aristotle’s ethics (Thompson (2008): 12, Engstrom (2009)). Interestingly, these recent attempts to reconcile Aristotelians and Kantians take a complex route through some major voices in analytic philosophy, such as Wittgenstein, who provides important integrations for the definition of ‘practice’ and ‘life-form’ (Williams (1981), McDowell (1995), Korsgaard (1996): 137–9, 208, Thompson (2008)). The ‘pure’ approach to moral psychology is thus alive and well. 2.2 The question of the normative relevance Philosophers who advocate the experimental method in moral psychology have important epistemological tasks. Their presumption is that empirical research has a direct impact on normative ethics. For instance, Knobe and Nichols hold that empirical investigation identifies what leads us to have the intuitions we do; ‘the ultimate hope is that we can use this information to help determine whether the psychological sources of the beliefs undercut the warrant for the beliefs’ (Knobe & Nichols (2008a): 7). According to these philosophers, the use of experiments has revolutionary effects both at the normative and epistemological levels, since it leads to questioning the grounds on which our individual moral convictions and traditional moral theories rest. To reach this important conclusion, experimental philosophers draw from four kinds of empirical research. First, they point at anthropological and sociological sources that document cultural variations in moral convictions (Prinz (2007): 223–9, 280–5). Second, they use surveys of people’s responses to verify the scope of moral convergence and agreement and track divergences and disagreements. Typically, these surveys concern responses in controlled circumstances to questions about moral dilemmas, such as ‘trolley cases’ dilemmas (Thomson (1976, 2008), Copp (2011)). These tests are designed to examine our intuitions about the distinction between killing and letting one die, as well as how sensitive we are to considerations about consequences versus deontological constraints. In particular, there are functional magnetic resonance imaging studies of brain activity conducted on subjects while they are answering questions about moral dilemmas (Greene et al. (2001), Cohen (2001)). A third source of data is



thus neurological research about the brain activity associated with moral deliberation and responses to moral dilemmas such as the trolley problem. General cognitive psychology provides a fourth source of empirical data. These empirical results are supposed to have direct relevance in moral epistemology. For instance, Doris and Stich point to the general conclusion that the responses to thought experiments ‘may be strongly influenced by ethically irrelevant characteristics of example and audience’ (Doris & Stich (2005): 139, 140, Haidt et al. (1993)). For Sinnott-Armstrong, the empirical evidence shows ‘ways in which our moral intuitions do not reliably track the truth’ (Sinnott-Armstrong (2008a): 62); and that ‘no moral intuitions are justified non-inferentially’ (Sinnott-Armstrong (2008a): 74). More specific conclusions concern the ontological and epistemological status of moral convictions and claims that ground traditional ethical theories. These conclusions are not only revisionary of our current practices and intuitions, but also have an impact on our ethical theories, insofar as their methodology appeals to such practices and intuitions. In particular, several psychologists and empirical philosophers have devised functional magnetic resonance imaging experiments to show that our deontological intuitions are not well grounded and should be disregarded or discounted. That is, empirical research is used to counter (or reinforce) some target ethical theory, such as deontology or consequentialism. Many are skeptical about the power of these experimental arguments, and question their validity by taking issue with their methodology. For instance, Joshua Greene’s argument starts with a hypothesis that there are two subsystems underlying our moral intuitions (Greene et al. (2001)). He speculates that the first makes use of more direct emotional neural processes, which generate moral judgments typically associated with deontological ethics; while the second makes use of more reflective cognitive neural processes and generates the sorts of judgments typically associated with consequentialism. From this speculation, further normative conclusions are derived about the sort of moral judgments we should make. Critics of this approach point out that the alleged experimental arguments make crucial use of normative assumptions. In this case, for instance, there is a complex conjecture about the brain at work, but the experiment also uses normatively laden terms, such as ‘consequentialism’ and ‘deontology’; and it appeals to intuitions about what counts as ‘morally irrelevant’ (Copp (2011)). Critics hold that when the arguments of this kind are closely scrutinized, the facts about brains are found to play a little role, or no role at all (Berker (2011)). This sort of critique leads to a more general problem about what counts as an ‘experimental argument’ ( Jones (2006), Kennett & Fine (2009)). Those unconvinced about the validity of the experiments do not exclude that empirical research may be indirectly useful for ethical theory, or that neuroscience in particular plays a role in shaping our normative positions. For instance, Selim Berker agrees that neuroscience may provide ‘clues for where to look when attempting to characterize the features to which distinctively deontological and distinctively consequentialist judgments respond’ (Berker (2010): }7). This is a significant role to play, but



again it rests on the assumption that our moral judgments respond to ‘some moral facts’ in the world. This is a problematic assumption in moral epistemology, and many would argue that standard moral and political methodology can be vindicated without answering certain kinds of skeptical epistemological worry, or taking positions about the ontological and metaphysical disputes about moral truths (Copp (2011)). This fervent debate about the status and method of moral psychology is particularly relevant for our study of the emotions and their role in morality. On the basis of evidence about brain activity associated with responses to moral dilemmas, these studies are supposed to show that differences in emotional responses explain people’s differing judgments about the cases (Greene (2001), Cohen (2001)). One key hypothesis is that ‘personal dilemmas’ tend to be associated with greater emotional engagement than ‘impersonal dilemmas’, and that these differences in emotional engagement affect people’s moral judgments. Other researchers have found that inducing emotions in their subjects can affect their moral judgments (Wheatley & Haidt (2005), Prinz (2007): 28). Finally, several experimental philosophers, such as Prinz and Nichols, argue that empirical findings point toward sentimentalism as the most adequate meta-ethical account of moral judgment (Prinz (2006, 2007), Nichols (2004)). This is not surprising, since the claim about the relevance of emotions in our life is traditionally used to counter rationalism. However, the wealth of new empirical data about the pervasiveness of emotions present novel challenges not only against rationalism, but also and more generally against the very possibility of normative ethics. 2.3 New challenges to the place of emotions in morality There is an emerging consensus that emotions are so pervasively present in the activity of reasoning that it makes little sense to think of reason as a cold and dry faculty, totally disengaged from sensibility. New challenges for the philosophical model of the ‘detached intellect’ come from studies based on neurobiological evidences (Churchland (1996), Damasio (1994), Gallagher (2005): 151, Lacewing (2005)). In one respect, these empirical findings have less disturbing effects than it may appear at first, since several philosophers have already questioned the dichotomy between passivity and activity, on the basis of distinctively philosophical arguments (Thalberg (1978), BenZe’ev (2000), de Sousa (1987), Frank (1988), Greenspan (1988), Oakley (1992), Raz (1997), Elster (1999), Solomon (2003): ch. 4). Indeed, some argue that the very idea of ‘practical reason’ is an attempt to overcome this distinction, by indicating that emotions represent the conditions of our receptivity to rational and moral considerations (Helm (2001, 2002, 2009), Bagnoli (2009)). However, there is another sense in which these empirical findings are indeed rather upsetting. Empirical research seems to indicate that reasoning includes automatic, nondeliberative, unintended, and unconscious mental processes. Likewise, intentional actions seem to involve substantive, automatic elements of this sort. Even in the case of what we typically regard as voluntary actions, the ‘readiness potential’ takes place



before the subject forms an intention or makes a decision (Libet (1987)).9 Other empirical studies conclude that emotion always precedes cognition (Zajonc (1980), Damasio (1999), LeDoux (1996)). These studies seem to establish that emotions intervene more substantially and diffusely in the mechanics of action than normative ethics assumes. From a philosophical point of view, however, what we derive from these empirical studies is not so much a positive assessment of the role of emotions in reasoning. Rather, these studies tend to emphasize that this pervasive interconnection between reasoning and emotions blurs the philosophical distinction between activity and passivity. The implication is that we are, in fact, more passive than the rationalist assumes—which indicates that moral theory assumes standards of agency that are simply beyond our reach. This line of argument has obvious consequences concerning the alleged standards for rational agency and responsibility, but it also supports some skepticism about the very status and authority of normative ethics, as a practical and theoretical enterprise. If emotions drive us independently of cognitions and decisions, what is the purpose of norms for holding people responsible for their emotions? It is noteworthy that skeptical conclusions about the legitimacy of normative ethics can be drawn only by adding a specific assumption about the emotions, namely, that we are indeed passive toward them. As shown in section 1, this simple view of emotions is under attack in the debates of the 1980s and 1990s, by arguments that vindicate the cognitive cores of emotions (Nussbaum (1990)), defend independent standards to assess their appropriateness (Greenspan (1988)), highlight steady emotional patterns (Frijda (1986)), or invoke the idea that emotions are complex and compound states, which are activated by paradigm scenarios (Rorty (1980a, 1998a, 1998b), de Sousa (1987)). Interestingly, then, the simple view of emotions resurfaces in the most recent debates about the empirical approach to psychology. For instance, if emotions are taken to be ‘neurological affect programs’, then questions of choice and responsibility that arise seem inappropriate. Paradoxically, this assessment of the pervasive role of emotions in reasoning agrees with the standard rationalist approach that emotions are passive and make us passive. The question of responsibility remains open for our contemporaries. To some, an important aspect in determining an answer to this question is whether emotions are educable and malleable. If so, then we would have some grounds for praising the jolly benevolent person and for judging that the grumpy and melancholic introvert has some responsibility for his state. Other scholars, on the other hand, think that the evidence that the emotions are malleable and adaptive is of no reassurance. Emotions have arisen through biological and social evolution, and it is precisely their contingent nature that

9 The difficulty in assessing the precise philosophical relevance of these findings is partly due to the difficulty of translating empirical concepts into ways useful for philosophical debates. For instance, in the case above, it is ambiguous whether such ‘readiness potential’ indicates a motive, or a disposition to select something as a reason.



casts doubts on their reliability as sources of moral knowledge (Crisp (2006): 24; cf. Prinz (2007, 2009); Greenspan (1988): Part ii). Likewise, it remains unsettled whether emotions should determine the normative standards for attributing moral agency. Answers to this question partly depend on the normative standards for rational agency and autonomy, as argued in section 1. For instance, one may claim that we can take an evaluative stance toward our natural endowments regardless of whether they are under our direct control (Solomon (2003): chs 1, 4, 12, Smith (2005)). Some may still want to argue that we are responsible for emotions even when they are not voluntary, but it would be inappropriate to punish or be punished for them (Sankowski (1977), Gordon (1987), Solomon (1980, 2003)). In any case, the question of responsibility for emotions cannot be solved solely by appealing to considerations about what an emotion is, but also by proposing a normative model of responsibility. The new empirical paradigm endorses an even less optimistic view then standard rationalist, as it points out that emotions drive us independently of our conscious deliberations. Taken to the extreme, the view that emotions precede cognition and decision invites skepticism about agential and cognitive autonomy, and undermines the very purpose of offering normative standards of justification and deliberation. Is ethical theory equipped to address these challenges? Empirical research documents and explains what we all have experienced: emotions interfere with our reasoning and sometimes defeat our decisions. Imagine, for instance, that you are planning your vacation to an exotic place to visit your friend, but fear of flying takes over. Although you really want to visit your friend, the fear impedes the realization of your plan. Suppose your friend asks that you reconsider your decision and shows that flying is actually statistically less dangerous than driving. You take her point and feel persuaded by her argument. So you are resolved to buy the ticket first thing in the morning; and yet, when you wake up you cannot bring yourself to do it. You know that you do not have any good reason to fear flying, but your fear is stronger than your evaluative judgment about the risk of flying. Cases like this one about ‘recalcitrant emotions’ seem to show that they work independently of the activity of reason (cf. Stocker (1996); D’Arms & Jacobson (2003); Tappolet (2003)). The normative question is whether, and on what grounds, we should attribute any authority to them. It is worth noticing that in the example above the emotion of fear interferes at several stages of your reasoning, and in different ways. First, fear makes the risk of flying very vivid, and thus sets the background of your reasoning by selecting the options available on the basis of their salience. The role of emotion here is evaluative: fear highlights options, and marks something as fearsome and undesirable. It thus provides considerations that may potentially be in competition with morality, e.g. your duties of friendship. Second, fear intervenes in the ordering of your priorities, by proposing that it is more important for you not to run the risk of flying than to see your friend. That is, it provides a motive and a pro tanto reason to disregard your original intention to visit a friend. Third, after you have reasoned through the options,



and considered other sources of evidence, fear still persists and defeats your decision to visit your friend. This defeat does not directly prove reason to be inert or emotions to be irrational. But it shows that emotions are independent sources of motives and reasons. This is a threatening discovery for morality. Insofar as emotions are independent sources of reasons, they can intervene in moral reasoning by countering the moral reasons. However, it is doubtful that all emotions interfere with practical reasoning in this way, by providing motives that directly defeat moral reasons, and the question partly relates to whether at least more complex emotions, such as respect, guilt, blame, and compassion, can be modified by reasoning. There is a general agreement that emotions can be managed and regulated, and that management and regulation are a sign of emotional intelligence (Zhu & Thagard (2002)). Emotion-regulation does not assure sensitivity to judgment, and it is possible even when emotions cannot be shaped according to some standards of correctness or habituated on the basis of patterns. However, emotion-regulation suffices to show that we are not completely passive in the face of emotions, and therefore we can be held responsible for them or for their education. A stronger claim is that emotions contribute to morality not only by concurring with its dictates, but also by directly driving and participating in moral reasoning. The acknowledgement of this more significant role of emotions is a central claim of many theories of practical reason.

3. Morality and the Emotions: New Assessments The call for moral psychology has led to important and surprising developments, such as the retrieval and elaboration of a variety of theories of practical reason that account for the relation that emotions bear to moral ontology and normative psychology. In this section, I show how the essays of this volume respond to the quest for an adequate moral psychology and contribute to current debates about the emotions. We can identify three main clusters of problems. The first concerns the place of emotions in practical rationality, and their role in the explanation of autonomous action and, more generally, of our capacity for rational normative guidance (3.1). The second concerns the relation of emotions to value and the expressive, normative, and epistemological roles that emotions play in morality (3.2). And the third concerns the standards of the assessment and accountability of emotions, and their relation to moral identity (3.3). 3.1 Emotions, practical reason, and moral agency The history of philosophy shows that the main source of hostility against the moral relevance of emotions comes from rationalist theories of morality. The basic rationalist claim is that emotions are not sensitive to reason, and this raises a cluster of issues about the basis and point of their moral assessment, their role in practical reasoning and in the account of autonomous and rational action. The rationalist’s reluctance to find a place for emotions in morality springs from two claims about the nature of emotions as unresponsive to reason, and the picture of reason as detached. Both of these claims have



been under attack in recent debates. We have seen in section 2.2 that the development of the cognitive sciences has given new impetus to the critique of the model of the detached intellect. This critique finds enthusiastic supporters among moral philosophers, especially those interested in working out a proper understanding of ‘practical reason’. There are a wide variety of theories of practical reason and of its relation to the emotions. Some admit that certain emotions or the capacity for them play an important epistemic enabling role, while denying that they qualify as a reliable source of moral knowledge (Crisp (2006): 23). Others argue for a more substantive and pervasive contribution on the part of emotions by construing practical reason as motivationally engaged (Helm (2001, 2002); cf. Deigh (1996), Korsgaard (2009): 18–19). To this extent, theories of practical reason identify a role for emotions in the formation of moral cognition. They recognize that emotions have the power to intrude in reasoning and upset or undermine its conclusions. But they take this evidence as the starting point for normative accounts of practical reasoning and norms that govern the appropriateness of emotions. The essays of Part I of the volume address directly these issues. In Chapter 1, Patricia Greenspan offers a novel account of the normative relevance of emotions and their role in practical reasoning. Greenspan holds a distinctive position, in that she seeks to find a significant place for emotions in deontological ethics. Traditional rationalist accounts doubt that emotions can be modified according to reason, and thus take morality to be in charge of silencing them. Others hold these emotional tendencies to be modifiable and take morality to be in charge of shaping, pruning, and managing them so that they become useful and conducive to some good. For instance, some hold that emotions are helpful in the process of decision-making because they are characteristically quick in detecting the morally salient features of the situation. By relying on emotions we gain in speed and efficiency, and exclude some options as immediately irrelevant (Nussbaum (2001); cf. Zhu & Thagard (2002): 27). Greenspan’s recent work shows that emotions participate in many more ways in practical reasoning. One important contribution of emotions is that they provide motivational support for our reasoning (Greenspan (1995)). In addition, they also explain the binding force of moral reasons (Greenspan (2005, 2007, 2010)). In Chapter 1, Greenspan further specifies the normative role of emotions, arguing that they reinforce moral reasons and work against their postponement. Greenspan frames the discussion in deontological terms, in contrast to the consequentialist understanding of moral requirements as a function of our weightiest reasons. Her account also importantly departs from the widespread assumption that privileges a Humean understanding of the role of emotion in ethics, and offers a more complex account of the processes of moral motivation, where there is a normative interplay of moral and non-moral considerations. One important area of the relation between practical rationality and rational agency concerns whether emotions express or undermine agential authority. We often refer to emotions in explaining our actions or in the attempt to make them intelligible. You cried because you received sad news. She smiled because she was glad to see you. He



shouted because he was angry. Since emotions play a role in acting, one may expect them to play a significant role also in moral agency. The rationalist does not dispute that most emotions dispose to act, and thus admits that they can explain action insofar as they participate in the generation and execution of action. However, the rationalist contends that emotions do not serve as moral motives for action. This argument appeals to autonomy as the mark of rational agency (Tappolet (2006)). The standard rationalist claim is that we passively experience the emotions, so we do not freely authorize actions driven by our emotions. This charge of passivity is a major obstacle to the recognition of the role of emotions in moral agency, and raises issues about our responsibility for them, as well as about their impact on our agential authority. In Chapter 2, Carla Bagnoli considers a related specific aspect of the normative relevance of emotions, which concerns the authority of moral reasons that yield requirements. Her contention is that moral reasons have categorical authority insofar as they are subjectively experienced in the guise of respect. Bagnoli argues for this claim by elaborating a Kantian account of practical reason, which takes respect as the emotional attitude constitutive of rational agency. This view of respect as the emotional aspect of practical reason avoids the dilemma between a rationalist view of the unconditional demands of morality, which has no grip on us, and a sentimentalist view of moral reasons that denies their categorical authority. She argues that both these views mischaracterize and misunderstand the relation between morality and the emotions. They mistakenly assume that emotions are separable from or only contingently related to practical reason. By contrast, the Kantian model takes this relation as structural: to undertake the practical standpoint requires us not only to act and think on principles that have the form of a law, but also to express a moral sensibility marked by respect. These requirements are constitutive of the practical standpoint, that is, they represent what is necessary for us to think and act as rational agents, together with other finite and interdependent rational agents. In Chapter 3, Edward Harcourt also argues for a constitutive relation between the emotions and practical reason. He starts with an investigation of the contrast between rationalist conceptions of integrity and the psychoanalytic approach, and he attempts to reconcile the two (cf. Deigh (1996), Lear (1998)). According to the rationalist view of the relation between emotions and practical reason, a central part of what it is for a human moral psychology to be properly formed consists in our being practically rational. By contrast, psychoanalysis addresses this very question by giving the notion of love a more prominent place. Harcourt intends to show that these conceptions are complementary, by focusing on ideals of psychological organization such as ‘autonomy’ and ‘individuation’ centered on self-love. He argues that there is a constitutive connection between love and practical rationality, thereby reconciling the two approaches. Focus on love makes us appreciate a distinctive feature of emotions, namely, their perspectival nature. Emotions originate in the special perspective of a subject, express his specific concerns or attitudes, target a specific object, and focus on some salient traits. When emotions function correctly, they draw attention selectively, highlighting



the salient aspects of the situation. Love is the paradigmatic example: it is an emotion grounded on exclusion; it sets apart the object of love from everything else. Medea’s love for Jason is unique and uniquely powerful: it overrides any other attachment she has. A mother’s love for her child is distinctively partial and exclusive. This partiality and exclusiveness clashes with morality’s request for impartiality, and with the claim that we should be concerned for others in virtue of what they are, rather than in virtue of the fact that they stand in a special relation to us. To resolve this contrast, moral philosophers have introduced a distinction in love. Love is the paradigmatic moral emotion when understood as agape and directed toward impersonal humanity (Badhwar (2003): 58, Helm (2009)); but not when it is informed by eros, as in romantic love. Romantic or erotic love is pathological and unreliable, while agapic love is steady and stable. This distinction between pathological and moral emotions may be of some help in highlighting the complexity of our motivational drives, but it underestimates the problem (Deigh (1991, 1995), Blum (1980, 1991), Frankfurt (2004)). The partiality of emotions cannot be neutralized by adding some conceptual distinctions (de Sousa (1987), Rorty (1980a, 1988), Helm (1994), Nussbaum (1990)). When one is in love, one sees the object as lovable under the description of love. Love blinds us, but not in the sense that it makes us irrational. Rather, it functions by silencing other aspects of the object of love, which become irrelevant to the subject. Insofar as they are perspectival, emotions shape our deliberative and normative horizon in ways that undermine the claims of impersonal morality. This thought has been developed in a variety of ways. Some argue that this perspectival feature of emotions is exactly what makes them morally relevant, since they convey moral knowledge and moral awareness of what is valuable (Blum (1980, 1986, 1991), Nussbaum (1990), Greenspan (1995)). That emotions are partial and perspectival does not necessarily show that they cannot be sources of categorical reasons, as Kantians hold. For instance, Harry Frankfurt regards love as an activity that bestows value onto its objects (Frankfurt (1999)). While personal, love commands as categorically as reason does (Frankfurt (1999, 2004)). By contrast, others argue that love is itself a moral emotion, which represents no threat to impartial morality (Velleman (1999)). In Chapter 4, Aaron Ben-Ze’ev offers a distinctive approach to romantic love. He addresses the normative dimension of romantic love and examines the nature and role of compromises in romantic loving relationships. In its traditional, literary, and idealized version, romantic love refuses any type of compromise. By contrast, Ben-Ze’ev argues for the importance of compromises, which acknowledge the existence of conflicts and represent one important way in which we attempt to cope with them. By analyzing the activity of complementing and compromising, Ben-Ze’ev provides an innovative normative model for romantic love relationships. 3.2 Emotions as modes of valuing In the last thirty years, moral philosophers have presented powerful arguments for treating emotions as highly discriminating responses to values. These arguments point



toward heuristic, epistemic, and expressive roles that emotions play in morality. For instance, Martha Nussbaum’s seminal work points out that emotions are intelligently informative, that is, they offer information as needed. Emotions are affective devices for detecting salient properties of the world. In some cases, the role of emotions is heuristic. In social contexts characterized by inequality and competitiveness, emotions such as envy, anger, resentment, and hatred may arise from a moral concern for fairness (Stocker (1996): ch. 10, La Caze (2001); cf. Ben-Ze’ev (2002)). While in itself envy can hardly be a positive trait, it might nonetheless be indicative of relations poisoned by unfairness and arbitrariness. To this extent, envy senses injustice. To argue that these emotions are diagnostic devices is not simply to say that they serve to highlight what goes wrong in social reality. Rather, these emotions present moral problems and demand solutions. That is, they are not only evaluative markers, but also sources of moral claims (Stocker (1996): 291). Suppressing them without addressing the problem they are signaling would be as dangerous as treating the hungry with hunger suppressants rather than feeding them. To accord emotions this heuristic role inevitably raises issues about the epistemological account of emotions as ‘perceptions of values’, and also about the status of values. As we saw in section 1.4, one prominent view about these issues is neosentimentalism, the view that to judge that something has an evaluative property is to judge that some affective or emotional attitude is appropriate, or fitting, with respect to it. Chapters 5 and 6, in Part II of this volume, directly contribute to the debate about the viability of perceptual sentimentalist models of emotions. Supporters of the perceptual model of emotions hold that emotional experiences can be reasons for evaluative judgments—in much the same way that sensory perceptual experiences can be reasons for our judgments about the non-evaluative world. In Chapter 5, Christine Tappolet defends a ‘descriptive’ version of the neo-sentimentalist account, which holds that appropriateness in emotions is a matter of correspondence to evaluative facts. This is a very specific sense in which emotions are said to be perceptions of values, in analogy with sensory perception. The plausibility of neo-sentimentalism is due to the fact that values and emotional responses, or at least their concepts, are closely related. Tappolet accounts for several varieties of neo-sentimentalism and argues that one has to distinguish between a normative and a descriptive version. She considers the main argument that can be given in favor of the normative version and shows that the descriptive version is far from excluded by this argument. Then, she offers two arguments in favor of the descriptive version, and defends it from the accusation of vicious circularity. By contrast, Michael Brady argues in Chapter 6 that there are significant differences between emotions and perceptions at the epistemic level. Emotions, unlike perceptions, often motivate the search for reasons that bear on their own accuracy, and hence on the correctness of the associated judgment. When emotions are reliable trackers of value, emotional experience, unlike perceptual experience, is at best a proxy for genuine justifying reasons. Since the point of the perceptual model is to provide an



adequate epistemology for our knowledge of value, then the fact that emotional experience doesn’t by itself provide reasons or evidence for evaluative judgment or belief would appear to be damaging. Perhaps these differences do not ultimately undermine the perceptual model, but defenders of the perceptual model owe us an explanation as to why they do not. In contrast to the perceptual model, Brady argues that emotional experiences play an important epistemic role, but they are not reasons for evaluative judgments. Chapters 7 and 8 represent two distinct approaches for evaluating the epistemological role and importance of emotions. Because of their methodological assumptions, these chapters take a stand about some of the methodological issues raised in section 2.2. In Chapter 7 Paul Thagard and Tracy Finn argue that emotions are both cognitive appraisals and somatic perceptions. In contrast to other authors, they invoke a distinct methodology, as they use a neural theory of emotional consciousness. Their aim is to develop a novel account of conscience and moral intuition. On their view, emotions are cognitive appraisals and somatic perceptions performed simultaneously by interacting brain areas. According to Thagard and Finn, conscience is a kind of moral intuition, which is a neural process that generates emotional intuitions that combine bodily reactions with cognitive appraisal concerning a special subset of goals. This account purports to explain how moral intuitions can be both cognitive and emotional, and why both moral agreement and disagreement are common phenomena. Thagard and Finn’s theory of conscience is descriptive and normative, and it offers important resources for evaluating the ethical and epistemological validity of intuitions. By contrast, in Chapter 8 Lawrence Blum argues that the normative, expressive, cognitive, and evaluative aspects of emotions are best understood from within a humanistic model of moral psychology. Blum takes issue with Shaun Nichols, a leading figure in empirical moral psychology within moral philosophy. Blum’s broader target, however, is the empiricist version of neo-sentimentalism. According to Blum, this form of neo-sentimentalism uses an impoverished view of moral emotions, especially of empathy and altruistic emotions, such as love and compassion. In contrast to neosentimentalism, Blum defends a richer account of altruistic emotions. On his view, emotions are intentional (rather than mere copies of feeling states of the other), and perceptual (involving ways of seeing the world). Secondly, emotions are cognitive insofar as they are ways of understanding others. Finally, emotions are motivational and expressive. Blum argues that although philosophy benefits from attention to empirical psychology, there are significant risks in abandoning the rich tradition of philosophical moral psychology, which can be carried out only with the distinctive humanistic methods of philosophy. Insofar as it focuses on altruistic emotions, Blum’s essay also represents a distinctive view in the debate over the partial and perspectival nature of emotions presented above. Blum’s past work has been crucial in proposing emotions as modes of moral discernment that call into question the requirements of impartiality.



3.3 Emotions, responsibility, and moral identity A strong argument for the moral relevance of emotions is that they are indispensable to morality understood as a system of norms, and to moral identity understood as a form of self-governance. Friedrich Nietzsche famously connects resentment and guilt to the very idea of what we owe to others, i.e. as the root moral obligation (Nietzsche (1968): 21ff. Foot (1994)). Williams calls morality ‘the blame system’ (Williams (1985): 177). Blame, respect, and resentment are the ‘deontic emotions’ typically associated with moral obligations (de Sousa (2006): 31, Skorupski (2010b)). The effect of linking moral obligation to such negative and reactive emotions is not necessarily that of debunking morality or exposing its all too human origins. On the contrary, for instance, P.F. Strawson’s essay on ‘Freedom and Resentment’ (1962) shows that reactive attitudes determine the conditions for the attribution of moral responsibility and are thus functional to our moral and normative practices. Strawson’s essay is a milestone in the literature on deontic emotions. Following Strawson’s lead, many argue that reactive attitudes, such as resentment and blame, implicitly address demands to somebody, hence presume the accountability of others. These attitudes figure prominently in current philosophical accounts of moral authority. They express moral claims and presuppose the normative expectation that such claims be recognized and also that we have the authority to expect and demand that they be recognized (Darwall (2006): 265). Two of the essays in Part III play Strawsonian themes. In Chapter 9 John Deigh reconsiders the grounds of Strawson’s account of the reactive attitudes. According to Deigh, Strawson endorses a non-cognitivist account of reactive attitudes, and this aspect plays a crucial role in Strawson’s argument that determinism posits no threat to moral responsibility. This reading is offered in contrast to recent proposals, such as those of Jay Wallace and Stephen Darwall, which pursue a cognitivist reading of reactive attitudes. One important consequence of Deigh’s argument is that Wallace’s and Darwall’s accounts of moral responsibility are only superficially similar to Strawson’s. Therefore they cannot avail themselves of Strawson’s solution to the problem of how we can hold agents accountable for their actions if determinism is true. Deigh’s argument directly contributes not only to current debates about the nature of moral responsibility, but also to debates concerning how to account for the emotional aspect of reactive attitudes. More importantly, the argument shows that these debates influence each other, and cannot be pulled apart. In Chapter 10, Bennett W. Helm also engages in the dispute about the nature of reactive attitudes and their relation to freedom and responsibility. Helm argues that we can best understand reactive attitudes by seeing them as individually presupposing and jointly constituting both our respect for persons and the dignity to which respect is responsive. Consequently, being both a proper subject and object of reactive attitudes is to be a member of the normative community of fellow persons within which one both takes responsibility and is held responsible for what one does. Helm’s position is thus in accord with neo-Kantian accounts of moral agency that



are centered on respect as recognition of the valid claims of others (Darwall (1977, 2006), Dillon (1992, 1997, 2010), Bagnoli (2003, 2007, 2009)). In Chapter 11, Angela Smith considers whether one is justified in feeling guilty for unexpressed attitudes. In these cases, the feeling of guilt attaches to the mere having these thoughts and attitudes, even when they remain private. This is paradoxical since these do not seem to be cases of moral transgression. Smith argues that a contractualist theory of morality is capable of accounting for such attitudinal wrongs. According to contractualism, to genuinely stand in appropriate relations to others, we must have certain attitudes toward them. Feelings of guilt over unexpressed attitudes reflect a judgment that we have failed to accord to others the basic respect and recognition they are due as fellow members of the moral community. Therefore, guilty feelings for unexpressed attitudes express our aspiration to live with others on mutually acceptable terms. It is often argued that one acts uncharacteristically when driven by emotions. Such cases are adduced to show not that emotions easily lead us astray, but that they undermine our integrity and character. In contrast to these cases, philosophers have noted that emotions play a significant role in self-government and in the development of individual moral integrity. Gabriele Taylor’s (1985) seminal work on pride, shame, and guilt shows that these emotions are modes of self-assessment, which signal status and directly contribute to building our moral integrity. Chapters 12 and 13 represent important contributions to some strands of this debate. In Chapter 12, Jacqueline Taylor investigates the sources of moral identity, and argues that pride and praise ground moral competence. Taylor’s argument starts with a critique of the current readings of sentimentalism, which focus solely on negative emotions that signal failure and transgression. She claims that this is an important difference in accounting for the character-building effect of emotions, and their capacity to nourish moral resilience. Her argument analyzes the relation between moral self-esteem and an active agency, which includes not only acting well, but also resisting evil. Taylor is right that the literature on emotions and moral responsibility centers on negative deontic emotions, such as blame or resentment.10 This is because this debate is focused on moral accountability, and arises from the view that emotions ensure normative conformity and signal defection. By shifting to other sorts of emotions, Taylor helps us appreciate other normative roles for emotions to play. For instance, emotions as diverse as compassion, indignation, anger, or resentment play a distinctive role, which is that of counteracting the perfunctory aspect of received morality. To insist on the varieties of contributions that emotions make to morality is particularly important in cases where there is a conflict between a code of norms recognized by the community and individual judgment. Some philosophers hold that in these cases, emotions are to be trusted as apprehensions of real moral values (Bennett 10 This is an area of inquiry that profited much from sociobiology and evolutionary psychology; see Plutchik (1980), Frank (1988), Gibbard (1990), Goldie (2001): ch. 4, Cosmides & Tooby (2000), Evans (2001), Prinz (2007): chs 2 and 7.



(1974), Starkey (2008)). Agents may not be able to coherently argue that their community is misguided, nor that its norms should not apply to this particular case, and yet feel that the norms to which the community appeals are deeply wrong. However, there are also cases where emotions complement and integrate positive norms. For instance, mercy makes norms more determinate and thus relevant to the individual case (Nussbaum (1999)). Compassion may mitigate the normative effects of moral norms or supply some normative guidance where there is none. In view of these cases, it is thus plausible to argue that the relevance of emotions is both moral and political (Rorty (1998a, 1998b), Dillon (1997)). The cases of disobedience informed by emotions indicate that emotions are key ways to express one’s personality in contrast to customary morality. As Ame´lie Rorty has shown, moral integrity and emotional coherence are psychological and political achievements, and their failures should be investigated from each of these perspectives, psychological and political. Talbot Brewer shows a similar concern and approach in Chapter 13. He presents a case where self-development is impaired by some processes of alienation from one’s own emotions. Brewer’s argument builds on sociological studies about the external pressure to alienate ourselves from our emotions, which depends on the widespread model of agency as performance and commodity. Emotions can give expression to the self’s pre-reflective evaluative posture towards the world. The aim of this essay is to examine different aspects of the phenomenon of alienation from emotions and to cast light on the conflict that can arise between the work of selfelaboration and the sort of ‘emotional labor’ required in the service economy. This critical overview is far from being a complete and exhaustive account of the complex relations that the emotions bear to morality, but it should provide the reader with a broad context in which to situate the essays of this volume. This volume has been conceived to reclaim emotions as a subject of investigation for moral philosophy. The suggestion is not that the philosophical investigation of emotions should be insulated from empirical research. On the contrary, our conviction is that philosophy should participate in the investigation of emotions alongside empirical research, in the spirit of fruitful dialogue where the distinctive value and contribution of all parties are recognized.11

References Abramson, Kate (2010) ‘A Sentimentalist’s Defense of Contempt, Shame, and Disdain’. In Goldie (2010): ch. 8. Anderson, Elizabeth (2007) ‘Emotions in Kant’s Later Moral Philosophy: Honor and the Phenomenology of Moral Value’. In Betzler (2007): ch. 1.

11 I am grateful to Selim Berker, Larry Blum, Ian Carter, David Copp, Luca Malatesti, Elijah Millgram, Patrizia Pedrini, Christine Tappolet, and Jackie Taylor for their helpful remarks on previous drafts of this Introduction. Special thanks to Richard Moran, who first got me interested in this subject.



Annas, Julia (2005) ‘Virtue Ethics: What Kind of Naturalism?’. In Gardiner (2005): ch. 1. Anscombe, G.E.M. (1958) ‘Modern Moral Philosophy’, Philosophy 33(124): 1–19. Antonaccio, Maria (2001) ‘Picturing the Soul: Moral Psychology and the Recovery of Emotions’, Ethical Theory and Moral Practice 4: 127–141. —— & Schweiker, W (1996) (eds) Iris Murdoch and the Search for Human Goodness. Chicago: University of Chicago Press. Appiah, Kwame Anthony (2008) Experiments in Ethics. Cambridge, MA: Harvard University Press. Badhwar, Neera K. (2003) ‘Love’. In LaFollette (2003): 42–69. Bagnoli, Carla (2000) ‘Value in the Guise of Regret’, Philosophical Explorations 3: 165–87. —— (2003) ‘Respect and Loving Attention’, Canadian Journal of Philosophy 33: 483–516. —— (2007) ‘Respect and Membership in the Moral Community’, Ethical Theory and Moral Practice 10(2): 113–28. —— (2009) ‘Practical Necessity: the Subjective Experience’. In Centi & Huemer (2009): 23–43. —— (2011) ‘The Exploration of Moral Life’. In Broakes (2011, forthcoming),193–221. Baier, Annette (1985) Postures of the Mind: Essays on Mind and Morals. University of Minnesota Press. —— (1991) A Progress of Sentiments: Reflections on Hume’s Treatise. Cambridge, MA: Harvard University Press. —— (2010a) Reflections On How We Live. Oxford: Oxford University Press. —— (2010b) The Cautious Jealous Virtue: Hume on Justice. Cambridge, MA: Harvard University Press. Baron, Marcia (1995) Kantian Ethics Almost without Apology. New York: Cornell University Press. Bedford, E. (1957) ‘Emotions’, Proceedings of the Artistotelian Society 57: 281–304. Ben-Ze’ev, Aaron (2000) The Subtlety of Emotions. Cambridge, MA: MIT Press. —— (2002) ‘Are Envy, Anger, and Resentment Moral Emotions?’, Philosophical Explorations 5 (2): 148–54. —— (2010) ‘The Thing Called Emotion’. In Goldie (2009): ch. 2. Bennett, Jonathan (1974) ‘The Conscience of Huckleberry Finn’, Philosophy 49: 123–34. Berker, Selim (2011) ‘The Normative Insignificance of Neuroscience’, Philosophy and Public Affairs (forthcoming). Betzler, Monica (2007) (ed.) Kant’s Virtue Ethics. New York and Berlin: de Gruyter. Blackburn, Simon (1984) Spreading the Word. Oxford: Clarendon Press. —— (1985) ‘Errors and the Phenomenology of Value’. In Honderich (1985): ch. 1. —— (1998) Ruling Passions. Oxford: Oxford University Press. Blum, Lawrence (1980) Friendship, Altruism, and Morality. London: Routledge & Kegan Paul. —— (1986) ‘Iris Murdoch and the Domain of the Moral’, Philosophical Studies 50: 343–67. —— (1991) ‘Moral Perception and Particularity’, Ethics 101: 701–25. Borges, Maria (2004) ‘What Can Kant Teach Us About Emotions?’, The Journal of Philosophy 4: 140–58. Brentano, Franz Clemens (1889/[1969]) The Origin of our Knowledge of Right and Wrong. London: Routledge & Kegan Paul. Broackes, Justin (2011) Iris Murdoch, Philosopher. Oxford: Oxford University Press. Brun G., Doguoglu, U., & Kuenzle, D. (2008) (eds) Epistemology and Emotions. Farnham: Ashgate Publishing.



Cagle, Randy (2005) ‘Becoming a Virtuous Agent: Kant and the Cultivation of Feelings and Emotions’, Kant-Studien 96(4): 452–67. Calhoun, Cheshire (2004) (ed.) Setting the Moral Compass: Essays by Women Philosophers. Oxford: Oxford University Press. Carr, David (2009) ‘Virtue, Mixed Emotions and Moral Ambivalence’, Philosophy 8(1): 31–46. Centi, Beatrice & Huemer, Wolfgang (2009) (eds) Value and Ontology. Frankfurt: Ontos-Verlag. Churchland, Paul (1996) The Engine of Reason, The Seat of the Soul: A Philosophical Journey into the Brain. Cambridge, MA: MIT Press. Cohen, D. (2001) ‘An fMRI Investigation of Emotional Engagement in Moral Judgment’, Science 293 (14 September 2001). Copp, David (2006) (ed.) Oxford Handbook of Ethical Theory. New York: Oxford University Press. —— (2011) ‘Experiments, Intuitions, and Methodology in Moral and Political Theory’, Oxford Studies in Metaethics (forthcoming). Cosmides, Leda & Tooby, John (2000) ‘Evolutionary Psychology and the Emotions’. In Lewis & Haviland-Jones (2000): 91–115. Crisp, Roger (2006) Reasons and the Good. Oxford: Oxford University Press. D’Arms, Justin & Jacobson, Daniel (1994) ‘Expressivism, Morality, and the Emotions’, Ethics 104: 739–63. —— (2000a) ‘Sentiment and Value’, Ethics 110: 722–48. —— (2000b) ‘The Moralistic Fallacy: On the “Appropriateness” of Emotions’, Philosophy and Phenomenological Research 61: 65–90. —— (2003) ‘The Significance of Recalcitrant Emotion’. In Hatzimoysis (2003): 127–46. —— (2006) ‘Sensibility Theory and Projectivism’. In Copp (2006): 186–219. —— (2010) ‘Demystifying Sensibilities: Sentimental Values and the Instability of Affect’. In Goldie (2010): ch. 26. Damasio, Antonio (1994) Descartes’ Error: Emotion, Reason, and the Human Brain. New York: G.P. Putnam’s Sons. —— (1999) The Feeling of what Happens: Body and Emotion in the Making of Consciousness. New York: Harcourt Brace and Co. Darwall, Stephen (1977) ‘Two Kinds of Respect’, Ethics 88: 36–49. —— (2006) The Second Person Standpoint. Cambridge, MA: Harvard University Press. —— (2011) ‘Bipolar Obligation’, Oxford Studies in Metaethics (forthcoming). de Sousa, Ronald (1987) The Rationality of Emotion. Cambridge, MA: MIT Press. —— (2006) ‘Moral Emotions’, Ethical Theory and Moral Practice 4: 109–26. —— (2010a) ‘The Mind’s Bermuda Triangle: Philosophy of Emotions and Empirical Science’, in Goldie (2010): ch. 4. —— (2010b) ‘Emotion’, Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.) http://plato. Deigh, John (1991) ‘Impartiality: A Closing Note’, Ethics 101(4): 858–64. —— (1994) ‘Cognitivism in the Theory of Emotions’, Ethics 104(4): 824–54. —— (1995) ‘Empathy and Universalizability’, Ethics 105(4): 743–63. —— (1996) The Sources of Moral Agency. Cambridge: Cambridge University Press. —— (2010) ‘Concepts of Emotions in Modern Philosophy and Psychology’. In Goldie (2010): ch. 1.



DePaul, Michael & Zagzebski, Linda (2003) (eds) Intellectual Virtue: Perspectives from Ethics and Epistemology. Oxford: Oxford University Press. Diamond, Cora (1988) ‘Losing Your Concepts’, Ethics 98(2): 255–77. —— (1996) ‘We are Perpetually Moralists: Iris Murdoch, Fact and Value’. In Antonaccio & Schweiker (1996): 79–109. Dillon, Robin (1992) ‘Respect and Care: Toward Moral Integration’, Canadian Journal of Philosophy 22: 105–32. —— (1997) ‘Self-Respect: Moral, Emotional, Political’, Ethics 107: 226–49. —— (2010) ‘Respect’, Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.) http://plato Do¨ring, S. (2003) ‘Explaining Action by Emotion’, The Philosophical Quarterly 53: 214–30. Doris, John M. & Stich, Stephen (2005) ‘As a Matter of Fact: Empirical perspectives on ethics’, in Jackson & Smith (2005): ch. 5. —— (2006) ‘Moral Psychology’, Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.) Elgin, Catherine (2008) ‘Emotion and Understanding’. In Brun, Doguoglu, & Kuenzle (2008): ch. 1. Elster, Jon (1999) Alchemies of the Mind: Rationality and the Emotions. Cambridge: Cambridge University Press. —— (2009) ‘Emotional Choice and Rational Choice’. In Goldie (2009): ch. 11. Engstrom, Stephen (2009) The Form of Practical Knowledge. Cambridge, MA: Harvard University Press. —— & Whiting, Jennifer (1997) (eds) Aristotle, Kant and the Stoics. Cambridge: Cambridge University Press. Evans, Dylan (2001) Emotions: The Science of Sentiment. Oxford: Oxford University Press. —— & Cruse, P. (2004) (eds) Emotion, Evolution, and Rationality. Oxford: Oxford University Press. Faucher, Luc & Tappolet, Christine (2006) (eds) The Modularity of Emotions, The Canadian Journal of Philosophy, supp. vol. 32. Flanagan, Owen (1991) Varieties of Moral Personality: Ethics and Psychological Realism. Cambridge, MA: Harvard University Press. —— & Rorty, Ame´lie O. (1990) (eds) Identity, Character, and Morality. Cambridge, MA: MIT Press. Foot, Philippa (1994) ‘Nietzsche’s Immoralism’. In Schacht (1994): 3–14. —— (2001) Natural Goodness. Oxford: Oxford University Press. Frank, Robert (1988) Passions Within Reason: The Strategic Role of the Emotions. New York: W.W. Norton. Frankfurt, Harry (1999) Necessity, Volition, and Love. Cambridge: Cambridge University Press. —— (2004) The Reasons of Love. Princeton, NJ: Princeton University Press. French, Peter & Wettstein, Howard (1999) (eds) Philosophy of Emotions, Midwest Studies in Philosophy, vol. 22. Frijda, Nico H. (1986) The Emotions. Cambridge: Cambridge University Press. Gallagher, Shawn (1997) ‘Mutual Enlightenment: Recent Phenomenology in Cognitive Science’, Journal of Consciousness Studies 4: 195–214.



—— (2000) ‘Philosophical Conceptions of the Self: Implications for Cognitive Science’, Trends in Cognitive Science 4: 14–21. —— (2005) How the Body Shapes the Mind. Oxford: Oxford University Press. Gardiner, Stephen M. (2005) (ed.) Virtue Ethics, Old and New. Cornell University Press. Gibbard, Allan (1990) Wise Choices, Apt Feelings. Cambridge, MA: Harvard University Press. Goldie, Peter (2001) The Emotions: A Philosophical Exploration. Oxford: Oxford University Press. —— (2004) ‘Emotion, Reason, and Virtue’. In Evans & Cruse (2004): ch. 13. —— (2010) The Oxford Handbook of Philosophy of Emotion. Oxford: Oxford University Press. Gordon, Robert (1987) The Structure of Emotions. Cambridge: Cambridge University Press. Greene, Joshua et al. (2001) ‘An fMRI Investigation of Emotional Engagement in Moral Judgment’, Science 293 (5537): 2105–8. Greenspan, Patricia S. (1981) ‘Emotions and Evaluations’, Pacific Philosophical Quarterly 62: 162–80. —— (1988) Emotions and Reasons: An Inquiry into Emotional Justification. New York: Routledge, Chapman and Hall. —— (1995) Practical Guilt: Moral Dilemmas, Emotions, and Social Norms. Oxford: Oxford University Press. —— (2000) ‘Emotional Strategies and Rationality’, Ethics 110: 469–87. —— (2004) ‘Practical Reasoning and Emotion’. In Mele & Rawling (2004): ch. 11. —— (2007) ‘Practical Reasons and Moral “Ought”’. In: Oxford Studies in Metaethics, vol. 2, R. Shafer-Landau (ed.), pp. 172–94. Oxford: Oxford University Press. —— (2010) ‘Learning Emotions and Ethics’. In Goldie (2010): ch. 24. Griffiths, Paul (1997) What Emotions Really Are: The Problem of Psychological Categories. Chicago: University of Chicago Press. Haidt, J., Koller, S., & Dias, M. (1993) ‘Affect, Culture, and Morality; or, Is it Wrong to Eat your Dog?’, Journal of Personality and Social Psychology 65: 617–19. Hatzimoysis, A., (2003) (ed.) Philosophy and the Emotions. Cambridge: Cambridge University Press. Helm, Bennett W. (1994) ‘The Significance of Emotions’, American Philosophical Quarterly 31: 319–31. —— (2001) Emotional Reason: Deliberation, Motivation and the Nature of Value. Cambridge: Cambridge University Press. —— (2002) ‘Emotions and Practical Reason: Rethinking Evaluation and Motivation’, Noûs 35: 190–213. —— (2009) ‘Love’, Stanford Encyclopedia of Philosophy, Edward N. Zalta (ed.) http://plato. —— (2010a) ‘Emotions and Motivation: Reconsidering Neo-Jamesian Accounts’. In Goldie (2010): ch. 13. —— (2010b) Love, Friendship, and the Self: Intimacy, Identification, and the Social Nature of Persons. Oxford: Oxford University Press. Herman, Barbara (1993) The Practice of Moral Judgment. Cambridge, MA: Harvard University Press. —— (1997) ‘Making Room for Character’. In Engstrom & Whiting (1997): ch. 2. —— (2007) Moral Literacy. Cambridge, MA: Harvard University Press. Honderich, Ted (1985) (ed.) Morality and Objectivity. Routledge & Kegan Paul. Hursthouse, Rosalind (1999) On Virtue Ethics. Oxford: Oxford University Press.



Hursthouse, Rosalind, Lawrence, G., & Quinn, Warren (1995) (eds) Virtues and Reasons. Oxford: Clarendon. Husserl, Edmund (1988) Vorlesung über Ethik und Wertlehre, Husserliana. Vol. XXVIII, U. Melle (ed.). Dordrecht: Kluwer. Jackson, Frank & Smith, Michael (2005) (eds) The Oxford Handbook of Contemporary Analytic Philosophy. Oxford: Oxford University Press. Johnston, Mark (2001) ‘The Authority of Affect’, Philosophy and Phenomenological Research 63 (1): 181–214. Jones, Karen (2004) ‘Emotional Rationality as Practical Rationality’. In Cheshire (2004): ch. 18. —— (2006) ‘Metaethics and Emotions Research: A Response to Prinz’, Philosophical Explorations 9(1): 45–53. Jost, L. & Wuerth, Julian (2011) (eds) Perfecting Virtue: New Essays on Kantian Ethics and Virtue Ethics. Cambridge: Cambridge University Press. Kant, Immanuel (1785) Grundlegung zur Metaphysik der Sitten, in Kants gesammelte Schriften, Preussische Akademie der Wissenschaften, Berlin, vol. 6, Groundwork of the Metaphysic of Morals, in Practical Philosophy, A. Wood (ed.). Cambridge University Press, 1983. —— (1797a) Metaphysical Principles of Virtue, Kants gesammelte Schriften, Preussische Akademie der Wissenschaften, Berlin, 1907, vol. 6, in Practical Philosophy, A. Wood (ed.), pp. 353–604. Cambridge: Cambridge University Press, 1996. —— (1797b) Anthropology from a Pragmatic Point of View, Kants gesammelte Schriften, Preussische Akademie der Wissenschaften, Berlin, 1907, vol. 7, ed. Robert B. Louden & Manfred Kuehn. Cambridge University Press, 2006. Kennett, Jeanette & Fine, Cordelia (2009) ‘Will the Real Moral Judgment Please Stand Up?’, Ethical Theory and Moral Practice 12(1): 77–96. Kenny, Anthony (1963) Action, Emotion and the Will. London: Routledge & Kegan Paul. Knobe, Joshua & Nichols, Shaun (2008a) ‘An Experimental Philosophy Manifesto’. In Knobe & Nichols (2008): ch. 1. —— (2008b) Experimental Philosophy. New York: Oxford University Press. Korsgaard, Christine M. (1996) Creating the Kingdom of Ends. New York: Cambridge University Press. —— (2008) The Constitution of Agency. Oxford: Oxford University Press. —— (2009) Self-Constitution: Agency, Identity, and Integrity. Oxford University Press. La Caze, Marguerite (2001) ‘Envy and Resentment’, Philosophical Explorations 4(1): 31–45. Lacewing, Michael (2005) ‘Emotional Self-Awareness and Ethical Deliberation’, Ratio 18(1): 65–81. LaFollette, Hugh (2005) (ed.) Practical Ethics. Oxford: Oxford University Press. Lauritzen, R. (1991) ‘Errors of an Ill-Reasoning: The disparagement of emotions in the moral life’, The Journal of Value Inquiry 25: 5–21. Lazarus, Richard (1991) ‘Cognition and Motivation in Emotion’, American Psychologist 46: 362–7. Lear, Jonathan (1998) Open Minded: Working Out the Logic of the Soul. Cambridge, MA: Harvard University Press. —— (2004) ‘Psychoanalysis and the Idea of a Moral Psychology: Memorial to Bernard Williams’ Philosophy’, Inquiry 47(5): 515–22. LeDoux, John (1996) The Emotional Brain. New York: Simon & Schuster. Levy, Neil (2006) ‘Cognitive Scientific Challenges to Morality’, Philosophical Psychology 9: 567–87.



Lewis, Michael & Haviland-Jones, Jeannette M. (2000) (eds) Handbook of Emotions. New York: Guilford Press. Libet, Benjamin (1987) ‘Are the mental experiences of will and self-control significant for the performance of a voluntary act?’, Behavioral and Brain Sciences 10: 783–6. Louden, Robert (2000) Kant’s Impure Ethics: from Rational Beings to Human Beings. New York: Oxford University Press. Lovibond, Sabina (2002) Ethical Formation. Cambridge, MA: Harvard University Press. McDowell, John (1985) ‘Values and Secondary Qualities’. In Honderich (1985): ch. 6. —— (1995) ‘Two Sorts of Naturalism’. In Hursthouse, Lawrence, & Quinn (1995): ch. 6. —— (1998a) ‘Projection and Truth in Ethics’. In McDowell (1998): ch. 8. —— (1998b) Mind, Value, and Reality. Cambridge, MA: Harvard University Press. Mackie, John L. (1977) Ethics: Inventing Right and Wrong. London: Penguin. Meinong, Alexius (1917) On Emotional Presentation. Evanston, Ill.: Northwestern University Press. Mele, Alfred & Rawling, Peter (2004) (eds) The Oxford Handbook of Rationality. Oxford: Oxford University Press. Moore, A.W. (2006) Philosophy As A Humanistic Discipline. Princeton, NJ: Princeton University Press. Moran, Richard (2002) ‘Vision, Choice, and Existentialism’, Notizie di Politeia 66: 88–101. —— (2011) ‘Iris Murdoch and Existentialism’. In Broackes (2011): ch. 4. Mulligan, Kevin (1998) ‘From Appropriate Emotions to Values’, Monist 81: 161–88. Murdoch, Iris (1956) ‘Vision and Choice in Morality’, Proceedings of the Aristotelian Society 30: 32–58. —— (1957) ‘Metaphysics and Ethics’. In Murdoch (1997): ch. 4. —— (1970) The Sovereignty of Good. London: Routledge. —— (1997) Existentialists and Mystics, Peter Conradi (ed.). London: Penguin Books. Nagel, Thomas (1970) The Possibility of Altruism. Princeton: Princeton University Press. Neu, Jerome (2010) ‘An Ethics of Emotion?’. In Goldie (2010): ch. 22. Nichols, Shaun (2004) Sentimental Rules: On the Natural Foundations of Moral Judgment. Oxford: Oxford University Press. Nietzsche, Friedrich (1968) On the Genealogy of Morals. New York: Modern Library. Nussbaum, Martha C. (1986) The Fragility of Goodness. Cambridge: Cambridge University Press. —— (1990) Love’s Knowledge. Oxford: Oxford University Press. —— (1993) ‘Equity and Mercy’, Philosophy and Public Affairs 22: 83–125. —— (1999) ‘Virtue Ethics: A Misleading Category?’, Journal of Ethics 3(3): 163–201. —— (2001) Upheavals of Thought: The Intelligence of Emotions. Cambridge: Cambridge University Press. O’Neill, Onora (1985) ‘Consistency in Action’. In Potter & Timmons (1985): ch. 14. Oakley, Justin (1992) Morality and the Emotions. London: Routledge & Kegan Paul. Panksepp, Jaak (1998) Affective neuroscience: the foundations of human and animal emotions. New York: Oxford University Press. —— (2000) ‘Emotion as a Natural Kind Within the Mammalian Brain’. In Lewis & HavilandJones (2000): ch. 9. Peacocke, Christopher (2004) Realm of Reason. New York: Oxford University Press. Picard, Rosalind (1997) Affective Computing. Cambridge, MA: MIT Press. Pinker, Steven (2002) The Blank Slate: The Modern Denial of Human Nature. New York: Viking. Plutchik, Robert (1980) Emotion: A Psychoevolutionary Synthesis. New York: Harper and Row.



Potter, Nelson & Timmons, Mark (1985) (eds) Morality and Universality. Dordrecht: Reidel. Price, A.W. (2010) ‘Emotions in Plato and Aristotle’. In Goldie (2010): ch. 5. Prinz, Jesse (2004) Gut Reactions: A Perceptual Theory of Emotion. New York: Oxford University Press. —— (2006) ‘The Emotional Basis of Moral Judgments’, Philosophical Explorations 9(1): 29–43. —— (2007) The Emotional Construction of Morals. New York: Oxford University Press. —— (2009) ‘The Normativity Challenge: Cultural Psychology Provides the Real Threat to Virtue Ethics’, Journal of Ethics 13: 2–3. —— (2010) ‘The Moral Emotions’. In Goldie (2010): ch. 23. Pugmire, David (1998) Rediscovering Emotion. Edinburgh: Edinburgh University Press. —— (2005) Sound Sentiments: Integrity in the Emotions. New York: Oxford University Press. Putnam, Hilary (1992) Renewing Philosophy. Cambridge, MA: Harvard University Press. —— (2002) The Collapse of the Fact/Value Distinction. Cambridge, MA: Harvard University Press. —— (2004) Ethics without Ontology. Cambridge, MA: Harvard University Press. Rawls, John (1971) A Theory of Justice. Cambridge, MA: Harvard University Press. Raz, Joseph (1997) ‘The Active and the Passive’, Aristotelian Society Supplementary Volume 71(1): 211–28. —— (2001) Value, Respect and Attachment. Cambridge: Cambridge University Press. —— (2010) ‘On Respect, Authority, and Neutrality. A Response’, Ethics 120: 279–301. Reath, Andrews (1980) ‘Moral Sensibility’, Kant Studien 80: 284–302. Roberts, Robert C. (2003) Emotions. Cambridge: Cambridge University Press. Robinson, Jenefer (2005) Deeper than Reason: Emotion and Its Role in Literature, Music, and Art. Oxford: Oxford University Press. Roeser, Sabine (2010) Moral Emotions and Intuitions. London: Palgrave Macmillan. Rorty, Ame´lie O. (1980a) (ed.) Explaining Emotions. Berkeley: University of California Press. —— (1980b) (ed.) Essays on Aristotle’s Ethics. University of California Press. —— (1987) ‘The Historicity of Psychological Attitudes: Love is Not Love Which Alters Not When It Alteration Finds’, Midwest Studies in Philosophy 10(1): 399–412. —— (1988) The Mind in Action. Boston: Beacon Press. —— (1998a) ‘Political Sources of Emotions: Greed and Anger’, Philosophical Studies 89(2–3): 21–33. —— (1998b) ‘Political Sources of Emotions’, Midwest Studies in Philosophy 12: 21–33. Ruddick, Sara (1980) Maternal Thinking: Toward a Politics of Peace. Boston: Beacon Press. Sabini, John & Silver, Maury (1987) ‘Emotions, Responsibility, and Character’, in Schoeman (1987): ch. 7. Sankowski, Edward (1977) ‘Responsibility of Persons for Their Emotions’, Canadian Journal of Philosophy 7(4): 829–40. Scanlon, Thomas (2008) Moral Dimensions: Permissibility, Meaning, Blame, Cambridge, MA: Harvard University Press. Schacht, Robert (1994) (ed.) Nietzsche, Genealogy, Morality. Berkeley and Los Angeles: University of California Press. Scheler, Max (1913–1916) Formalism in Ethics and the Non-Formal Ethics of Values. Evanston, Ill.: Northwestern University Press. Schoeman, Frederich (1987) (ed.) Responsibility, Character, and the Emotions. New York: Cambridge University Press. Schroeder, Mark A. (2007) Slaves of the Passions. Oxford University Press.



Sherman, Nancy (1990) ‘The Place of Emotions in Kantian Morality’. In Flanagan & Rorty (1990): ch. 7. —— & White, Heath (2003) ‘Intellectual Virtue: Emotions, Luck, and the Ancients’. In DePaul & Zagzebski (2003): ch. 2. Sinnott-Armstrong, Walter (2008a) ‘Framing Moral Intuitions’. In Sinnott-Armstrong (2008b): ch. 2. —— (2008b) Moral Psychology: Volume 2, The Cognitive Science of Morality: Intuition and Diversity. Cambridge, MA: MIT Press. Skorupski, John (2010a) ‘Sentimentalism: Its Scope and Limits’. In Ethical Theory and Moral Practice, 13: 125–36. —— (2010b) ‘Moral Obligation, Blame, and Self-governance’, Social Philosophy and Policy 27(2): 158–80. Slote, Michael (2006) ‘Moral Sentimentalism and Moral Psychology’. In Copp (2006): ch. 8. Smith, Angela (2005) ‘Responsibility for Attitudes: Activity and Passivity in Mental Life’, Ethics 115(2): 236–71. Solomon, Robert (1980) ‘Emotions and Choice’. In Rorty (1980): ch. 10. —— (1984) The Passions: The Myth and Nature of Human Emotions. New York: Doubleday. —— (1999) ‘The Philosophy of Emotions’. In Lewis & Haviland-Jones (1999): ch. 1. —— (2003) Not Passion’s Slave. Oxford University Press. Stark, Susan (2001) ‘Virtue and Emotion’, Noûs 35: 440–55. —— (2004) ‘A Change of Heart: Moral Emotions, Transformation, and Moral Virtue’, Journal of Moral Philosophy 1: 31–50. Starkey, Charles (2008) ‘Emotion and Full Understanding’, Ethical Theory and Moral Practice 11: 425–54. Stocker, Michael (1994) ‘Emotions and Ethical Knowledge: Some Naturalistic Connections’, Midwest Studies in Philosophy 19(1): 143–58. —— with Elizabeth Hegeman (1996) Valuing Emotions. Cambridge: Cambridge University Press. Strawson, Peter (1962) ‘Freedom and Resentment’. In Strawson (1974): ch. 1. —— (1974) Freedom and Resentment. London: Methuen. Tappolet, Christine (2000) Emotions et Valeurs. Paris: Presses Universitaires de France. —— (2003) ‘The Emotions and the Intelligibility of Akratic Action’. In Tappolet & Stroud (2003): ch. 4. —— (2006) ‘Autonomy and the Emotions’, European Journal of Analytic Philosophy 2: 45–60. —— (2010) ‘Emotion, Motivation, and Action: The Case of Fear’. In Goldie (2010): ch. 14. —— & Stroud, Sarah (2003) (eds) Weakness of the Will and Practical Irrationality. Oxford: Clarendon Press. Taylor, Gabriele (1985) Pride, Shame, and Guilt: Emotions of Self-Assessment. Oxford: Oxford University Press. Teroni, Fabrice (2007) ‘Emotions and Formal Objects’, Dialectica 61: 395–415. Thalberg, Irvin (1978) ‘Mental Activity and Passivity’, Mind 87: 376–95. Thompson, Michael (2008) Life and Action: Elementary Structures of Practice and Practical Thought. Cambridge, MA: Harvard University Press. Thomson, Judith Jarvis (1976) ‘Killing, Letting Die, and the Trolley Problem’, The Monist 59: 204–17.



Thomson, Judith Jarvis (2008) ‘Turning the Trolley’, Philosophy and Public Affairs 36: 359–74. Velleman, J. David (1999) ‘Love as a Moral Emotion’, Ethics 109: 338–74. —— (2006) Self to Self. Cambridge: Cambridge University Press. Wallace, R. Jay (1994) Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. —— (2005) ‘Moral Psychology’. In Jackson & Smith (2005): ch. 4. Wheatley, T. & Haidt, John (2005) ‘Hypnotically Induced Disgust Makes Moral Judgments More Severe’, Psychological Science 16: 780–4. Wiggins, David (1987a) ‘A Sensible Subjectivism?’. In Wiggins (1987c): ch. 5. —— (1987b) ‘Truth, Invention, and the Meaning of Life’. In Wiggins (1987c): ch. 3. —— (1987c) Needs, Values, Truth. Oxford: Blackwell Publishers. Williams, Bernard (1966) ‘Morality and the Emotions’. In Williams (1973): ch. 13. —— (1973) Problems of the Self. Cambridge: Cambridge University Press. —— (1981) Moral Luck. Cambridge: Cambridge University Press. —— (1985) Ethics and the Limits of Philosophy. Cambridge, MA: Harvard University Press. —— (1995) ‘What Does Intuitionism Imply?’. In his Making Sense of Humanity, pp. 153–92. Cambridge: Cambridge University Press. —— (2000) ‘Philosophy as a Humanistic Discipline’. In Moore (2006): ch. 16. Wollheim, Richard (1999) On The Emotions. New Haven: Yale University Press. Wood, Allan (1999) Kant’s Ethical Thought. Cambridge: Cambridge University Press. Wuerth, Julian (2011) ‘Moving Beyond Kant’s Account of Agency in the Grounding’. In Jost & Wuerth (2011): ch. 6. Zagzebski, Linda (2003) ‘Emotion and Moral Judgment’, Philosophy and Phenomenological Research 66(1): 104–24. Zajonc, Robert (1980) ‘Feeling and Thinking: Preferences Need No Inferences’, American Psychologist 35: 151–75. Zemach, Eddy M. (2001) ‘What Is Emotion?’, American Philosophical Quarterly 38(2): 197–207. Zhu Jing & Thagard, Paul (2002) ‘Emotion and Action’, Philosophical Psychology 15(1): 19–36.


This page intentionally left blank

1 Craving the Right Emotions and Moral Reasons1 Patricia Greenspan

I first began working on emotions as a project in philosophy of action, without particular reference to moral philosophy. My thought was that emotions play a role in rationality that tends to be under-appreciated by philosophers. Bringing this out was meant to counter a widespread tendency to treat emotions just as ‘blind’ causes of action (for the general picture, see Greenspan (2009)). Instead, I thought they could be seen as providing reasons. I took their significance as moral motivators to be hard to miss. Of course, philosophers and others sometimes rightly insist that we need to put emotions aside in order to formulate satisfying moral principles, but I would have been surprised to hear anyone deny that moral motivation typically rests on emotion and that we need that basis in early life in order to get to the stage of acting on moral principles. However, I have since come to think that none of the main philosophical approaches to ethics fully appreciates the role of emotion, in part because of a misconception of practical reasons. Reasons for action are commonly taken as prima facie requirements, so that moral reasons would yield requirements just insofar as they outweigh competing reasons such as reasons of simple self-interest. Someone who recognizes a moral reason as holding, all things considered, would thus be irrational not to act on it. But I argue in recent work (starting with Greenspan (2005)) that even all-things-considered reasons may in one sense be optional: a rational agent can legitimately discount them, canceling their deliberative weight and their force for motivation. What keeps us from setting aside reasons of the sort that underlie moral requirements—what explains the ‘binding’ force of moral obligation, on the deontological approach to ethics I favor—is not that moral reasons are necessarily our weightiest. Instead, we lack authority to discount reasons that rest on criticism from the standpoints of other agents (see Greenspan (2007 and 2010b)). 1 Let me thank Karen Jones for extended discussion of the issues in this paper on my visit to Melbourne in July 2008, supported by Australian Research Council Discovery Project Grant DP0557651. Drafts of the paper also benefited from comments from Samuel Kerstein, Stephen Leighton, Christopher Morris, David Wasserman, and the Ethics Reading Group at Boston University.



This argument makes no mention of emotions, though it does leave room for them, by weakening the force ascribed to non-emotional reasons. It rests on what I call the critical conception of practical reasons, which interprets reasons as offering or answering criticism. Reasons in favor of something are seen in the first instance as responses to a more basic sort of reason that counts against an act or other practical option. The argument I put forth earlier on emotions (Greenspan (1988): ch. 6; cf. Greenspan (2004)) took them as supplying us with reasons capable of reinforcing our nonemotional reasons for doing what we ought, even in some cases where emotions diverge from warranted judgment. Anger, for instance, might augment the reasons in favor of a timely response to an insult, where weighing one’s non-emotional reasons favors restraint. My aim was to show how occurrent emotions can be factors in practical reasoning. But my argument intertwined normative and motivational conceptions of reasons that later work has distinguished more clearly. I sometimes spoke of emotions as themselves constituting reasons, which suggests an interpretation of reasons as states of mind rather than objective facts. I illustrated my argument, moreover, with a case of anger in response to an insult that did not bring out the moral significance of emotional reasons. Indeed, since I interpreted their force for action in terms of a need to improve the agent’s psychological state, my argument gave rise to questions about how a foundation in self-concern could be compatible with genuinely moral motivation. So I now want to reformulate the argument in terms of my later interpretation of reasons, with more attention to relevant distinctions in the area, and with particular focus on how emotions can reinforce moral reasons. In general, I want to say that emotions can serve as rational barriers to discounting reasons. Within the moral sphere, they supply reasons against discounting either moral reasons themselves or the supplementary reasons we have for taking the most reliable means to satisfying applicable moral reasons. Discounting moral reasons of the sort that yield requirements would exceed our rational authority and hence would be illicit, whereas discounting the instrumental reasons that supplement moral reasons would be licit but inadvisable, insofar as it involves taking a risk of moral failure. Later I focus particularly on the licit but inadvisable form of discounting, since I think it brings out more fully the rational significance of moral emotions. As barriers to the illicit discounting of moral reasons, emotions can provide important motivational backing to reasons sufficient to require action on their own; but as barriers to the inadvisable discounting of reasons serving to supplement moral reasons, emotions can actually supply further normative reasons needed to require action of us now. In slogan form: emotions help to make the case against moral delay. These claims will be explained more fully as they emerge in my argument below. For the moment, though, let me outline briefly my general approach to reasons as I take it to bear on emotions, at this point focusing on the simpler role of emotions as barriers to illicit discounting. Essentially, my current approach makes out emotions as a source of higher-order reasons: reasons against setting certain first-order reasons aside, or declining to take them into account in deliberation, in the way I think we sometimes



can do without compromising rationality. But what that role amounts to will take different forms, with different implications for moral motivation, depending on what sorts of first-order reasons are in question. The critical conception of practical reasons makes a basic distinction between reasons that yield requirements and reasons that merely justify or commend (cf. Gert (2004) and Horgan & Timmons (2010)). Within the moral sphere, this amounts to a distinction between reasons of the sort that can generate moral requirements and reasons for morally good but optional acts such as ideal instances of virtue. I call reasons of the sort that yield requirements negative or critical reasons, since they count against alternatives to the required act, offering criticism of alternatives as in some way objectionable. The contrast is to (purely) positive or favoring reasons, which cite valuable features of an act as answers to potential criticism but imply no significant criticism of alternatives and therefore may be discounted at will, without citing any higher-order reasons. Now, emotions clearly have a motivational role to play in getting us to satisfy moral favoring reasons that we otherwise might discount. Any independent reasons we recognize for an act of generosity, for instance, will be harder to ignore if they are backed up by love or sympathy for the person benefited by it. The story is more complicated for critical reasons, but also more interesting, if we take them as sources of binding requirements. All-things-considered critical non-moral reasons, despite their role as sources of non-moral requirements, will sometimes be subject to discounting by appeal to higher-order reasons such as those based on decisions to set certain priorities. For instance, I might decide to place a priority on single-minded devotion to research that excludes attention to reasons for service on university committees, even if those reasons sometimes happen to be weightier than my reasons for avoiding distractions from a particular research project. However, while a rational agent might sometimes be inclined to discount critical moral reasons—to prioritize self-interest, whatever its weight relative to competing interests—doing so would indicate a failure to appreciate fully the status of those reasons as based on criticism from the standpoints of other agents.2 We are authorized to discount only our own potential criticism, in short. So whatever emotional qualms we feel about wrongful treatment of others provides a kind of back-up reason, against exceeding our rational authority with regard to applicable moral reasons. Insofar as they reinforce critical moral reasons, then, emotions serve to block a sophisticated rational maneuver from being misapplied to the moral sphere. That, in highly condensed form, is the basic point I want to defend in this paper, as ascribing a normative role to emotions that is compatible with both rationality and genuinely moral motivation. The simpler point about emotions as motivating action on moral

2 My account of moral reasons here is in broad sympathy with the ‘second-personal’ view in Darwall (2006); it originally was based on reflection on Thompson (2004). I also take others’ criticisms (e.g. administrators’ potential criticisms of my failure to serve on university committees) to give rise to nonmoral reasons, but only by virtue of their relation to the agent’s standpoint (e.g. because I have something to gain from pleasing administrators).



favoring reasons, adding inducements to virtue, seems obvious enough. Perhaps that is why assigning importance to emotions in ethics usually tends to be interpreted as support for virtue ethics. A number of authors understand emotions as ways of perceiving values (see especially de Sousa (1987)), which fits well with an Aristotelian approach, and even recent Kantian arguments against taking Kant as dismissive of emotions (e.g. Sherman (1990)) defend him by bringing out what he has to say about virtue. But I have wanted to make out emotions as significant factors in deontological ethics as well. My interpretation of practical reasons in terms of criticism is itself intended to support a deontological approach: favoring reasons, however many and however strong, need not add up to a requirement on my account, so the binding force of moral obligation cannot be explained just by the weight of reasons in favor of satisfying it or by the features making a certain option best. What I want to say of emotions, then, is not just that they let us perceive values, and in that sense ‘register’ favoring reasons, but also that they give rise to further critical reasons. The latter are objective normative reasons, not just motives, though awareness of them (along with habits formed on the basis of that awareness) is the source of morality’s motivational force. Their weight may be relatively trivial—as we need to grant in order to avoid ‘bootstrapping’ problems (taking action on misdirected emotions to be rational in light of them)—but they can add crucial reinforcement to moral and other requirements. My discussion in what follows will be framed in deontological terms, but my central points should also apply to Aristotelian theories that eschew talk of obligation, as long as they have room for a threshold of basic moral worth that would demand a modicum of virtue of all moral agents, whether or not they attain an ideal of virtue or the status of virtuous persons. Some of the more recent versions of virtue ethics instead seem to be offshoots of the ‘positive thinking’ movement; I briefly take issue with a related approach to emotions toward the end of this paper. But the intended contrast to my understanding of moral requirements in terms of critical reasons is a consequentialist account of them simply as a function of our weightiest favoring reasons. Taking certain acts as inappropriate on a virtue-ethical account should be enough to yield a notion of requirement, though my treatment of moral requirement would have trouble accommodating virtue-ethical approaches that fail to accord a special place to the demands of moral virtue. In this way and others, I also depart from the widespread assumption that a stress on the role of emotion in ethics is essentially Humean. Morality on my account does have something like a Humean basis in childhood development. Elsewhere (in Greenspan (2010a); cf. Greenspan (1995 and 1998)) I argue that the role of emotions in early moral learning is what initially supplies moral judgments with motivational force and that the need for this basis constrains the content of morality, so that emotions figure indirectly among the grounds for moral judgment. In this paper I instead turn back to my earlier argument for the role of emotions as sources of reasons in adult life, but now formulated more explicitly in connection with moral reasons, which need not retain a basis in occurrent emotion at that stage. First, in sections 1.1–1.2, I extend the argument in light of later work on



reasons and spell it out in fuller detail to make better sense of the claim that emotional reasons can be factors in rational and moral justification, even if relatively trivial in terms of weight. Then, in sections 1.3–1.4, I show why it is important that their independent weight is minimal and that they serve merely to reinforce moral and other non-emotional reasons. My fuller statement of the argument in terms of the critical conception of practical reasons should at that point help me answer two persistent objections to my 1988 argument: an objection from rational ‘bootstrapping’ on the basis of misdirected emotions, and an objection from self-concern as the ‘wrong sort of reason’ for moral action. I think that both of these rest on confusions in or about the argument that require a step-by-step reformulation.

1.1 Reinforcing Reasons I understand an emotion as a compound of affect and evaluation: pleasant or unpleasant feelings about valued or disvalued features attributed to something that counts as the object of emotion. The evaluative component of an appropriate emotion, while itself a state of awareness, can register a fact that amounts to an objective practical reason: a reason holding independently of anyone’s awareness of it. The fact, for instance, that someone has been treated unjustly is a reason against remaining silent, whether or not anyone recognizes the injustice or is upset about letting it go unchallenged. My claim about emotions is that a further reason is added to this if one does feel upset—not because being upset about something implies awareness of it, but rather as a fact about one’s own well-being. This is a further fact with normative implications insofar as it attributes to an agent a bad psychological state: a state of emotional discomfort. The fact that one is uncomfortable about something counts in itself as a reason—against failing to act as needed to prevent the feeling from continuing—whether or not one happens to be motivated by awareness of it. So an appropriate emotion, besides having an evaluative component that reflects a practical reason, can add as a further reason a criticism, from the agent’s standpoint, of failure to act to improve her own state of feeling. Where the evaluative component of emotion reflects a moral reason, though, one might ask why it should need any normative reinforcement. What could be weightier than a moral reason, after all? First, let me say that I do not take it for granted, as some philosophers do, that moral considerations necessarily outweigh all others. In any case, moral reasons sometimes compete among themselves. Reinforcement, even by a reason comparatively trivial—the need to improve one’s own state of feeling—might sometimes serve to tip a precarious balance. But secondly, I question whether weight is all that is at issue in comparing reasons. Another factor that came into my 1988 argument was the ‘pressuring’ aspect of emotions: the tendency of emotional discomfort to demand more or less immediate action. An emotion is slanted toward present satisfaction to the extent that action is needed to prevent continuation of its element of discomfort. This is not to say that the weight of the reason or reasons supplied by an emotion necessarily adds up over time to something non-trivial, but just that post-



ponement imposes a cost that undermines the effectiveness of action on it. There is a time constraint on satisfying an emotional reason. My argument here should indicate how its position in our overall structure of reasons may give an emotional reason a significance that is not just a matter of weight. Let us work with an everyday example: Brad, a department head, unfairly blocks a salary increase for Carl. Simply being aware of an injustice may not give Alice decisive reason to speak out on her colleague’s behalf, if others could do so instead while she held back. She could save her response until later, perhaps just in order to phrase her remarks better. But by then Brad’s action would have had its effect. Let us grant that a response is called for that might be rendered ineffective by postponement. Now, Alice might have a reason, possibly even a weightier reason, to remain silent, at least for the time being, leaving any immediate response to others. If we suppose, though, that she is upset at so far failing to respond, that will give her a reason against postponement. The need to avoid further emotional discomfort may have little weight in itself, but it adds an element of pressure to act now. My 1988 argument briefly presented a non-moral version of this sort of case, involving a response to a personal insult, with the emotional reason summed up (1988: 155) as I am uncomfortable at the thought that I ought to get back at X.

Before reformulating this for Alice’s case, some clarification is in order. Despite disclaimers, my representation of the reason by a first-person judgment may have given the impression of an attempt to capture a step in a passage of reasoning on the part of an agent contemplating action. But the aim was just a rational reconstruction of practical reasoning—or, more precisely, of the grounds for action—of the sort exemplified by Aristotle’s practical syllogism: putting into the form of a first-person argument what in psychological terms may just have been an automatic response to a tacitly accepted reason. Nor was the self-attribution of emotion meant just to capture a reason in some subjective sense, dependent on the agent’s state of belief about her emotional state, possibly mistaken. The indented statement attributes to the agent a subjective state of emotion, but the fact that someone is in that state can still be said to supply an objective reason, in the sense of a reason holding independently of her awareness of it. The ought judgment represented as an object of thought in the indented statement—‘I ought to get back at X’—asserts what I think of as the evaluative content of the relevant emotion. The emotion in my 1988 example was characterized as personal anger, and the parallel for Alice’s case would be moral indignation (though discomfort at so far failing to act on that emotion might be seen as a form of guilt; cf. Greenspan (1995)). The indented statement is somewhat abbreviated: the object of discomfort is not precisely the requirement to get back at X, but rather the fact that I have yet to satisfy that requirement; it might be spelled out further as ‘[the thought that] I still ought to get back at X’. My claim was that, assuming the appropriateness of the emotion, the fact of emotional discomfort could reinforce independent reasons for action



on the ought judgment. I now want to ask how my later work on critical reasons and discounting might help make sense of this claim, as applied to reasons supporting a moral ought judgment, of the sort now illustrated in the third person by the case of Alice. Let me proceed slowly, at the risk of belaboring a few points. I did not undertake the project on reasons with any thought of bringing it to bear on emotions, so my task entails combining two somewhat complex arguments, each of which turns on some claims at variance with common assumptions. I was surprised myself to find that a recent extension of my argument on reasons to the notion of imperfect obligation (Greenspan (2010b)) yields a way of explaining how a need to relieve emotional discomfort can add something significant to an agent’s reasons for action despite its own relatively trivial weight. My general claim is that emotion can supply, not just a motivational influence, but also a normative reason. It could do so even where inappropriate, but in that case its force for action would be negligible: it would provide just as much reason for various ways of alleviating discomfort that do not satisfy the ought judgment, in the way that a headache provides reason for taking an aspirin. I have more to say later about such cases, but for the moment we may assume that Alice’s emotion in the case described is appropriate. On my 1988 account this means that Alice has adequate reason to hold in mind its evaluative component, the ought judgment, by making it an object of affect (assumed to be proportionate in degree). Let us now even assume that a corresponding judgment, the judgment that an injustice has been committed and deserves a timely response, would be both warranted and true. The question is why it should matter that Alice actually feels discomfort at standing idly by. For it is the affective element of emotion that supplies a further normative reason on my account. An understanding of emotions as perceptions of value might see our feelings in such cases simply as serving to make us aware of relevant non-emotional reasons, so that emotions have epistemic rather than direct practical significance. On that account emotions would not supply, but rather just ‘track’, reasons that in principle could become known to us in other ways, even if affect serves to make them more salient and thereby affords us more immediate or reliable access to them (see Jones (2003 and 2004); cf. Rorty (1980), de Sousa (1987)). My alternative proposal will incorporate elements of a perceptual account, but I mean to add something further. Nor is it just that emotions have moral significance as signs of good character or praiseworthy motives. I should also spell out that my talk of ‘appropriate’ emotion in Greenspan (1988) was not meant to imply or rest on a judgment of moral appropriateness (cf. D’Arms & Jacobson (2000)), but just of something akin to evidential warrant; the term I used was ‘representational rationality’. Similarly, in my later work on reasons I extend ‘criticism’ beyond moral criticism to any claim that something is bad, possibly just for a given agent. So what I want to say about emotions is that, besides being ‘reasontrackers’, they also are ‘reason-providers’. Besides alerting us to moral reasons by registering them in affect—and thereby backing them up with a further motive for action, a practical ‘push’ of the sort suggested by psychologists’ talk of the ‘valence’ of



emotions (cf. e.g. Frijda (1986): 207)—emotional discomfort reinforces moral reasons with non-moral criticism of failure to act that yields a further reason. On this account emotions play a role within practical rationality, rather than merely serving as external inducements to it in the way that feelings of curiosity or worry about evidence might promote theoretical rationality. Emotional discomfort can sometimes add a layer of criticism to moral reasons that are not critical themselves. Consider again Alice’s reason to challenge Brad. A moral reason against failing to respond to an injustice essentially offers a criticism of failure to respond, but from a standpoint other than the agent’s—in the first instance, that of the victim of injustice. So Alice could not legitimately discount such criticism on her own, though she might counter it with weightier reasons for remaining silent, at any rate for the time being. For one thing, there presumably are other members of the department to whom she might leave the job of responding, assuming that the requirement plus other facts of the case do not pick out her in particular. It would be good for her to make some response, though also permissible for her not to. But the thought that she ought to may still be appropriate for her to hold in mind with emotion, even if the corresponding judgment is not strictly warranted in the terms that apply to belief. Insofar as the thought is an object of discomfort, moreover, the fact that her failure to respond would sustain that feeling constitutes a further, non-moral criticism of inaction: that it would prolong her discomfort, unless and until someone else happens to respond. In general, then, in a case of collective obligation, emotion can add a critical reason to what amounts merely to a favoring reason for action on the part of an individual agent, with no criticism specifically of her failure to act. It would not just add some slight weight to a pre-existing reason, but rather would introduce a critical reason capable of getting a general moral requirement to devolve on a particular agent.3 I now want to apply this model to what might be called ‘cross-temporal’ cases, in which the agent gets to choose when to act to satisfy an obligation.

1.2 Rational Pressure What we have in both collective and cross-temporal cases is an imperfect obligation: there are reasons, but merely favoring reasons, for various particular ways of satisfying the obligation (see Greenspan (2010b)), with moral criticism restricted to a failure to satisfy it at all. A collective obligation can be satisfied by different individuals, whereas in crosstemporal cases there are several different times at which one might act to satisfy an obligation. I now want to use this model to explain the sense in which emotions introduce

3 Note that I do not hold that critical reasons necessarily outweigh favoring reasons. Someone with a reasonable memory might still have an all-things-considered reason for taking a memory-enhancing tonic, say, despite the fact that the tonic tastes slightly bitter. My claim is just that only critical reasons can yield requirements. Indeed, the division into critical and favoring reasons in itself counts against a weighting model; cf. Gert (2007).



a slant toward the present into individual practical reasoning. In Alice’s case, even if a response to Brad’s comments were required of her in particular, an immediate response might not be specifically required. Delay would be risky—besides the threat of postponement until too late, it might weaken the impact of the response—but any of several times during a certain period would be acceptable, let us suppose. However, the prospect of emotional discomfort introduces a further criticism of even slight postponement. In that sense it adds an element of urgency or pressure to Alice’s practical reasons. Now, taking action right away would be advisable, not just commendable, in this crosstemporal case. Even without considering emotional factors, that is, we have to take Alice as subject to rational criticism for failing to respond immediately, if she would thereby risk compromising the effectiveness of her response. But all she is morally required to do is to fulfill the obligation somehow, not necessarily to take the most reliable means to fulfilling it. She can be said to have a moral reason to take the most reliable means, but in my terms this amounts to a favoring rather than a critical reason. However, instrumental considerations do suffice to yield a critical non-moral reason that might be said to supplement the moral reasons in play here, apart from any emotional reinforcement. Even where emotion is not needed to supply a critical reason, though, it may still add reinforcement to one, in the form of a further critical reason appealing to the agent’s psychological state as a barrier to discounting other non-moral reasons. Discounting, remember, amounts to setting a reason aside, simply canceling its deliberative weight, while still acknowledging it as a reason: deciding not to be moved by a certain kind of consideration. For instance, in assigning grades, you ignore a student’s need for a certain grade in order to graduate this term, or during last-minute preparations for a class, you put out of mind various other tasks pressing for attention. Discounting involves a refusal to take certain reasons into consideration, but not necessarily a low estimate of their importance in relation to competing reasons, or their pre-existing weight. In some such cases you might even take the discounted reason to be weightier, all things considered, but exercise an option you have, rationally speaking, of imposing your own structure on the relevant practical reasons.4 In short, you decline to ‘consider all things’ and instead set priorities. 4

This might be seen as an element of voluntarism applicable to the weights of reasons; however, I take reasons themselves as facts, in contrast to Chang (2009). Though your act of will in deciding to discount certain reasons gives you a higher-order reason to follow through, I understand the reason as the fact that you so decided. Let me stress, moreover, that I do not take discounting reasons to mean declining to take them as reasons—or as reasons for you, reasons you have, or the like. Even in cases like grading, where you are required to discount certain considerations, you can still recognize them as points for or against an envisioned act— ‘counting’ for or against it, as I sometimes say, following the common parlance, though ‘counting’ sounds odd for considerations you have decided not to count! A discounted reason still ‘counts’ as a reason for you in the sense of being available to justify a failure to follow through on your decision to discount it. Chang instead gives a voluntarist account of the nature of certain reasons, but subject to limitation by nonvoluntarist ‘given’ reasons, with weights that cannot be reversed by acts of will. But it is important that she actually recognizes rather few ‘given’ reasons with comparable or fixed weights; her voluntarism comes in only where other reasons ‘give out’. Some of the cases I have in mind as involving weight-reversal might instead be thought to involve a fixing of indeterminate weights, and my overall view here could accommodate that interpretation. But



Discounting on this basis involves appeal to higher-order considerations: reasons for setting certain priorities, favoring some first-order reasons above others.5 Of course, there may also be higher-order reasons against setting those priorities, and it will sometimes be irrational, in the sense of unwise, to discount particularly weighty non-moral reasons such as serious considerations of health. But the basic maneuver here is rationally legitimate. By contrast, though you might sometimes question the weight of a critical moral reason, you cannot legitimately discount it, since that would mean ignoring criticism from another standpoint that you do recognize as a source of reasons bearing on your action. To the extent that we sometimes might be inclined to discount moral reasons, emotions yielding higher-order reasons that make this harder to manage have a role to play in moral reasoning: they keep it on course. However, my treatment of emotional reasons in cases involving imperfect obligation extends the role of emotions beyond blocking an error in moral reasoning. In cross-temporal cases emotions provide both a further reason for action and a further higher-order reason against discounting the reasons that supplement critical moral reasons, bringing them to bear on a specific moment of action. Discounting is legitimate where it is applied to either non-critical or non-moral reasons: either reasons that favor some particular way of satisfying an obligation but imply no criticism of alternatives, as in cases of individual action to satisfy a collective obligation, or reasons that imply a rational but not a moral criticism of alternatives, as in cross-temporal cases. So besides making it harder to shift responsibility, emotions reinforce reasons against moral delay. This amounts to more than simply spurring or goading us to act, as I often put the matter in Greenspan (1988), though it does have a motivational component. Emotional pressure on my account is meant to be part of the justification for immediate action, not just part of its cause. Alice’s emotion, for instance, gives her further reason for an immediate response to Brad, rather than merely increasing the probability of an immediate response that may be justified on other grounds. I do not take this point to be incompatible with perceptual or other accounts of emotions, but I doubt whether existing accounts can accommodate it as they stand. I explain the legitimacy of discounting critical reasons in terms of higher-order reasons—‘exclusionary’ reasons, in Raz’s (1990) term, which I apply to reasons based

I see no reason to insist on it, as essential to taking decisions to modify the weight of one’s reasons as rational. It also seems plausible that some reasons would be more important than others, absent the decision to discount them, but that the general value of imposing one’s own order on deliberation—the ‘active agency’ that both Chang’s view and mine are meant to make room for—justifies appeal to higher-order reasons in favor of discounting. 5 This may sometimes just involve lessening the deliberative weight of opposing reasons rather than completely canceling it: degrees of discounting, on the model of economists’ talk of a ‘discount rate’ for value over time (see especially Ainslie (1992)), or a bias toward the present that sometimes prompts us to give inadequate weight to future harms. But for simplicity’s sake, in my treatment of reasons I rely on an all-ornothing notion of ‘discounting’. I should note, though, that my argument here for moral reasoning might have a prudential analogue, in which emotions serve as a check on the tendency to discount future harms, if we now feel anxiety about them (e.g. about ill health or indigence in old age).



on an agent’s decision to exclude certain first-order reasons from consideration. That decision in turn may be made on the basis of further reasons—but possibly just favoring reasons, where discounting is optional rather than required.6 In Alice’s case, for instance, perhaps she could phrase her comments better (though let us assume they would be no more effective), if she took more time to respond. This reason for discounting should not be confused with the exclusionary reason she would have if she does discount, which amounts to a critical reason. For a sketch of some of the reasons in play in a case where Alice discounts reasons against postponement, see Figure 1.1. Alice’s primary moral reason appears in the figure in capitals. A second figure incorporating Alice’s emotional reasons will later be set up for contrast. Discounting means putting on blinders of a sort, making oneself insensitive to a certain first-order reason, as indicated by the line struck out in the figure. While still acknowledging it as a reason, Alice would decline to take that consideration into account in her current deliberation—as distinct from taking it to be outweighed by opposing first-order reasons, reasons in favor of postponement. If anything, her decision would be to put greater relative weight on her opposing reasons, by discounting the reasons they oppose, rather than simply acting in accordance with their pre-given weights. Unless she leans on the scales, as it were, by discounting (which the case assumes to be rationally as well as morally permissible, but not required), she might be said to have less reason to phrase her response better than to make sure it comes in time. She is within her rights, though, to accept a certain risk of waiting too long for the sake of abiding by a commitment to verbal clarity or the like, as long as she does manage to respond in time. As thus described, discounting amounts to a legitimate maneuver within practical reasoning, not a failure to abide by its conclusions, or an instance of weakness of will, of the sort that is often involved in procrastination. This is not to deny that there are limits to how much risk Alice should accept—and of course there also might be reasons for revising a commitment in some cases (cf. especially Bratman (1987)). Let me also acknowledge that, if emotions serve as barriers to discounting, they can sometimes have the effect of undermining rationality. One may also have reason to discount emotions themselves, or the practical reasons based on them, as evidenced by many familiar cases.7 One common example is jealous rage—usually assumed to be inappropriate, though there also may be reason to discount it in some cases where the reaction itself is warranted, but action on it would make matters worse. My claim is not that

A simple example, on the level of first-order reasons, may help motivate the view of favoring reasons that I take for granted in this paper (for extended discussion see Greenspan (2005 and 2007)). A good movie is showing on TV now, so I have a reason to turn the TV on; but I do not need a strong enough competing reason to justify not turning it on. That I have no need to be entertained right now would be adequate justification, even if there were nothing else I had reason to do, and I had no objections to watching the movie. 7 Note, however, that on my account one would indeed need a reason to justify discounting the sort of critical emotional reason that is at issue in my argument here. That would not be so, I take it, if emotions served merely as motivational barriers to discounting, without supplying reasons. 6



Discounting reasons against postponement would let Alice phrase her remarks better.


[Alice decides to discount reasons against postponement.]


Alice has decided to discount reasons against postponement.

Postponement would risk making Alice’s response ineffective.



¯ [Alice challenges Brad later.]

Figure 1.1 Discounting a reason against postponement. Bracketed entries amount to acts, rather than reasons. Same-level reasons are separated by ‘+’. The diagonal arrow indicates the effect of discounting. For simplicity’s sake the figure omits first-order reasons in favor of postponement, which are assumed to be outweighed.

emotions are always, or even mainly, supportive of rationality, but rather just that they can play a unique supportive role. Let me now try to spell out step by step for the cross-temporal case a third-person analogue to the first-person reconstruction of practical reasoning in my 1988 argument. The following abbreviations should help keep straight the several sorts of reasons involved in my current discussion: R(em) for ‘emotional reason’, R(mor) for ‘moral reason’, R(sup) for ‘supplementary reason’, and R(disc) for ‘reason to discount’ (understood as grounding the decision to discount, rather than directly grounding



action in accordance with that decision). We can now replace the statement indented earlier with R(em) Alice is uncomfortable at the thought that she ought to challenge Brad.

Framing this in the third-person should prevent misinterpretation of its attribution of an emotional reason as a report of the agent’s thought processes. It is meant to convey something Alice might acknowledge as her reason, though perhaps she never reflects on it as such. R(em) reinforces critical moral reasons that Alice might or might not be motivated to act on, the sorts of reasons that stand behind the ought judgment in R(em), on the order of R(mor)

Brad’s treatment of Carl deserves challenge.

But R(em) also reinforces supplementary critical non-moral reasons incorporating rational advice about how to satisfy the ought judgment, as summed up in R(sup)

Postponement would risk making action on R(mor) ineffective.

Even if R(mor) is true, warranted, and acknowledged by Alice, that is, she might be inclined to put off acting on it, or on the ought judgment in R(em), by discounting R(sup)—perhaps quite legitimately, on the basis of reasons such as R(disc)

Discounting reasons against postponement would let Alice phrase her remarks better.

The legitimacy of discounting need not reflect the comparative weights of Alice’s preexisting reasons: a well-phrased response may be less important all things considered than ensuring a timely response. But Alice may still have the option of assigning less importance to R(sup), and if she decided to exercise that option, the fact that she so decided would be a reason against failing to act accordingly. This is the higher-order exclusionary reason against attending to R(sup), whereas R(disc) just gives a reason in favor of making the decision to discount it. The threat of continuing discomfort until Alice acts, besides reinforcing the ought judgment in R(em), or the reasons given for it in R(mor), figures as a reason against discounting R(sup) whose significance does not depend solely on weight. Here is where we can make use of the perceptual theorists’ point that emotions affect salience. Given the tendency of emotional discomfort to absorb attention, emotional reasons are not so easily set aside. So while discounting R(em) may also be an option, its exercise would be hard to manage. We may sometimes be required to discount emotional reasons, where they pull against other practical or moral requirements, but presumably whatever reasons we have for discounting emotional reasons would not be decisive in a case where they add support to moral requirements, as they do indirectly here by reinforcing R(sup). By hypothesis, neither R(sup) nor R(disc) is decisive either, so even weak reinforcement may carry the day.



In short: it is the position of emotional reasons within an overall structure of reasons, rather than their independent weight, that gives them a normative role on my account. Emotions yield reinforcing reasons on two levels.8 Besides reinforcing first-order moral reasons with a motivational barrier to illicitly discounting them, they also yield higherorder reasons, against a form of discounting that is licit but inadvisable: discounting the supplementary reasons such as reasons against delay, that narrow our options for fulfilling moral reasons. A rough picture of the role of emotional reasons as reasons against discounting is given in Figure 1.2. We should note that their role also depends on theoretical reasons that do not appear in the figure: the fact that emotional discomfort tends to absorb attention is a reason for thinking it would be hard to abide by a decision to discount reasons against postponement. Motivational factors do come in here, but since difficulty imposes a cost, the upshot is a normative practical reason against making the decision—in effect, against an act of self-legislation that would yield an unworkable law. The third-person formulation of Alice’s case makes it clear that its conclusion is not just a claim about what is likely to motivate Alice or how she is likely to assess her reasons while in the grips of an emotion. Anyone in a position to attribute the emotion to Alice can attribute to her the same reasons for action. One might object that this would give other members of the department a reason not to speak up themselves: ‘Let Alice address the issue, since she feels so strongly about it.’ In fact, that does make sense, though it ignores the possibility that others ought to share Alice’s emotion. In any case, we should bear in mind that even Alice’s reinforced reason need not be decisive: it might be countered, say, by reasons against allowing herself to be exploited for her heightened sensitivity. Alternatively, if others did share Alice’s reaction, there might be reasons against too many people responding. But let me bypass further possible complications in order to see how the general idea this case is meant to illustrate can deal with some major objections.

1.3 Misdirected Emotions The general idea is that of emotions as cravings. It is important that we can capture their force for action without assigning significant weight to the reasons they give rise to, for like cravings they can often be misdirected. Cravings also yield normative reasons against postponing action: the fact that you crave a certain snack counts at least to some extent against passing it up, or even saving it till later, just insofar as leaving the craving unsatisfied tends to be uncomfortable and distracting. But whether this reason has any real weight depends on what other reasons apply. A craving for a healthful snack would add a reason for having it now to the fact that it is good for you. But rather than acting to satisfy a craving for junk food, say, you would have a stronger reason in light of its distracting aspect to do what you can to get rid of the craving without satisfying it. Similarly for an inappropriate emotion.


Cf. the account of decision in Raz (1990): 37–45.



Alice is uncomfortable at the thought that she ought to challenge Brad. ¯

Discounting reasons against postponement would be hard to manage.

Discounting reasons against postponement would let Alice phrase her remarks better.



[Alice decides to discount reasons against postponement.]

Postponement would risk making Alice’s response ineffective.



¯ [Alice challenges Brad now.]

Figure 1.2 An emotional reason against discounting. The higher-order reason given on the upper left amounts to R(em), Alice’s emotional reason; the diagonal arrow indicates its bearing on a decision to discount, with R(disc), the reason for making that decision, at the upper right. Among the first-order reasons given below on the left, the topmost corresponds to R(sup), which gives non-moral advice about how best to satisfy the moral reason, R(mor), appearing in capitals. Though not indicated on the chart, R(em) also counts as a first-order reason for immediate action but adds little in terms of weight to R(mor) and R(sup).

In Alice’s case, for instance, suppose that her thought that Brad has treated Carl unjustly amounted to an unjustified suspicion. If we took her feelings of outrage to provide serious reason for action in and of themselves, that would threaten to ‘bootstrap’ into rationality a response that intuitively seems irrational.9 Her emotion itself would be irrational, in the sense of inappropriate, but action on it might seem to be 9 Cf. Bratman (1987). A difference from Bratman’s worries about bootstrapping intentions, however, is that cravings do seem to rationalize acts that are otherwise rationally indifferent.



rationally justified simply as a way of alleviating discomfort. However, on the present account whatever reason Alice might have for venting her feelings in such a case would be so weak that it would be reasonable to act on only in the absence of serious reasons to the contrary. The very fact that an emotion is inappropriate would itself seem to count as a reason against expressing it, though. In some situations it might be overridden, but here it would be backed up by reasons against an unjust response. Moreover, even if we imagine a case where the reason based on an inappropriate emotion faces no serious opposition—where no one is around, say, so that venting presumably can do no harm—the critical conception would leave Alice the option of discounting this non-moral reason. So action on an inappropriate emotion would not come out as rational in the sense of rationally required. The most anyone could bootstrap into on the basis of an inappropriate emotion would be a rational permission. It would be ‘within reason’ to vent one’s feelings in cases where there was no serious reason not to. Surely that claim sits perfectly well with intuition. It might seem odd for me to suggest that one can have a reason for discounting an emotional reason even where one has no reason not to act on it. Emotional discomfort is supposed to supply a critical reason, after all, and discounting a critical reason involves appeal to a higher-order exclusionary reason on my account. But here is where it is important to distinguish between a reason for discounting and the reason that discounting would give rise to. The fact that an emotion is inappropriate would seem to be a reason to discount whatever reasons it yields—but a favoring reason, so that discounting would merely be an option. Note, too, that, with reasons understood as objective facts, the higher-order optional reason to discount would not be limited to emotions the agent recognizes as inappropriate. So we can say both that an agent has the option of discounting reasons based on an inappropriate emotion and that the emotion gives her a reason against failing to express it, albeit of only minimal weight. In the rare case where she has no weightier opposing reasons, she can take her choice. This assumes that discounting is something she can manage to do. Discounting requires control of attention, but in a practical sense that in this case involves, not awareness of the emotion per se, but rather control over its influence on deliberation: attending to it as a reason. The analogy to cravings might help here. Even if you cannot get rid of your craving for a Big Mac and fries when you pass McDonald’s, you might still refuse to assign it any weight as a reason to stop for lunch there. My argument in the previous section appealed to the difficulty of discounting emotional discomfort insofar as it tends to absorb attention, and the transition from awareness to practical attention is natural and hard to resist. But it is not impossible. You can refuse to be moved by an urge you cannot ignore. That is what it would mean to discount the reason based on an inappropriate emotion that persists as a state of discomfort. Instead, as with a food craving, emotional discomfort would seem to yield a reason for doing something else to get rid of the emotion. Similarly, the discomfort of a headache is a reason to take an aspirin. However, we need not conclude from this that an emotion is reason-giving only in the way that a headache is. We can take an



emotion as yielding a reason specifically for action to satisfy its evaluative component, but a reason that will be easily overridden, and can always be discounted, in cases where the emotion is inappropriate. The analogy between emotions and cravings needs careful interpretation in light of the concern in recent literature with the question whether desires can be reason-giving. Cravings are desires, of course, but they involve more than what contemporary philosophers typically have in mind by desires. A desire in the philosopher’s sense— covering what my 1988 argument calls ‘affectless desire’—amounts just to another propositional attitude figuring alongside belief in the explanation of action. It is often said to differ from belief in ‘direction of fit’: whereas beliefs are supposed to fit the world, the world is supposed to fit our desires. But it should be evident from the role of emotional discomfort in my argument that this is not all I think there is to an emotion, or to practical reasons based on emotion. (For that matter, I argue in Greenspan (1995) that emotions exhibit a mixed direction of fit, insofar as their evaluative elements are belief-like.) An affective element is essential to a craving, as an instance of the common-sense (and earlier philosophical) conception of a desire as involving discomfort if left unsatisfied. This reference to discomfort provides a basis for emotional reasons in facts that may be independent of desires in the philosopher’s sense. My 1988 argument interpreted discomfort as a state an agent would naturally want to escape, but even on that reading it would remain open whether an agent on a particular occasion actually does want to escape some instance of discomfort. I take discomfort to be a state of feeling that is picked out as such, but is not therefore constituted, by the desire people have on typical occasions to escape it. We can see it as yielding a reason against letting it continue, even if we would deny that the mere fact of wanting something, without discomfort at not yet having it, would provide a reason for action to attain it. In the absence of discomfort, what provides a reason for satisfying a desire might instead be the value attributed to its object, or to the experience of attaining its object, rather than the present state of desire. Something emotions and cravings have in common, in contrast to standard states of desire (cf. Pettit & Smith (1990): especially 576f.), is that they yield reasons only for as long as they continue. My current desire (in the philosopher’s sense) to finish this paper by the deadline is a desire that I meet the deadline whether or not I retain the desire. If desires were reason-giving, my desire would presumably yield a reason I now have for meeting the deadline, even if I should change my mind as the deadline approaches. But worry about missing the deadline, considered just as a state of discomfort, apart from any independent reasons why I ought to meet the deadline, would yield only a reason conditional on my continuing to worry. So in the case of an inappropriate emotion, reason-giving force would not extend beyond the emotion’s element of discomfort— just as a craving for a Big Mac and fries while walking past McDonald’s would give you no reason to plan to go back there later, if you knew that the craving would dissipate once you walked by.



This limitation applies in the absence of further reasons for retaining an emotion or craving. In some cases there may be such reasons, despite the intrinsic disvalue of discomfort. An emotion or craving may be valued for its motivational effect or its tendency to heighten the enjoyment of attaining the desired object. Consider curiosity of the ‘driven’ sort, involving an itch to know that would be frustrating not to satisfy, rather than just pleasure at satisfaction. There also may be non-instrumental reasons, most notably moral reasons, for emotion—for guilt in response to wrong action (or even just wrong-making features of action that are serious enough to hold in mind with affect) or for grief or sympathetic sorrow or some other instance of the kind of concern for others that is essential to moral virtue. A different sort of problem besides rational bootstrapping might seem to be raised by the possibility that appropriate emotions (in the sense, once again, of rationally appropriate) sometimes reinforce contra-moral reasons. In a conflict between moral and self-interested reasons, could a stronger emotion on the side of self-interest manage to tip the balance in its favor, when moral reasons otherwise would have been decisive? The relatively trivial weight of emotional reasons counts against this in general, but there might be cases where weighty enough self-interested considerations, such as reasons against abandoning one’s life project, have enough weight as supplemented by emotion to carry the day against moral considerations. However, the kinds of moral reasons for emotion that were just mentioned presumably include some that yield requirements to feel strongly enough about fulfilling our more serious moral obligations. The fact that the agent in a particular case might happen not to care whether she harms others, say, would not mean that she ought to favor self-interest at their expense. But let me now address another common objection to my account, focusing on the role of self-concern in moral cases.

1.4 Non-Moral Motivation If emotions motivate by way of a need to alleviate one’s own discomfort, that might be taken to mean that agents acting out of emotion to satisfy a moral obligation are acting on the basis of ‘ulterior’ motives of a sort that undermine moral worth. Emotions may merely serve to reinforce moral reasons, but cases where they are essential to motivation—if Alice, for instance, would not have spoken up in time without pressure to improve her own state of feeling—might seem analogous to familiar examples of apparent moral virtue that really manifests morally unworthy motives: acts of honesty or charity motivated by a desire to look good in others’ eyes, say, rather than concern with truth-telling or with others’ welfare for its own sake. Or at best the agent would have mixed motives, some of them moral and morally worthy, but compromised by others that bring in self-interest. So even if emotions provide normative practical reasons, they are the wrong sort of reasons for moral action. One response available to us immediately, on the basis of the preceding discussion, is an appeal to the slight weight of emotional reasons in relation to the moral reasons that



primarily motivate the agent in such cases. It would seem to take an over-scrupulous concern with purity of moral motive to object to a moral agent’s placing even minimal weight on her own state of feeling. Kant might object, but I think that even many current Kantians would agree that his insistence on ‘doing the right thing for the right reasons’ is among the features of the historical view that could benefit from a less severe interpretation. In any case, we can also see that the emotions in play in such cases are moral motives, even though they supply us only with non-moral reasons on the account I have defended. R(em) above should make it clear how integrally the affective and evaluative components of emotion are connected: the affective state of discomfort is about a moral evaluation that amounts to the content of the emotion. The result is a moral emotion—an emotion with a moral content—that motivates via self-concern. The mechanism is complex: what we have is an emotion incorporating a moral reason but reinforcing it with a non-moral reason—the fact of emotional discomfort—that provides pressure for immediate action. But this is not to say that an agent thus motivated will be focusing attention on her own state of feeling rather than on moral considerations (cf. Pugmire (2007)). Part of the point of emotional affect is to convey attention to what it is about: in this case, a moral reason. The analogy to cravings may be useful here too. An unsatisfied craving tends to absorb attention and thus provides a reason for action to satisfy it, reinforcing other reasons for thinking its object worthwhile. That I crave the taste of oats in the morning gives me a further reason against postponing breakfast, if only to get the craving off my mind. But if I act in part to alleviate discomfort, action shifts my primary focus to its object. An unsatisfied craving absorbs attention, but only as long as one leaves it unsatisfied. Similarly, it is important that emotional discomfort has an intentional object, so that something alluding to the reasons it reinforces is the likely object of thought when one acts to get rid of it. This is why my third-person reformulation of an emotional reason remains in the present tense: my argument concerns occurrent emotion, action out of emotion. An emotion one acts to avoid (or attain) is another story. Anticipation of it at some temporal distance can of course provide a reason, but without the same tendency to convey attention to its object. The intentional aspect of emotional discomfort also distinguishes it from discomfort that merely accompanies a perceived need for action. A buzzer that kept going off in my head, giving me a headache, until I fulfilled a moral requirement, would not play the role of a moral emotion, if in acting to get the annoying sound to stop I would indeed be aiming primarily at my own relief. I see emotions as the way we build awareness of moral reasons into our individual motivational resources. An agent deserves moral credit for being open to the emotions in question—for indeed feeling discomfort about unsatisfied moral requirements, instead of being emotionally unfazed or taking steps to get rid of any discomfort without satisfying them, in the manner appropriate to a headache. This depends on acknowledging that we have a degree of control over what we feel, but I think that



claim can be defended (see Greenspan (2000)). The point for present purposes is just that moral credit does not depend on keeping our motives unpolluted by self-interest, but rather on being ready to harness our complete stock of motivational resources, including self-interest, to moral ends. So the problem raised for moral motivation by the self-interested nature of emotional reasons seems to evaporate on closer analysis. We should bear in mind how emotional discomfort reinforces moral reasons rather than substituting for them on my account.10 This is not to say that the moral reasons in question would be insufficient on their own to motivate the agent. Without my craving for oats in the morning I might still choose Cheerios over Pop-Tarts. Instead, my claim is that, in order to yield a requirement of immediate action, a moral reason may need to be supplemented by a further, non-moral reason, represented above by R(sup). In the first instance, the normative role of the moral emotion is to back up this non-moral reason, by providing a higher-order reason not to discount it. In short, then, while an emotional reason is indeed the wrong sort of reason to be a moral reason, or to substitute for a moral reason as the source of moral action, that is not the role I have assigned it. One might think, instead, that a different problem about emotions as motivators of moral action arises in cases of conflict between moral and emotional reasons, of the sort that came up briefly at the end of my previous section, but now viewed from the opposite angle. That is, some might say I have conceded too much to the moralists in trying to explain the moral relevance of emotional reasons without assigning them more than minimal weight in comparison to moral reasons. A ‘bolder’ view would allow for cases where strong-enough feelings rightly rule the day against at least some non-trivial moral requirements. I am reminded particularly of an example in McFall (1987) of an adulterer who is said to be acting out of integrity—personal, rather than moral, integrity—insofar as she acts on principle rather than impulsively. Her principles, that is, demand sacrificing even moral virtue for the sake of a great love. However, I think one could grant the point—and even that such an agent might be acting on all-things-considered objective reasons—without assigning any great weight to emotional reasons per se. What might be held to justify her act would presumably not just be the fact that she would undergo discomfort if she failed to follow her principles. It is the value the principles attribute to romantic relationships that bears the weight here. Alternatively, one might cite some ill effects of romantic frustration—a debilitating sense of personal conflict, say, possibly even undermining the agent’s marital relationship—as serious enough to warrant

10 We also should bear in mind that my argument is limited to particular cases of appropriate emotions, here taken as registering genuine moral reasons. I do not claim that emotions ‘track’ moral reasons in the sense of hooking onto them reliability on repeated occasions. As I noted for questions of rationality above, my point is not that emotions always or even mainly support rather than undermine moral motivation, but just that they are capable of playing a special supportive role as sources of normative reasons.



compromising moral integrity. But again, this goes beyond the fact of current discomfort, which is the source of the emotional reasons at issue in this paper. We also sometimes place non-instrumental value on emotions themselves, considered not just as aspects of our well-being or as accompaniments of valued relationships, but also as manifesting a valued sensitivity. Feelings of love, even where painful, might be encouraged as signs of a form of vulnerability to others that we regard highly: being ‘a person who needs people’, in the words of the popular song. Or in the case of ‘driven’ curiosity mentioned earlier, the craving for knowledge might be seen as an ennobling trait, quite apart from its effects. Moral emotions as states involving discomfort also have general moral value just insofar as they are directed at the right objects. We ought to feel something for those in need, say, even where there is nothing we can do to help them. But the ennobling quality of such feelings should not prompt us to maintain them by withholding acts that could help others. There will be enough occasions for sympathy later on. Some who value emotions as aspects of moral virtue apparently mean to favor only the more pleasurable manifestations of love or sympathy for others, without the element of emotional discomfort that my argument turns on. How odd to argue on behalf of emotions and speak incessantly of discomfort, ignoring emotional rewards! But my aim is not to enthuse about the joys of emotional life. Of course, the emotions one might enthuse about supply us with ends of action, and as occurrent states they ground expressive action. But I take it that in both cases they provide only favoring reasons. What I have tried to bring out instead is the way emotions also make practical demands. Those emotions that we tend to think of as unalloyed states of pleasure are actually multilayered, and their affective manifestations vary with our changing circumstances. Love, for instance, involves a craving for the love object that with distance or other barriers to closeness becomes a yearning. Our loving concern for the good of others often means sharing their pains. In the way that active love of learning involves a craving to find out, the practical impact of such feelings depends on an element of discomfort, felt (but perhaps overshadowed by pleasures, even just at fantasized satisfaction) as long as one postpones acting on them. As with ordinary cravings, they might not entail actually feeling discomfort in situations where one can satisfy them easily, without thought or effort (including moral debate). But in normal cases they would seem to amount to unalloyed states of pleasure only where they are passive states, whereas a propensity to act is surely an essential part of moral virtue. What emotional discomfort does on my account is to back up the propensity with a felt need. If a kind of self-concern helps push us beyond passivity to immediate action, I do not see that as morally compromising. Instead, it adds a naturalistic motivational base to conceptions of morality framed in terms of reasons or other abstract normative notions. Part of what motivates a behaving organism is the betterment of its own state, simultaneously with a focus on changing the external situation. We are beings complex enough to combine inner and outer sources of action. I have argued accordingly that emotions add to moral motivation what amount to cravings for the right: moral



motives reinforcing moral reasons with weak but crucially positioned non-moral reasons against delay.

References Ainslie, George (1992) Picoeconomics: The Strategic Interaction of Successive Motivational States Within the Person. Cambridge: Cambridge University Press. Bratman, Michael E. (1987) Intention, Plans, and Practical Reason. Cambridge, Mass.: Harvard University Press. Chang, Ruth (2009) ‘Voluntarist Reasons and the Sources of Normativity’. In: Reasons for Action, D. Sobel & S. Wall (eds), pp. 243–71. New York: Cambridge University Press. D’Arms, Justin & Jacobson, Daniel (2000) ‘The Moralistic Fallacy’, Philosophy and Phenomenological Research 61: 65–90. Darwall, Stephen (2006) The Second-Person Standpoint: Morality, Respect, and Accountability. Cambridge, Mass.: Harvard University Press. de Sousa, Ronald (1987) The Rationality of Emotions. Cambridge, Mass.: M.I.T. Press. Frijda, Nico H. (1986) The Emotions. Cambridge: Cambridge University Press. Gert, Joshua (2004) Brute Rationality: Normativity and Human Action. Cambridge: Cambridge University Press. —— (2007) ‘Normative Strength and the Balance of Reasons’, Philosophical Review 116(4): 533–62. Greenspan, Patricia S. (1988) Emotions and Reasons: An Inquiry into Emotional Justification. New York: Routledge, Chapman, and Hall. —— (1995) Practical Guilt: Moral Dilemmas, Emotions, and Social Norms. Oxford: Oxford University Press. —— (1998) Moral Responses and Moral Theory: Socially-Based Externalist Ethics’, Journal of Ethics 2: 103–22. —— (2000) ‘Emotional Strategies and Rationality’, Ethics 110: 469–87. —— (2004) ‘Practical Reasoning and Emotion’. In: The Oxford Handbook of Rationality, A. Mele & P. Rawling (eds), pp. 206–22. Oxford: Oxford University Press. —— (2005) ‘Asymmetrical Reasons’. In: Experience and Analysis: Proceedings of the 27th International Wittgenstein Symposium, M.E. Reicher & J.C. Marek (eds), pp. 387–94. Vienna: oebv&hpt. —— (2007) ‘Practical Reasons and Moral “Ought”’. In: Oxford Studies in Metaethics, vol. 2, R. Shafer-Landau (ed.), pp. 172–94. Oxford: Oxford University Press. —— (2009) Interview. In: Philosophy of Action: 5 Questions, J.H. Aguilar & A.A. Buckareff (eds), pp. 59–69. London: Automatic Press/VIP. —— (2010a) ‘Learning Emotions and Ethics’. In: The Oxford Handbook of Philosophy Emotions, P. Goldie (ed.), pp. 539–59. Oxford: Oxford University Press. —— (2010b) ‘Making Room for Options: Moral Reasons, Imperfect Duties, and Choice’, Social Philosophy and Policy 27: 181–205. Horgan, Terry & Timmons, Mark (2010) ‘Untying a Knot from the Inside Out: Reflections on the “Paradox” of Supererogation’, Social Philosophy and Policy 27: 29–63.



Jones, Karen (2003) ‘Emotion, Weakness of Will, and the Normative Conception of Agency’. In: Philosophy and the Emotions, A. Hatzimoysis (ed.), pp. 181–200. Cambridge: Cambridge University Press. —— (2004) ‘Emotional Rationality as Practical Rationality’. In: Setting the Moral Compass: Essays by Women Philosophers, C. Calhoun (ed.), pp. 333–52. New York: Oxford University Press. McFall, Lynne (1987) ‘Integrity’, Ethics 98: 5–20. Pettit, Philip & Smith, Michael (1990) ‘Backgrounding Desire’, Philosophical Review 99: 565–92. Pugmire, David (2007) Sound Sentiments: Integrity in the Emotions. New York: Oxford University Press. Raz, Joseph (1990) Practical Reason and Norms. Princeton: Princeton University Press. Rorty, Ame´lie O. (1980) ‘Explaining Emotions’. In: Explaining Emotions, A.O. Rorty (ed.), pp. 103–26. Berkeley: University of California Press. Sherman, Nancy (1990) ‘The Place of Emotions in Kantian Morality’. In: Identity, Character, and Morality, O. Flanagan & A.O. Rorty (eds), pp. 149–70. Cambridge, Mass.: MIT Press. Thompson, Michael (2004) ‘What is it to Wrong Someone? A Puzzle about Justice’. In: Reason and Value: Themes from the Moral Philosophy of Joseph Raz, R.J. Wallace, P. Pettit, S. Scheffler, & M. Smith (eds), pp. 333–84. Oxford: Oxford University Press.

2 Emotions and the Categorical Authority of Moral Reason Carla Bagnoli

Few deny that emotions can be motivating. Many recognize that emotions may account for the motivational appeal of moral reasons, and some argue that emotions provide moral reasons for action. In this chapter, I consider a specific aspect of the normative relevance of emotions, which concerns the authority of moral reasons that yield requirements. This is a question that arises at the meta-normative level of ethical theory, an area of inquiry that investigates the nature of moral normativity. Answers to this question have important consequences about the content of moral reasons, but my primary task is to examine the role of emotions in explaining that and how moral reasons are authoritative. On a standard rationalist model, moral reasons apply to all rational agents as such, and bind us with categorical authority because they are intrinsically normative facts, whose authority does not depend on anything contingent, such as our conventions, beliefs, and emotions (Clarke (1705); Prichard (1912); Falk (1986)). By contrast, the standard sentimentalist model holds that moral reasons spring from emotions, and denies that they exhibit the sort of categoricity that rationalism attaches to them (Hume (1740); Wiggins (1987a); Blackburn (1998); McDowell (1985); D’Arms & Jacobson (2000); Johnston (2001)). Both views face serious difficulties. The rationalist model makes mysterious or fails to account for how moral reasons command with genuine authority in the presence of competing concerns and interests. The sentimentalist model clashes with the common view that moral reasons have special authority and importance, in contrast to reasons that spring from our idiosyncratic preferences, individual interests, and personal plans. My argument is that an adequate explanation of moral authority requires a different philosophical treatment of the role of emotions, and of their relation to practical reason. The problem with these standard views is that they take emotions to be either completely separable from or only contingently related to reason. By contrast, I argue that the experience of moral emotions is constitutive of the exercise of practical reason. The categorical authority of moral reasons does not depend upon, but constitutively implies, moral emotions. I argue for this claim by drawing from a Kantian account of practical



reason, which takes respect as the emotional attitude constitutive of rational agency. My contention is that moral reasons have categorical authority insofar as they are subjectively experienced in the guise of respect.

2.1 The Problem of the Categorical Authority of Moral Reasons Morality is normative in that it provides reasons for action. If it is morally wrong to deceive others to further one’s interests, then there is a normative reason for Amadeus not to deceive Boris in order to become chapel master. According to rationalism, moral reasons provide normative reasons for action that apply to all rational beings in all relevantly similar situations. If it is morally wrong to deceive others, then in any situation where an agent morally ought not to deceive others, there will be a normative reason for her not to do so. This requirement of universality indicates that to act on moral reasons is to act on principles. In deliberation, moral reasons are taken to provide normative reasons that are conclusive, that is, such that they override any other sorts of considerations that are surveyed in deliberation; I shall call this feature of moral reasons overridingness. This latter feature of moral reasons is problematic, since in rational deliberation we are required to take into account all sorts of considerations, including particular desires, interests, and projects that are in competition with morality. How can moral reasons drive us when we have competing concerns and motives that make legitimate claims on us? This is the question of subjective authority. Bernard Williams argues that moral reasons cannot be shown to be rationally overriding, hence they lack authority (Williams (1981): 20–40, 114–32; Williams (1985): chs 4 and 10). For instance, it is hard to explain why and how (the fictional) Paul Gauguin would be rationally driven by moral reasons to take care of his family when such reasons undercut the project that grounds and gives meaning to his life, that of becoming a painter (Williams (1981): 23).1 Williams argues that the (Kantian) rationalist view attributes to morality a sovereignty that morality does not actually possess. For Williams we have in fact deep and persistent reason to be grateful that morality does not have the features that Kantians attribute to it (Williams (1981): 23). To attach overridingness to moral reasons is to say that they govern other projects, subtract or attribute authority to them, prioritize, subordinate, and ground other reasons. Gauguin’s case shows that when moral reasons conflict with other projects that give meaning to his life, it does not make sense for the agent to take moral claims as overriding, even on the assumption that the claims of 1 Williams uses Gauguin’s example to draw a number of important claims about the nature of morality, its susceptibility to luck, and the retrospective nature of its justification; see Williams (1981): 20–39. My discussion here is limited to his claim that moral reasons debunk all other concerns one has, and that this view of moral reasons makes the rationalist account of moral motivation unreasonably demanding, hence inadequate as a model of practical rationality and practical identity (see also Wolf (1986, 1997)). I focus on Gauguin’s example as supporting Williams’ conviction that morality should not be given the sort of overridingness that it has in the Kantian model.



others have a hold on him. Unlike the amoralist who is indifferent to moral claims, Gauguin is ‘concerned about these claims and what is involved in their being neglected’ (Williams (1981): 23; see also p. 38). The point is that, despite his moral qualms and pangs of conscience, Gauguin cannot be rationally required and expected to take care of his family, if this means that he must renounce painting—the only project that makes his life worth living. The upshot of Williams’ argument is that moral reasons are not unconditionally authoritative, as the rationalist claims. The canonical problem with the rationalist account, as the case above illustrates, is that it fails to explain how agents are driven by moral reasons.2 The sentimentalist may seem to have a ready answer to this problem since sentimentalism holds that moral reasons originate in emotions, which have direct motivating power. For instance, Amadeus has a reason not to deceive Boris, insofar as he loves him as a friend. Amadeus’ love for Boris provides a motive to refrain from deceiving him, even if deception would further Amadeus’ interest. But love also explains how Amadeus is driven by the consideration that one should not deceive others, even if one is interested in becoming a chapel master.3 The question is whether this counts as an explanation of the authority of moral reasons. The sentimentalist explains the motivational force of moral reasons by making them arise from the emotions. Is he thereby providing an account of their normative authority? This is, recognizably, a rationalist rejoinder. The rationalist wants the normative force of moral reasons to be intrinsic to those reasons, hence not reducible to their motivational force. When normative authority is reduced to motivational power, it looks arbitrary and spurious.

2.2 Reactive Emotions and the Internalization Hypothesis One way to deflate the question about the alleged distinction between the motivational force and the normative force of moral reasons is to deny that moral reasons have a distinctive status. J.S. Mill famously takes this route in contrast to the rationalist claim that moral obligations are ‘intrinsically binding’.4 According to Mill, the apparent 2

Rationalism can, of course, explain why agents abide by morality by adding a further premise, a desire for being rational, but this is not a solution to the issue I am highlighting, as it emerges in section 2.8 below. 3 This is not to say that love directly provides a general or universal reason not to deceive others. In the case described, love for a particular person makes the agent responsive to considerations against deception. Reasons of love are sufficient to qualify the authority of his interest in becoming a chapel master, even though they may be insufficient to undermine the force of this interest. Love for a particular individual may also have a role to play in making the agent realize that there is a universal reason for not deceiving others; but it may not be enough to motivate the agent to act on the basis of the universal reason against deception. This is not a worrisome result, though, because reasons against deception do not spring from love, but from respect. In the example, the reasons of love undermine the authority of the agent’s interest, and thus work only indirectly against deception. 4 Mill (1861): 73, and more generally, see chapter 3. Mill’s internalization hypothesis is grounded on his associationism, that is, the view that moral responses are acquired through processes of association; see Wilson (1998): 216–17, and Skorupski (1989): 263. However, the internalization hypothesis can be defended independently of psychological associationism; see footnote 10.



compelling nature of moral requirements can be fully accounted for by referring to the education of our sensibility. Through social conditioning and education, our minds become accustomed to associate some types of actions with some emotional reactions. Actions externally sanctioned by society become internally sanctioned thanks to the work of reactive emotions, such as guilt, self-reproach, and shame.5 I shall call this the internalization hypothesis. This hypothesis is supposed to supply the ultimate explanation of the halo or mystic character of moral obligation. Its authority is nothing but ‘a subjective feeling in our mind, attendant on violation of duty, a pain more or less intense, which in properly cultivated minds rises, in the more serious cases, into shrinking from it as an impossibility’ (Mill (1861): 74). It is important to remark that the argument based on the internalization hypothesis is not designed to debunk morality. While Mill denies that moral reasons exhibit categorical authority, he does not deny that morality provides us with normative reasons for action. Rather, he purports to anchor moral obligations on firmer grounds by showing that moral reasons acquire their binding force via social and psychological processes, hence dissipating the mystery of their authority.6 Current moral psychology heavily concurs with J.S. Mill that reactive emotions, such as guilt and blame, are the basis of our conformity to moral norms.7 Many also agree that this explanation shows that morality does not have special authority and importance (Foot (1978a); Blackburn (1998); Crisp (2006): 20–36). Reactive emotions are required in order to make moral norms efficacious, by providing them with motivational support. For instance, Allan Gibbard argues that moral emotions have an immediate motivating power; they serve as incentives to comply or deter future defection. By and large they are ‘reactions against threats to one’s place in cooperative schemes’.8 This is because ‘norms for guilt can attach a bad feeling to things bad feelings can move us to avoid’ (Gibbard (1990): 297).9 In a similar vein, Bernard Williams remarks that ‘remorse or selfreproach is the characteristic first-person reaction within that [moral] system, and if an These emotions are also called ‘deontic norms’; see de Sousa (2006): 31. For this reason, Mill stresses the importance of a system of education whose ‘main and incessant ingredient is restraining discipline’ (Mill (1963–1991): 133), Mill holds that impulses such as vengeance that propose themselves as immediate sources of moral concern should not be trusted as moral guides on their own. They need to be scrutinized, disciplined, and educated (Wilson (1998): 218). 7 Allan Gibbard openly appeals to Mill; see Gibbard (1990): 41–2, 67–8; Blackburn (1998): 17. Railton argues that the peculiar nature of moral normativity can be captured by focusing on ‘a distinctive set of emotions, such as guilt, pride, shame and reproach’ that attend the violation of moral norms (Railton (2005): 13). In a broader sense, the internalization hypothesis also underlies philosophical projects that appeal to psychodynamic approaches that reduce the issue of authority to complex processes of idealization and internalization; cf. Deigh (1996); Velleman (2006): 110–29, 129–56. As Williams notices, the internalized gaze that these moral emotions convey ‘is not just a screen for one’s own ethical ideas but is the locus of some genuine social expectation’ (1993: 98, 103). 8 ‘Emotions, in evolutionary terms, cash out in action: in the action to which they lead and in the actions they elicit from others’ (Gibbard (1990): 139). 9 The category of reactive deontic emotions is quite heterogeneous and admits of a complex phenomenology. For instance, Gibbard argues that bad motivations elicit anger while other inadequacies elicit disdain; they each sanctions different failures and call for different remedies (1990: 139). 5 6



agent never felt such sentiments, he would not belong to the morality system or be a full moral agent in its terms’ (1985: 177). Likewise, Simon Blackburn emphasizes the punitive character of morality and notices that ‘moral judgment is indeed used to coerce, and cajole, and to judge: when it is internalized, its victims may walk around under the burden of guilt and anguish’ (1998: 3; Gibbard (1990): 297).

2.3 Residual Emotions and the Enforcement of Moral Norms The internalization hypothesis purports to explain how moral reasons acquire motivational force without invoking any dubious metaphysics.10 This is an advantage over the standard rationalist proposal, insofar as the rationalist’s claim about the categorical authority of moral reasons is associated with a special kind of practical necessity, which differs from causal and logical necessity. For some, the Kantian category of practical necessity cannot claim any place in the scientific conception of the world, and thus it commits ethical theory to a peculiar metaphysics. But this argument in favor of the internalization hypothesis does not secure the soundness of sentimentalism. In fact, several Kantians focus on reactive emotions, even though they regard moral reasons as categorical (Nagel (1970): 80; Korsgaard (1996): 151–6; O’Neill (1997): 92–3; Scanlon (2008); Darwall (2011)). For instance, Christine Korsgaard thinks that we experience emotions of this kind precisely because moral reasons have independent and categorical authority. Emotions are backward-looking responses to reasons: ‘A person’s own mind does indeed impose sanctions on her: that when we don’t do what we should do, we punish ourselves, by guilt and regret and repentance and remorse’ (1996: 151). These reactive emotions are residual, in that they sanction reasons that have been disregarded.11 By feeling them, ‘we pay the price of unmet demands’ (1996: 151 n. 25). Rationalists and sentimentalists agree that the role of residual emotions is mainly auxiliary and corrective.12 Such emotions assist us because of the shortcomings of

10 J.L. Mackie moves the objection of queerness to the intuitionist claim that there are intrinsically normative entities (1977: 39ff.). But Mill can be interpreted as formulating a similar objection against Kant’s view that moral obligations are intrinsically binding; see section 2.2. 11 The claim that reactive emotions are residual—insofar as they stand for moral claims that have been disregarded in deliberation—is prominent in the literature on moral dilemmas. Philosophers disagree as to the deontic valence that we should attach to these emotions. For instance, Williams argues that agent-regret signals a remainder, which is an ‘item not acted upon’ (1963: 173–4, 183). If the item is an ‘ought’ or a ‘moral claim’, this sort of regret indicates that conflicts of obligations are possible. By contrast, Hare argues that the residual emotions such as guilt or regret do not stand for obligations, but are associated to prima facie duties, which are coherent with the denial of conflicting obligations (Hare (1980); cf. Bagnoli (2000, 2007b)). Some argue that the deontic valence and significance of residual emotions differ according to whether we are talking of regret, remorse, or guilt; see Barcan Marcus (1980); Greenspan (1995). 12 Crisp (2006): 23. However, neo-Kantians try to make room for a more complex interplay with sensibility, for example adopting a more Aristotelian view of moral psychology, or simply rejecting the claim that emotions are mere feelings or tastes (Sherman (1990); Baron (1995); Herman (1993, 1997); Korsgaard (2009): 19). I am developing an independent argument in this direction.



reason. Emotions either rule when, or because, reason is incapable of guiding us. The two philosophical proposals differ in their respective diagnosis of this failure: for the sentimentalist, reason fails to guide us because it is inert; for the rationalist, reason provides us with standards that we are unable to put into practice. In both cases, however, the role of residual emotions is reparative and disciplinary, if not utterly castigatory.13 These are important normative roles of emotions in deontic contexts, which the sentimentalist takes as the proof that reason is motivationally inert. Because of this diagnosis, sentimentalists seem to have a significant advantage over the rationalist. It is open to sentimentalists to argue that emotions play a more significant role, by broadening the category of moral emotions so as to include ‘positive’ emotions that push us toward the endorsement of moral norms.14 This integration shows that the normative relevance of emotions is not simply corrective, but also reparative and motivating. Is not this kind of integration all we need to account for the authority of moral reasons?

2.4 Emotions as Auxiliary Motives and Contributory Reasons Section 2.3 ended with the question of whether proving the motivational appeal of moral reasons by grounding them on emotions amounts to proving their authority. It may not be obvious that we need an account of the subjective authority of moral reasons if we have a plausible account of their motivational force. An example may help clarify why we need to distinguish the issue of motivation from the issue of authority. Suppose Bessy judges that it is best not to voice her concern about salary compression, but then she decides, instead, to pursue the issue publicly, out of resentment, anger, or indignation. I think we can offer two readings of the role of these emotions in accounting for Bessy’s action. On the first reading, anger simply tips the balance in favor of something that is not supported by rational deliberation. In this case, the role of the emotion is psychological: anger does not provide any new normative reason, but it makes Bessy act on some considerations that she already had surveyed in deliberation, and discounted. Moreover, she does not regard such considerations more important than she previously found them to be, hence she does not accord to them a different normative status. She acts on those considerations not because she thinks that she ought to, but because she is angry. That is,

13 Following Melanie Klein, Williams distinguishes sharply between unproductive and uncreative guilt, which has simply punitive or persecutory functions, and reparative guilt; see Williams (1973): 222. Williams responds to a utilitarian variation of the argument. His strategy proves successful against either of these variations of the argument. Interestingly, Williams imputes the ‘comparative neglect of this basic moral phenomenon’ to the utilitarian reaction against the destructive emphasis placed on coercive education, and the ‘uncreative aspects of guilt’. 14 Taylor argues for this strategy in Chapter 12, in this volume. A broader list of positive emotions includes benevolence and respect; see Gibbard (1990): 255–73. Virtually any emotion may provide a source for motivation and all emotions contribute to valuing in complex ways; see de Sousa (2001): 109–26.



anger provides a ‘driving motive’ and makes such considerations motivating, but it does not make them right, and thus it does not qualify them as reasons. On the second reading, Bessy’s anger provides her with a new reason that adds to the considerations she already had, and prompts a new deliberation, which leads to a change of mind about what she ought to do in the situation. As a consequence of feeling anger, her previous considered judgment is reversed: she now thinks that she ought to speak up about salary compression. It is not that she ought to do so because she is angry. Rather, her anger makes her appreciate some aspects of the situation that she had not noticed before. For instance, through anger she vividly realizes that salary compression not only affects her family but it also poisons the relations with her colleagues. On the basis of these further elements, she concludes that the matter should be addressed publicly. In this scenario, anger alerts the agent to morally relevant considerations, which were ignored or inadequately attended and factored in previous deliberation. On both readings, anger has an impact on Bessy’s practical reasoning, but only on the second reading does it provide a contributory reason for action, as I shall call it. On the first reading, anger changes the relative motivational force of the conflicting motives at stake, but it does not change the normative status of the reasons the agent had. To make sense of the role of anger in these two scenarios, it is necessary to distinguish the normative and the motivational dimension of the emotions’ impact on practical reasoning. This distinction is useful in locating the disagreement between the rationalist and the sentimentalist proposals, and it helps us to identify their respective strengths and weaknesses. Whether one endorses sentimentalism or rationalism, it is hard to deny that emotions play a significant role in rational deliberation by affecting our motivational set. It is exactly because emotions have an impact on our motivation that their relation to morality is problematic. Emotions posit a threat to morality insofar as they are likely to produce motives that compete with moral reasons. In fact, rationalists do not deny that emotions supply us with motives that add or reinforce the psychological force of moral reasons, as in the first reading of Bessy’s case. Some rationalists also accord emotions an enabling role. For instance, they may agree that feelings of anger and resentment enable us to appropriately react against social injustice, that love disposes us favorably to others and encourages us to enter cooperative schemes, and that compassion fosters social stability by sensitizing us toward the needs of others. However, the rationalist would not cite these examples of auxiliary motives as cases where emotions make a normative contribution. The sentimentalist discounts the rationalist position by pointing out that it postulates intrinsically normative entities. But the rationalist’s resistance to taking emotions as sources of contributory reasons does not require her to invoke intrinsically normative entities, distinct and disconnected from our sensibility. In the next section, I introduce Kant’s argument in support of the distinction between the auxiliary and the normative role of emotions. My point will be that this distinction introduces another and, to my mind, more adequate characterization of the normative relevance of emotions, as constitutive of the exercise of practical reason.



2.5 Emotions and the Efficacy of Moral Reasons: Kant’s Argument against External Motives Kant acknowledges that emotions are key auxiliaries to moral motives (Kant (1788): 152ff.; Kant (1797b): 402, 456–8). For instance, he lists parental love, compassion, and sense of honor as emotions that concur with moral duty. Kant is also aware that the appeal to natural emotions is quite pervasive in the practice of morality and thus agrees with the sentimentalist to an important extent. In stark contrast to the sentimentalist, however, Kant regards these natural emotions as mere ‘surrogates for the motive of duty’. First, he denies that the emotions that concur with moral duty have the status of ‘moral emotions’ simply because they favor moral action; by contrast, he takes them as ‘merely analogous to moral feelings’.15 Second, even though Kant recognizes that such emotions play a motivational role in driving agents to conform to duty, he denies that these are genuine moral motives that account for the efficacy of moral reasons ((1788): 152). Third, and as a consequence, he denies that emotions provide direct normative support for morality when they are represented as surrogates for moral motives. Auxiliary incentives cannot forestall immoral motives or reinforce moral ones. In fact, emotions that work vicariously as auxiliary motives detract from morality, insofar as they undermine its genuine authority (Kant (1784a), (1788): 152ff.). While this position seems inimical to the recognition of the moral relevance of emotions, I hope to show that it follows from a conception of practical reason that requires moral sensibility. Famously, Kant proposed such a conception of practical reason as ‘entirely new’ (1788: 153) and ‘more necessary than ever’ (1788: 157) to overcome the impracticality of moral philosophy. What makes Kant’s conception distinctive and novel is that it aims at representing our responsiveness to moral reasons as part of their objectivity. His task is ‘to provide the moral law with influence on the human heart’ (1788: 156). One difficulty in elucidating Kant’s account of the interplay between practical reason and the emotions is that he does not use a single term to refer to emotions. In fact, he allows for a complex taxonomy of the modes of our sensibility, which covers quite disparate phenomena of the mind: affects, moral feelings, inclinations, and passions.16 The relevant issue for Kant is how these phenomena relate to rational agency; and his suggestion is that different affective states bear different relations to rational agency depending on whether they belong to the faculty of desire, or they are simply modifications of the feelings of pleasure and pain.17

15 On the analoga instinctorum moralium, see Kant (1762–1763): 77, 200/449. I owe this reference to Bacin (2006): 19. 16 Inclinations (Neigungen), moral feelings (moralische Gefühlen), affects (Affekten), passions (Leidenschaften). See Kant (1797a): 211–14, (1797b): 7. 17 Some affects, such as anger, are like urges, not under our control and contribute nothing to moral agency. We are passive in respect to them, and they are obstacles to morality insofar as they undermine reflection and rational deliberation (Kant (1797b): 408; (1797a): 252–3). Passions, such as love, relate to the



In their assessment of Kant’s treatment of the emotions, most scholars focus on the argument of the Grounding for the Metaphysics of Morals }1, which is directed against inclinations as original and blind sensations. Kant’s point seems to be that emotions are pre-cognitive feelings that do not convey any moral insight, and thus should not be taken as inarticulate perceptions of moral reasons.18 Given the random, unruly, incorrigible, and unstable nature of emotional states, Kant argues that they have no moral value, even when they conform to duty (1785: 398). This argument seemingly fails to do justice to the complexity of emotions and the pervasive role they play in moral life. In order to acknowledge such a role it may seem necessary to discard the simple conception of emotions as unresponsive to judgment, and opt for the cognitivist conception of emotions, which takes emotions to be cognitive modes of discerning values.19 However, this move eliminates the distinction between Kant’s peculiar form of rationalism and the Moral Sense School along with the contemporary sensibility theories typically associated with it.20 For Kant it is important that moral feelings remain ‘subjective relations’, because in his view they neither afford moral knowledge nor play any foundational role as the ground of moral obligations ((1788): 22–5, 76–9, (1797b): 399). Therefore, they are not perceptions of moral value in the way some contemporary sensibility theories hold. Their role is not epistemological in the sense of conveying knowledge of some normative properties.21

faculty of desire, and are directed to objects and reveal a more complex relation to rationality. They can upset our capacity to reason (1797a: 265), but they can also make us more clever and apt to identify the adequate means, hence improving our instrumental rationality (1797b: 625). Vengeance as a passion manifests some sort of reason (1797a: 270). 18 See also Kant (1797b): 211–12; Korsgaard (2009): 18–19. Some scholars attribute to Kant a very simple conception of emotions as passive bodily sensations (Sabini & Silver (1987)). The simple model of emotions as sensations does not even explain the emotions animals have, because some of their states are reactive and not purely subjective: they react against a threat of some sort (e.g. against life and integrity) (Korsgaard (1996): 150–6; Borges (2004): 147). 19 Recent attempts to rehabilitate Kant’s ethics highlight the complexity of emotions, and show that moral emotions, such as love and compassion, enable us to fulfill our moral duties (Baron (1995); Cagle (2005)). Some Kantians take emotions themselves as ‘moral responses’ that determine what is morally relevant and, in some cases, what is required (Sherman (1990): 2). In contrast to these views, the proposal I defend in sections 2.6–2.8 does not focus on emotions as sources of reasons or direct modes of moral discernment, but on their pervasive normative role in accounting for the ‘susceptibility’ (Empfänglichkeit) to reason. 20 Francis Hutcheson had a significant influence on Kant’s early conception of moral thinking; see Wood (1996): xiv–xv. In some early writings, Kant admits a universal and uniform capacity for moral feeling, distinct from reason, which contains the foundations of practical principles. But he subsequently distanced himself from this theory: first he questioned the uniformity of feelings; then, he argued that, independently of their uniformity, sentiments cannot provide morality with an objective justification; see Kant (1762–1763): 116, 117, 120, 149; (1785): 408; (1788): 26. 21 Korsgaard notices that a way out of this problem is to deny that only a standard form of realism can vindicate the claim that emotions are perceptions of value; see Korsgaard (2009): 19. Korsgaard thinks that it is in the nature of every animal to have ‘normative perceptions’, that is, to see the world under the relevant descriptions dictated by their interests and concerns, or their values (2009: 19). My argument goes in the direction of making sensibility constitutive of practical rationality, hence intrinsic to our understanding of moral reasons. However, I find the talk of perception and the claim that emotions are ‘perceptive of values’



Nonetheless, emotions do play a significant normative role, which I hope to elucidate by reviewing a different argument that Kant puts forward at various times. The thrust of this argument is that if moral reasons are motivating only in virtue of concurrent emotions, they do not command with genuine authority ((1762–1763): 77, (1788): 72, 83, 85, 152–3, 156–7ff.) To make moral reasons obligating via something external to the obligation itself is self-defeating (contradictio in adjecto), because it undercuts the very concept of obligation.22 This argument bears some resemblance to the rationalist argument reviewed above, that if we reduce authority to motivation, we fail to account for the idea of the authority of moral obligations. The standard rationalist argument is that to vindicate the nature of obligations, they should be acknowledged as intrinsically normative entities (Clarke (1705): 193–4). Kant also thinks that obligations are intrinsically normative, but the explanation he offers importantly differs from the standard rationalist one, exactly because of the role he ascribes to moral sensibility. Contrary to the understanding of Kant’s view of emotions based on the usual reading of the first section of the Grounding, Kant’s concern in this argument is not the blind nature of emotions, but their normative role as surrogates and external aids. The problem is that when taken as auxiliary motives, emotions show that the agent’s will is caused externally; hence, it is not a fully autonomous will. It is noteworthy that the same argument can be used for any external surrogates, such as the application of sanctions and rewards (Kant (1784b): 1326).23 This extension is significant, since it points out that what is objectionable in the sentimentalist approach is not that they appeal to emotions, but that they appeal to emotions that are external to practical reason. That is, the disagreement with sentimentalism and sensibility theories does not stem from Kant’s account of emotions as subjective feelings. Instead, it is a disagreement that concerns the role of emotions as external aids to reason (cf. Kant (1785): 441). The same argument works, mutatis mutandis, for obligations that derive their normative force from uncritical reliance on beliefs, conventions, and ideologies (Kant (1784a): 36–7). This shows that what Kant finds problematic in the sentimentalist account is neither that emotions lack cognitive cores, nor that they are episodic and adventitious, but that they are taken as uncritical surrogates of reason, which undermine the agent’s autonomy.

misleading. In sections 2.6–2.8, I hope to show that emotions are morally relevant by another route, which emphasizes the emotional aspect of practical reason. 22 ‘The maxim must not get its legal character from anything outside of itself. For, if there were an outside source of legal character, then that source, rather than the legal character itself, would be what makes the action right. Instead, the maxim’s legal character must be intrinsic’ (Korsgaard (1996): 61). 23 Kant’s target is narrowed to the obligatio per poenas; but the argument holds for any case where the authority of obligations rests on an external normative expectation, were it a punishment or a reward. As Wood observes: ‘From a Kantian standpoint, any use whatever of social coercion in any form to enforce ethical duties (whether through private blame, or public opinion, or associations of moral education to shape people’s feelings) must be regarded as a wrongful violation of individual freedom by corrupt social customs’ (1997: 9).



This argument bears three important consequences. First, both external sanctions and auxiliary motives undermine the very idea of obligation. Second, and as a consequence, the integration of positive moral emotions undermines the genuine authority of moral reasons. This strategy, which had seemed to be an advantage of sentimentalism, is no advantage at all, because it is vulnerable to the objection to externality. Third, moral reasons are genuinely authoritative only if their authority can be explained solely in terms of the motivational states that are alleged to constitute them. This is the most important upshot of Kant’s argument. The key question is, then, whether there is an emotion that can figure as the appropriate moral motivational state of rational agents. That emotion is respect.

2.6 Normative Authority and Subjective Authority To place respect at the core of Kant’s account of authority requires some argument. His attack against external sanctions and auxiliary incentives is one step in the larger argument meant to show that moral obligations are rationally binding: their contents are requirements of practical reason (Kant (1788): 42, 72, 83, 85–6); and they apply to all rational beings as such. Gauguin’s case posits a challenge that any rationalist account of moral reasons must meet to prove its practical relevance. Kant in fact agrees with Williams that moral theory needs to be ‘put into practice’.24 Kant’s claim is that moral reasons command with genuine authority only if they are self-legislated ((1785): 431–2, 438, (1788): 25–40, 33). Any attempt to ground their authority on external foundations is self-defeating (1785: 441–5). Kant appeals to selflegislation to specify the basic requirements of practical reason. As a specification of the demands of practical reasoning, the claim about self-legislation grounds the categorical imperative. It tells us that to reason correctly amounts to conforming to universal principles. However, the appeal to self-legislation is also supposed to show how the requirements of reason are binding. That is, the argument directly addresses the issue of normative authority. The Kantian argument locates the source of subjective authority in the very deliberative process that leads to the determination of moral reasons. How is it that moral reasons are compelling? The Kantian answer is that moral reasons are chosen among considerations that the agent already finds motivating. The ‘good will’ works as the basic normative standard for action. The motive that drives the person of good will is the motive of duty. The justifying reason for action is a maxim (a subjective principle of action) with legal character, that is, which has the form of a law. For the maxim to

24 This is the general purpose of the Doctrine of Method, which aims to prove that (i) the laws of pure practical reason have access to the human mind and influence on the maxims (Kant (1788): 151); (ii) one can make objective practical reason subjectively practical as well (ibid.: 151, 153, and 157); and (iii) it is possible to produce not only mere conformity to moral duty (legality of actions), but also genuine ‘morality of dispositions’ (ibid.: 151). See Bagnoli (2009a); Bacin (2010).



have genuine normative authority its legal character must not derive from anything outside itself; that is, the maxim must be self-legislated. But how does this account of normative authority explain the compelling character of moral reasons? It seems that there is still a gap between finding moral reasons authoritative (insofar as they are self-legislated) and taking them as conclusive and compelling. Kant’s claim that moral reasons are authoritative insofar as they are selflegislated concerns the appeal that such reasons have on rational beings as such. This leaves open a more specific question concerning the subjective appeal of these reasons. In other words, the claim about self-legislation offers only a partial answer to the question of authority. It answers the question of normative authority when we consider the matter objectively, that is, from the mere standpoint of pure practical reason. From this standpoint the question of normative authority is straightforward, since purely rational agents are determined by practical reason. There is, however, a further aspect of the same question, which concerns ‘animals endowed with reason’, such as human beings. For human beings, the issue of moral authority is complicated by the fact that they are also animals driven by a variety of interests and concerns. To show that moral reasons are objective, Kant thinks that we have to show that they are authoritative for people like Gauguin, that is, for any one of us. This is the aspect of the authority of moral reasons that Kant calls ‘subjective’ (1788: 42; cf. 151–3, 156–7).25 His distinctive proposal is that there must be a subjective ground for moral reasons, or they would not guide us. On the basis of the argument reviewed in section 2.6, such a subjective ground cannot be external to morality, an incentive to morality, ‘but morality itself, subjectively considered as an incentive’ (Kant (1788): 76).26 This subjective aspect is named respect. Respect is the subjective experience of autonomy, and shows our capacity for practical reason.27 It accounts for how we are actually driven by the thought that something is right. Kant thought this to be part of our common experience, and a proof of our general propensity to reason, and therefore to morality (1788: 76, 154). The point is established via a thought experiment. In the face of a tragic choice between rendering false testimony and facing death, the moral agent feels that telling the truth is compelling even though, eventually, he may not take it as an operative reason when faced with the real prospect of losing life (Kant (1788): 30). Supposedly, the agent feels a natural attachment to life; thus, love of life figures prominently among the motives that he reviews in deliberation. In addition, let us assume that the man in question thinks very highly of himself, and he considers his own life worthier than the

The term ‘subjective’ stands for the dimension of finite subjects; see Kant (1788): 38, 72–5, 81, 88, 117. ‘Respect for the law is not the incentive to morality; instead it is morality itself subjectively considered as an incentive inasmuch as pure practical reason, by rejecting all the claims of self-love in opposition with its own, supplies authority to the moral law, which now alone has influence’ (1788: 76). 27 See Kant (1788), part II; Beck (1960): 233–6. 25 26



life of the innocent man against whom he is asked to render false testimony. Yet, he enters rational deliberation about what to do. The case is meant to show that respect, which is the name for our ‘receptivity to pure moral interest’, becomes ‘the most powerful incentive to moral good’ (Kant (1788): 152). That is to say that respect serves as a constraint on deliberation. This normative role is captured by linking reverence for the law to respect for others as independent sources of legitimate claims.28 In this case, for instance, respect outranks the agent’s natural attachment to life, and limits the false pretenses of self-love. This natural attachment to life is still present as a maxim of rational self-love, but it occupies a lower position in the final ranking of the incentives. One other effect of respect understood as the recognition of the equal standing of others is that it cancels the arrogant thought that the agent’s life is worthier than the life of others; this maxim does not have any place in the final ranking of incentives. In short, respect undercuts the pretenses of natural self-love and annuls the false claims of self-conceit. Insofar as respect works as a deliberative constraint, its phenomenology is partly similar to that of sanctions, as it involves painful feelings of frustration. This similarity explains the widespread attention to reactive and residual emotions. But there are some important differences between the effects of constraining one’s maxims by respecting others and the effects of applying sanctions. First, the pain of respect does not derive from punishment or mere frustration. Rather, respect involves a painful feeling of having some of your claims discounted because they were advanced as merely yours. It exposes our limitations and vulnerability in many more ways than frustrating our desires; it is, more basically, ‘the feeling of an incapacity to attain to an idea that is a law for us’ (Kant (1788): 57). Second, like sanctions, respect works as a deliberative constraint, but unlike sanctions the outcome of deliberation constrained by respect is a form of self-control and self-discipline, which are specific forms of reflective self-government. In other words, respect is an expression of our freedom. These resulting forms of self-government do not aim at curbing natural tendencies, although they may also have such an effect. Their defining task is to provide genuine moral motives, that is, motives that are grounded on mutual respect and recognition. Unlike sanctions, then, respect does not point to any ulterior reason for acting morally. Rather, respect is the motive to act morally (Kant (1788): 76). This second aspect of respect accounts for a third, which concerns its expressive role and distinctive phenomenology. The experience of respect is also the positive and 28 To justify the derivation of duties of respect, Kant supplies an argument for the equivalence between reverence for the law and respect for others. I do not need to address this argument here, since my task is not to investigate the normative determination of respect. My main focus is the recognitional aspect of reverence for the law. The recognitional aspect of respect has a cognitive core, but this is not to say that it has a specific object. It is directed to the idea of self-legislation, and indirectly to others insofar as they are constitutively implied in the practice of self-legislation. As I have argued elsewhere, Kant’s self-legislation constitutively entails reference to others because the sort of reflexivity that pertains to it is dialogical; see Bagnoli (2007a, 2009b).



elevating experience of being capable of understanding and acting on moral reasons. This experience fosters self-esteem (Kant (1788): 78–9).29 As for Gauguin and the agent who faces the choice between losing his life or rendering false testimony against an innocent man, to attend to our moral duty might prove just too demanding. But the point is that even when the stakes are high, these agents nonetheless feel the pull of moral duty, and if they end up acting against such duty, they know they have done wrong. This shows that moral reasons have a compelling force, which they do not derive from natural inclinations or external sanctions. Respect is the subjective mode in which moral reasons are felt compelling. These examples elicit ‘respect for ourselves’ (Kant (1788): 161), and reinforce our consciousness of moral life. That we find moral reasons authoritative thus confirms their objectivity.

2.7 Respect as the Constitutive Attitude of Rational Agency Respect is not merely one incentive among the many that rational agents happen to have. The question of moral authority is not simply how this moral incentive outranks others, as Williams suggests in casting his case against the unconditional authority of morality. The normative significance of respect is more basic and pervasive, insofar as it is the emotional attitude that is constitutive of rational agency. It sets the standard of moral competence and reciprocity. That is, it is constitutive of the stance of practical reflection. It thus turns out that in the most paradigmatic rationalist model, the mark of rational agency is an emotional capacity.30 Since moral agency amounts to rational agency, the normative role of respect is pervasive: the whole practice of rational justification is based on respect. It is in virtue of respect that we are not only susceptible to reasons, but, more fundamentally, capable of forming reasons at all. Respect governs not only the exchange of reasons in the practice of justification of our action to others, but also the very formation of reasons. In determining the reasons that we have to do something, we represent ourselves as members of an ideal community of agents having equal standing.31 In judging morally, we constitute ourselves as representatives of such an ideal community, and we judge according to the standards of mutual respect and recognition that identify the moral community (Bagnoli (2007a)). 29 ‘Those who think of Kant’s moral doctrine as one of law and guilt badly misunderstand him. Kant’s main aim is to deepen and to justify Rousseau’s idea that liberty is acting in accordance with a law that we give to ourselves. And this leads not to a morality of austere command but to an ethic of mutual respect and self-esteem’ (Rawls (1971): 256). 30 In contrast to other prominent accounts, I take respect to be more fundamental than other emotions such as blame, and resentment; cf. Strawson (1962); Skorupski (1998, 2010); Scanlon (2008); Darwall (2006, 2011). Respect is more fundamental in the sense that it is a structural principle of morality and accounts for the conditions upon which blame or resentment can be appropriately expressed. It serves the purpose of governing blame and other excluding or redeeming attitudes; see Bagnoli (2007a). 31 That is, it requires universality in both form and scope; see O’Neill (2004); Korsgaard (1996): 98, 135–7.



However, the pervasiveness of respect in the practice of rational justification should be sought at the structural level of practical reasoning. Respect works as a limiting condition for something to count as a reason for action, but it does not fully determine the content of reasons for action, nor does it command uniformity. To clarify this point, it is useful to look at the relation between respect and the categorical imperative. The categorical imperative accounts for the legal character of reasons, so that what counts as a reason is something that all rational agents, concerned with organizing their thoughts and actions publicly, could choose. The categorical imperative defines the practical standpoint, objectively considered. It specifies what considerations stand out as rationally valid and thus qualify as reasons for action. Respect defines the practical standpoint, subjectively considered, that is, from the specific perspective of finite, separate, but interdependent individuals, concerned with organizing their thinking and doing publicly. It is this emotional capacity that explains how individual agents are bound by principles that they find rationally valid. This emotional capacity does not add anything to the rational validity of moral reasons. Instead, it explains why we cannot ignore their authority (Korsgaard (1996): 151). Respect is not the normative source of moral reasons, but it explains how moral reasons drive us to action. It does so not by adding to the many incentives that rational agents review in deliberation, but by constraining and ranking such incentives (Kant (1788): 76). On the Kantian view, respect as reverence for the law amounts to respect for others. This equivalence is problematic, and its normative consequences are debatable. But these questions lie outside the scope of this essay. The focus of my argument is narrowed to the normative role that respect plays as the moral incentive. To elucidate this role, it is not necessary to consider how it grounds duties we have toward humanity; this is part of normative ethics. Instead, the meta-ethical question addressed here concerns the role of emotions in explicating the categorical authority of moral reasons. My argument has been that on the Kantian view, respect defines the standpoint of practical reflection, where agents take responsibility for what they do. At the structural level of practical rationality, respect does play a normative role, but it does not determine or qualify the many specific relations that we entertain toward others. For this same reason, the normative role of respect is not limited to actions that are directed toward or affect others, such as cases of wronging or benefiting someone.32 Rather, its role as the moral incentive is to establish the subjective authority of any moral reason. As the moral incentive, it affords the (subjective) conditions of the possibility of mutual intelligibility and coordination among mutually affecting agents. 32 Kant’s self-legislation constitutively entails reference to others because the sort of reflexivity that pertains to it is impersonal; see O’Neill (2004); Bagnoli (2007a). However, to insist on the constitutive role of others does not imply that this notion of respect as reverence is directed to or addressed to others: this is a further normative determination of respect that I am not considering here. I believe that this latter distinction accounts for the difference between the dialogical model I am proposing and Darwall’s second-personal model, and better explains why respect does not govern only reasons for action that yield bipolar obligations; cf. Darwall (2011).



It is the subjective counterpart of our need for principles, and that is the reason why it is best understood as the psychological specification of the claim about self-legislation.33

2.8 Respect as the Emotional Aspect of Practical Reason This account of respect as the attitude constitutive of rational agency gives us important resources for rethinking the relation between emotions and practical reason. The problem of normative authority is that of understanding how reasons compel us to act. My argument has been that Kant approaches this question from two complementary standpoints. From the point of view of pure practical reason, moral reasons are binding for all rational beings insofar as they are rational. This account of moral authority fails to adequately account for Gauguin’s case, and many follow Williams in thinking that this failure shows that moral reasons lack distinctive authority. This conclusion is too hasty, though, since the Kantian model of moral reasons does not rest solely on the objective standards of pure practical reason. On the contrary, it is precisely designed to address cases such as Gauguin’s. As the man in the example of the false testimony, Gauguin feels the authority of the moral demands, even if he does not properly attend to such demands. His moral failure is both normative and motivational, since he fails to be guided by his own reasons.34 Gauguin’s failure to act on reasons does not show that his reasons lack authority; it only shows that reasons do not have irresistible power (Korsgaard (1996): 104). To explain the phenomena of the authority of moral reasons, Kant attributes to ‘pure practical reason’ an emotional or subjective aspect. It is in virtue of this subjective and emotional aspect that we are capable of normative guidance. I take this to be the core of Kant’s distinctive form of rationalism. On the standard rationalist account, the question arises about how to close the gap between reasons and motives (Kant (1788): 391). Some hold that the gap is closed by a causal mechanism that disposes us to act on moral reasons. But this implies that moral reasons are not normative reasons, and that they drive action only if there is a concomitant emotion, desire, or interest that activates a disposition to act. By contrast, the Kantian view is that there is no extra factor required to close the gap between reason and action. Or, more accurately, on

33 In my previous works, I have defended a dialogical view of respect as mutual recognition, which takes the reference to others as constitutive of respect (Bagnoli (2007a, 2009b)). The dialogical view of respect is both psychological and modal. It is modal, insofar as it accounts for self-legislation as the requirement that we justify our actions on the basis of reasons that all rational agents could coherently adopt. It is psychological, insofar as it identifies respect as the subjective condition of our receptivity or responsiveness to practical reasoning. I owe this characterization to Onora O’Neill. 34 I have argued elsewhere that these are failures of agency, but they are not reducible to incoherence; see Bagnoli (2009b). In contrast to the incoherence account (Korsgaard (1996): chs 3–4), I argue that immoralists do not fully enjoy autonomous agency because they are not capable of engaging in the proper form of practical reflection, which requires relating to others as having equal standing. The dialogical account I propose has the distinctive merit of identifying the internal costs of disregarding moral reasons, and of showing that immoralists may become susceptible to practical reason.



this view there is no gap between moral reasons and motives. Respect as the constitutive emotional aspect of rational activity warrants a conceptual connection between morality and normative reasons. Contrary to the standard rationalist and sentimentalist models that take the relation between practical reason and sensibility to be contingent, the Kantian view makes it structural. To be a rational agent, one must have the motive to comply with moral reasons, and this motive is respect. Practical reason is thus not merely procedural or computational, but also an emotional capacity.35

2.9 Conclusion This view of respect as the emotional aspect of practical reason avoids the dilemma between the rationalist view of the objective demands of morality that have no grip on us, and a sentimentalist view of moral reasons that denies their categorical authority. Both these views mischaracterize and misunderstand the relation between morality and the emotions. They mistakenly assume that emotions are separable from or only contingently related to practical reason. By contrast, the Kantian model takes this relation as constitutive and structural: to undertake the practical standpoint requires us not only to act and think on principles that have the form of a law, but also to display a moral sensibility marked by respect. These requirements are constitutive of the practical standpoint, that is, they represent what is necessary for us to think and act as rational agents, together with other finite and interdependent rational agents.36

References Bacin, Stefano (2006) Il senso dell’etica: Kant e la costruzione di una teoria morale. Bologna: Il Mulino. —— (2010) ‘The meaning of the Critique of Practical Reason for moral beings: the Doctrine of Method of Pure Practical Reason’. In Reath & Timmerman (2010): pp. 197–215. Bagnoli, Carla (2000) ‘Value in the Guise of Regret’, Philosophical Explorations 3: 165–87. —— (2007a) ‘Respect and Membership in the Moral Community’, Ethical Theory and Moral Practice 10(2): 113–28.

35 This shows that ‘emotions must play an essential role in moral life even on the most rationalistic theory’ (Korsgaard (1996): 151 n. 26). In support of a deep convergence between the Aristotelian and Kantian on this point, see Korsgaard’s remarks that ‘The implication for rational beings is that the development of rationality requires the acquisition of a second nature—a set of emotional responses and an accompanying normative view of the world that conforms to the demands of reason. The acquisition of virtue, a condition of the receptive faculties that makes us sensitive to the demands of reason, is therefore essential to the perfection of our moral nature, and to the integrity that makes agency possible’ (2009: 19). 36 A distant ancestor of this essay was presented at the International Conference on Emotions at the Universities of Neuchatel and Berne in 2005, and then delivered as the E.J. and Rosa Lee Audi Memorial Lecture in Ethics at Colgate University in 2007. I should like to thank these audiences, and especially Scott Anderson, Maudemarie Clark, Josep Corbı`, Stefaan Cuypers, Ronald de Sousa, and Christine Tappolet. I am also grateful to Stefano Bacin, Ian Carter, Bennett Helm, Oliver Sensen, and Steven Winkelman for comments on more recent drafts of this chapter.



—— (2007b) ‘Phenomenology of the Aftermath. Ethical Theory and the Intelligibility of Moral Experience’. In Tenenbaum (2007): pp. 183–221. —— (2009a) ‘Practical Necessity: the Subjective Experience’, in Centi & Huemer (2009): pp. 23–43. —— (2009b) ‘The Mafioso Case: Autonomy and Self-Respect’, Ethical Theory and Moral Practice 12(5): 477–93. Barcan Marcus, Ruth (1980) ‘Moral Dilemmas and Consistency’, The Journal of Philosophy 77: 121–36. Baron, Marcia (1995) Kantian Ethics Almost without Apology. Ithaca, NY: Cornell University Press. Beck, Lewis White (1960) A Commentary on Kant’s Critique of Practical Reason. Chicago: University of Chicago Press. Blackburn, Simon (1998) Ruling Passions. Oxford: Oxford University Press. Borges, Maria (2004) ‘What Can Kant Teach Us About Emotions?’, The Journal of Philosophy 4: 140–58. Cagle, Randy (2005) ‘Becoming a Virtuous Agent: Kant and the Cultivation of Feelings and Emotions’, Kant-Studien 96(4): 452–67. Centi, Beatrice & Huemer, Wolfgang (2009) (eds) Value and Ontology. Frankfurt: Ontos-Verlag. Clarke, Samuel (1705) Boyle Lectures. In: British Moralists 1650–1800, D.D. Raphael (ed.). Hackett Publishing Company, 1991. Crisp, Roger (2006) Reasons and the Good. Oxford: Oxford University Press. D’Arms, J. & Jacobson, D. (2000a) ‘Sentiment and Value’, Ethics 110: 722–48. —— (2000b) ‘The Moralistic Fallacy: On the “Appropriateness” of Emotions’, Philosophy and Phenomenological Research 61: 65–90. Darwall, Stephen (1977) ‘Two Kinds of Respect’, Ethics 88: 36–49. —— (2006) The Second Person Standpoint. Cambridge, MA: Harvard University Press. —— (2011) ‘Bipolar Obligation’, Oxford Studies in Metaethics (forthcoming). de Sousa, Ronald (2001) ‘Moral Emotions’, Ethical Theory and Moral Practice 4: 109–26. Deigh, John (1996) The Sources of Moral Agency. Cambridge: Cambridge University Press. Engstrom, Stephen & Whiting, Jennifer (1997) (eds) Aristotle, Kant and the Stoics. Cambridge: Cambridge University Press. Falk, D.W. (1986) Oughts, Reasons, and Morality. Ithaca, NY: Cornell University Press. Flanagan, Owen & Rorty, Ame´lie O. (1990) (eds) Identity, Character, and Morality. Cambridge, MA: MIT Press. Foot P. (1978a) ‘Morality as a System of Hypothetical Imperatives’. In Foot (1978b): pp. 158–73. —— (1978b) Virtues and Vices. Berkeley: University of California Press. Gibbard, Allan (1990) Wise Choices, Apt Feelings. Oxford: Clarendon. Greenspan, Patricia S. (1995) Practical Guilt. Moral Dilemmas, Emotions, and Social Norms. Oxford: Oxford University Press. Hare, Richard M. (1981) Moral Thinking: Its Levels, Method and Point. Oxford: Oxford University Press. Harman, Gilbert (2007) ‘Guilt-free Morality’, Oxford Studies in Metaethics 4: 26–35. Herman, Barbara (1993) The Practice of Moral Judgment. Cambridge, MA: Harvard University Press. —— (1997) ‘Making Room for Character’. In Engstrom & Whiting (1997): pp. 36–63. Honderich, Ted (1985) (ed.) Morality and Objectivity. London: Routledge & Kegan Paul.



Hume, David (1740) A Treatise of Human Nature, P.H. Nidditch (ed.). Oxford: Oxford University Press, 1978. Johnston, Mark (2001) ‘The Authority of Affect’, Philosophy and Phenomenological Research 63(1): 181–214. Kant, Immanuel (1762–1763) Reflexion, Kants gesammelte Schriften. Preussische Akademie der Wissenschaften, Berlin, 1907, vol. 19, 6560. —— (1784a) An Answer to the question: What is Enlightenment? Kants gesammelte Schriften, Preussische Akademie der Wissenschaften, Berlin, 1907, vol. 8; in Practical Philosophy, A. Wood (ed.), pp. 11–22. Cambridge: Cambridge University Press, 1996. —— (1784b) Naturrecht Feyerabend, Kants gesammelte Schriften, Preussische Akademie der Wissenschaften, Berlin, 1907, vol. 27, 1319–94. —— (1785) Grundlegung zur Metaphysik der Sitten, in Kants gesammelte Schriften, Preussische Akademie der Wissenschaften, Berlin, 1907, vol. 4; in Practical Philosophy, A. Wood (ed.), pp. 37–108. Cambridge: Cambridge University Press, 1996. —— (1788) Critique of Practical Reason, Kants gesammelte Schriften, Preussische Akademie der Wissenschaften, Berlin, 1907, vol. 5; in Practical Philosophy, A. Wood (ed.), pp. 133–272. Cambridge: Cambridge University Press, 1996. —— (1797a) Metaphysical Principles of Virtue, Kants gesammelte Schriften, Preussische Akademie der Wissenschaften, Berlin, 1907, vol. 6; in Practical Philosophy, A. Wood (ed.), pp. 353–604. Cambridge: Cambridge University Press, 1996. —— (1797b) Anthropology from a Pragmatic Point of View, Kants gesammelte Schriften, Preussische Akademie der Wissenschaften, Berlin, 1907, vol. 7, Robert B. Louden & Manfred Kuehn (eds). Cambridge University Press, 2006. —— (1907) Nachlass. Kants gesammelte Schriften, Preussische Akademie der Wissenschaften, Berlin, 1907, vols 14–22: Notes and Fragments, trans. Paul Guyer, Curtins Bowman, & Frederick Rauscher. Cambridge and New York: Cambridge University Press, 2005. Korsgaard, Christine Marion (1996) The Sources of Normativity. Cambridge: Cambridge University Press. —— (2009) Self-Constitution: Action, Identity and Integrity. Oxford: Oxford University Press. Laslett, P. & Ranciman, G. (1962) (eds) Philosophy, Politics, and Society. Oxford: Oxford University Press. McDowell, John (1985) ‘Values and Secondary Qualities’. In Honderich (1985): pp. 110–29. Mackie, John L. (1977) Ethics: Inventing Right and Wrong. London: Penguin. Mill, John Stuart (1861) Utilitarianism. Cambridge: Cambridge University Press. —— (1963–1991) The Collected Works of John Stuart Mill, John Robson (ed.). Toronto: University of Toronto Press. Nagel, Thomas (1970) The Possibility of Altruism. Princeton: Princeton University Press. O’Neill, Onora (2004) ‘Self-Legislation, Autonomy, and the Form of Law’. In: Recht, Geschichte, Religion: Die Bedeutung Kants für die Gegenwart (Sonderband der Deutschen Zeitschrift für Philosophie), Herta Nagl-Docekal & Rudolf Langthaler (eds), pp. 13–26. Berlin: Akademie Verlag. Prichard, Henry A. (1912) ‘Does Moral Philosophy Rest on a Mistake?’, Mind 21: 21–37. Railton, Peter (2005) ‘Normative Guidance’, Oxford Studies in Metaethics 1: 3–33. Rawls, John (1971) A Theory of Justice. Cambridge, MA: Harvard University Press. Reath, Andrews (1989) ‘Kant’s Theory of Moral Sensibility’, Kant Studien 80: 284–302.



—— & Timmerman, Jens (2010) (eds) Kant’s Critique of Practical Reason. A Critical Guide. Cambridge: Cambridge University Press. Sabini, John & Silver, Maury (1987) ‘Emotions, Responsibility, and Character’. In Schoeman (1987): pp. 165–75. Scanlon, Thomas (2008) Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Harvard University Press. Schoeman, Frederich (1987) (ed.) Responsibility, Character, and the Emotions. New York: Cambridge University Press. Sherman, Nancy (1990) ‘The Place of Emotions in Kantian Morality’. In Flanagan & Rorty (1990): pp. 149–70. Skorupski, John (1989) (ed.) John Stuart Mill. London: Routledge. —— (1998) (ed.) The Cambridge Companion to Mill. Cambridge: Cambridge University Press. —— (2010) ‘Moral Obligation, Blame, and Self-governance’, Social Philosophy and Policy 27(2): 158–80. Strawson, Peter (1962) ‘Freedom and Resentment’. In Strawson (1974): pp. 1–25. —— (1974) Freedom and Resentment. London: Methuen. Tenenbaum, Sergio (2007) (ed.) New Trends in Moral Psychology. Amsterdam: Rodopi. Velleman, David (2006) Self to Self. Cambridge: Cambridge University Press. Wiggins, David (1987a) ‘A Sensible Subjectivism?’. In Wiggins (1987b): pp. 185–214. —— (1987b) Needs, Values, Truth. Oxford: Blackwell Publishers. Williams, Bernard (1962) ‘The Idea of Equality’. In Williams (1973): pp. 230–49. —— (1963) ‘Ethical Consistency’. In Williams (1973): pp. 166–86. —— (1973) Problems of the Self. Cambridge: Cambridge University Press. —— (1981) Moral Luck. Cambridge: Cambridge University Press. —— (1985) Ethics and the Limits of Philosophy. Cambridge, MA: Harvard University Press. —— (1993) Shame and Necessity. Los Angeles: University of California Press. Wilson, Fred (1998) ‘Mill on Psychology and the Moral Sciences’. In Skorupski (1998): pp. 203–54. Wolf, S. (1986) ‘Moral Saints’, reprinted in R. Crisp & M. Slote (eds), Virtue Ethics: pp. 79–98. Oxford: Oxford University Press, 1997. Wolf, S. (1997) ‘Meaning and Morality’, Proceedings of the Aristotelian Society 97: 299–314. Wood, Allen (1996a) ‘General Introduction’. In Wood (1996): pp. xiii–xxxiii. —— (1996b) (ed.) Practical Philosophy. Cambridge: Cambridge University Press. —— (1997) ‘The Final Form of Kant’s Practical Philosophy’, The Southern Journal of Philosophy 36: 7–19.

3 Self-Love and Practical Rationality Edward Harcourt

3.1 What does our moral psychology—our character—look like when it develops as it should? How, more specifically, should our various faculties of mind—reason, appetite, and so on—be related to one another in maturity? The dominant answer in the philosophical tradition has tended to be a rationalist one at least in the following sense, that—as Aristotle has it in the Nicomachean Ethics—a central part of what it is for character to be properly formed consists in our being practically rational. These very same questions are, in my view, also addressed by psychoanalysis. But psychoanalytic writers have often given the notion of love a very much more prominent place in their conception of what it is for character to develop as it should than those in the philosophical mainstream. Thus the psychoanalyst Hans Loewald speaks of ‘[the] erotic dimension of [man’s] . . . becoming what may properly be called a self ’: ‘a human being’s becoming a person’—which I take to be the same thing as character’s developing as it should—is ‘[a matter of the] development of our love-life’ (Loewald (1978): 32). I believe that this conception and the rationalist conception of the philosophical tradition complement one another. I am encouraged in this belief by the prominence given, in some versions of psychoanalysis, to ideals of psychological organization such as ‘autonomy’ and ‘individuation’ and, in attachment theory, to secure attachment. For I take it that the lesson of these theories is that these dispositions are formed in us by the right kind of loving nurture. And not only that: though this is controversial, I take it also that these dispositions overlap in their manifestations of, even if they do not exactly coincide with, the practical rationality of the philosophers. If these beliefs are right, then, a correct account of our moral psychology should show love and practical rationality to be closely connected. However, I record these beliefs not because I propose to defend them here, but rather in order to motivate the investigation I shall be pursuing. The connection between love and practical rationality I have already mentioned—that love of the right kind is formative of practical rationality—is a causal claim, about the development of character. But it would be that much more interesting if



the fact that practical rationality typically has such a history shows up in what practical rationality is, that is, if a second, constitutive claim were true. It’s a constitutive claim of this kind that I think Jonathan Lear has in mind when he says that ‘[without love] we cannot understand the psychic structure which constitutes the individual’. But what does it really mean to say that love is part of the structure of the (properly formed) character? I approach the question this way. Supposing—as my first claim has it— that a properly formed character comes about via receiving the right kind of other-love, what feature might we expect such a psychology to have? A plausible answer is that we might expect it to exemplify self-love: the properly formed self might be expected to have the same attitude to itself such that receiving it from others has helped to form it. So if properly formed characters—as the rationalist philosophical tradition has it—exemplify practical rationality, we should expect practically rational individuals to be self-lovers. I propose to argue for a constitutive, that is, more than merely causal, connection between love and practical rationality by trying to show that this expectation is well-founded.

3.2 Before I go further, I want to note some ambitious claims which have been made for the connections which Aristotle undoubtedly makes in some shape or form between love, self-love, and practical reason. Pauline Chazan, for example, has written that [A] certain kind of self-love is foundational for moral agency. [Moreover] that virtue requires self-love . . . is a central Aristotelian claim . . . Aristotle holds that the relation the self has to itself has priority over its relations to others: virtuous conduct depends on the relation to self which Aristotle terms ‘love of self ’ . . . [A] virtuous person is one with the right kind of self-love. (Chazan (1998): 63)

Now if Aristotle really thinks that self-love is the foundation of virtue, then given that (for Aristotle) virtue (excellence) is what the proper development of our distinctively human natures consists in, Aristotle must also hold that self-love is the foundation of the proper development of our natures. So, because practical rationality is essential to virtue (excellence) on Aristotle’s account, he must hold that self-love is, if not the foundation of, then at least intimately connected to, practical rationality. But if he thinks all this, any argument I might present to the same effect is just so much reinventing of the wheel. So before I argue for a connection between practical rationality and self-love, I want to spend a little time arguing that this is not a connection that Aristotle makes. I want to argue that, on the contrary, Aristotle’s conception of love and friendship presupposes his conception of virtue. And since virtue is what the proper development of our distinctively human natures consists in (in Aristotle’s view), his conception of love and friendship presupposes his conception of that too, and so cannot tell us what constitutes it. From Aristotle’s initial remarks in Nicomachean Ethics book VIII, my reading seems obviously right: true friendship is possible only between virtuous people, and so only



between people whose natures are properly formed. So we need to be able to identify the virtuous people before we can identify which friendships are the true ones. So the concept of virtue—and so Aristotle’s conception of what it is for character to be properly formed—is prior to that of friendship or love. Aristotle appears to say something much closer to the constitutive claim I’m pursuing myself, however, when he says that love of others ‘proceeds from the good man’s relation to himself ’, namely self-love. The relations which a good man has to his friends (such as desiring their good and their existence for their sakes, grieving and rejoicing with them, etc.) turn out to be the same as the relations he has to himself. And though the initial list of relations does not include friendship itself, Aristotle adds a little later that ‘friendship too is thought to be one of ’ the relations the good man has both to his friends and to himself (Aristotle (1954): 228). So is Aristotle not saying that as it were the true name of the set of relations to oneself which constitute being a good (read: properly formed) man is ‘self-love’, and thus that being a good (properly formed) man is constituted by self-love? I think not. The evidence is the two separate shots Aristotle has at spelling out the relation between virtue and self-love, which are slightly different from one another. At one point he says that bad people don’t love themselves because, insofar as they are bad, there is nothing lovable about them (ibid.: 229). This gives the conceptual priority to virtue, and self-love is to be seen as the intermediate term that explains why other-love presupposes virtue, as Aristotle initially maintained: other-love proceeds from self-love and self-love presupposes virtue. One should ‘strain every nerve to endeavour to be good’, ‘for so and only so can one be either friendly to oneself or a friend to another’ (ibid.: 230). The second attempt, which begins with Aristotle’s question whether one should be praised or reproached for loving oneself, takes a slightly different route. The man, he argues, who ‘gratifies the most authoritative element in himself ’, i.e. his rational or thinking part, is (rightly) ‘most of all a lover of self ’ because he assigns to himself the things that are best (ibid.: 235). To condense what is not in fact a single line of argument, things then go roughly as follows. The good man loves the good, and good things are things done with one’s rational part in control, so the good man loves his rational part. But the self should be identified with the rational part (‘Just as a city . . . is most properly identified with the most authoritative element in it, so is a man’). So the good man (rightly) loves himself. The characterization of virtue and of the relation among our mental faculties in which it consists—the authority of the rational part—again proceeds without mention of self-love, while the self-love of the good man is explained by his goodness. So again the priority, for Aristotle, of the concept of virtue over the concept of self-love is confirmed. Before I leave Aristotle, I want simply to note that Aristotle foreshadows not, indeed, the constitutive claim about the relation between practical rationality and self-love, but the intuitive claim that those whose natures are formed by love will be self-lovers. For, Aristotle says, several of the ‘marks of friendship’ which ‘proceed from a man’s relation to



himself’ (i.e. self-love)—e.g. ‘wish[ing] and do[ing] what is good . . . for the sake of his friend’—are also things ‘mothers do to their children’, and one of these marks—‘grieving and rejoicing with [one’s] friend’—is ‘found in mothers most of all’ (Aristotle (1954): 227). If the right kind of love from others makes us both practically rational (as I claim psychoanalysis teaches) and self-lovers (as Aristotle says), might self-love not be a part of practical rationality, rather than an independent effect of the same cause?

3.3 Established usage has it that ‘self-love’ is a synonym of ‘self-interest’. It’s also well established that practical rationality includes self-interest—indeed it’s sometimes thought to include only that. If we leave matters there, that there’s a constitutive connection between self-love and practical rationality is scarcely an exciting conclusion. However, the established usage of ‘self-love’ does not give us a theory of selflove, and in particular, established usage is compatible both with affirming and with denying that to have the attitude of self-love is to stand in the same relation to oneself as one stands in to another when one loves them. If one denies this, then connecting selflove and practical rationality adds nothing to what’s already accomplished by the connection between practical rationality and self-interest. If one affirms this, on the other hand, connecting self-love and practical rationality becomes a more demanding business, and one which needs argument. I want to begin arguing for the connection by taking issue with some remarks on this subject by the late Oswald Hanfling. According to Hanfling, self-love is precisely not a matter of standing in the same relation to oneself as one stands to another when one loves them, since—much as ‘promise’ does in ‘promising oneself ’—‘love’ occurs in the compound ‘self-love’ only by a linguistic accident: ‘self-love’ means ‘self-interest’ but, Hanfling goes on, acting in somebody’s interest (including one’s own) is not the same thing as acting out of love. My reply to this is that self-love is indeed, in part at least, constituted by self-interest, but that this is compatible with self-love’s being a matter of standing in the same relation to oneself as one does to another when one loves them; for short, compatible with self-love’s satisfying the same-relation constraint. But the connection between selfinterest and practical rationality comes, I take it, for free. Given the latter connection, then, to show that self-love can both be (partly) self-interest and satisfy the samerelation constraint should go quite some way to establishing a connection between selflove and practical rationality. Hanfling gives three arguments against the claim that self-love, as the term is usually meant, satisfies the same-relation constraint. First, he maintains that a person who acts in his own interest, say by investing some money, is not acting from love . . . I do not promote my own interests or happiness because of self-love [sc., because I bear the same relation to myself as I bear to others when I love them], or from any other motive.



The argument for this seems to be that acting in one’s own interest does not require a motive because it, like self-preservation, is ‘inevitable’. But this, the argument continues, just goes to show that self-interest is not a matter of loving oneself since love of oneself, like love of others, is not inevitable at all—we can hate ourselves sometimes. Secondly, Hanfling claims that there are relations to loved others such as being charmed or entertained that it is impossible to have to oneself. Finally, one can make sacrifices for love, but it makes no sense to speak of making sacrifices for oneself. Only the first argument seems to speak to the identification of self-love with selfinterest; the other two relate directly to the same-relation constraint. All three arguments, however, are questionable. As regards the second—‘entertaining oneself ’ and so on—people do, and sometimes without self-satisfaction, take pleasure in their own sense of humour, or their ability to entertain, their accomplishments, appearance, and so on. Of course if one takes pleasure in these features of oneself, it seems more likely that others will too. But, more importantly for present purposes, one can come to take pleasure in a feature of oneself—one’s sense of humour, for example, or one’s looks— by finding that it is enjoyed (in the right way) by others. As regards the third, one can make sacrifices for oneself, insofar as one can choose— and this may be difficult—to sacrifice one thing one holds dear (say, a promise to a friend) for the sake of some goal of one’s own. I don’t see why a politician should not sacrifice a friendship for the sake of political power, and suffer for it, in just the same way in which he might sacrifice power for the sake of a friendship. It’s important to the example that the politician in question shouldn’t see his life solely in terms of service to ideals, for if he does, sacrificing the friendship for the ideal won’t be a case of sacrificing something for himself. He also mustn’t be too venal, or else there will be no sacrifice because no real sense of loss. But the following kind of example seems perfectly imaginable: a politician wants to reform the National Health Service for the sake of the National Health Service, or for the sake of the ill, but also wants to do it himself: if another politician were able to take his place and do it all just as well, and his friendship be saved, this wouldn’t be good enough. That is why our politician wants power—in order to be the agent of the changes he wants. Though it is indecent to say (thinking of his shattered friendship) ‘I sacrificed a lot for National Health Service reform’ if he gets something very coveted thereby, such as becoming Prime Minister, it could still be true—he could look back on his life and say truly ‘political life cost me a lot’, having his friendship in mind. What of Hanfling’s first argument? It is a matter of common observation that selfinterest is far from inevitable: witness the diabetic man who dies of gangrene because he ‘cannot be bothered’ to seek medical attention before it is too late, the mother who leads an unhappy and frustrated life because she can never put herself before her husband and children, and other similar cases. Inattention to one’s own long- or even medium-term interests can be a standing feature of a person’s approach to life, as well as a matter of (say) local weakness of will. A good explanation for at least some forms of self-neglect, or inattention to one’s own interests, would appear to be



self-hatred, or at least insufficient self-love. Hanfling’s argument may seem to be persuasive because we are inclined to confuse self-interest, which is as far from being inevitable as other-love, with a different capacity or range of capacities which have a much better claim to inevitability. These are the distinct, and lesser, capacities which Aristotle calls ‘cleverness’ (e.g. as manifested by the weak-willed alcoholic getting himself to the nearest place he can buy a drink (Wallace (2001)): this isn’t self-interest, because self-interest would require him to stick to what he has resolved is best for him, namely to give up drink. Rather it is a much more localized instrumentality which survives the destruction of the mechanisms which translate one’s deliberated strategic goals into action. It’s the capacity which Hume, perhaps rightly, claims to operate independently of reason, but in thinking he has shown this is non-rational, wrongly concludes that the whole idea of practical rationality has been debunked. And it’s the capacity which Kant, rather as Hanfling does, (wrongly) inflates into the capacity to look after our own ‘happiness’, but of which he (rightly) says that if its goals were the only goals we had, our endowment with practical reason would be superfluous or indeed harmful. By contrast self-interest proper, which corresponds much better to the capacity to look after our own ‘happiness’, and includes the capacity to envisage goals that are distinctively one’s own, to organize them properly, and to act on them, is something many people seem to lack to a greater or lesser degree. It is precisely self-interest in this sense that Butler, to whom Hanfling refers, means by ‘self-love’ (‘If we act conformably to the economy of man’s nature, reasonable selflove must govern’ (Sermons 2, }217)), and this usage has been continued by various writers in his wake. R.M. Adams, for example, argues that self-love is not the same as selfishness (roughly, Hanfling’s but not my ‘self-interest’), but corresponds to the ‘desire for one’s own long-term happiness or good on the whole’ (1998: 501). Also it is a rational achievement—it doesn’t come for nothing. And J.D. Mabbott explains Butlerian self-love as the rational organization of desires (Mabbott (1953): 118). The more we think of self-interest as a rational achievement and the less we think of it as like Aristotle’s or Hume’s or Kant’s sub-rational cleverness, the less substance there is to the claim that because ‘self-love’ means (approximately) ‘self-interest’, ‘love’ occurs in ‘self-love’ only thanks to a linguistic accident, and the more acceptable it is to maintain that self-love, when it’s used to mean this, satisfies the same-relation constraint. Indeed, self-love satisfies the same-relation constraint in part because it overlaps in meaning with ‘self-interest’, when self-interest is properly understood. But I take it that it does not require much argument to show that what I am calling self-interest proper overlaps, even if it does not coincide with, practical rationality. So self-love—that is, standing in the same relation to oneself that one stands in to another when one loves them— overlaps, even if it does not coincide, with practical rationality. This conclusion chimes with the causal claim I made (though did not argue for) in section 3.1, that practical rationality is fostered by the right kind of loving nurture. As Adams puts it



None of us invented for ourselves the concept of our own happiness or good. Teaching children to conceive of, and care for, their own good is one of the main ways in which one cares for their good. Conversely children who are undervalued by those who bring them up are apt to find it harder to adopt their own good with clarity and firmness as a project of their own. (Adams (1998): 509)

We become able to identify and organize our goals for ourselves by first of all being shown what they are and how to order them by others. Because they are our goals (and not just our carers’ or educators’ which they have failed to disentangle from ours), at least if we make the further (but I think undemanding) assumption that love of the right kind involves just this kind of disentangling of self from other, this teaching manifests another’s love for us of the right kind. It’s no surprise, then, that—as I have argued on independent grounds—the relation we stand in to ourselves when we come to be able to identify and organize our goals relatively unaided (etc.), that is, when we are practically rational, is that same relation—love—but with different relata (no longer other and self, but self and self ).

3.4 I now want to look at a different problem that arises for the thought that to be a selflover is (in part at least) to be practically rational, which lurks in the background of the above account. This problem is visible in Marcia Homiak’s account of Aristotle on selflove: the eudaimon life is, for us, the rational life, so insofar as we act out of the motivations, rationally chosen, that constitute eudaimonia, we act out of self-love— the person who is eudaimon is the truest self-lover. However well or badly this fares as an interpretation of Aristotle, there is a separate worry here too. Homiak says that ‘we can explain what virtuous [or again as I would prefer to say, excellent or properly formed] character is by explaining what practical wisdom is’ (Homiak (1981): 634). Now add—as she does and as I have suggested above that we do also—the claim that a practically rational (sc., wise) person is a self-lover. The problem is that this additional claim doesn’t seem to amount to much if all it is to be a self-lover is to pursue the ends that practical wisdom commends. Self-love sounds like an attitude to an object, but it looks as if we can give a complete inventory of the attitudes the practically rational person has without mentioning self-love, because no distinctive attitude is contributed by being a self-lover. We don’t further our understanding of the link between practical wisdom and properly formed character by introducing the idea of self-love, since the notion adds almost nothing to the following pair of claims: one’s character is properly formed only if one is practically wise, and a practically wise person is a person who pursues the ends commended by practical wisdom. The danger for the account I have recommended above is that we can affirm the letter of a connection between practical rationality and self-love at the cost of denying it any real substance. Let us call this problem the transparency problem.



Of course the apparent transparency of self-love isn’t a problem if there is a good reason for it, and according to Harry Frankfurt—who embraces the transparency of self-love with open arms—there is. On Frankfurt’s account, the self is transparent to the self-lover in the following sense, that self-love does not consist in having the attitude of love towards the self: Someone who loves himself displays and demonstrates that love just by loving what he loves,

where ‘what he loves’ is to be understood as a set of objects other than the self. Indeed this isn’t just an accidental fact about self-lovers: for one to be a self-lover at all, there must be something else—not himself—for the self-lover to love. So ‘self-love is derivative from or constructed out of, the love that people have for things that are not identical with themselves’ (Frankfurt (2004): 85). But can self-love, so understood, satisfy the same-relation constraint? It is (says Frankfurt) a feature of love quite generally that the lover is selflessly devoted to the true interests of the object of love, for their sakes (‘the lover cares about the good of his beloved for its own sake’ (ibid.)). So what would manifest my self-love—i.e. my love when the loved object is myself—if not my being selflessly devoted to pursuing my own true interests, for the sake of those interests themselves? Conversely, if one is not devoted to anything beyond oneself, there are no ‘true interests’ to count as one’s own, and so nothing for a lover of oneself—whether or not that lover of oneself is oneself— to promote, and so (if the lover of oneself is indeed oneself) no room for self-love. Similarly, a familiar manifestation of lack of self-love is not negative reflexive attitudes, but rather the absence of any things beyond the self to which one is devoted, or cares about. Frankfurt’s account provides a ready solution to the transparency problem because one and the same set of features that make self-love look like practical rationality—that is, that make the self-lover’s attitudes consist of all and only those of the practically rational person, and so render it transparent—also serve, on Frankfurt’s account, to explain why it meets the same-relation constraint. So self-love can be transparent without being insubstantial. The unargued causal thesis of section 3.1, to return to it once more, also fits Frankfurt’s account of self-love. Let’s assume, as before, that ‘secure attachment’ labels a disposition which is typically formed by the right kind of loving nurture and whose manifestations overlap with practical rationality. According to the psychiatrist and attachment-oriented psychotherapist Jeremy Holmes, there is a reciprocal relation between secure attachment and creative or playful exploration; only when attachment needs are [satisfied] can the individual turn away from her attachment figure towards the world. (Holmes (1996): 4, my italics).

Another’s love (of the right kind) promotes not only practical rationality but also an interest in things beyond the self. But that is the mark, on Frankfurt’s account, of self-love. So another’s love of the right kind promotes both practical rationality and self-love. If self-love



and practical rationality have as much to do with one another as I have argued they have, this should come as no surprise.

3.5 If Frankfurt’s account of love (and therefore of self-love) were the whole story about it, it looks as if the connection between self-love and practical rationality would stand up well. Unfortunately, however, I don’t think it is the whole story. In this section I am going to raise two separate sets of reservations about Frankfurt’s account and show how they can be met without upsetting the conclusions already reached. The first set of reservations stems from the fact that we are all familiar with people who both seem to be secure in their self-love, and take a minimal interest in themselves, that is, whose focus is almost entirely outwardly directed. (Quine’s autobiography—or perhaps ‘autobiography’—is said to be an example of this.) But not everyone is like that, and attitudes towards oneself are surely among the possible manifestations of self-love, even if not everyone’s self-love manifests itself this way (and even if everyone’s does manifest itself at least in part in the ‘transparent’ way). Consider the ability to accept others’ positive attitudes to oneself—be it their love, or their compliments on one’s good looks, beautifully designed garden, or whatever it might be. To the worry that this might just be vanity, remember that the self-lover is not being pictured as spontaneously reflecting on his or her good looks, beautiful garden and so on: it’s enough to entertain these thoughts when they are suggested by another. But if someone does entertain such thoughts, even if only suggested by another, the thoughts surely manifest attitudes which are self-focused in a way not budgeted for on the transparency account (‘my garden is beautiful’, ‘I look nice in this’, etc.). Now this shortcoming in the transparency account of self-love must stem from a parallel shortcoming in Frankfurt’s account of love in general on pain of violating the same-relation constraint, for the transparency account of self-love satisfied the constraint just because it portrayed self-love as a special case of love, no matter what its object. But a parallel shortcoming in Frankfurt’s account of love in general is ready to hand. It is that Frankfurt’s general account of love concentrates too one-sidedly on ‘protecting and pursuing’ the loved one’s interests—or perhaps, better, too one-sidedly on the loved one’s interests, the sorts of things that can be protected and pursued. This language fits well the case where one is (for example) a campaigner for the preservation of a piece of wilderness: Frankfurt’s account of love implies that those who love one will disinterestedly pursue the preservation of the wilderness too, or at least care about its preservation, and this (let’s allow for now) seems right. But what about, for example, a person’s appearance? A fashion model might perhaps have her appearance as an interest, occupying her thoughts in the same way as the wilderness occupies the thoughts of the wilderness campaigner. If that were so, her lover’s concern with her appearance could properly take the form of protecting or promoting something. However, though people often focus on the appearance of those they love, it is rare



for them to focus on it in this way. This objection carries over to the case of self-love. I’ve said that a mark of self-love can be a positive attitude to one’s own appearance. But it is rare for this to take the form of ‘promoting and protecting’, because few people see their appearance as an ‘interest’. However, though we have certainly come across something Frankfurt’s account of love (of self or other) overlooks, filling the gap does not bring us into conflict with the transparency account of self-love. The self-directed attitudes I have mentioned that normally go with self-love are hedonic attitudes—pleasure in one’s own appearance, or garden. (Again, there needn’t be anything reflective, still less gloating, about this: one can see that someone enjoys their good looks simply from the way they move, for example.) This complements rather than conflicts with the transparency account because that account rules out only that the self be the object of a first-order attitude of love: it doesn’t rule out that the self be the object of first-order attitudes of other sorts, including taking pleasure. Coming to see the role of pleasure in self-love also helps to show how this expanded account of self-love fits the same-relation constraint. As Neera Badhwar has rightly said If [the look of delight and tenderness on the mother’s face] is the primordial experience of being loved and the first lesson in learning to love, then one would expect delight or tenderness to be present as a strand in different sorts of loving relationships . . . Some form of pleasure in (the thought of ) the loved object’s existence . . . is central to the most general and basic expression of the emotion of love. In love of persons or animals, this basic emotion of love also includes pleasure in the well-being . . . of the loved object. (Badhwar (2003): 43)

Love, in short, involves taking pleasure in the object of love. The same-relation constraint implies that where the object of love is oneself, one will take pleasure in oneself—that is, it predicts that the self-lover will have precisely the self-directed hedonic attitudes I have mentioned. My second set of reservations about Frankfurt’s account of love has to do with his deployment of the idea of caring about something ‘for its own sake’. One worry is that, in emphasizing the lover’s care about the loved one’s interests for the sake of those interests, there is as it were an argument-place—that is, a ‘for the sake of . . . ’, where the dots are to be filled in by a designation of the loved one—that Frankfurt has overlooked. Suppose I am landed with a cat to look after. For all that cats can—as we may assume— be objects of love, my caring for this cat might be an act of love towards its deceased owner, who was a dear friend, not towards the cat itself. And yet I might be said to care for the cat’s true interests disinterestedly (i.e. look after it properly, without any thought of further gain, and so on). So caring for somebody’s (or something’s) true interests disinterestedly is not sufficient for loving them (or it). But is caring for someone’s or something’s true interests disinterestedly even necessary for loving them? For one thing, my various loved ones’ true interests might themselves be very various—the preservation of a piece of wilderness, playing the violin, and the restoration of the French monarchy, for instance. Do I really have to care



about them all in order to count as loving the people in question? So far this worry has merely to do with my own finite capacity to care about things. But what if one loved one yearns for the restoration of the French monarchy and another is an ardent republican? If what I am meant to do as a lover is care about both disinterestedly, I will be saddled with conflicting attitudes. However, consider the reality of loving mothers whose children have opposing political commitments. The mothers need not be subject to conflicting political attitudes of their own, and may be even-handedly delighted by the triumph of either child’s party. The explanation would seem to be that what love for their children requires of them is not (as Frankfurt would have it) care for their children’s true interests—here, the triumph of this or that party—for the sake of those interests, but rather care for those interests for the sake of the children. As Bradley puts it, ‘the self rises and falls with its world’ (Bradley (1962): 285). But the bit of the world this mother’s self rises and falls with is not the rise and fall of her children’s political idols, but rather her children’s selves’ rising and falling with their respective idols’ rise and fall. How do these objections to Frankfurt’s account play out where lover and loved are one and the same? As to the first objection, it might be said that, where the lover is the loved, it is incoherent to suppose that one’s own true interests might be pursued for anyone’s sake but one’s own. Whether this is in fact so is going to depend in part on what account one gives of a person’s true interests, and I cannot explore that properly here. But suppose that I have a great talent for the violin. On some conception of my true interests, it might be my true interest to develop my talent, but nonetheless the case that I develop it not for my own but for my parents’ sake. If that possibility is coherent, the same-relation constraint—which thanks to the first objection now requires that the lover cares about the loved one’s true interests for the loved one’s sake—implies that the violinist is not a self-lover. But when one imagines someone going through all that violin-playing not for his own sake, that seems like the right thing to say. If on the other hand the possibility is incoherent, then the same-relation constraint will be satisfied come what may. According to the second objection, if the same-relation constraint is to be met, it is no longer a necessary condition of self-love that one care about one’s own true interests for their own sake. The possibility of the self-lover’s not doing so may seem puzzling, but here again reflection on the notion of pleasure can help us to see why it need not be. We would do well here to remember two thoughts of Peter Winch’s, apropos of a remark by Simone Weil. In Ibsen’s The Master Builder, Mrs Solness welcomes Hilda Wangel to her house, as Winch puts it, ‘in splendid Kantian tones’ with the words ‘I’ll do my best for you. That is no more than my duty’. ‘How very differently we would have regarded her’, Winch goes on, ‘if she had said: “Do come and see your room. I hope you will be comfortable and enjoy your stay”. ’ Here Winch wants to draw attention to how one obvious candidate for the label ‘action for the sake of duty’ is not a manifestation of a will that is good without qualification. But for our purposes this is just scene-setting. Of greater interest is what Winch has to say about the contrast case



(i.e. ‘Do come and see your room’, etc.). Switching the example of warm spontaneity to Weil’s case ‘of a father playing with his child—not out of a sense of duty but out of pure joy and pleasure’, Winch says I also want to resist a suggestion that many philosophers would make here: namely that the father is doing what he is doing [i.e. playing] ‘for its own sake’. The trouble with this locution is that it makes his behaviour too like that in which a man does what he does for the sake of something else—as if the father thought that a situation in which a father plays with his child has positive value in itself and played with the child for that reason, rather than because he thought that doing so would be conducive to further consequences which he positively valued. (Winch (1972): 181, 183)

‘For its own sake’ misdescribes the father’s reasons for playing with the child just as much as ‘for the sake of further consequences’. But the father’s playing with his child is a manifestation of love: certainly of his love for the child (as Badhwar has helped us to see), but also surely of his self-love, because his relation with the child is among his true interests. Generalizing, we both satisfy the same-relation constraint and have a credible psychological picture of the self-lover if the self-lover’s pursuit of his own true interests fails to be for their own sake just in the case where he pursues them ‘out of pure joy and pleasure’. I said at the outset that there is reason to expect a correct moral psychology to show love and practical rationality to be closely connected, and undertook to demonstrate this by arguing for a constitutive connection between practical rationality and self-love. The connection was seen to depend on maintaining that self-love is an instance of the same relation one stands in to another when one loves them, that is, to respecting the same-relation constraint. Notwithstanding the objections I have raised to Frankfurt’s account of love, I have argued that this constraint can be respected if we take due account of how thoroughly pleasure is involved in love.

References Adams, Robert M. (1998) ‘Answers to Prayer and Conditional Situations’, Faith and Philosophy 15: 41–51. Aristotle (1954) Nicomachean Ethics, tr. W.D. Ross. Oxford: Oxford University Press. Badhwar, Neera K. (2003) ‘Love’. In: The Oxford Handbook of Practical Ethics, Hugh LaFollette (ed.), pp. 42–69. Oxford: Oxford University Press. Bradley, F.H. (1962) Ethical Studies. Oxford: Oxford University Press. Chazan, Pauline (1998) The Moral Self. London: Routledge. Eagle, Morris N. (1993) Recent Developments in Psychoanalysis. Cambridge, MA & London: Harvard University Press. Fonagy, Peter (1999) ‘Psychoanalytic Theory from the Point of View of Attachment Theory and Research’. In: Handbook of Attachment, J. Cassidy & P. Shaver (eds), pp. 595–624. New York & London: The Guilford Press. Frankfurt, Harry (2004) The Reasons of Love. Princeton, NJ & London: Princeton University Press. Hampton, Jean (1993) ‘Selflessness and the Loss of Self’, Social Philosophy and Policy 10(1): 135–65.



Hanfling, Oswald (1993) ‘Loving my neighbour, loving myself ’, Philosophy 68: 145–57. Holmes, Jeremy (1996) Attachment, Intimacy, Autonomy. Northvale, NJ & London: Jason Aronson Inc. Homiak, Marcia (1981) ‘Aristotle on Self-Love’, Canadian Journal of Philosophy 11(4): 633–51. ——(1997) ‘Aristotle on the Soul’s Conflict: towards an understanding of virtue ethics’. In: Reclaiming the History of Philosophy: Essays for John Rawls, A. Reath, B. Herman, & C.M. Korsgaard (eds), pp. 7–35. Cambridge: Cambridge University Press. Hume, David (1896) A Treatise of Human Nature, L.A. Selby-Bigge (ed.). Oxford: Clarendon Press. Kant, Immanuel (1784) Groundwork of the Metaphysics of Morals, in Kant’s Critique of Practical Reason and other Works, tr. T.K. Abbott. London: Longmans, 1909. Lear, Jonathan (1992) Love and its Place in Nature. London: Faber. Loewald, Hans (1978) Psychoanalysis and the History of the Individual. New Haven & London: Yale University Press. Mabbott, J.D. (1953). ‘Reason and Desire’, Philosophy 28: 113–23. Murdoch, Iris (1970) The Sovereignty of Good. London: Routledge & Kegan Paul. Suttie, Ian (1963) The Origins of Love and Hate. Harmondsworth: Penguin. Wallace, R. Jay (2001) ‘Normativity, Commitment, and Instrumental Reason’, Philosophers’ Imprint 1: 4, Winch, Peter (1972) ‘Moral Integrity’. In his Ethics and Action, pp. 171–92. London: Routledge & Kegan Paul.

4 The Nature and Morality of Romantic Compromises Aaron Ben-Ze’ev

Compromises are of great value in many circumstances of human environment; compromises can help us to avoid disputes and fights and enable us to live peacefully with each other. However, compromises also have a negative aspect in that they might require us to surrender our principles. These two contrasting views of compromise are based upon the two opposing aspects implicit in compromise: the necessity to be sensitive to the other’s needs and the willingness to relinquish something of value. In the act of compromise, a sort of yielding occurs in which we give in to what we no longer can, or no longer wish to, oppose or resist. The issue of compromise becomes even more complex when we are dealing with romantic love, the ideal characterization of which seems to refuse any type of compromise. However, since romantic love involves an intimate and intense relationship between two people, compromises are required in many romantic circumstances. The test of genuine love is whether lovers can accommodate these compromises in a manner that prevents, or significantly reduces, the negative aspects, and allows the positive ones to prevail. Compromises acknowledge the existence of conflicts and are one way in which we attempt to cope with them. Accordingly, I will begin my discussion by describing the major types of conflict resolution in personal relationships and will then discuss the nature and role of compromises in romantic relationships. The last section will discuss the activity of complementing, which together with the activity of compromising, can provide a novel model for romantic relationships in our modern dynamic society.

4.1 Types of Conflict Resolution The map of activities that we use for coping with conflicts ranges from one-sided activities, such as coercion and surrender, to the most interactive ones, such as compromise and complementation. Although the focus of the article is on the latter


A A RO N B E N - Z E ’ E V

(and in particular their expression in the romantic realm), understanding the former is of great explanatory value. Here are a few major activities for coping with conflicts: 1. Coercing—coercing the other into accepting something negative. 2. Unilateral concession—accepting something negative without getting anything in return. 3. Tolerating—accepting something negative while maintaining a strong negative attitude toward it. 4. Compensating—something is given in order to compensate for loss, suffering, or injury (which has often been inflicted by some kind of coercing). 5. Compromising—accepting something negative in order to obtain something positive in return. 6. Complementing—combining characteristics from various parties to form a complete whole enhancing each party. 7. Accommodating—the process by which the positive aspects gain greater weight to the extent that the negative ones almost (or even entirely) disappear. 8. Settling—the process by which the negative aspects gain greater weight to the extent that the positive ones almost (or even entirely) disappear. The first four activities, coercing, unilateral concession, tolerating, and compensating, are essentially one-sided activities; as such they are not helpful in resolving most conflicts. The next two activities, compromising and complementing, are the major methods of seeking a positive solution to conflicts, and my discussion will center upon them. The last two activities, accommodating and settling, are processes by which we cope with a given conflict resolution. The first two activities, coercing and unilateral concession, are the most one-sided; they do not involve any consideration of the other and even display a profound wish to disassociate from the other. Sometimes, when no other type of interaction is feasible, disconnecting oneself from the other is the only remaining option. Unilateral concession seems to be the more moral way, as it offers the other greater freedom, but it is unhelpful in the long run as it demeans the giver and resentment is likely to accumulate. When intimate personal relationships involve submissiveness, they are morally unjust and problems are bound to occur, since such relationships need to express the value of each person in the relationship. Tolerance and compensation are still one-sided, but less so than coercing and concession. Tolerance involves concession, but it is a more temporary and limited concession. It can be described as a tactical concession compared with the strategic one associated with unilateral concession. In tolerance we accept the negative recourse, not because we think it is good, but because we believe it is the best solution in the present circumstances (which could change in the future). Compensation is defined as ‘something, such as money, given or received as payment or reparation, as for a service or loss or injury’. In compensation we do not



change whatever it is that we did, which is typically a unilateral decision or action that we took without consulting the other. Compensation is often not a solution to the conflict, but a measure that might facilitate the healing of the wounds caused by an unfavorable resolution. In compensation there is no real engagement between the two parties; there is only a one-off action done in recompense for having taken a one-sided resolution. In some cases, our moral obligation is to make either the best possible compensation or the best possible compromise. Resolution by compromise has priority over resolution by compensation. The major means for genuine conflict resolution, especially in love, are compromising and complementation. Compromise is typically not a one-time state, but rather an ongoing process. In compromise there is the need to accept something that has a certain negative value. There can be various reasons for accepting something of negative value, but the major one in the case of compromise is that something positive is gained in return. In compromise one gives up something of value in order to get something else of greater value. Compromise can be characterized as an attempt to cope with a situation in which one experiences a gap between the desired and the available by giving up some aspects of the desired while keeping others of them. In this sense, resolution by compromise has moral priority over resolution by coercing, concession, toleration (Day (1991)). Complementation is defined as ‘completing something or making two things into one whole; being complements of each other’. ‘Complement’ is derived from the Latin word for ‘fill up’ or ‘complete’. When something complements another thing, it does more than complete it; it enhances it. Certain wines complement particular dishes; certain condiments complement specific foods; and certain colors might complement your hair or eyes. A complementary arrangement is one that is compatible with both parties. Note the difference between complement and compliment. A compliment is a ‘remark of praise’; complement refers to ‘something that completes’ (V2 Vocabulary Building Dictionary (2008)). The linguistic affinity between complement and compliment might suggest the value in the process of complementation. Complementation is a relational structure by which two independent people enhance each other’s strength and improve their common experience. The opposite of ‘to complement’ is ‘to clash’. Complementation requires both people to be constructively active. More than any other means of conflict resolution, it requires relational and interactive activities from both partners. Whereas coercion and concession are one-sided activities involving no reciprocity whatsoever, tolerance and compensation have some aspects of limited reciprocity. But quite often both tolerance and concession are typically chosen in order to keep a distance from the other person. Due to the lack of genuine reciprocity, both tolerance and compensation involve inequality. Complementation involves the most profound sense of reciprocity and equality, both of which are indispensable elements in the


A A RO N B E N - Z E ’ E V

newly created circumstances. In compensation the reciprocity and equality exist as well, but they are not so profound since they involve some negative elements. Once the above conflict resolutions are in use, certain processes will begin to occur. One obvious process is that of maintaining the initial state that has been established by the conflict resolution. If attitudes do change, they can do so in two major manners: by accommodating or by settling. In accommodation, the positive aspects of the conflict resolution gain greater weight, to the extent that the negative aspects almost (or even entirely) disappear. Thus, if a woman loves opera and her husband likes soccer, they might compromise by occasionally—or even regularly—accompanying the other to their preferred pastime. However, because they enjoy each other’s company, over time they might both begin to enjoy the pastime to which they accompany their spouse. This is a process of accommodation in which the negative aspect becomes a positive one. Another example of accommodation is that of women who do not like oral sex, but who compromise and perform it because they want to please their partner. In time, they might begin to enjoy it, but if they continue to dislike it, or their dislike turns to disgust, the situation is likely to become problematic, or to require the partner to relinquish his preference for this sexual activity. Settling can be characterized as a gradual sinking to a lower level. Settling, which can be indicated by a readiness to suffer in order to make the other person suffer more, is illustrated in the story about the angel that granted a farmer any wish that he wanted, but on one condition: his neighbor would receive twice as much of the same. In that case, said the farmer, please remove one of my eyes. In one blog discussing settling, a woman has suggested the following distinction between settling and compromise: ‘Settling would be having no clothes on. A compromise would be having a towel on.’ Settling is evident in those cases in which two people remain in an intimate relationship ‘for the sake of the children’. Their initial compromise is that although they do not want to stay together, they do so as they believe it will benefit their children. As time goes by, it becomes clear that the new situation is even worse for the children who now have to witness constant fights between their parents. In this instance, settling involves a lose–lose situation. Accommodation is more typically the outcome of complementation, while settling is often the result of compromise. Having delineated the map of possible conflict resolutions in this section, in the rest of the article I will focus on compromise and in particular romantic compromises. I will then devote a special discussion to romantic complementation and its possible integration with compromise.

4.2 The Nature of Compromise Compromise entails accepting something negative in exchange for the prospect of gaining or maintaining something positive. There are two major aspects in compromise: (a) the settlement of differences by consent reached by mutual concessions; and



(b) a concession to something disagreeable. The first aspect entails sensitivity to the other’s needs; in this sense, compromises enable people to live together peacefully. The negative aspect of compromise involves the possibility that one might have to relinquish some of one’s principles, values, or merely one’s likes, preferences, or habits. The second aspect can entail the loss of something of considerable value. Unlike politics, which is often characterized as the ‘art of compromise’, in ethics, compromise is usually regarded as a sign of weakness or lack of integrity (Benjamin (1990); Goldin (1979); Van Willigenburg (2000)). The term ‘collaboration’ has somewhat similar dual positive and negative meanings: (a) the act of working together with one or more people in order to achieve something; (b) the act of working with an enemy to betray one’s own (particularly applied to working with an occupying force against one’s own people). Both terms, ‘compromise’ and ‘collaboration’, refer to some involvement with people; the involvement itself is positively perceived as long as it is not done at the price of one’s and others’ fundamental values. Interaction with other people, which is so essential and valuable for human life, often involves the agent in making significant unjust sacrifices (in compromise) or unjust gains (in collaboration). The negative meaning of both terms has been compared to betrayal—either of one’s values or those of others. There is in compromises the recognition of the complexity of life and human interaction. In Hebrew the word ‘compromise’ stems from the same word as ‘understanding’. To compromise is to understand other people and their realities. One is not alone in the world, hence one must take into account the existence of other people; something can be positive for one purpose and negative for another. There is hardly anything that is good or bad in every respect. We always have to take the bitter with the sweet. ‘Doing the right thing’, which is typically what is reasonable and possible, often requires compromise. Doing the daring and innovative, which typically goes beyond the reasonable and possible, usually involves a refusal to compromise. Accordingly, it can be claimed that in many cases of compromise, there is a sense of mediocrity—a recognition that one cannot excel, cannot expect everything, and cannot have the very best, so one settles for what is available and comfortable in the current circumstances. However, compromise does not necessarily imply mediocrity or indifference (Biddle (1957)). On the contrary, it is often an ongoing process of trying to accommodate in the best possible manner to the dynamic circumstances. Although the choice that is made in a compromise is not final and not optimal, it may nevertheless be the best available choice. At the basis of compromising is our ability to hold various perspectives at the same time. When we see a situation from various perspectives, we become aware of the value of these perspectives and hence that our own initial attitude is not the only game in town; other perspectives might be preferable in certain circumstances. One of the risks in taking on multiple perspectives is that the boundaries between good and bad can become interrelated and blurred, causing choices to become confusing and decision-making difficult.


A A RO N B E N - Z E ’ E V

Rational behavior should take account of all possible perspectives and make an optimal decision accordingly. Underlying our intellectual needs is the curiosity that helps us to enlarge our knowledge and be aware of various perspectives. Curiosity, which is an essential feature in intellectual attitudes, is expansive and seeks out more and more perspectives. In this sense, rational thinking, which pays greater attention to our limitations and impediments in reality, tends to make compromises. Emotions are the opposite—they typically take a very partial perspective. Emotions are partial in that they focus on a narrow target, such as one person or very few people, and express a personal and interested perspective. Emotions direct and color our experience by selecting what attracts and holds our attention; they make us preoccupied with some things and oblivious to others. Not everyone and not everything is of emotional significance to us. The intensity of emotions is possible due to their focus upon a limited group of objects. The partial and personal nature of the emotional perspective means that it is much harder to make compromises while we are under the influence of intense emotions (Ben-Ze’ev (2000)). The need for compromise is then compatible with the need to grasp the perspectives of others. Understanding the perspectives of others involves at least an appreciation, and to a certain extent an acceptance, of some of their aspects. One does not have to adopt all types of compromises in order to maintain a viable personal relationship: personality and specific circumstances are relevant in determining the most fitting types and the extent of their implementation. It is obvious, nonetheless, that in principle, some types of compromises are urgently needed. Very few are fortunate enough to have maintained intimate personal relationships without having invested any effort or made any compromises. Furthermore, there are those who have been unable to find or to maintain long-term romantic love, either because they were not ready to invest and compromise or because their investment and compromises were to no avail. Romantic compromises also require one not to take oneself too seriously. In compromising, one is ready to give up some of one’s values and attitudes for the greater value of being with other people. By doing this, one is recognizing that one’s attitudes or preferences have limited importance, a recognition that in itself brings with it some advantages. As someone once said, ‘Angels can fly since they take themselves lightly’. Accordingly, Julia, a married woman, said: ‘As I think more profoundly about life and love, I think that we take it all too seriously. Life should be taken sincerely and wholeheartedly, but sometimes we are just too full of it and ourselves.’ From another perspective, compromising by taking our stands less seriously may be one of the most serious (in the sense of profoundness) activities of our lives. We need some distance in order to achieve a perspective that encompasses the multiple aspects of the object and thereby to see it more fully. This perspective is also required in love; without it, our perception of our partner would be fragmented. However, keeping our distance is contrary to the involved and intimate perspective typical of intense love. In many cases, the lover is unable to detach herself from the beloved in such a way that a more complex perspective can be achieved.



In compromise we accept a negative aspect in order to maintain a positive one. Thus, people are ready to sacrifice monetary gain to ensure that they will not subsequently experience greater loss. Compromises are typically risk-avoiding choices. The tendency to minimize risk, and hence to minimize sadness, prevails among many but not all people; some people are motivated to maximize benefits, and hence to maximize joy. The latter appear to be less likely to compromise. As people cannot completely avoid the risk of losing the positive aspect, they try to minimize it. They do it by compromising, by making provisions against unforeseeable difficulties, as a kind of insurance. They believe that by accepting the negative aspect of a person, they are increasing their chances of not losing this person. Life is full of trade-offs and sometimes the expected benefits are greater than the expected costs. One reason why battered wives do not leave their husbands is that they do not know what the future holds— they do not want to take a risk by stepping into the unknown. Similarly, a woman may compromise and remain with her partner who is not profoundly intellectual, rather than trying to look for a new partner who may be, if found at all, more intellectual but not as kind as her current partner. In some cases compromises can deliver a kind of insurance while in others it fails to do so. The need to compromise is even more pronounced in our flexible modern society, which is populated with so many alternatives from which to choose. When there are hardly any alternatives to the present situation, there is no need to compromise or to ponder our optimal way of behaving. The issue of compromise becomes more critical when there are many tempting alternatives. Compromises can be of a limited or profound nature. Limited compromises are small in scope or value; they are confined within certain boundaries. Those can be temporal ones, in which case the compromise is temporary and will last until external circumstances change; they can be limited by their value, in which case a person gives up something of little value to him, such as sleeping late, in order to enhance his romantic relationship with his partner, who loves to be with him in the early hours of the morning. A profound compromise is one in which we give up a profound value in exchange for something of lesser value. An example is when I hold a profound value that I want to fulfill, such as finishing writing a book, but I fail to do so because of some superficial values that are easier to fulfill, such as enjoying a television comedy that is more enjoyable in the short run.

4.3 Compromises in Love Speaking about compromising in love seems odd, as ideal love is perceived to be so intense that it has no need to compromise. Thus, no one would say to his beloved, ‘I love you darling, although this love is a compromise for me’. Compromising about people can easily be perceived as degrading to these people; however, compromising about a specific behavior of these people might be perceived in a somewhat better light, as the compromise refers not to the person as a whole, but to certain of his specific qualities


A A RO N B E N - Z E ’ E V

that do not reduce his lover’s overall perception of him. Even when the entire value of a partner is evaluated and considered to be a compromise, we must still point out the framework in light of which the person is a compromise. He might be a compromise compared with the perfect prince riding on the white horse, but not compared to most, or even all, other men. The opposition to romantic compromises is evident in the Romantic Ideology in which ideal love is total (in the sense of comprehensive), uncompromising (in the sense of inflexible), and unconditional (i.e. unaffected by the conditions of reality) (Ben-Ze’ev & Goussinsky (2008)). The uncompromising aspect of ideal love refers to the lover’s own attitude toward the beloved: this attitude does not accept ‘maybe’, ‘to a certain extent’, or ‘gradually’ as terms that apply to such love. Moreover, the beloved is considered to be perfect, and committing suicide because of unrequited love is not an unusual story; it is even regarded as a perfect illustration of true love. The great availability of prospective romantic partners enhances the feeling that one should not compromise in love. Nevertheless, compromises are quite frequent in the romantic realm. There are various types of romantic compromises; two such major types concern the state of one’s loving heart. The first type refers either to (a) not activating one’s loving heart by staying within a romantically dull relationship, or (b) repressing one’s loving heart by not pursuing one’s genuine love. The second type involves cases in which following one’s loving heart violates normative romantic boundaries, such as that of strict exclusivity. In all such cases, the decision to make the compromise is often complex as it depends upon taking account of factors outside the romantic realm and according them a specific weight. It seems that people in our fluid modern society do not have to compromise and pay the price for committed romantic relationships. They want ‘to eat the cake and have it, to cream off the sweet delights of a relationship while omitting its bitter and tougher bits’ (Bauman (2003): ix). In committed relationships, where compromises are made, paying a price for being in a relationship implies paying a price for leaving it. Unlike virtual relationships, it is not that easy to enter or to exit from them; there is no ‘enter’ or ‘delete’ key to press (Ben-Ze’ev (2004)). In the next sections, I will discuss important aspects of various compromises: compromise as crossing boundaries, compromising the ideal, compromise as an ongoing process, compromise as waiting for the best, and compromise as getting less.

4.4 Compromise as Crossing Boundaries Desires are broad in their scope (we want more than what we can have or believe we should have), not very organized (they lack a clear order of priority and are frequently incompatible), and often unrelated to external constraints. Desires reflect our wish to overcome basic human limitations and inadequacies. Managing our desires requires drawing boundaries, setting ideals, and making compromises. The normative boundaries that prevent us from engaging in desirable activities are typically contrary to our spontaneous inclinations that express our momentary desires;



if the norms were in accordance with these, no boundaries would be necessary. In this sense, boundaries are highly inconvenient and maintaining them is a kind of compromise. But boundaries also protect us and hence generate pleasant feelings of comfort and security. Boundaries are necessary for leading a normal everyday life. They clarify the zone in which we are supposed to be. Ideals are like positive beacons, signaling the direction in which we should head; boundaries indicate the lines we should not cross. While animals exhibit desires, only humans exhibit the desire to have certain desires but not others. This self-reflexive aspect of emotional life inevitably generates conflicts among our various desires as well as between our desires and other values. Coping with such conflicts requires compromises. Compromises are related also to another basic human feature, namely, ambition. We want more than we can get and more than we are morally entitled to have. The desire to transcend boundaries is inevitable since moral rules often forbid what people desire most. As our normative boundaries are constitutive of our personality, ignoring them is tantamount to a failure to uphold our intrinsic dignity at a time when our central values are abused. The risk in such behavior is not limited to the possibility of being caught and punished, but also includes the risk of an internal crisis, which threatens to ruin the foundations of our self-identity. Accordingly, people are often proud of their ability to maintain normative boundaries when confronting obvious temptations. Ideals and boundaries imply that some things are more significant than others; in this manner, they determine the kinds of compromise we are supposed to make. We can surmise that setting boundaries forces us to compromise on what we really want. However, it can also be argued that neglecting to draw boundaries, set ideals, and make compromises actually means being enslaved to one’s present desires and demonstrates an inability to direct our life in light of our ideals and long-term considerations. The dark side of being free from compromises is that it makes people neglect some of their profound values. The extent that we prize our fundamental values is exhibited in our readiness to compromise other values and needs, which we consider to be of lesser weight. Accordingly, we ought not to perceive self-control or the adherence to boundaries as negative compromises in which we surrender to external pressures that are at odds with our desires. It follows that some of our deepest conflicts are not at the intersection of external boundaries and our desires; rather, they are situated between some of our most profound values. Coping with these conflicts requires compromises. People in love occasionally feel chained by external constraints that prevent them from acting in accordance with their wild passions, yet they are ready to compromise their autonomy and let their beloved rob them of their liberty. They are ready to be chained to the beloved because they consider acting in accordance with their loving heart to be the greatest expression of freedom. Since it is our values that construct, and yet contain, the way in which we draw our boundaries, our autonomy is both expressed and constrained in this process. People must compromise in the sense that they have to give up, at least to some extent, some of their values.


A A RO N B E N - Z E ’ E V

The need to prioritize our values implies both the establishment and the violation of values and the extensive use of compromises. Prioritization is an expression of the rules we employ in deciding by which values we should abide and which we can compromise. In this sense, we habitually cross boundaries and make compromises. This, however, does not mean that all compromises involve compromising oneself. Compromises are frequently perceived to involve giving up pleasurable activities in order to maintain more profound factors. But compromises also occur in the other direction—when we give up profound factors in order to pursue pleasurable activities. In both cases we cannot have it all; we must compromise one valued factor for another. In compromises, we do not abandon our values all together. We maintain both our wish to keep our values and the desire to violate them. The act of compromise implies the ability to distinguish between the significant and the insignificant. If everything was significant to the same degree, there would be no need for compromises and there would be no rationale whatsoever in making choices about one’s behavior. A marked characteristic of human behavior is the attempt to make reasonable choices when faced with many valued possibilities. As the distinction between the significant and insignificant is less pronounced in children, and many of their wishes are of equally great significance, compromises are less dominant in children’s behavior. When everything is of great importance, we are unable to establish an order of priority and to choose what is most suitable in the given circumstances. We cannot have everything and trying to have everything is a sure way of disregarding reality and generating constant dissonance. The unattainable wish to have everything and the necessity of making compromises are expressed in the pain of making choices, which is an unavoidable part of human life. As we have very limited ability to change external reality, the issue of when to compromise our values and when to try to change external circumstances in order to make them fit our values better is a central issue in life. Compromising too quickly is often as harmful as fighting hopelessly against external circumstances. Compromises are sometimes similar to bargaining, but they are usually more creative than simply ‘splitting the difference’, as they refer to aspects that could be different in nature. Measuring the value of a compromise is difficult as the positive and the negative might not be commensurable; furthermore, any attempt to compare a present value against a future one is bound to be flawed given that the future is unpredictable. Crossing boundaries can be advantageous, in the sense that it enhances creativity and the ability to adopt new perspectives and insights. Globalization, which has great advantages to offer us in many aspects of our lives, is essentially a process of boundary crossing.

4.5 Compromising the Ideal Lovers describe their love as boundless, as an experience that makes them feel as if they are flying, knowing they will never fall. Such an uncompromising attitude is common, as it stems from the human wish to overcome our basic limitations. Nevertheless,



compromises are possible, and even necessary, when it comes to the implementation of pure love. The Western self-help culture is obsessed with the idea that there is a difference between the myth of love and its reality and that the implementation of true love must involve compromise. Many people suffer because of the gap between their ideal notion of a lover and their real-life partner. We can cope with this gap by, for instance, revising our ideal, changing our partner, or changing ourselves. Compromises are instrumental in all these. When compromises fail, the only way left is to replace the ideal or the partner. Romantic compromises are the last means before replacing everything and starting afresh. Sometimes compromises are of great value; sometimes they are not worth the effort. Can we predict in advance the value of romantic compromise? In many cases, we cannot. Considering the beloved as perfect, or approaching such perfection, does not necessarily mean that lovers are blind to their beloved’s flaws, but just that such flaws have little, or greatly reduced, significance. At the basis of romantic love there is a profound positive evaluation of one or a few of the beloved’s qualities. This qualitative evaluation is typically associated with a comprehensive quantitative evaluation, whereby the positive evaluation is extended to include many additional characteristics. In this case, love is not based primarily upon the value of the specific characteristic, but upon its association with what we love. By giving significant weight to various characteristics of the beloved, lovers do not necessarily distort reality, nor are they completely blind to the beloved’s faults; they just consider such faults to be insignificant. Indeed, in Fisher’s survey of people in love, about 60% agreed with the statement that they love everything about their beloved and that even though the beloved has some faults, these don’t really bother them (Fisher (2004): 8). This view of romantic love, as basically attributing significance to one or a few of the beloved’s characteristics, enables ideal lovers to avoid the feeling that they have to make compromises and hence provides their love with a better chance of lasting longer. The act of assigning specific weight to a quality is not a cognitive process that can be appraised as true or false; rather, it is an evaluative process that has in it a significant subjective element that refers to what the lover wants and needs. Needless to say, this evaluative task is less sensitive to being refuted by reality and hence it is less likely that it will be considered to be a compromise. Avoiding perceiving ourselves as making compromises by attributing various weights to different properties is less efficient in the case in which we compare our lover not to an ideal lover, but to specific people around us. Here the impact of reality is greater and the comparison is more specific. However, those whom we do not know so well are often perceived idealistically. Staying within a committed relationship has become a choice that requires us to constantly re-examine its value in light of, among other issues, the presence of love. When we have an easy way out of our current relationship and into a new one, the need for a romantic compromise increases. However, using the term ‘compromise’ in


A A RO N B E N - Z E ’ E V

the same breath as ‘love’ could appear a contradiction in terms—love has been profoundly idealized and it is not easy to relinquish such an ideal.

4.6 Compromise as an Ongoing Process A distinction can be made between a one-time compromise and an ongoing one. The one-time compromise can be an outcome of a previous process in light of which the compromise was made, or it can be the starting point for a further ongoing process of accommodating oneself to the compromise or making the circumstances that led to the initial compromise even worse. Although reference to ‘the one-time compromise’ expresses the momentary nature of the situation, a sort of snapshot of that moment, ‘the ongoing compromise’ is a more accurate reflection of one’s overall situation. Time is indeed an important factor in our ability to evaluate and accept compromises. The greater the significance of the compromise and the longer it is required to be accepted, the stronger the opposition to this compromise will be. Profound compromises might be accepted if they are perceived as provisional. The assumption is that as time goes by the person will somehow accommodate to the compromise and will not see it as such. If this does not happen, it is likely that the compromise will be viewed with even greater negativity. Accepting a compromise does not mean, of course, abolishing the negative aspects of the compromise, but just seeing other advantages to it. The initial situation of compromise is that of a win–lose situation; there is a complex situation and you win some things and lose others. As time goes by, the win–lose situation might remain so and the whole process could be described as a prudential compromise. Alternatively, the initial compromise might turn into a win–win situation, characterized as accommodation. It might also turn into a lose–lose situation, characterized as a process of resignation or of settling. In prudential compromise the initial weight of positive versus negative aspects remains more or less similar, although the nature of these aspects might change. Thus, if the husband wants to have a dog and his wife hates dogs, she might agree to have a dog for the sake of her husband. After five years of having a dog, she might never grow to like the idea; the negativity of the original situation might have retained its initial weight. Although this situation remains a win–lose situation, it is likely to become an ever-more problematic one, since it introduces and then maintains negativity into the relationship. In accommodation, the positive aspects of the conflict resolution gain greater weight, to the extent that the negative aspects almost (or even entirely) disappear. Settling can be characterized as a gradual sinking to a lower level. An example of settling can be found in Kayser’s study of disaffected marriages. This study found that the major events responsible for the deterioration of love involve one partner’s controlling behavior, in particular behavior that consists of unilateral concessions and compromises. This study noted that the spouses of the controlling partners tried, with little success, to stop the process of disaffection by seeking to please their partners even



more. These women (and women are typically the subject of control, while men are typically the controllers) were making more and more concessions and compromises in order to keep their marriages alive, but the results were the opposite—the marriages continuously deteriorated (Kayser (1993)). As compromises are reactions to complex situations, it is likely that each compromise is an ongoing process and so is the experience of learning to live with it.

4.7 Compromise as Waiting for the Best A typical example of limited compromises is a temporal compromise, which involves delaying gratification or postponing a desire. While this involves endurance or patience in order to achieve the desired goal, profound compromises involve more than enduring the delay or postponement of pleasure; they also require us to lower our standards and values. One might not consider a temporal compromise to be a genuine compromise, since it is merely a way in which to achieve our goals. Indeed, there are many love stories that praise the postponement of implementing love or sexual gratification till the right moment. True love, it is said, is able to wait and can prevail even when the circumstances are unfavorable. Lines such as ‘I’ll be waiting for you till the sun don’t shine’ and ‘I will patiently wait for you till the end of time’ echo the sentiments of many lovers. Such an attitude leads people to compromise on the less significant temporal aspect in order to avoid compromising on the more significant aspect: the beloved. Lovers are prepared to be patient and to provisionally relinquish some of their wishes if this is the only way to achieve their desired end. However, when the unpleasant situation is permanent, it can no longer be considered a temporary means and it becomes a genuine, painful compromise. Although in a temporal compromise time is of less significance and its sacrifice is not considered to be a profound one, there are other circumstances in which time is a significant factor in making sacrifices. The ideals that people had when they were younger are often compromised when they grow older. Modern consumer culture opposes the value of waiting, just as it opposes the value of other kinds of compromise. As the commercials introducing credit cards put it, we can now ‘take the waiting out of wanting’. Like other consumer goods, people are forced to believe that romantic relationships should also be available on the spot and be disposable, with little or no regret and hesitation, the moment a new and improved version can be found. From such a perspective, there seems no sense in investing time and effort in repairing the old model. As in stocks, the only hesitation concerns timing—is now the right moment to sell and buy? Romantic regrets are mainly about compromises in which our heart has been compromised. The wish to give up compromises and regrets is tantamount to the wish to relinquish concern for the other. In light of the lack of concern for the other, there is little one can do when one is being disposed of by one’s partner (Bauman (2003): 12–16).


A A RO N B E N - Z E ’ E V

4.8 Compromise as Getting Less In compromising we have to live with some negative aspects. In some cases, the negative aspect is that we just cannot get everything we want and must be satisfied with getting much less than we want. This aspect of getting less can take various forms, such as, for example, getting a reduced amount of the desired thing, making do with an imitation of the desired thing, and merely imagining the desired thing. Getting quantitatively less When we compromise, we do not get whatever we want, but less, and sometimes much less, than what we want. In this sense, compromise is different from the circumstances of ‘the winner takes all’; in compromise, you take some and lose some. In my view, the complex emotional experience of romantic love involves two basic evaluative patterns referring to (a) attractiveness—that is, an attraction to external appearance, and (b) praiseworthiness—that is, positive appraisal of personal characteristics. Romantic love requires the presence of both patterns. An attractive woman might want to be loved not merely for her beauty but also for her actions and personal traits. An unattractive woman might wish the contrary: that her beloved would value her external appearance as much as he does her kindness or wisdom. In contrast to romantic love, where both evaluative patterns are essential, in sexual desire attraction is far more dominant. Sexual desire is a simpler attitude based largely on spontaneous and non-deliberative evaluations, whereas romantic love often requires deliberative evaluations. Sexual desire is typically focused on limited aspects of external appearance; romantic love is more comprehensive. Romantic relationships involve both acute sexual passion and valued emotional intimacy. A relationship can be officially pronounced dead when no sexual passion exists or if there is no valued emotional intimacy. In most cases of long-term romantic relationships, sexual desire is considerably reduced with time, while the valued emotional intimacy could even improve. It is harder to keep sexual desire very intense, as this desire depends a lot on the presence of change and novelty. However, the valued intimacy usually benefits from time spent together and hence it might improve. Romantic compromises can result in lowering the weight of each component. A prevailing romantic compromise is accepting the decline in sexual desire while maintaining or even increasing emotional intimacy. Thus, in a loving relationship, sexual desire might decrease as time goes by, while caring and the desire to spend time together might increase. Accepting that the frequency of sexual intercourse is not the only measure of romantic love—nor even the most important one—means attributing less significance to it within the entire romantic framework. Recognizing that each feature of love can be weighted differently and that the significance of a single feature might shift over time can even help to rekindle the flames of romance. This approach is different from the one that advises us that we should ‘expect less from marriage’ over time. Lowering our level of expectation can reduce the risk of disappointment and



temper our excitement, yet it does not offer a solution to the issues that arise in longterm romantic relationships; it merely indicates one way of escaping from them. The compromise suggested here is to expect less in one domain but more in others; in such a way, people are still able to experience all aspects of love. Nevertheless, there are certain features that are necessary for romantic love: relationships in which there is a total absence of passion cannot be considered romantic love. In this regard, we can distinguish between the sentiment of love—i.e. the dispositional, attitudinal complex of love—and the specific, acute emotion of love. The sentiment is usually realized by various eruptions of acute passionate emotion. When love is in its initial stages, the intervals between each eruption are typically brief; as time goes by, the hiatus between each eruption will become longer. A relationship in which intensity has declined is by no means the same as a disaffected relationship. Disaffection refers to the absence of loving feeling and not merely to declining intensity (Kayser (1993)). In a loving romantic relationship, sexual desire, which is an important component of romantic love, can decline in intensity, but still remain alive. This is an important characteristic of compromise—in compromise you do not give up everything; you give up something in order to get something else. There are various types of compromises concerning sexual exclusivity that are based upon getting less, on having some kind of sexual interaction that is limited in some essential aspects. Such compromises are expressed, for example, in rules such as the ‘doesn’t count’ rule, which allows for oral sex, one-time sex, out-of-town sex, phone sex, and even mental infidelity; as well as rules such as ‘anything goes—except love’, ‘sex and nothing more’, ‘no couple-like behavior outside the bedroom’, and ‘anything above the waist isn’t cheating’ (Kipnis (2003): 12). Such rules make romantic boundaries more flexible and compromises concerning the exclusivity of sexual relationships more acceptable. The type of romantic compromise one ought to adopt depends on various complex contextual and personal variables. Thus, it matters whether one’s main concern is with the passionate aspect of love, which is of shorter duration, or the caring aspect, which is of longer duration. In other words, the weight attributed to the various aspects of love will help determine what compromises are chosen. Different people need different types of compromises. This is similar to the choice you make in deciding how to invest your money; your choice depends on when you want to enjoy its fruits. If it is needed in the long run, you might want to invest in more stable bonds, but for short-term goals, stocks might be of greater interest despite their risky nature. No doubt, there is no one-size-fits-all compromise for maintaining love; there are, however, certain general patterns that are evident in many cases. Getting the imitation of the real thing A prevalent compromise is one in which we do not get the real thing, but something that is close to it. Our imagination plays a crucial role in this regard; dreaming about


A A RO N B E N - Z E ’ E V

another person, looking at pornography, masturbating, and having cybersex are such examples. Imagination plays an important role in our life. It is impossible to conceive of human beings without their imaginations. Imagination fulfills cognitive and affective functions crucial for human existence. It would be harmful for human existence if we were to evade our imaginations and in any case, it is impossible to do so. The imaginative compromise has many advantages: it brings people closer to their desired experience, in a way that would be impossible for them otherwise, while saving them the costs of actually risking that experience. Consider the case of a married man who indulges in sexual fantasies about a movie star, while having sex with his wife. Since his wife is not as exciting as the movie star, being with her is a kind of compromise for him—though it is essentially a psychological compromise, as in reality he cannot have sex with the movie star. We can distinguish various manners in which the motivational component in emotions can be connected to actual behavior: (a) a full-fledged desire, which is expressed in actual behavior; (b) want, which is not expressed in actual behavior because of external constraints; and (c) a mere wish, which is not intended to be translated into actual behavior. A most typical compromise is the one in which our wants are not expressed in actual behavior. In this sense, the motivational component of emotions is more strictly related to action tendency than to actual behavior. The consumer culture of modern society encourages people to transfer wants and even wishes into a full-fledged desire. This culture undermines the importance of compromises.

4.9 Fighting Against Compromises In this section, after discussing the main aspects of various types of compromises, I will briefly discuss some major ways of fighting against compromises. The negative aspects of compromises have led people to fight against them. I discuss here a few such attempts to fight compromises and in particular the ‘zero-tolerance’ policy and the ‘slippery slope’ argument. After showing their futility, I will discuss in the last section the policy of complementation, which ought to be used in conjunction with compromises. The zero-tolerance policy A common method of fighting romantic compromises is expressed in the ‘zerotolerance’ policy, which strictly prohibits any type of boundary violation. This policy, which essentially disregards reality, is extremely harsh and difficult to adopt. It is easier to draw clear boundaries than to keep them. Although normative boundaries are supposed to guide our behavior, reality is more complex than what can be prescribed by simply drawing a line between dos and don’ts. Guiding principles should provide general directions, such as ‘drive safely’, rather than specific rules like ‘don’t exceed 100 kilometers per hour’. What constitutes safe driving can vary considerably, depending



on several factors, such as the competence of the driver, the conditions of the road, and how other people drive (Averill et al. (1990): 34). Similarly, compromises vary considerably, depending on personal and contextual features. People use specific rules to help them cope with their chaotic environment, but there is no golden rule to tell them what constitutes an optimal compromise. A more prevailing expression of the zero-tolerance policy is that which isolates us from reality. We often build fences around our boundaries in order to prevent ourselves from being dangerously close to these boundaries. Thus, various religions demand that women adopt a modest appearance, in order to prevent temptation. Such fences, while intended to stop people before they step across the line into forbidden behavior, do more than just protect people from overstepping boundaries; they also prevent them from engaging in pleasant activities that do not cross any normative lines. We should guard against building such high fences and walls since they prevent the healthy beams of the sun from entering the tight space in which we have enclosed ourselves. In certain other cases, fences can have the opposite effect by increasing desire. This accords with the saying that ‘Stolen waters are sweet’, as well as with the successful tactic of ‘Playing hard to get’. The attempt to prevent compromises in one realm often generates compromises in other realms. The opposition to the zero-tolerance approach, which attempts to disallow any type of compromises, is evident in the many types of violations of strict romantic exclusivity. Alongside the individual compromises that each of us needs to make within our particular relationship, there are also some more general changes that have occurred in societal norms and behavior that facilitate romantic compromises. Examples include postponing romantic gratification, accepting declining sexual intensity, reducing the exclusive nature of the romantic relationship by enlarging the scope of activities that are not restricted to the beloved, serial monogamy, and loving more than one person at the same time. The change in social norms regarding such compromises has generated new terminology. Thus, instead of the highly negative terms of ‘adultery’ and ‘betrayal,’ some people use the more neutral term of ‘parallel relationship’. These compromises, which make love less bounded, entail the violation of prevailing normative boundaries and as such might not be suitable for everyone in all circumstances. A major difficulty stemming from such compromises is that they can lower the level of commitment to the primary relationship. This is a highly problematic development as it has been found that members of a less committed couple are more vulnerable to negative partner characteristics than are highly committed members. Less committed individuals lack the motivation to ignore negative partner information. Thus to some extent, certain romantic compromises in themselves can lead to lowering commitment levels. The slippery slope argument Our romantic life is made more complex by the many alternatives available to us. Some of these alternatives are clearly good for us, while others can prove harmful. One of the easiest ways to avoid the harmful alternatives is to steer clear of those situations in


A A RO N B E N - Z E ’ E V

which the bad alternatives are more prone to tempt us. In other words, we can protect ourselves by refusing to go anywhere near the brink of the slippery slope. The desire to avoid choosing a bad alternative and to compromise in general is expressed in the argument that says that once you take the first step on a slippery slope, you are bound to slide all the way down the hill. This argument maintains that once one is in such a situation, taking any course of action is liable to lead to unhappy results. In this view, a relatively small first step inevitably leads to a chain of related events, culminating in a disastrous event. However, I would argue that this slippery slope argument can be fallacious when there are intermediate events that can prevent the agent slipping further or when the agent can avoid such a negative outcome by the strength of her own will. Extreme religious and conservative societies prohibit lustful emotional imagination since they assume that such imagination is highly likely to lead to immoral behavior. However, it is realistically impossible to avoid lustful fantasies. Even Jimmy Carter admitted that, although he was very religious, he had lusted after a woman in his heart. Sometimes, trying to repress a certain image—for example, by imagining that a pink elephant is standing in the corner—simply makes that image even more irrepressible. It is also doubtful whether prohibiting lustful imagination is morally just. Such prohibition assumes the validity of the slippery slope argument. The slippery slope argument is flawed since our lives are full of slopes and hills and avoiding all of them is tantamount to ceasing to live. We must make compromises without giving into the extreme pole. We can take a few steps down the slope without necessarily falling all the way to the bottom. Living involves taking risks, but these can be calculated risks that have certain safety lines attached. Drawing lines is an inevitable everyday activity, which should take into account the given context. Indeed, prevailing moral norms draw more flexible lines concerning the use of imagination. These norms do not prohibit—at least not completely—lustful fantasizing; it is the enactment of such fantasies that can be morally wrong. Morality is highly complex and sensitive to contextual and personal subtleties. The very fact that morality recognizes the presence of an order of priority between our values indicates that morality accepts violation of less important values when this is necessary. Considering minor breaches as morally acceptable recognizes the value of compromises. Contrary to the popular saying, ‘Once a cheater, always a cheater’, human beings are able to moderate their activities while not necessarily slipping down the slippery slope. In contrast to the zero-tolerance approach, a Scottish proverb says: ‘Better bend than break’. Bending, which is a kind of compromise, is the flexibility that enables the ideal to be maintained for a long time. People who refuse to compromise their ideals often abandon them.

4.10 Complementation Once we realize that there is no use in fighting compromise, we should look for ways in which we can not merely cope with compromises, but use them to enhance our



lives. Accommodation and complementation are major means of doing this. In previous sections, I have discussed the process of accommodation in which the agent accommodates to the compromise to the point that it is no longer regarded as a compromise. In this section, I will discuss the process of complementation. Complementation (or integration) involves adding something else to complete something, so that it forms a whole. Complementation is a positive process that adds something of value to something else. In a successful process of complementation, the combination of the two is greater than the sum of the value of their elements (Graham (1998)). Applying the notions of compromise and complementation to the romantic realm, we can say that when one compromises, one is accepting the negative in order to prevent harmful consequences; during complementation the focus is upon further nurturing the relationship. Accordingly, we might say that compromise is part of a Preventative Approach to love, while complementation is part of a Nurturing Approach to love. In reality we need both approaches, and the quality of the relationship is determined by the weight and the overall balance between them both. One meaning of ‘to nurture’ is to promote and sustain the growth and development of someone. Nurturing often refers to how we help a person, usually a child, to grow and develop. In raising our children, for example, we place emphasis on their nurturing capacities, talents, tolerance, and friendships. However, we can also nurture ourselves and our own intimate relationship. At the basis of our Nurturing Approach to romantic love is the conviction that satisfied people, who are able to further develop and flourish within a romantic relationship, are those who are most likely to stay in love. Although romantic love has to do with giving to others, such giving can best be done by a person who is growing and flourishing within the relationship (Ben-Ze’ev & Goussinsky (2008)). The Nurturing Approach adopts a self-validated model of romantic relationship over the other-validated model that prevails in the Preventative Approach. Paying a lot of attention to and investing a lot of resources in the other’s validation of one’s behavior could prevent a crisis in the relationship, but is not sufficient for nurturing the relationship in a more qualitative manner. In order to achieve the latter, one’s own personal flourishing is a crucial aspect. Accordingly, in the Nurturing Approach, intrinsically valuable activities, in which the value of the activity lies in the activity itself, become more important than extrinsically valuable activities, which aim to achieve a certain external goal. An extrinsically valuable activity is always incomplete; as long as the external goal has not been achieved, the activity is incomplete, and the moment the goal has been achieved, the activity is over. Pursuing merely external goals becomes a kind of unfinished business involving a never-ending, unsuccessful struggle to reach the unreachable. Genuine love does not consist of craving after ready-made complete external products, but of ongoing, mutual intrinsically valuable activities. The former, which is self-destructive, may provide immediate superficial pleasure, while the later, which is self-perpetuating, offers ongoing profound satisfaction.


A A RO N B E N - Z E ’ E V

The emphasis upon self-worth is also expressed in the central role played by promoting, rather than preventing, types of behavior. Accordingly, caring has more significance in love than prohibiting various types of sexual practices. In the Nurturing Approach to romantic love, uniqueness is more significant than exclusivity. Uniqueness focuses on nurturing ourselves and others, while exclusiveness entails preventing the other from engaging in certain actions or forms of behavior. In conclusion, compromises cope with basic human predicaments, such as whether to evaluate an event positively or negatively, whether to aspire for the desired possible or accept the disappointing actual, and the difference between the way we feel and the way we actually behave. As these predicaments are intrinsic to human life, there are many circumstances in which compromises are necessary and valuable. We are condemned to make compromises. This is also true in the romantic realm.

References AAVV (2008) V2 Vocabulary Building Dictionary. http://free-online-dictionary-websites.blogspot. com/2008/08/v2-vocabulary-building-dictionary-free.html Averill, J.R., Catlin, G., & Chon, K.K. (1990) Rules of Hope. New York: Springer-Verlag. Bauman, Zygmunt (2003) Liquid Love: On the Frailty of Human Bonds. Cambridge: Polity Press. Benjamin, Martin (1990) Splitting the Difference: Compromise and Integrity in Ethics and Politics. Lawrence: University Press of Kansas. Ben-Ze’ev, Aaron (2000) The Subtlety of Emotions. Cambridge, MA: MIT Press. —— (2004) Love Online: Emotions on the Internet. Cambridge: Cambridge University Press. —— & Goussinsky, Ruhama (2008) In the Name of Love: Romantic Ideology and its Victims. Oxford: Oxford University Press. Biddle, Francis (1957) ‘Necessity of compromise’. In: Integrity and Compromise: Problems of Public and Private Conscience, R.M. MacIver (ed.), pp. 1–8. New York: Harper. Day, T.J. (1991) ‘Moral dilemmas, compromise and compensation’, Philosophy 66: 369–75. Fisher, Helen (2004) Why We Love? New York: Holt. Goldin, M.P. (1979) ‘The nature of compromise: A preliminary inquiry’. In: Compromise in Ethics, Law and Politics, J.R. Pennock & J.W. Chapman (eds), pp. 3–25. New York: New York University Press. Graham, Pauline (1998) ‘Saying “no” to compromise; yes to integration’, Journal of Business Ethics 17: 1007–13. Kayser, K. (1993) When Love Dies: The Process of Marital Disaffection. New York: Guilford. Kipnis, Laura (2003) Against Love: A Polemic. New York: Pantheon. Pennock, Roland J. & Chapman, J.W. (1979) (eds) Compromise in Ethics, Law and Politics. New York: New York University Press. Van Willigenburg, Theodor (2000) ‘Moral compromises, moral integrity and the indeterminacy of value ranking’, Ethical Theory and Moral Practice 3: 385–404.


This page intentionally left blank

5 Values and Emotions Neo-Sentimentalism’s Prospects* Christine Tappolet

5.1 Introduction Neo-sentimentalism is the view, roughly, that to judge that something has an evaluative property is to judge that some emotional response is fitting or appropriate with respect to it. Such an account of value concepts, sometimes also called fitting-attitude (FA) analysis, has made a recent comeback.1 Part of the plausibility of neo-sentimentalism is due to the fact that it is difficult to deny that values and emotional responses, or at least that their concepts, are closely related. It is quite obvious that concepts such as admirable or disgusting are interconnected with the concepts of emotions, such as admiration and disgust. As has been often noted, the main attraction of this approach is that it promises to account for two features of evaluative judgements that are notoriously difficult to combine: their action-guidingness and their cognitive character (Darwall, Gibbard, & Railton (1992);

* I’m grateful to Justin D’Arms, Julien Deonna, Je´roˆme Dokic, Luc Faucher, Kent Hurtig, Daniel Jacobson, Rae Langton, Ste´phane Lemaire, Pascal Ludwig, Colin MacLeod, Olivier Massin, Kevin Mulligan, Ruwen Ogien, Jonas Olson, Christian Piller, Huw Price, Andrew Reisner, Gideon Rosen, Tim Schroeder, Ronald de Sousa, Gopal Sreenivasan, Sarah Stroud, Fabrice Teroni, David Velleman, and Catherine Wilson for questions and discussions. Ancestors of this paper have been presented at the Edinburgh Will and Moral Psychology conference, the Conference for the memory of Nicolas Kaufman at the University of TroisRivie`res, the Markt, Wert, Gefühle Conference of the SFB ‘Moderne’ of the University of Graz, the Congress of the Socie´te´ de Philosopie du Que´bec in Rimouski, the Congress of the Canadian Philosophy Association in Halifax, at the conference on values and value judgements organized by Ste´phane Lemaire and Pascal Ludwig in Rennes, and at the Philosophy Departments of Uppsala and Geneva. 1 Neo-sentimentalism can be traced back to Clarke (1706); Hutcheson (1725); Hume (1740); Brentano (1889); Husserl (1988); Scheler (1913–1916); and Meinong (1917). More recently, versions of it have been advocated by Broad (1930); Brandt (1946); Ewing (1947, 1959); Wiggins (1976, 1987); Chisholm (1981, 1986); Blackburn (1984, 1998); McDowell (1985); Falk (1986); Gibbard (1990); Anderson (1993); Lemos (1994); Tappolet (1995, 2000); Mulligan (1998); Sainsbury (1998); Scanlon (1998); Skorupski (2000); D’Arms and Jacobson (2000a, 2000b); Zimmerman (2001); Helm (2001); Oddie (2005); Danielsson and Olson (2007); Prinz (2007); and Olson (2009). For a historical survey, see Rabinowicz and RnnowRasmussen (2004).



D’Arms and Jacobson (2000a)). Though neo-sentimentalism need not be committed to internalism, it promises to explain the tight connection between evaluative judgement and action, insofar as the invoked responses are related to motivations. It would account for the cognitive character of evaluative judgements because such judgements would be truth-assessable and could possibly be known to be true or false. Moreover, insofar as the responses at stake can be grounded in reasons, neo-sentimentalism would also make room for the intuition that evaluative judgements are sensitive to reasons (D’Arms and Jacobson (2000a)). According to many, neo-sentimentalism would have still another virtue: it would involve no ontological commitment to independent values (Brentano (1889: 60)). The difficulty in assessing neo-sentimentalism is that it allows for a great many different versions. My aim here is to spell out what I take to be its most plausible version and to try and defend it by comparing it to what I take to be its main contender. In the first section, I lay out the main varieties of neo-sentimentalism and argue that one has to distinguish between a normative and a descriptive version. In the next section, I consider the main argument that can be given in favour of the normative version and show that the descriptive version is far from excluded by this argument. After this, I offer two arguments in favour of the descriptive version. The first one turns on the question of normative action explanation, while the second develops the Wrong Kind of Reason Argument. I end with a discussion of the accusation that the kind of account I favour involves vicious circularity. Before I start, I should say that I will consider only judgements that involve concepts such as admirable, disgusting, shameful, and frightening. These concepts are a kind of thick evaluative concepts, which can be called ‘affective concepts’.2 These are the best candidates for neo-sentimentalism, for they wear their response-dependence on their sleeves, to use David Wiggins’ expression (Wiggins (1987)). More general evaluative concepts, such as good or bad, are likely to admit a neo-sentimentalist treatment as well, but given that they share many features with determinable concepts, such as coloured, the story is bound to be more complicated.3

5.2 Two Versions of Neo-Sentimentalism According to neo-sentimentalism, evaluative concepts such as admirable and disgusting are response-dependent, in the sense that they are related to the concepts of specific responses—admiration and disgust in this case.4 Something counts as admirable if and only if admiration is an appropriate or fitting attitude, where this is taken to be a 2 For such concepts, see Anderson (1993) and Tappolet (2004). And see D’Arms & Jacobson (2003) for the claim that these concepts are the best candidates for neo-sentimentalism. This is also Rabinowicz & Rnnow-Rasmussen’s way of framing FA-analyses (2004: 402). 3 For the claim that good is a determinable, see Mulligan (ms); Tappolet (2004); and Oddie (2005). 4 Some, like Wright (1992) and Johnston (2001), only consider dispositional or projective accounts to be response-dependent. For a more liberal take on response-dependence see D’Arms & Jacobson (2000a: n. 20).



conceptual truth. More generally, the relation between evaluative concepts and the corresponding responses can be spelled out as follows (where V is an affective value and E the corresponding attitude): (NS) x is V if and only if x is such that feeling E is appropriate in response to x.5

A question that is crucial for understanding neo-sentimentalism is that of knowing what it is for a response to be appropriate. To start with, however, let me briefly address the question of what kind of states are involved. Neo-sentimentalists all agree that the responses in question are affective, in contrast with judgements, beliefs, or types of actions. Neo-sentimentalist accounts thus form a subclass of fitting-attitude analyses. Even so, concentrating on affective responses leaves many possibilities open—one could think of ‘occurrent, object-laden, affect-laden mental states’ (D’Arms & Jacobson (2000a): 723), but also of emotional dispositions (Prinz (2007)), for instance. When affective concepts are considered, the corresponding responses clearly include states such as disgust, admiration, and fear. Since these are paradigmatic cases of emotions, it makes sense to claim that the unitary type of response at play in neo-sentimentalism is emotion. There are two main ways to understand the concept of appropriateness at stake. The first, which is now standard, is to take this concept to be normative.6 An appropriate emotion is one that satisfies a normative requirement; the emotion ought to be felt, in some sense of ought.7 More precisely, we would have the following claim: (NS-normative) x is V if and only if x is such that feeling E is required with respect to x.

An alternative conception, which has been left mostly unexplored, consists in denying that the concept of appropriateness at stake is normative. There are different ways of spelling out this idea. The suggestion that I would like to make is that the appropriateness of emotions is a matter of representing things as they are. In the relevant sense, appropriate emotions are emotions that are correct from an epistemic point of view.8 (NS-descriptive) x is V if and only if x is such that feeling E is correct in response to x.9

5 One might want to require that the response is only appropriate if one were to contemplate x, and also that the intrinsic features of x be properly grasped (Chisholm (1981, 1986); Lemos (1994); Zimmerman (2001); Bykvist (2009)). 6 ‘Normative’ is used in its narrow sense, which is equivalent to ‘deontic’ and excludes the evaluative. If one takes the normative to encompass both the deontic and the evaluative, this would make for two subpossibilities, one of which being that appropriate is evaluative (see Williams (1971)). Given the circularity that would be involved, this dose not seem a very tempting suggestion. 7 One could also say that the attitude is one that there is reason to have, where the concept of reason is taken to be normative. See Scanlon (1998): 95; Skorupski (2000); and Anderson (1993). 8 The terminology is inspired by Brentano, who claimed that something is good if and only if loving it is correct (‘richtig’) (1889). Another possibility is to claim that being appropriate for an emotion is for it to be justified, while this is not taken to be a normative claim (Mulligan (1998)). The problem with this suggestion is that it is not clear that it can satisfy the Normativity Requirement (see next section). 9 This suggestion is close to Danielsson and Olson’s claim that x is good means that x has properties that provide content-reasons to favour x, where content-reasons for an attitude are reasons for the correctness of the attitude, a notion which they claim is analogous to truth (2007). One difference between our accounts is



The claim is that something is disgusting just if feeling disgust towards this thing were correct from an epistemic point of view—it would represent the thing as it is, evaluatively speaking. This suggestion is grounded in an account of emotions which underlines the numerous analogies between emotions and perceptual experiences. According to the so-called Perceptual Account, emotions are a kind of perception: they represent their objects in certain ways.10 What is specific about emotions is that they represent things as having certain evaluative properties. To use the medieval jargon Anthony Kenny favoured (1963), the emotions’ formal objects are evaluative properties.11 Thus, an emotion of admiration with respect to a friend will be correct just in case the friend really is admirable. This account differs from the normative version of neo-sentimentalism, for at least according to a plausible interpretation to say that an emotion is correct is not yet to make a normative judgement. It simply amounts to saying that such an emotion is one that corresponds to how things are evaluatively speaking. For instance, amusement is correct just if its object is amusing. And this is not, arguably, a normative claim.12 It might help to compare appropriateness with truth. At least according to a number of important conceptions of truth—correspondence theories, deflationary theories, and possibly coherentist theories—to say that a proposition is true is to make a cognitive assessment, but it is not, as such, to make a normative judgement. In particular, it would not amount to saying that the proposition is good in a way, or that it is required. As such, that a proposition corresponds to the facts, for instance, is certainly not a normative fact in itself. It is only insofar as truth is our goal that requirements follow. One might object that truth is a goal that is constitutive of belief, so that normative requirements follow necessarily from the claim that a belief is true. This might well be so. But the suggestion that truth is the constitutive goal of belief can be understood as the claim that truth is a good at which beliefs necessarily aim. So, that they would resist the idea that correctness is a matter of representing correctly (2007: 516). Another difference is that they take correctness to be normative (2007: 512). 10 This account goes back to the moral sense theorists Shaftesbury (1711), Hutcheson (1725), and arguably Hume (1740), and to the turn-of-the-century philosophers Scheler (1913–1916) and Meinong (1917). More recently, see McDowell (1985); Wiggins (1987); de Sousa (1987 and 2002); Tappolet (1995, 2000, and forthcoming); D’Arms & Jacobson (2000a and 2000b); Johnston (2001); Wedgwood (1994, 2001); Prinz (2004, 2006); Deonna (2006); and Do¨ring (2007). 11 According to Kenny, the formal object of a state is the object under that description which must apply to it if it is possible to be in this state with respect to it (1963: 132). He claims that the description of the formal object of an emotion involves a reference to belief: one has to believe that something is dangerous in order to feel fear. In recent times, however, it has become common to claim that the formal object of an emotion is a property. Thus, de Sousa claims that ‘The formal object of fear—the norm defined by fear for its own appropriateness—is the Dangerous’ (2002: 251). I shall leave aside the fact that the intensity of the emotion has to fit the degree of the value in question. See Broad (1954: 293); Tappolet (2000); D’Arms & Jacobson (2000b); and Jones (2004). 12 The difference between this account and D’Arms & Jacobson’s is possibly quite small. They readily embrace the presentation jargon I favour. Thus, they claim that ‘[e]motions present things to us as having certain evaluative features’, where the ‘fittingness of an emotion is like the truth of a belief’ (2000b: 72).



having a true belief would amount to having a belief that meets the requirements set by the constitutive goal of that kind of state. It is a belief that has met its success conditions. However, this claim does not entail that true itself is an evaluative or a normative concept (Horwich (2000)). In the same way, it might well be true that correct representation is a constitutive goal of emotions.13 This goal would ground epistemic norms pertaining to emotions, such as the norm that, all things being equal, we ought to have correct emotions. Given these norms, it would be a good thing for an emotion to be correct; a correct emotion would be one that happens to satisfy the relevant epistemic norms. But this does not entail that correct is a normative concept. If you are looking for a tall person, Anna, who happens to be tall, meets your requirement, but this does not mean that ‘tall’ is itself a normative term. As I formulated it, both versions of neo-sentimentalism are claims about value concepts. However, they are naturally taken to go hand in hand with ontological claims. In contrast with NS-descriptive, NS-normative is naturally taken to aim at ontological simplification: it purports to explain evaluative judgements in terms of norms that apply to emotions.14 This suggests that there being a norm, of a kind to be specified, that requires us to feel shame or admiration with respect to something is what it is to be shameful or admirable. Values themselves could thus be said to be reflections or projections of required emotions. Or else, values could be claimed to be constituted by required emotions. This is quite a different account from NS-descriptive, which has no eliminative or reductive ambitions. Contrary to the NS-normative, NS-descriptive makes the normative requirements on emotions merely derivative. If we ought to have a certain emotional response with respect to something it is because it has a certain value, and it is a constitutive goal of emotions to match evaluative facts. Although it is a claim about concepts, NS-descriptive appears to be incompatible with certain kinds of anti-realism. The choice between the two versions of neo-sentimentalism would seem to amount to choosing whether the priority should be given to norms governing emotions or to evaluative facts. While I shall examine an argument for the normative version in the next section, it is with an eye to motivating the descriptive version.

5.3 The Open Question Argument and the Normativity Requirement The main argument for the claim that the concept appropriate as it appears in the neosentimentalist biconditional is normative consists in an application of G.E. Moore’s Open Question Argument (Wiggins (1987): 187; Darwall, Gibbard, & Railton (1992): 13

I am indebted to Justin D’Arms for reminding me of this point. But see Wedgwood (2009): 516–17 for the claim that ‘fitting-attitude’ equivalences are best understood as stating an identity relation between facts, but no conceptual asymmetry. 14



116–18; D’Arms & Jacobson (2000a): 726–7). As Darwall, Gibbard, and Railton explain, this argument can be read as entailing a constraint on any analysis of evaluative and normative concepts. That constraint is that the analysans ought to maintain the action-guidingness, or more generally the normative force, of the analysandum. What is wrong with an analysis of good in terms of biological fitness, for instance, resides in the fact that biological fitness has no particular normative force; it does not, as such, involve any requirement on what to desire or on what to do. The question whether we ought, other things being equal, to devote ourselves to bringing about biological fitness is wide open. Similarly, it would be wrong to say that something falls under an evaluative concept just if it is such as to cause some particular attitude; for to judge that something causes an attitude simply lacks normative force. From this it is tempting to infer that the analysans should be spelled out in terms of responses that are appropriate in some normative sense. Something would be shameful, for instance, just in case shame ought to be felt with respect to it. Thus, and only thus, would the normativity of the evaluative judgement be preserved by the analysans. This is the train of thought that leads D’Arms and Jacobson to claim that ‘to think a sentiment appropriate in the relevant sense is a normative judgement, of a type yet to be explicated, in favour of feeling it’ ((2000a: 729); Chisholm (1986: 53); and Rabinowicz & RnnowRasmussen (2004: 391)). No doubt that NS-normative satisfies the Normativity Requirement that follows from the open question argument. However, it would be wrong to believe that this is the only way to satisfy this requirement. NS-descriptive also satisfies it. To judge that an emotion is correct is to be committed to the claim that its object really has the corresponding evaluative property. Thus, whatever normative force the evaluative judgement has is implicitly preserved by the analysans. This will seem too circular an account for many. But as I shall argue below, there is reason to believe that the circularity at stake is not vicious. While agreeing that the normative force of the evaluative judgement is preserved, one might worry about the fact that NS-descriptive does nothing to explain or elucidate the normative force of evaluative judgements. Rabinowicz and RnnowRasmussen suggest that what is welcome about an account like NS-Normative is that ‘it removes the air of mystery from the normative “compellingness” of values. There is nothing strange in the prescriptive implications of value ascriptions if value is explicated in deontic terms’ (2004: 391–2). It has to be acknowledged that NS-descriptive does not attempt to throw any light on the normative force of value ascriptions. It just takes it as given. As will become apparent in the next section, however, it is not clear that the way NS-normative attempts to explain the normativity of evaluative judgements is in fact better off. The first argument in favour of NS-descriptive, to which I now turn, concerns the normative character of action explanation.



5.4 Values and Normative Action Explanation We have seen that NS-normative satisfies the Normativity Requirement; evaluative judgements are claimed to be judgements stating that certain emotional responses are required. As a result, NS-normative can make room for the action-guidingness of evaluative judgements, that is, for the fact that evaluative judgements involve claims about what we have reason to do. This is so because many emotional responses are intimately connected to motivational states and actions. However, if the analysans preserves the action-guidingness, it is only indirectly, via the relation between emotions, motivational states, and action. It is only to the extent that the emotional response involves a motivational state that a requirement on a response comes with a requirement on motivation and action.15 The problem is that this account of action-guidingness excludes what would seem to be an important kind of normative action explanations, namely explanation couched in evaluative terms instead of normative requirements on emotions. Suppose I meet a brown bear in the woods. If I ought to curl on the ground, it is because the bear is fearsome, a feature that is plausibly taken to supervene on dangerousness.16 If I ought to do this, it is not because there is a norm that requires me to feel fear, so that, given that fear involves a desire to escape what one is afraid of, and curling is the best way to do so, it follows that curling is required. The reason for curling has to do with the thing I am afraid of and its properties, and not with whatever feelings cum motivational states are required. Indeed, it would seem that the feelings are required because the bear is fearsome.17 If some time later, when I’ve reached safety, I am asked why I curled, a natural answer would be ‘because there was a fearsome bear’. Though I could also reply that I was feeling fear, it would be odd if I said ‘I curled because fear was required and that means that a desire to act the way I did was required’. Or, to switch to another example, suppose I feel shame because I told a lie to a friend. What would have explained why I should have refrained from telling the lie is not that shame and some related desire were required. The reason that speaks against telling the lie has to do with the lie itself. Its being a shameful thing is a reason to refrain from it. The reason is not that I would violate some norm regulating my feelings. The point, it seems, is that values give us reasons to act. This is a controversial claim and I will come back to it. But in any case, there is a problem insofar as NS-normative is committed to the claim that reasons for actions are based on normative requirements on an emotional response involving a motivational state. The problem with NS-normative appears to be of the same kind as the problem Talbot Brewer (2002) identifies in his argument against reason internalism. On such 15

See D’Arms & Jacobson (1994: 762). One might want to add that this is only part of the story; on an internalist account, reasons depend on motivational states, such as a standing concern for one’s safety. 17 This is at least what a realist would reply to Euthyphro’s question. See also Johnston’s missing explanation argument (1991). 16



accounts, ‘the justificatory reason one might have to avoid cruel actions is not the fact that the actions would be cruel but rather the fact that one is disposed to count the actions as cruel ’ (2002: 450). The problem is that this involves a reversal of the ‘direction of gaze’ that is appropriate to the sound deliberative search for reasons: ‘When we are in search of such reasons, we generally do not and ought not to look inward at our dispositions to evaluate actions in various ways, but rather outward at the values we are disposed to find in proposed actions or their expected outcome’ (ibid.). In the same way, it would consist in a reversal of the ‘direction of gaze’ to explain an action in terms of our emotional responses and the norms that apply to them, instead of looking to the world and its evaluative features. The advocate of NS-normative is likely to object that her view has been misrepresented. In any case, she can reply that what grounds the requirement to feel the emotional responses is some non-evaluative, and in all likelihood, natural feature of its intentional object (Scanlon (1998); Dancy (2000); Olson (2006)). Thus, it would be something about the object, and not a feature of the emotional responses, that gives us reasons to have an emotional reaction involving a motivational state. One could also add that the non-evaluative features ground two different normative requirements: one bearing on the emotion and one on action (Scanlon (1998): 95–100).18 Thus, one would avoid making normative action explanations indirect; the explanation would not take us to the action by moving from the required emotion to the motivation involved in the emotion. The fact that the bear is about to attack would be both the reason why one ought to feel fear and the reason one ought to curl into a ball. The problem, again, is that this view excludes normative explanations in terms of values. Some, following Thomas Scanlon, will be happy with the conclusion that values do not give reasons for action.19 However, it has to be acknowledged that this view consists in an important revision of ordinary thought. We usually take it that the fact that something has such and such evaluative property is at least part of what explains why we have to do certain things. This is particularly obvious when one thinks of thick evaluative concepts such as cruel or generous. It is also plausible in the case of the subclass of thick concepts I have called ‘affective concepts’. That an action is shameful is naturally taken to be a reason to refrain from performing it, while its being admirable can be taken to be a reason to perform it. It might be objected that whether or not this reflects ordinary thought, the view that values give us reason is misguided. The fact that the bear might injure and kill you gives you reasons to fear it and to curl. Does the fact that the bear is fearsome give an additional reason to fear it and to curl? 20 Certainly not, it will be claimed.


This was suggested by Justin D’Arms. Also see Ewing (1947, 1959); Scanlon (1998: 95–100); Dancy (2000); and Olson (2006). For the contrary view, see for instance Crisp (2005). 20 Thanks to Jonas Olson for pressing this point. 19



The question is whether natural facts can give you reasons, or whether when we refer to a natural fact and claim that it is a reason, we tacitly refer to some evaluative facts. As far as I can see, the fact that the bear might injure and kill you is a reason to do anything only insofar as being injured or killed is something undesirable and bad. The natural facts ground and partly constitute its negative value, but as such they are insufficient to explain why you ought to do anything. Note that a reason to think that reference to values is indispensable to explain why we ought to do certain things is that there are cases in which we have reasons for action, and are aware of these reasons, although we have no clue as to the underlying natural properties. When experiencing shame at a course of action, we can sense that what we consider doing is shameful, but we might have little insight into what makes it so. What we will say, then, is that the reason we avoided this option is that we thought that it was a shameful thing to do. Let us turn to another problem with NS-normative.

5.5 The Wrong Kind of Reason Objection The Wrong Kind of Reason Objection aims at criticizing all current versions of neosentimentalism (D’Arms & Jacobson (2000a); Rabinowicz & Rnnow-Rasmussen (2004)). It starts with the observation that emotions can be assessed in terms of quite different dimensions. They can be claimed to be morally, aesthetically, or prudentially adequate or inadequate, for instance. The second step is to argue on the basis of examples that some of these considerations do not bear on whether the object of the emotion has the corresponding evaluative property. It might well be morally objectionable to be amused by a cruel joke, but this does not entail that the joke is not amusing. Or it might be ill-advised to envy a rich friend’s possessions, though they are genuinely enviable (D’Arms & Jacobson (2000a): 731). As D’Arms and Jacobson write, ‘only certain good reasons for or against having a response bear on the associated evaluative judgement’ (ibid.). Their claim, then, is that a satisfactory account has to ‘offer [ . . . ] resources to differentiate (and hence to preclude conflating) moral and prudential reasons for feeling a sentiment [ . . . ] from reasons bearing on whether [x is V]. Hence, until the relevant notion of appropriateness is specified, the theory is incomplete’ (2000a: 732). In other words, a neo-sentimentalist has to give an account of the notion of appropriateness that allows us to distinguish considerations that are relevant from those that are not with respect to the question whether something has the corresponding value. D’Arms and Jacobson discuss the accounts of the major contemporary neo-sentimentalists—John McDowell, Simon Blackburn, David Wiggins, and Alan Gibbard—and conclude that they all fail what could be called the ‘Conflation Test’. As far as I can see, the account I favour passes the test, but, somewhat ironically, it is not clear that D’Arms and Jacobson’s own account passes it. More precisely, it is not clear that an account that analyses evaluative judgements in terms of normative



requirements on emotions can avoid the Wrong Kind of Reason Objection—I shall leave it open whether this is really the account D’Arms and Jacobson embrace. First, it is quite obvious that NS-descriptive passes the test. Since appropriateness is not considered to be a normative feature, there is no risk of conflating it with moral or prudential considerations. Thus, there should be no risk of conflating the reasons to believe that an emotion is appropriate or inappropriate, in the given sense, with moral or prudential reasons for the emotion. That a joke is cruel and hence the amusement morally objectionable, for instance, has no bearing on whether the amusement you experience fits the evaluative fact. The latter depends strictly on whether the joke is amusing or not. It might be thought that this is not sufficient to meet the Wrong Kind of Reason Objection. What we need, it might be claimed, is an account which tells us whether a consideration is one that is relevant or not with respect to a response’s appropriateness. Thus, when discussing David Wiggins’ account, D’Arms and Jacobson complain that this account fails ‘to point out something about emotions such as shame which we can use in deciding when they are and are not appropriate’ (2000a: 736). They agree that part of the reply is that (to switch to the case of amusement) ‘there is no way to specify what is funny, for instance, except by reference to amusement’ (ibid.). But they claim that more is needed: what we would need to know is what sorts of consideration about amusement can be appealed to in order to determine whether something is funny or not. According to them, the conflation problem looms until we know what the relevant class of reasons is. Compare the following considerations: that being amused by this joke is morally objectionable, that your sister is the target of the joke, or that you have smoked pot. It is natural to think that the two last considerations, but not the first, count against the amusement being appropriate in the relevant sense. Some considerations refer to circumstances that by analogy to the perceptual case can be said to count as defeaters for the emotions, in the sense that they indicate that the emotion in question is likely to fail to present things as they are, while others mention facts that are irrelevant. Thus, the question D’Arms and Jacobson ask is what the principled way of distinguishing between these two kinds of considerations would be. I think that the correct reply to D’Arms and Jacobson is to say that one should not put too heavy a burden on the neo-sentimentalist’s shoulders. Quite generally, most of the truths about what counts as a defeater cannot be known a priori. Consider again the perceptual case. To know that looking at things in the dark tends to interfere with our colour perception or that things plunged into water appear to have a different shape compared to their real one, it is not sufficient to learn colour or shape concepts. What we know a priori is that looking red or looking square is correct on condition that no defeater interferes. In the same way, the neo-sentimentalist can claim that what we know a priori is that being amused is appropriate on condition that no defeater interferes with the amusement. But she can add that when we acquire an evaluative concept, we do not learn the list of circumstances that are likely to make our responses inappropriate.



Indeed, given the open-ended character of such lists, this is not something that could be known a priori (Pettit (1991): 603). As Hume would put it, it is experience that teaches us what circumstances are likely to interfere with our responses (see Hume (1740): III, 3, i). It is experience that teaches us that looking at things in the dark tends to interfere with our colour perception or that things plunged into water appear to have a different shape compared to their real one. Again, it is experience that teaches us that having smoked pot or being the target of a joke is not particularly helpful in assessing whether something is genuinely amusing. An important insight of recent discussions of this question is that in order to determine what counts as a defeater, we need to take into account not only our own experiences at different times and in different conditions, but also the experiences of other persons. Thus, when we want to determine whether or not something should count as a defeater, we have to look to our shared practice of discounting certain conditions as likely to interfere with our responses, a practice that aims at making sense of intrapersonal, but also interpersonal discrepancies (Pettit (1991): 600–1). What about the NS-normative? As it stands, it is certainly threatened by the Wrong Kind of Reason Objection, for it does not exclude moral or prudential considerations. The obvious move is to specify the kind of norms at stake. If one wants to resist the claim that it is because x is V that the emotional response is required, the only possibility is to claim that there is a special kind of norm, to be specified, that avoids the conflation problem. It seems that there are two ways to go here. One is to claim that the norm in question is sui generis.21 The problem with this is that it would make the biconditional quite unattractive. The analysis would appeal to a concept that seems much more obscure than our familiar evaluative concepts. Unless we are told more about this sui generis norm—how do we know that it applies, and how does it relate to other sorts of norms in case of conflict?—we are left with nothing but an empty promise. The other possibility is to deny that the norm is sui generis. However, as long as the norm in question is not specified, it will be quite unclear whether this is merely wishful thinking or not.22 In any case, what should be underlined is that an account of the norms in question is in order. Note that this will also involve taking a stance with respect to the ontological status of norms. That means that there might be an ontological price to pay, but as things stand, its cost is not known.


See Ewing (1947) and Danielsson & Olson (2007). This might also be Brentano’s view. In any case, he would reject NS-descriptive, since he denies that correctness is a matter of identity between the content of the attitude of love and hate and something which is external to the mind (1889: 60). 22 It is not clear where D’Arms & Jacobson stand. They write: ‘To call a sentiment appropriate in this sense is to give it a specific and limited form of endorsement, which is neither a judgment of rightness, prudence, or warrant, nor an all-in endorsement of the sentiment as what to feel’ (2000a: 746). Since they go on to claim that to judge an emotion to be fitting constitutes taking its object to have the evaluative property, it might well be that they endorse NS-descriptive. But they do not adopt this view explicitly.



5.6 Circularity Threats The upshot so far is that NS-descriptive is better placed than NS-normative. But few will be tempted by NS-descriptive. What appear to be obvious and decisive objections threaten it. The aim of this section is to try and dispel the main ones. The most important worry is that NS-descriptive is much too circular to be of interest (Blackburn (1998); Sosa (2001)). In contrast with NS-normative, NS-descriptive avoids Blackburn’s Charybdis (1998: 108), namely, the danger of incorporating normative concepts into the analysans. And it also avoids Blackburn’s Scylla, that is, the danger of going naturalistic or empirical, which is in fact simply the danger of not meeting the Normativity Requirement. But this might well seem a meagre consolation. The reason is that it might seem that the biconditional merely claims that something has an evaluative property just in case it really has the evaluative property in question. Quite true, but quite unhelpful. This is to forget that NS-descriptive claims that there is a relation between evaluative concepts and emotion concepts. As David Wiggins noted in an early defence of neo-sentimentalism, the aim is to ‘elucidate the concept of value by displaying its actual involvement with the sentiments. One would not [ . . . ] have sufficiently elucidated what value is without that detour’ ((1987: 189); see also Pettit (1991): 604). But why make that detour? To paraphrase Wiggins, the important point is that when we try to find out whether something is admirable or shameful, for instance, there is nothing more fundamental to appeal to than our responses of admiration and shame. What this means is that the biconditional makes an epistemic point. Clearly, the expertise in affective concepts involves the ability to feel the relevant emotions. The biconditional points toward the epistemic indispensability of our emotional responses.23 But how can it be claimed that emotions are epistemically indispensable?24 Given NS-descriptive, it might seem that in order to make evaluative judgements, it is sufficient to know that the relevant emotion represents things correctly, something one can know without ever having experienced any emotion. The reply to this objection consists in explaining why NS-descriptive is true. NS-descriptive does not aim to give the application conditions of evaluative concepts—it would be quite hopeless for such a task. Instead, it aims at stating what a theorist can infer from our evaluative practice. The claim is that the a priori relation between evaluative and emotional concepts expressed in the biconditional follows from the way these concepts have been formed and are presently used. Whatever way the details of this story are filled out, it is plausible that the fact that we have such concepts is intimately related to our

23 See Johnston (2001): 181: ‘Seeing the utterly specific ways in which a situation, animal or person is appealing or repellent requires an appropriate affective engagement with the situation, animal or person. Absence of appropriate affect makes us affect blind.’ 24 I owe this question to Sarah Stroud.



emotional capacities.25 More precisely, it seems plausible that primary or canonical attributions of evaluative concepts of the kind I am considering here are done on the basis of our emotional responses (Pettit (1991): 600ff.). When something amuses me, I am inclined to think that the thing in question is amusing. And it would seem that our practice with respect to such concepts is such that if I have no reason to believe that I got things wrong, then I am justified in thinking the thing amusing. Though fallible, our emotional responses would thus ground our evaluative judgements. Understood this way, the biconditional aims at explaining the possession conditions of evaluative concepts, something which is claimed to involve experiencing emotional responses and grounding one’s evaluative judgements on these responses. There is another charge of circularity that concerns the relation of emotions to evaluative judgements. In its strongest version, the objection claims that emotions are evaluative judgements. No wonder, then, that the biconditional is true. It is about as interesting as the claim that x is a proton if and only if the judgement that x is a proton is correct. This would be viciously circular, since ‘is a proton’ appears on both sides of the biconditional. And this is so even if one takes the analysis to state possession conditions instead of application conditions. It is true that the concepts used by the theorist need not be possessed or used by the agent whose concept possession is explained (Menzies (1998)). However, the problem is that the concept in question appears within the content of a mental state. To explain the possession of the concept, one would have appealed to a state whose content involves the concept at stake. Moreover, if emotions are evaluative judgements, the epistemic claim is also threatened. It would amount to the claim that what justifies an evaluative judgement is the evaluative judgement itself. Whatever one’s favourite epistemology, self-justification of this kind is a non-starter. One way to go here is to bite the bullet and claim that there is no way of understanding what the emotional response is independently of what the matching evaluative concept is and vice versa, that the evaluative concept has to be elucidated in terms of the corresponding response (Wiggins (1987)). I think that this is part of the story. However, there is something that should be added. To feel an emotion like amusement, fear, or disgust, it is not necessary to possess the corresponding evaluative concepts. It’s quite wrong to think that to be amused by something is to judge, believe, or even think or imagine that the thing in question is amusing. Quite generally, as cases of so-called recalcitrant emotions show, emotions are not and do not necessarily involve evaluative judgements. Actually, there are reasons to think that they involve a pre-judgemental or non-conceptual representation of values.26 This is not the place 25 My preference goes to Philip Pettit’s ‘ethnocentric’ genealogy: we start with our various emotional responses to things, which we thus tend to find similar and we end with our evaluative concepts because, in order to make sense of the differences in our responses, we assume that certain conditions interfere with our responses (Pettit (1991); see also Wiggins (1987)). 26 See Tappolet (1995, 2000, and forthcoming); Charland (1995); Johnston (2001); D’Arms & Jacobson (2003); and Prinz (2007). See also Griffiths (1997) for the related claim that certain emotions are cognitively encapsulated.



to give the full argument for this claim. But if it is true, then the biconditional need not involve vicious circularity: the evaluative judgement is not analysed in terms of a mental state that involves the corresponding evaluative concepts in its content.27 Even if it is true that in order to identify an emotion, it is necessary to state which evaluative property it is supposed to track—its formal object—experiencing the emotion does not amount to applying this very concept. A last worry is that by stipulating that the emotion’s correction conditions are satisfied, NS-descriptive is trivial. It would not be better than the alleged analysis of red in terms of what looks red in whatever conditions it takes to ensure that something looks red when it is red. But this would seem to be entirely trivial (Wright (1992)). Philip Pettit’s reply to this challenge (1991, 1998) in the case of colours is that the conditions in question have to be specified as those that are fit to survive the practice of discounting certain responses in an attempt to make sense of intra- and interpersonal discrepancies, assuming colour stability. According to this suggestion, something is red if and only if it looks red in conditions that are fit to survive our discounting practice (1998). In the same way, then, one can suggest that NS-descriptive should be amended in the following claim: (NS-descriptive-amended) x is V if and only if x is such that E in response to x is felt in conditions that are fit to survive our discounting practice.

The question is how to understand what it is for a condition to be fit for surviving our discounting practice. In the case of colours or shapes, it is natural to assume that the practice in question aims at discovering an objective reality. It differs thus from the practice underlying ‘U’ as used by the Sloanes, a practice which depends on a group of people having ‘an authoritative, dictating role in regard to the concept’ (Pettit (1991): 611). But even if we take a practice to aim at knowledge in a given area, and we also assume that there is a reality out there to be known, there is no guarantee that the conditions which are robust enough to pass the test of time are the ones in which things are seen as they are. At most, we can hope that the conditions which pass the test are the ones in which there are no distortions. And this brings us back to NS-descriptive, for the conditions we hope will survive our discounting practice are the ones in which we get things right, so that these are the ones which have to figure in the analysans. Understood this way, the amended version is equivalent to NS-descriptive. Note again that it would be wrong to ask for more. As I argued above, most of our knowledge of what counts as unfavourable conditions is not a priori. That means that NS-descriptive only gives the general framework of what has to be done to determine if something has an evaluative property of the kind considered here. It reminds us that we have to start with our affective responses. This may seem a frustratingly small step, but it is better than nothing.


See Peacocke (1992): 89 for the same type of suggestion with respect to perceptual concepts.



5.7 Conclusion The upshot is that NS-descriptive has to be preferred to NS-normative. NS-descriptive easily satisfies the Normativity Requirement. It does better than NS-normative with respect to the Wrong Kind of Reason Objection and contrary to NS-normative it allows for straightforward normative action explanation. Moreover, understood properly the account avoids vicious circularity. Given that it bears testimony to the epistemic indispensability of our emotional responses, one could dub it ‘epistemic neo-sentimentalism’. What about the ontological commitments of this approach? In fact, it leaves the question of the ontological status of values quite open. We have seen that it appears to rule out certain kinds of value anti-realism. Epistemic neo-sentimentalism is naturally read as claiming that an emotion with respect to something is appropriate or correct because the thing in question has some mind-independent evaluative property. It is compatible with robust value realism. But epistemic neo-sentimentalism does not entail that there are objective values. The story could be true and there could still be no value out there. Epistemic neo-sentimentalism is thus consistent with an error theory (Mackie (1977)). In addition to this, an argument would be needed to show that our evaluative concepts are used in the way epistemic neo-sentimentalism claims. I believe this task is not impossible, but note that as far as I can see, it has to be done piecemeal, concept by concept, for it might turn out that some evaluative concepts are underlined by a practice of discovery, while the practice corresponding to the others is one of invention. If what I have called ‘epistemic neo-sentimentalism’ is on the right track, one could conclude that the recognition of evaluative facts involved in emotions could indeed have an impact on our action. This is so when the emotion itself comes with a motivational state.28 In such cases, emotions would have the dual role of tracking values and motivating us accordingly. Thus, epistemic neo-sentimentalism, as expressed in NS-descriptive, preserves the main attractions of neo-sentimentalist accounts.

References Anderson, Elizabeth (1993) Values in Ethics and Economics. Cambridge, Mass.: Harvard University Press. Blackburn, Simon (1984) Spreading the Word. Oxford: Oxford University Press. —— (1998) Ruling Passions. Oxford: Clarendon Press. Brentano, Franz Clemens (1889) The Origin of our Knowledge of Right and Wrong. Trans. Roderick M. Chisholm and Elisabeth H. Schneewind. London: Routledge & Kegan Paul, 1969. Brandt, Richard B. (1946) ‘Moral Valuation’, Ethics 56: 106–21.

28 In fact, the relation between emotions and motivation is more complicated than usually supposed. See Tappolet (2009).



Brewer, Talbot (2002) ‘The Real Problem with Internalism about Reasons’, Canadian Journal of Philosophy 32(4): 443–73. Broad, C.D. (1930) Five Types of Ethical Theory. London: Routledge & Kegan Paul. Broad, C.D. (1954) ‘Emotion and Sentiment’. In: Broad’s Critical Essays in Moral Philosophy, H.D. Lewis (ed.), pp. 283–301. London: Allen & Unwin, 1971. Bykvist, Krister (2009) ‘No good fit: Why the fitting attitude analysis fails’, Mind 118: 1–30. Charland, Louis C. (1995) ‘Feeling and Representing: Computational Theory and the Modularity of Affect’, Synthese 105: 273–301. Chisholm, Roderick M. (1981) ‘Defining Intrinsic Value’, Analysis 41: 99–100. —— (1986) Brentano and Intrinsic Value. Cambridge: Cambridge University Press. Clarke, Samuel (1706) A Discourse Concerning the Unchangeable Obligations of Natural Religion. In: The British Moralists, D.D. Raphael (ed.). Oxford: Oxford University Press, 1969. Crisp, Roger (2005) ‘Value, Reason and the Structure of Justification: How to Avoid Passing the Buck’, Analysis 65(1): 80–5. Dancy, Jonathan (2000) ‘Should we Pass the Buck?’. In: Philosophy The Good, the True and the Beautiful, Anthony O’Hear (ed.), pp. 159–73. Cambridge: Cambridge University Press. Danielsson, Sven & Olson, Jonas (2007) ‘Brentano and the Buck-Passers’, Mind 115: 511–22. D’Arms, Justin & Jacobson, Daniel (1994) ‘Expressivism, Morality, and the Emotions’, Ethics 104: 739–63. —— (2000a) ‘Sentiment and Value’, Ethics 110: 722–48. —— (2000b) ‘The Moralistic Fallacy: On the “Appropriateness” of Emotions’, Philosophy and Phenomenological Research 61: 65–90. —— (2003) ‘The Significance of Recalcitrant Emotions: Or Anti-QuasiJudgmentalism’, Proceedings of the Royal Institute of Philosophy, supp. vol. 8: 127–46. Darwall, Stephen, Gibbard, Allan, & Railton, Peter (1992) ‘Toward Fin de sie`cle Ethics: Some Trends’, The Philosophical Review 101: 115–89. de Sousa, Ronald (1987) The Rationality of Emotions. Cambridge, Mass.: MIT Press. —— (2002) ‘Emotional Truth’, Proceedings of the Aristotelian Society, supp. vol. 76: 247–63. Deonna, Julien (2006) ‘Emotion, Perception and Perspective’, Dialectica 60: 29–46. Do¨ring, Sabine (2007) ‘Affective Perception and Rational Motivation’, Dialectica 61: 363–94. Ewing, A.C. (1947) The Definition of Good. London: Hyperion Press. —— (1959) Second Thoughts in Moral Philosophy. London: Routledge & Kegan Paul. Falk, W.D. (1986) ‘Fact, Value, and Nonnatural Predication’. In his Ought, Reasons and Morality, pp. 99–122. Ithaca, N.Y.: Cornell University Press. Gibbard, Allan (1990) Wise Choices, Apt Feelings. Cambridge, Mass.: Harvard University Press. Goldie, Peter (2001) The Emotions: A Philosophical Exploration. Oxford: Oxford University Press. Griffiths, Paul E. (1997) What Emotions Really Are. Chicago: Chicago University Press. Helm, Bennett (2001) Emotional Reason: Deliberation, Motivation and the Nature of Value. Cambridge: Cambridge University Press. Horwich, Paul (2000) ‘Norms of Truth and Meaning’. In: Philosophy, the Good, the True and the Beautiful, Anthony O’Hear (ed.), pp. 19–34. Cambridge: Cambridge University Press. Hume, David (1740) A Treatise of Human Nature, second edition, L.A. Selby-Bigge (ed.). Oxford: Oxford University Press. Husserl, Edmund (1988) Vorlesungen über Ethik und Wertlehre, 1908–1914. Ullrich Melle (ed.) (Husserliana, vol. xxviii). Dordrecht: Kluwer Academic.



Hutcheson, Francis (1725) An Inquiry into the Original of our Ideas of Beauty and Virtue. Hildensheim: Georg Olms, 1971. Johnston, Mark (1989) ‘Dispositional Theories of Values’, Proceedings of the Aristotelian Society, supp. vol. 63: 139–74. —— (1991) ‘Explanation, Response-Dependence and Judgement-Dependence’, Response-Dependent Concepts, Working Papers in Philosophy 1: 122–83. —— (2001) ‘The Authority of Affect’, Philosophy and Phenomenological Research 53: 181–214. Jones, Karen (2004) ‘Emotion, Weakness of Will, and the Normative Conception of Agency’. In: Philosophy and the Emotions, A. Hatzimoyis (ed.), pp. 181–200. Cambridge: Cambridge University Press. Kenny, Anthony (1963) Action, Emotion and the Will. London: Routledge & Kegan Paul. Lemos, Noah M. (1994) Intrinsic Value: Concept and Warrant. Cambridge: Cambridge University Press. McDowell, John (1985) ‘Values and Secondary Qualities’. In: Morality and Objectivity: A Tribute to John Mackie, Ted Honderich (ed.), pp.110–29. London: Routledge & Kegan Paul. Mackie, John L. (1977) Ethics. Inventing Right and Wrong. London: Penguin. Meinong, Alexius (1917) On Emotional Presentation, trans. with introduction by Marie-Luise Schubert Kalsi. Evanston, Ill.: Northwestern University Press. Menzies, Peter (1998) ‘Possibility and Conceivability: A Response-Dependent Account of Their Connections’, European Review of Philosophy 3: 255–77. —— & Pettit, Philip (1993) ‘Found: the Missing Explanation’, Analysis 53: 100–9. Mulligan, Kevin (1998) ‘From Appropriate Emotions to Values’, The Monist 81: 161–88. —— (manuscript) ‘Wie verhalten sich Normen und Werte zueinander?’. Oddie, Graham (2005) Value, Reality, and Desire. New York: Oxford University Press. Olson, Jonas (2006) ‘G.E. Moore on Goodness and Reasons’, Australasian Journal of Philosophy 84 (4): 525–34. —— (2009) ‘Fitting Attitude Analyses of Values and the Partiality Challenge’, Ethical Theory and Moral Practice 12: 365–78. Peacocke, Christopher (1992) A Study of Concepts, Cambridge, Mass.: MIT Press. Pettit, Philip (1991) ‘Realism and Response-Dependence’, Mind 100: 587–626. —— (1998) ‘Terms, Things and Response-Dependence’, European Review of Philosophy 3: 55–66. Prinz, J. Jesse (2004) Gut Reactions: A Perceptual Theory of Emotion. New York: Oxford University Press. —— (2006) ‘Is Emotion a Form of Perception?’. In: The Modularity of Emotions, The Canadian Journal of Philosophy, Luc Faucher & Christine Tappolet (eds), supp. vol. 32: 137–60. —— (2007) The Emotional Construction of Morals. New York: Oxford University Press. Rabinowicz, Wlodek & Rnnow-Rasmussen, Toni (2004) ‘The Strike of the Demon: on Fitting Pro-attitudes and Value’, Ethics 114(3): 391–423. Sainsbury, R. Mark (1998) ‘Projections and Relations’, The Monist 81: 133–60. Scanlon, T.M. (1998) What we Owe to Each Other. Cambridge, Mass.: Harvard University Press. Scheler, Max (1913-1916) Formalism in Ethics and the Non-Formal Ethics of Values, trans. Manfred S. Frings and Roger L. Funk (fifth revised edition). Evanston, Ill.: Northwestern University Press. Shaftesbury, Cooper, A.A. Earl of (1711) An Inquiry concerning Virtue, or Merit, in Characteristics of Men, Manners, Opinions, Times. 1711, revised edn in 1714. New York: Bobbs-Merrill, 1964.



Skorupski, John (2000) ‘Irrealist Cognitivism’. In: Normativity, Jonathan Dancy (ed.), pp. 436–59. Oxford: Basil Blackwell. Sosa, David (2001) ‘Pathetic Ethics’. In: Objectivity in Law and Morals, Brian Leiter (ed.), pp. 287–330. Cambridge: Cambridge University Press. Tappolet, Christine (1995) ‘Les e´motions et les concepts axiologiques’. In: La Couleur des pense´es, Raisons Pratiques 6, Patricia Paperman & Ruwen Ogien (eds), pp. 237–57. —— (2000) Emotions et Valeurs. Paris: Presses Universitaires de France. —— (2004) ‘Through Thick and Thin: Good and its Determinates’, Dialectica 58(2): 207–20. —— (2009) ‘Emotions, Action, and Motivation: the Case of Fear’. In: The Oxford Handbook of Philosophy of Emotion, Peter Goldie (ed.), pp. 325–45. Oxford: Oxford University Press. —— (forthcoming) ‘Emotions, perceptions, and emotional illusions’. In: Perceptual Illusions. Philosophical and Psychological Essays, Clotilde Calabi (ed.), Houndmills: Palgrave-Macmillan. Wedgwood, Ralph (1994) Principle and Sentiment. An Essay in Moral Epistemology. PhD Thesis (Cornell University). —— (2001) ‘Sensing Values?’, Philosophy and Phenomenological Research 53: 215–23. —— (2009) ‘The ‘Good’ and the ‘Right’ Revisited’, Philosophical Perspective 23: 499–519. Wiggins, David (1976) ‘Truth, Invention and the Meaning of Life’. In his Needs, Values, Truth, pp. 87–138. Oxford: Basil Blackwell, 1987. —— (1987) ‘A Sensible Subjectivism?’. In his Needs, Values, Truth, pp. 185–214. Oxford: Basil Blackwell, 1987. Williams, Bernard (1971) ‘Morality and the Emotions’. In his Problems of the Self, pp. 207–29. Cambridge: Cambridge University Press, 1973. Wright, Crispin (1992) Truth and Objectivity. Cambridge, Mass.: Harvard University Press. Zimmerman, Michael (2001) The Nature of Intrinsic Value. Lanham: Rowman & Littlefield.

6 Emotions, Perceptions, and Reasons Michael S. Brady

One of the most interesting developments in the theory of emotions has been the perceptual model. Supporters of this model maintain that emotional experiences represent evaluative properties, in much the same way that perceptual experiences represent non-evaluative properties. Because of this, supporters of the perceptual model hold that emotional experiences can play a similar epistemic role to that played by perceptual experiences: they can constitute justifying reasons with respect to beliefs or judgements. In this paper I want to raise some doubts about this picture of the epistemic role of emotional experiences, doubts which are grounded in a significant disanalogy between emotions and perceptions. In the first section of the paper I outline the perceptual model of emotion. In the second and third parts I raise some problems for the account, suggesting that emotions are not themselves reasons for evaluative judgements. And in the final part of the paper I respond to a possible objection, and explain how emotional experiences can still play a role in the justification of our evaluative judgements even if they are not reasons for those judgements.

6.1 The perceptual model maintains that emotions involve, or at least are analogous to, perceptions of values.1 On this account, fear involves perceiving something as dangerous, anger involves perceiving something as insulting, shame involves perceiving something as shameful, and so on for other central cases of emotion. Proponents of perceptual models support their views by listing a number of important ways in which emotional reactions are similar to sensory perceptions, where sensory perceptions are taken to be the paradigm of perceptual experiences.2 Thus, they claim that both 1 Supporters of perceptual models of the emotions include Elgin (1996): ch. 5 and (2008); Do¨ring (2003); Johnston (2001); de Sousa (1987); Prinz (2004); Roberts (2003); Tappolet (2000 and forthcoming); Zagzebski (2004). 2 If we think that sensory perceptions exhaust the class of perceptual experiences, then we will deny that emotions literally involve perceptions of value. But even if we accept this and thus reject a more liberal understanding of perception, we can still call the relevant models of emotions ‘perceptual models’ if they


M I C H A E L S . B R A DY

emotions and sensory perceptions possess phenomenal properties; both are essentially perspectival; both are ‘passive’ responses; both are (typically) caused by features of the subject’s environment; and both can diverge or come apart from their associated judgements or beliefs, and as a result are not to be thought of as involving judgements or beliefs.3 Instead, perceptual experiences and emotional responses are states which are weaker than judgement or belief. There is a further point of similarity claimed by supporters of the perceptual model, which is that perceptual and emotional experiences are both representational states. For according to a standard position in the philosophy of perception, a perceptual experience is a conscious mental state with intentional or representational content. To say that experiences have representational content is to say that they inform us or tell us about something. On the standard view, perceptual experiences have ‘naı¨ve content’: they tell us about and thus represent things in the external world, rather than items which are intermediate between our experiences and the world, such as sense data.4 Thus, as Tim Crane writes, ‘An intentional mental state is normally understood . . . as one which is about, or represents, something in the world’ (Crane (2005)). Or as Christopher Peacocke claims: ‘A perceptual experience represents the world as being a certain way’ (Peacocke (1992): 61). The representational content of a perceptual experience is typically held to be a proposition—for instance, the proposition that there is an iMac computer on the desk in front of me—and as a result the content of a perceptual experience is something which has conditions of correctness: my experience represents the world as being a certain way, and my experience is accurate or veridical if the world is indeed how I represent it as being.5 So perceptual experiences, like beliefs, have conditions of correctness; and experiences, like beliefs, are accurate when the associated propositions which form their content are true. But it is a commonplace that emotions are intentional as well: they are, in other words, about or directed at objects and events in the world, which constitute their ‘target’. As Sabine Do¨ring puts it: ‘you hate your rivals, grieve over your mother’s death, or are afraid of the aggressive-looking woman: your rivals, your mother’s death and the aggressive-looking woman are the targets of your hatred, grief and fear respectively’ (Do¨ring (2003): 221). It is also a commonplace that emotions involve, or are partly constituted by, an evaluation or an appraisal of the relevant target.6 Here is

propose that there are important and interesting similarities between emotions and sensations, and hence (given the assumption) between emotions and perceptions. 3 Perceptions diverge from the relevant perceptual beliefs in the case of known visual illusions; emotions diverge from the relevant evaluative beliefs in the case of ‘recalcitrant’ emotions—for instance, in cases of fear where someone knows that the object of fear is harmless. 4 As a result, representationalists think that perception is direct. In writing this section I have learned much from Kathrin Glüer’s excellent (2009) paper. 5 On the view that the representational content of perceptual experience is a proposition, then perceptual experience counts as a propositional attitude. But this view of content is not essential to the representationalist theory. 6 On this view, ‘evaluations are not independent of emotions. Feeling fear is an evaluation and not a reaction to an evaluation’ (Teroni (2007): 408).

E M O T I O N S, P E R C E P T I O N S, A N D R E A S O N S


Do¨ring again: ‘Hating your rivals implies that you are seeing them as awful people; grieving over your mother’s death implies that you are regarding her death as a sad event; being afraid of the aggressive-looking woman implies that you are thinking of her as dangerous’ (ibid.). So emotions, like perceptions, can be viewed as having representational content, although in the case of emotions this content is evaluative: in hating your rivals, you represent them as awful; in grieving over your mother’s death, you represent this as a sad event; in being afraid of the woman, you represent her as dangerous. So whereas perceptual experiences are about material objects and their sensible properties, emotional experiences are about objects and events and their evaluative properties. The similarity between emotional and perceptual experience at the representational level enables the former to play a role in the justification of our evaluative beliefs which mirrors the role played by the latter in the justification of our empirical beliefs—or so it is claimed. In particular, perceptual theorists maintain that emotional experiences are akin to perceptual experiences insofar as they constitute reasons or evidence, at least absent defeaters, for the relevant evaluative beliefs or judgements. Now the idea that perceptual experiences are reasons or evidence for empirical beliefs seems relatively uncontroversial.7 After all, we often appeal to our perceptual experiences in order to explain why we believe what we do, and in order to justify our believing as we do.8 As John McDowell writes, ‘suppose one asks an ordinary subject why she holds some observational belief, say that an object within her field of view is square. An unsurprising reply might be “Because it looks that way”’ (McDowell (1994): 165). Here the subject cites her perceptual experience to explain and to justify her empirical belief. Supporters of the perceptual model claim that emotional experiences enjoy a similar epistemic status. Thus, Catherine Elgin writes that ‘Fear is evidence of danger; trust is evidence of reliability’ (Elgin (2008): 33). And: ‘something’s looking blue is ordinarily evidence that it is blue. If the analogy [between perceptions and emotions] holds, emotional deliverances are indicators, but not always accurate indicators of aspects of their objects. Just as my experiencing something as blue is evidence . . . that it is blue, my being frightened of something is evidence . . . that it is dangerous (ibid.: 37). In a similar vein, Sabine Do¨ring writes: ‘An emotion . . . resembles a sense-perception in having an intentional content that is representational. As a consequence, an emotion can justify a belief. Like a perception, it can do so by its representational content . . . justifying the content of that belief’ (Do¨ring (2003): 215). And Christine Tappolet states: ‘If we accept the claim that emotions have contents of this sort [i.e. they present the world as being in a certain way], then it becomes natural to claim that emotions are See, for instance, Brewer: ‘Perceptual experiences provide reasons for empirical beliefs’ (Brewer (1999): 18). This suggests an internalist view of justification. At least, it suggests that reasons and evidence are considerations to which we can appeal in order to justify our beliefs, and so are factors which are suitably internal to our perspective. I’ll assume, in what follows, that this way of understanding reasons and evidence is correct. It is, of course, compatible with the idea that external factors—such as the reliability of our perceptual mechanisms—can also play a role in the justificatory story. 7 8


M I C H A E L S . B R A DY

like sensory perceptions in that they allow us to be aware of certain features of the world, namely values. They do so, at least, under favourable circumstances, that is, when nothing interferes with them’ (forthcoming: 7). An analogy between the epistemic role of perceptual and evaluative experiences is also drawn, in meta-ethics, by Graham Oddie. Oddie maintains that in normal circumstances, ‘the visual experience of a bright red rose—that is to say, the rose’s appearing bright red to me—gives me a reason to believe that the rose really is bright red’ (Oddie (2005): 40). He now proposes that ‘[i]f there are genuine experiences of value, they could stand to values as ordinary perceptual experiences stand to the objects of perceptual experience. An experience of the goodness of P, say, would be the state of P’s seeming (appearing, presenting itself as) good, where this seeming is an experiential, non-doxastic take on the value of P. If there is such a state as the experience of the goodness of P, then, by analogy with the perceptual case, it would give me a reason to believe that P is good’ (ibid.). So Oddie thinks that experiences of value in general constitute defeasible reasons for evaluative beliefs, in much the same way that perceptions constitute defeasible reasons for empirical beliefs. (Oddie differs from supporters of the perceptual model of emotions because he thinks that the relevant experiences of value are desires.) Similar ideas can be found in the psychological literature. Gerald Clore and Karen Gaspar suggest ‘that beliefs are adjusted to be compatible with internal evidence in the form of feelings, just as they are adjusted to be compatible with external evidence from perceptual experience . . . Evidence from the sensations of feeling may be treated like sensory evidence from the external environment’ (Clore & Gaspar (2000): 25). There are, therefore, a number of philosophers and psychologists who claim that emotional experience can justify evaluative belief in much the same way that perceptual experience can justify empirical belief. Such theorists maintain that an emotional experience can constitute a reason or evidence for an evaluative belief: that my emotional experience of some object as evaluatively thus-and-so can constitute a reason or evidence for the belief or judgement that the object is evaluatively thusand-so. Moreover, those who favour perceptual accounts of the emotions would seem to be committed to the claim that the same kind of justificatory story can be told in both the perceptual and emotional cases.9 In the following section I’ll raise some doubts about this view of the epistemic role of emotions, grounded in a significant disanalogy between emotional and perceptual experience.

9 Do¨ring, Elgin, and Oddie are explicit about this. For example, Do¨ring (2003) maintains that in certain circumstances a perceptual experience ‘entitles’ one to take the representational content of experience at face value and form the relevant empirical belief. Here the justificatory story is non-inferential. She thinks something similar applies with respect to emotions: in certain circumstances, an emotional experience likewise entitles a subject to take the representational content of this experience at face value and form the relevant evaluative belief.

E M OT I O N S, P E R C E P T I O N S, A N D R E A S O N S


6.2 An obvious difference between perceptions and emotions is that at least some emotions can be assessed for rationality as well as accuracy. For instance, it is common to think of ‘recalcitrant’ emotions as in some sense irrational: it is irrational to be afraid of house spiders that one judges to be harmless, or to feel guilt when one believes that one has done nothing wrong. This marks a contrast with perceptual experiences. Even though perceptual experience can diverge from belief, it is not irrational to perceive the lines in the Müller-Lyer illusion as having different lengths even when one knows that they are the same length. Now one explanation of this difference in assessment is that, intuitively at least, emotions can be responses to features of objects or events which constitute reasons for those responses, and hence to which a subject will respond emotionally insofar as they are rational. It makes little sense, however, to talk of features of objects or events as reasons to literally see or hear something in a particular way. So whereas both perceptual and emotional experiences can (at least according to the perceptual model) constitute reasons for their respective beliefs, supporters of the model should accept that only emotional experiences are themselves responses to reasons. However, this fact, when allied with a central feature of our emotional lives, casts doubt upon the claim that perceptual and emotional experiences provide reasons or evidence, or at least reasons or evidence of the same kind. Consider, to illustrate, my perceptual experience as of a red car parked on the pavement outside my flat. In normal circumstances, and in the absence of epistemic defeaters (e.g. that my eyesight is poor, or that I am given to hallucinations), it is plausible to suppose that my perceptual experience constitutes very good evidence for—indeed, constitutes a conclusive reason for—my perceptual belief. At least, we would find it very strange if, after citing the fact that we can see what looks like a red car, we are then asked for, or feel the need to discover, further reasons or evidence for our belief. We might be puzzled as to what more we could do to justify the belief in this instance. If perceptual evidence won’t suffice, why should additional evidence, in the form of remembering a red car being there yesterday, or testimony from my partner who also claims to see what looks like a red car? We therefore have a strong intuition that, in the absence of defeaters, perceptual experience suffices for justification. In contrast, we often feel the need to discover reasons or evidence in support of our emotional experience, even in the absence of defeaters. Consider the experience of fear, when trying to get to sleep at night, upon hearing a noise downstairs. In such circumstances we are motivated to seek out and discover additional reasons or evidence. In particular, we are motivated to seek out and discover considerations that have a bearing on whether our initial emotional appraisal—namely, that we are in danger— is accurate. We strain our ears to hear other anomalous noises, or rack our brains trying to think of possible non-threatening causes for the noise, and so forth. It is unlikely that, in these circumstances, we would regard our feeling of fear as a conclusive reason to judge that we are in danger. Or, to take a different example, suppose that I feel guilty,


M I C H A E L S . B R A DY

upon waking, about my behaviour at a colleague’s party the previous evening. Again, we might not think that the feeling by itself is a conclusive reason for me to believe that I did anything wrong at the party. In such circumstances, I am typically motivated or inclined to find reasons or evidence that bear on the question of whether I am right to feel guilty, reasons which confirm or disconfirm my sense that I behaved badly. In cases like these, and in contrast with cases of normal perceptual experience, we do not rest content with our emotional appraisal, but instead seek out features and considerations which bear on the correctness of that appraisal. The fact that emotions, but not perceptions, can themselves be responses to reasons—and reasons which we are motivated to bring to awareness—suggests that the justificatory story we tell with respect to evaluative beliefs will be rather different from the justificatory story we tell with regard to empirical beliefs. It seems to me that the above phenomenon is not an accidental feature of our emotional lives, but instead is closely related to what emotions are and how emotions work. For it is often the emotions themselves which facilitate the search for and discovery of reasons and evidence, by modifications of our attention. And there are good reasons why this should be so, at least given a plausible understanding of the nature of emotion. To see this, note first that there is considerable evidence that emotion and attention are closely linked. Emotions typically direct and focus our attention on to objects and events that are potentially significant for us. As Aaron Ben-Ze’ev puts it, ‘like burglar alarms going off when an intruder appears, emotions signal that something needs attention’ (Ben-Ze’ev (2000): 13). The need for direction and focus of attention stems from the fact that as human beings we are presented with vast amounts of information about the state of the world and the state of ourselves, only some of which will be relevant to our concerns. Given that we have limited mental resources, we face a problem of efficiently locating or identifying which information in the environment is potentially important. We therefore have a need to pick out potentially significant stimuli from the mass of irrelevant stimuli that impinge upon our senses. Our emotional systems are thought to have been set up—by evolution and by socialization—to satisfy this need.10 For instance, a fear response quickly directs and focuses a subject’s attention on to a potential threat, and mobilizes a subject’s resources to enable her to react appropriately. These include, importantly, behavioural and motivational resources: fear directs and focuses attention, and at the same time primes the subject for the appropriate fight-or-flight response.

10 See Vuilleumier, Armony, & Dolan: ‘from an adaptive-evolutionary perspective, it can be assumed that emotion has a privileged role in biasing the allocation of attentional resources toward events with particular significance for an organism’s motivational state’. They continue: ‘Given a limited processing capacity, the brain must meet the challenge of detecting and representing only those stimuli most relevant for ongoing behaviour and survival. It is likely that attentional mechanisms evolved to enable the brain to regulate its sensory inputs so as to afford such selective perceptual processing and goal oriented action’ (Vuilleumier, Armony, & Dolan (2003): 419).

E M OT I O N S, P E R C E P T I O N S, A N D R E A S O N S


But emotions such as fear and shame do not just automatically and reflexively direct and focus attention; they also capture and consume attention. To say that attention is captured and consumed by emotional objects and events is to say that such objects and events hold sway over us, often making it difficult for us to disengage our attention and shift focus elsewhere. (This highlights another difference between emotions and perceptions, since even novel perceptual experiences do not capture or consume our attention for long, at least absent a related emotion.) Emotions such as fear and anger stay with us; they are not simply short-term interruptions to our mental life, but persist and dominate that life. We remain focused on and attentive to danger, infidelity, slights. Now the point of the persistence of attention is not to alert us to potentially significant objects and events in our environment; as we have seen, we are alerted to and focused on the presence of such objects and events automatically and reflexively, and hence prior to the persistence of attention. Instead, we might think that the point (or one of the points) of attentional persistence is to enhance our representation of potentially significant emotional objects and events, by enabling us to discover reasons which bear on the accuracy of our emotional appraisals.11 Again, there seems to be a clear need for emotional governance of our attention in this direction, at least given the widespread and plausible view of ‘basic’ emotions as fast, frugal, and relatively indiscriminate responses to objects and events of potential significance.12 There is a clear need, that is, for an evaluative system which acts as a ‘check-and-balance’ on our relatively indiscriminate emotional-appraisal system, so as to ensure that emotional appearance really does match evaluative reality. This suggests that the epistemological status of perceptual and emotional experiences, and the respective accounts of the justification of beliefs, are rather different from that suggested by supporters of the perceptual model of emotions. Perhaps it is true, as Do¨ring suggests, that (absent defeating evidence) we automatically take the representational content of perceptual experiences at face value when forming the relevant empirical beliefs. And perhaps it is true that, except in very unusual circumstances, we are entitled to do so. Perceptual experiences, after all, are not plausibly regarded as quick-and-dirty responses to external stimuli. But if emotional responses are often indiscriminate, and if it is a common feature of our emotional lives that we search for reasons that bear on the accuracy of our emotional experiences, then it is implausible to suggest that in normal circumstances we take the representational content of emotional experiences at face value when forming the relevant evaluative beliefs—at least if this suggests that we take such experience at face value in the absence of an awareness of those reasons which bear on the accuracy of our emotional responses. Given that emotions capture and consume attention, and given that this facilitates the search for

11 See Ronald de Sousa (1987): 196. For convincing empirical evidence that emotions enhance our representational capacities by capturing and consuming attention, see LeDoux (1996). 12 For an influential account of basic emotions as fast-and-frugal ‘affect program responses’, see Ekman (1977): 38–84.


M I C H A E L S . B R A DY

considerations that confirm or disconfirm our initial, relatively indiscriminate emotional appraisal of our circumstances, we might think instead that it is normal for us to endorse the content of our emotional experience only after we have discovered (or on some occasions invented or fabricated) reasons to think that our emotional appearance is veridical. Absent the discovery (or invention) of such reasons, it is by no means obvious that we regard ourselves as entitled to take the content of our emotional experiences at face value. It is by no means obvious, in other words, that we regard our emotional experiences as themselves reasons or evidence for our evaluative beliefs. At this point the defender of the perceptual model might respond as follows. We do indeed search for reasons in support of our emotional appraisals, and typically do not with respect to our perceptual experiences. But this merely reflects the fact that, as Elgin puts it, our emotional experiences are less reliable than perceptual experiences, and thus ‘need more collateral support in order to be [fully] tenable’. On this view, emotional experiences ‘start out with less initial tenability’ than perceptual experiences. Nevertheless, Elgin claims that ‘to have less initial tenability is not to have none. The very fact that [emotional experiences] present themselves as indicators of how things stand gives them some degree of initial tenability’ (Elgin (2008): 40). It is therefore compatible with the search for reasons that emotions provide some, albeit not conclusive, support for our evaluative beliefs.13 Emotions can still constitute reasons for evaluative beliefs, in other words, despite the fact that we are often motivated to look for additional evidence when it comes to emotional experience. If so, then the perceptual theorist can maintain that a difference in how attention is governed in perceptual and emotional experience does little to undermine the claim that perceptions and emotions play a similar epistemic role. In the following section, however, I will raise doubts about the plausibility of this weakened version of the perceptual model.

6.3 The fact that emotions, but not perceptions, are themselves responses to reasons might still ground scepticism about the idea that emotional experiences can constitute justifying reasons of any strength for evaluative beliefs. To see this, note first that if emotional experiences are reasons, then they should presumably fit easily into the class of considerations which clearly do function as reasons or evidence for evaluative beliefs. These considerations will include facts about what we might call the ‘emotion-relevant features’ of some object or event. For instance, the fact that the large dog has sharp teeth, a short temper, is off its lead, and is advancing rapidly towards me are all good reasons for me to believe that it is dangerous. Or, to take another example, the fact that Jones keeps changing his story under questioning, refuses to meet his interlocutor’s


Elgin is explicit about the fact that emotions might not provide conclusive reasons. See Elgin (2008: 34).

E M OT I O N S, P E R C E P T I O N S, A N D R E A S O N S


eyes, and stands to gain financially from testifying against the defendant are all good reasons to believe that he is untrustworthy as a witness for the prosecution. In each case, the considerations in question are the non-evaluative features of the object or event which seem directly relevant to the correct ascription of the evaluative property; they are precisely the kinds of consideration that we seek out when assessing the accuracy of our initial emotional appraisals. So should we be tempted to add to these considerations the respective facts that we are afraid of the dog or that we feel that Jones is untrustworthy? In other words, is it plausible to maintain that our emotional perceptions of the dog and of Jones are additional reasons to believe or judge that the dog is dangerous and that Jones is untrustworthy? The answer to both of these questions is ‘no’. To see why, note again that our emotional appraisals can themselves be responses to the properties or features which constitute genuine reasons for the relevant evaluative judgements (Goldie (2004): 254). That is, our suspicion of Jones (and thus our emotional appraisal of Jones as untrustworthy) can be a response to the fact that Jones keeps changing his story, refuses to meet his interlocutor’s eyes, and stands to gain financially from testifying against the defendant. By the same token, our fear of the dog (and thus our emotional perception of the dog as dangerous) can be a response to the fact that it has sharp teeth, a short temper, and is off its lead. However, to repeat a point made earlier, it is not just that our emotions can be responses to these reasons; our emotions should be responsive to these kinds of reasons, since these are the emotion-relevant features of the object or event. This highlights the fact that considerations which constitute reasons for evaluative judgements are equally reasons for the relevant emotional response. So reasons to think that something is dangerous are equally reasons to fear that thing; reasons to think that one has just been insulted are equally reasons to be angry; reasons to think that one has done something wrong are equally reasons to feel guilt; and similarly for other central cases of emotions. But this suggests that an emotional experience cannot constitute a reason for the relevant evaluative judgement, for it might then appear to be capable of justifying itself. That is, my fear of the dog cannot be a reason to judge that the dog is dangerous, since then we would have to conclude, from the fact that I am afraid of the dog, that I have good reason to be afraid of the dog. And fear, we might think, cannot justify itself in this way. We might make the same point in a different way, by focusing on the concepts that form the content of the relevant evaluative judgements. A plausible position in metaethics is that evaluative concepts like ‘dangerous’, ‘insulting’, ‘disgusting’, ‘amusing’, and so forth are best understood along sentimentalist lines. At least, many prominent philosophers have maintained that these kinds of evaluative concept can only be understood in terms of particular human emotions or sentiments. Sentimentalist accounts propose that we are to understand what it is for something to be dangerous or shameful, let’s say, in terms of its ‘eliciting or meriting’ certain emotional responses, namely fear and shame. Now the most promising versions of sentimentalism about value are second-order accounts, according to which ‘to apply a response-dependent


M I C H A E L S . B R A DY

concept Ö to an object X (i.e. to think that X is Ö) is to think it appropriate (merited, rational, justified, warranted) to feel an associated sentiment F towards X’ (D’Arms (2005): 3). These are the kinds of accounts favoured by ‘sensibility theorists’ such as Simon Blackburn, Allan Gibbard, John McDowell, and David Wiggins, all of whom reject dispositionalist theories of evaluative concepts.14 Sensibility theorists deny, that is, that an evaluative concept Ö is to be understood in terms of the sentiments that people are disposed to feel under ‘normal’ conditions. Instead, they maintain that we must understand evaluative concepts in terms of appropriate or fitting or merited emotional responses. On this account, then, to judge that X is Ö is to judge that it is rational or appropriate to feel F in response to X; to say that some object is dangerous is, therefore, to say that it merits fear, or that fear in this instance would be rationally appropriate, correct, warranted, or fitting. Now sensibility theorists face difficulties in spelling out just what it is for an emotional response to be appropriate, merited, or fitting. Despite this, the approach has a good deal of initial plausibility when it comes to evaluative concepts like ‘dangerous’, ‘insulting’, and ‘shameful’.15 If this account is correct, however, then our emotional responses cannot be reasons or evidence for the associated evaluative judgements. My fear of the dog, for instance, cannot be a reason to judge that the dog is dangerous, for then my fear would be a reason to judge that fear in these circumstances is appropriate or merited or fitting—and we have good reason to doubt that fear can justify itself in this way. The very fact that I am afraid of the dog cannot, by itself, be evidence that it is fitting or appropriate to be afraid of the dog. On a sentimentalist account of evaluative concepts, therefore, emotions cannot provide evidence or reasons for evaluative judgements.16 Note that this form of argument cannot be used to show that perceptions cannot be reasons for perceptual beliefs. If a response-invoking or response-dependent account is to be plausible with respect to sensory or perceptual properties like colours, then in all likelihood it will be a dispositionalist account. On this view, we are to understand colour concepts, let us say, in terms of the colours that people are disposed to see under ‘normal’ conditions. We should not be tempted here to propose that we should understand the concept ‘red’ in terms of the rational appropriateness of what people see, or in terms of responses to reasons to see, simply because we do not assess seeing red in terms of its rational appropriateness; to see something as red is not a response to 14

See Blackburn (1998); Gibbard (1990); McDowell (1998a and b); Wiggins (1987). See, for instance, the following important papers by D’Arms and Jacobson: (2000a and b), (2003), (2006). 16 It might be argued that it is the content of the emotional experience—viz. the representation of the dog as dangerous—that constitutes the reason, rather than the fact that I am afraid. However, the representational content of fear on sentimentalist lines is the proposition that fear of the dog is fitting. Since fear involves this representational content, then the content cannot be a reason for the relevant judgement and hence for fear; if it were, then the representational content of fear would justify fear. But we might think that the representational content of my fear cannot justify my being afraid, any more than my being afraid can justify my being afraid. 15

E M OT I O N S, P E R C E P T I O N S, A N D R E A S O N S


reasons. As a result, seeing something as red can be a reason to believe that it is red; but since reasons for perceptual beliefs are not, in and of themselves, reasons to perceive things, then seeing something as red does not constitute a reason for itself.

6.4 In the previous section I argued that perceptual and emotional experiences differ in terms of their epistemic value or role. Whilst perceptual experiences can constitute reasons for empirical beliefs, it seems to be the features of an object or event to which emotions respond, rather than emotions themselves, that constitute reasons for evaluative beliefs. One strand of this argument was that emotions, unlike perceptions, focus, capture, and consume attention, the point of which is to alert us to potentially significant objects and events and to facilitate an enhanced appraisal of our initial (and often relatively indiscriminate) evaluative construal: to facilitate, in other words, the search for those features and considerations which bear on the accuracy of this construal. Since this is so, the emotional response cannot be one of this class of reasons. The fact that I am afraid cannot itself be a consideration which has a bearing on the question of whether I am right to be afraid; the fact that an object appears dangerous cannot itself be a consideration that has a bearing on whether this appearance is veridical. Perhaps the defender of the perceptual model will agree that emotions sometimes motivate the search for reasons that bear on the accuracy of the emotions themselves, and hence the search for reasons for evaluative beliefs. They might also agree that in such cases it is doubtful that emotions are themselves to be understood as reasons for evaluative beliefs. Nevertheless, the perceptual theorist might still maintain that emotions can be perfectly good reasons for evaluative beliefs in other circumstances, and in particular in those circumstances in which emotions do not motivate the search for reasons. For instance, we might doubt that young children are often, if ever, motivated to seek out reasons that bear on the accuracy of their emotional appraisals. Rather, young children would seem to form evaluative beliefs—such as the belief that they are in peril—directly as a result of experiencing fear. Here it is surely tempting to cite the child’s feeling of fear as the reason why she believes that she is in danger.17 Consider now cases where we do not search for reasons because we trust in the reliability of our emotional responses. After all, if emotions can play the role of alerting us to objects and events of potential significance, then emotions can do this more or less reliably. It hardly needs pointing out that our emotions can be trained or calibrated so that they function more efficiently in bringing to our attention important aspects of our environment. Moreover, it seems true that we can learn to associate emotional responses with the relevant values. As a result, and as Elgin claims, we commonly ‘take ourselves


I owe this example, and objection, to Tim Bayne.


M I C H A E L S . B R A DY

to be able to reliably correlate emotions with circumstances’, at least given ‘a sophisticated understanding of when and to what extent they are trustworthy’ (Elgin (2008): 37). That is, ‘we can often tell which emotional reactions reflect the presence of emotional [evaluative] properties. So under certain recognizable circumstances, an emotional reaction affords epistemic access to such properties’ (Elgin (2008): 40–1). Consider, finally, cases where we are not motivated to search for reasons because the features to which our emotions are responses are unlikely to be available to us at the level of conscious thought, and hence unlikely to be identified by us as reasons. I might judge, for instance, that the piece of music is beautiful on the basis of aesthetic feelings, even though I cannot tell—and will be unable to tell even after much contemplation— precisely what it is about the music that generates my emotional response or justifies my judgement. I might, nevertheless, rightly judge that the music is beautiful on the basis of my emotional response. In all of these cases, it seems to be our emotional responses themselves that play the role of justifying reasons with respect to our evaluative judgements or beliefs. I agree that there are cases where we are not motivated to search for the reasons that bear on the accuracy of our emotional experiences, and not just because we have already identified such reasons. And I agree that in certain cases people take their emotional responses to be reasons for evaluative beliefs. But there is still room to ask whether people are right to do so. On the one hand, if a person’s emotional response is an unreliable ‘tracker’ of reasons, then such a response cannot be a justifying reason for the relevant belief—even though the person might take their response to be a reason. At best, such experiences can be mere causes of, or motivating reasons for, evaluative beliefs. Indeed, we might think that this is the most plausible interpretation of the situation where a child believes that he is in danger because he feels afraid, given that fear in young children is typically unreliably correlated with danger. On the other hand, even if we are aware that our emotional experiences in certain circumstances are reliably correlated with values, it is not obvious that such experiences are anything more than proxies for the underlying reasons: emotions can be useful substitutes for reasons in those circumstances where we cannot access the reasons that our emotions reliably track. As a result, it can sometimes be helpful (and natural) for us to treat emotional responses as if they were reasons; but this should not blind us to the fact that even in these circumstances, it is the reasons that emotions track, rather than the emotions themselves, that are doing the genuine justificatory work. For an argument in support of this picture of emotional responses as mere proxies, consider what happens when one becomes aware of the reasons that one’s emotional response tracks. In such a situation, it seems to me that any justificatory power that the emotional response has with respect to the relevant evaluative belief disappears. This is not the claim that the putative justifying power of the emotional experience is trumped by better reasons if we become aware of these. Rather, the point is that our emotional response to such features cannot be an additional justifying reason for evaluative belief in these circumstances, since this would involve ‘double counting’: the emotion-relevant

E M OT I O N S, P E R C E P T I O N S, A N D R E A S O N S


features of the object or event that constitute reasons would enter the justificatory story directly and via our emotional response. That is, if we have already taken the relevant features into account as considerations which justify our evaluative belief, we are debarred from taking these features into account again by treating our emotional response as an additional reason for the evaluative belief. Since we should avoid double counting, we should reject the idea that emotional experience retains its capacity to justify when we are aware of the reasons that such experience reliably tracks.18 But this just means that emotional experience isn’t a genuine reason after all, on the assumption that genuine reasons do not lose their justifying power when we become aware of other reasons for the same belief. Emotional experience is, therefore, at best a proxy for genuine reasons: useful in those cases where genuine reasons are unavailable to us at the reflective level or in situations where a search for such reasons would be inappropriate, but not in itself reason or evidence for evaluative belief or judgement. We can therefore acknowledge the important role that emotional experience can play in the justification of our evaluative beliefs—either by motivating the search for reasons, or by constituting proxies for such reasons—without succumbing to the inclination to conflate a proxy with the genuine article.

6.5 I have argued that there are significant differences between emotions and perceptions at the epistemic level. In particular, I have argued that emotions, unlike perceptions, often motivate the search for reasons which bear on their own accuracy, and hence on the correctness of the associated judgement. I have pointed out that there is a clear need for such search and discovery in the case of emotional but not perceptual experience, given that a central class of emotions are plausibly regarded as quick-and-dirty responses to objects and events of potential significance. And I have claimed that even when emotions are reliable trackers of value, emotional experience, unlike perceptual experience, is at best a proxy for genuine justifying reasons. It is unclear to what extent these differences threaten the perceptual model of emotion. If, for instance, the account of how emotions enable us to access values is a relatively small part of the perceptual model, then the perceptual theorist need not be overly concerned with the criticisms raised above. But if, as I suspect, part of the point of the perceptual model is to provide an adequate epistemology for our knowledge of value, then the fact that emotional experience doesn’t by itself provide reasons or 18 We can make the same point with regard to belief. If you believe that p on the basis of evidence, it would be a mistake to take your believing as you do to be an additional piece of evidence in favour of p. As Thomas Kelly writes, ‘The fact that you believe as you do is the result of your assessment of the probative force of the first-order evidence: it is not one more piece of evidence to be placed alongside the rest. That is, you do not treat the fact that you believe [p] as a further reason to believe that [p], above and beyond the firstorder considerations that you take to rationalize your belief.’ Kelly thinks that treating the fact that one believes that p as a further reason is to engage in illicit double counting (Kelly (2005): 167–96).


M I C H A E L S . B R A DY

evidence for evaluative judgement or belief would appear to be more damaging. In either case, defenders of the perceptual model owe us an explanation as to why the differences I’ve highlighted between perceptual and emotional experiences don’t, ultimately, make a difference to their case.19

References Ben-Ze’ev, Aaron (2000) The Subtlety of Emotions. Cambridge, MA: MIT Press. Blackburn, Simon (1998) Ruling Passions. New York: Oxford University Press. Brewer, B. (1999) Perception and Reason. Oxford: Clarendon Press. Clore, G. & Gaspar, K. (2000) ‘Some Affective Influences on Belief ’. In: Emotions and Beliefs, N. Frijda, A. Manstead, & S. Bem (eds), pp. 10–44. Cambridge: Cambridge University Press. Crane, Tim (2005) ‘The Problem of Perception’, in The Stanford Encyclopedia of Philosophy, Edward Zalta (ed.), D’Arms, Justin (2005) ‘Two Arguments for Sentimentalism’, Philosophical Issues 15: 1–21. D’Arms, Justin & Jacobson, Daniel (2000a) ‘The Moralistic Fallacy’, Philosophy and Phenomenological Research 61: 65–90. —— (2000b) ‘Sentiment and Value’, Ethics 110: 722–48. —— (2003) ‘The Significance of Recalcitrant Emotion’. In: Philosophy and the Emotions, A. Hatzimoysis (ed.), pp. 127–46. Cambridge: Cambridge University Press. —— (2006) ‘Anthropocentric Constraints on Human Value’. In: Oxford Studies in Metaethics, vol. 1, R. Shafer-Landau (ed.), pp. 99–126. New York: Oxford University Press. de Sousa, Ronald (1987) The Rationality of Emotion. Cambridge, MA: MIT Press. Do¨ring, S. (2003) ‘Explaining Action by Emotion’, The Philosophical Quarterly 53: 214–30. Ekman, P. (1977) ‘Biological and cultural contributions to body and facial movement’. In: The Anthropology of the Body, J. Blacking (ed.), pp. 38–84. London: Academic Press. Elgin, Catherine (1996) Considered Judgement. Princeton, NJ: Princeton University Press. —— (2008) ‘Emotion and Understanding’. In: Epistemology and Emotions, G. Brun, U. Doguoglu, & D. Kuenzle (eds). pp. 33–50. Farnham: Ashgate Publishing. Gibbard, Allan (1990) Wise Choices, Apt Feelings. Cambridge, MA: Harvard University Press. Glüer, K. (2009) ‘In defence of a doxastic account of experience’, Mind and Language 24: 297–373. Goldie, Peter (2004) ‘Emotion, Reason, and Virtue’. In: Emotion, Evolution, and Rationality, D. Evans & P. Cruse (eds), pp. 249–67. Oxford: Oxford University Press. Johnston, Mark (2001) ‘The Authority of Affect’, Philosophy and Phenomenological Research 53: 181–214. Kelly, Thomas (2005) ‘The Epistemic Significance of Disagreement’. In: Oxford Studies in Epistemology, vol. 1, J. Hawthorne & T. Gendler Szabo (eds), pp. 167–96. New York: Oxford University Press. LeDoux, J. (1996) The Emotional Brain. New York: Simon & Schuster.

19 Versions of this paper were presented at the universities of Geneva, Hull, Lancaster, Rijeka, Osnabrück, and Stirling, and at a workshop in Weggis, Switzerland, organized by the University of Fribourg. I would like to thank the audiences and participants for their feedback, and I am particularly grateful to Carla Bagnoli, Fiona Macpherson, Cain Todd, and anonymous referees for this volume for their helpful comments on previous drafts.

E M OT I O N S, P E R C E P T I O N S, A N D R E A S O N S


McDowell, John (1994) Mind and World. Cambridge, MA: Harvard University Press. —— (1998a) ‘Values and Secondary Qualities’. In McDowell (1998c): pp. 131–50. —— (1998b) ‘Projection and Truth in Ethics’. In McDowell (1998c): pp. 151–66. —— (1998c) Mind, Value, and Reality. Cambridge, MA: Harvard University Press. Oddie, Graham (2005) Value, Reality, and Desire. Oxford: Oxford University Press. Peacocke, Christopher (1992) A Study of Concepts. Cambridge, MA: MIT Press. Prinz, Jesse (2004) Gut Feelings. New York: Oxford University Press. Roberts, R. (2003) Emotions. Cambridge: Cambridge University Press. Tappolet, Christine (2000) Emotions et Valeurs. Paris: Presses Universitaires de France. —— (forthcoming) ‘The Irrationality of Emotions’. In: Philosophical Perspectives on Irrationality, D. Weinstock (ed.). Oxford University Press. Teroni, Fabrice (2007) ‘Emotions and Formal Objects’, Dialectica 61: 395–415. Vuilleumier, P., Armony, J., & Dolan, R. (2003) ‘Reciprocal Links between Emotion and Attention’. In: Human Brain Functions (second edition), R. Frackowiak et al. (eds), pp. 419–44. San Diego: Academic Press. Wiggins, David (1987) ‘A Sensible Subjectivism?’. In his Needs, Values, Truth, pp. 185–214. Oxford: Basil Blackwell. Zagzebski, Linda (2004) Divine Motivation Theory. Cambridge: Cambridge University Press.

7 Conscience What is Moral Intuition? Paul Thagard and Tracy Finn

Conscience is the internal sense of moral goodness or badness of one’s own actual or imagined conduct. The products of conscience are moral intuitions, which are the feelings that some acts are right and others are wrong. This paper offers a new theory of conscience and moral intuition as a particular kind of emotional consciousness, produced by brain processes that combine cognitive appraisal with perception of bodily states. We will discuss the significance of the new theory of conscience for philosophical issues about the role of intuitions in evaluating moral claims. To evaluate its plausibility, we compare the emotional consciousness theory of moral intuition with previous philosophical and psychological theories of conscience. From a theological perspective, conscience is the God-given ability to distinguish right from wrong. Non-theological accounts of conscience range from the Platonic theory that moral intuitions can be true a priori to the Humean view that they are just emotional reactions based on previous experience. We argue that our emotional consciousness theory provides a better explanation of conscience and moral intuition than these alternatives.

7.1 The Emotional Consciousness Theory Here is an example of conscience experienced by one of the authors of this paper. When Apple announced the iPhone early in 2007, he knew he wanted one, but was taken aback by the high projected price. Then it occurred to him that perhaps he could put the iPhone on his research grant. Immediately, however, he got the bad feeling that this would be wrong, a misuse of government funds for personal purposes. His conscience said not to do it by generating the moral intuition that charging an iPhone to a research grant would be wrong. This reaction was a conscious experience marked by negative emotion, tied to a mental representation of the possibility of getting the iPhone in an inappropriate manner.

C O N S C I E N C E : W H AT I S M O R A L I N T U I T I O N ?


Introspection supports the claim that moral intuitions are a kind of emotional consciousness. The intuitions that killing and eating people is wrong and that aiding needy people is right are judgments of which we usually have conscious awareness. Further, their emotional status is evident, as most people feel bad about cannibalism and good about helping. Obviously, there are other kinds of emotional consciousness besides moral intuitions, as when someone feels happy that the sun is shining without any moral judgment about it. But if moral intuitions are indeed conscious and emotional, a theory of emotional consciousness should be a good start to understanding them, provided that an account can be given of how moral intuitions differ from other kinds of emotional experiences. Thagard & Aubie (2008) have proposed a neural theory of emotional consciousness that we will draw on here, although it will not be possible to revisit the full range of neurological and psychological evidence adduced for it. Philosophers and psychologists have long debated the nature of emotion, with two primary views contending for dominance. Since the Greek Stoics, many theorists have argued for the view that emotions are cognitive appraisals concerning the extent to which something aids or hinders our goals (e.g. Nussbaum ( 2001); Scherer, Schorr, & Johnstone (2001)). A positive emotion such as happiness is an evaluation that something is contributing to our goals, as when the sun shining enables us to play outside. In contrast, a negative emotion such as sadness or anger is an evaluation that something is preventing the accomplishment of our goals. Since the nineteenth century, the major alternative to the cognitive appraisal theory of emotions has been the somatic perception theory, according to which emotions are perceptions of bodily states (e.g. James (1884); Damasio (1994); Prinz (2004)). On this view, emotions are not judgments, but rather internal representations of bodily states. For example, fear is a perception of the bodily changes such as heart rate, respiration, and cortisol levels that respond to external stimuli. According to Thagard & Aubie (2008), emotions are both cognitive appraisals and somatic perceptions, performed simultaneously by interacting brain areas. The interactions are captured by the EMOCON model of emotional consciousness, shown in Figure 7.1. When a scene is presented to the senses, for example a sunny day or a gruesome corpse, this information is simultaneously conveyed to brain mechanisms for somatic reactions and their perception as well as to mechanisms for appraisal of the significance of the scene. Somatic perception is accomplished by brain areas such as the amygdala and insula that integrate information about bodily states. Cognitive appraisal is performed by brain areas such as the dorsolateral prefrontal cortex and anterior cingulate that evaluate the relevance of a situation for the goals of the perceiver. As shown in Figure 7.1, these two mechanisms interact using many neural connections between the relevant sets of brain areas such as the prefrontal cortex and the amygdala. The EMOCON model of emotional consciousness hypothesizes that positive and negative experiences are generated by ongoing rapid interactions between cognitive appraisals and somatic perceptions. For example, your reaction of disgust and fear to a



emotional feeling

anterior cingulate external stimulus

external sensors


dopamine system






internal sensors

bodily states

Figure 7.1 The EMOCON model of emotional consciousness. Abbreviations are PFC for prefrontal cortex, DL for dorsolateral, OF for orbitofrontal, and VM for ventromedial. The dotted line represents emotional experience. Source: Thagard & Aubie (2008).

mangled body results from your brain’s integrated representation of your bodily response and your evaluation of a possible threat to your survival goals. Thagard & Aubie (2008) argue that this theory can explain many aspects of emotional consciousness, such as the varying intensity, change, and positive and negative valence of emotions. From the perspective of the theory of emotional consciousness, conscience is not a special mental faculty, but rather a neural process that involves both cognitive appraisal and somatic perception. The role of bodily reactions explains why moral intuitions are marked by ‘gut reactions’: the disgust most people feel at torture and the joy we feel on a sunny day are indeed visceral. But moral intuition, like emotional consciousness in general, is not just visceral, because our overall interpretation of a scene is colored by cognitive appraisal as well. For arguments that there is more to emotions than somatic perception, see Rolls (2005: 26–9) and Barrett (2006).

C O N S C I E N C E : W H AT I S M O R A L I N T U I T I O N ?


There is ample evidence from brain scanning experiments that the brain areas specified by the EMOCON model are involved in moral judgments (see, for example, Greene & Haidt (2002); Moll, Zahn, et al. (2005); Raine & Yang (2006)). But what makes moral intuitions different from other non-moral kinds of emotional reactions? For somatic perception, there seems to be nothing special about the brain processes involved in moral intuition compared to emotional consciousness in general: the relevant areas such as the insula, amygdala, and dopamine system are the same. For example, Moll, de Oliveira-Souza, et al. (2005) report that both moral and non-moral disgust recruit remarkably similar brain areas. Hence, the crucial difference must involve cognitive appraisal, with moral intuitions activating a narrower set of goals than general emotional consciousness. From philosophical and religious writings on ethics, it appears that the relevant moral goals might include the following:  Increase the well-being of people in general.  Act in accord with abstract moral principles such as fairness and respect for autonomy.  Satisfy the expectations of social groups such as family.  Comply with religious standards or other moral code. Violation of these goals produces negative emotions such as guilt and shame, whereas compliance produces positive emotions such as happiness and relief. Thus, when Thagard felt bad at the thought of paying for an iPhone with his research grant, it was the result of brain activity that integrated perception of his bodily states with cognitive appraisal that the considered action would violate his goal of using his research funds responsibly. A highly detailed account of the role of cognitive appraisal in emotions is due to Sander, Grandjean, & Scherer (2005), summarized in Table 7.1. They identify four appraisal objectives concerning the major types of information that an organism needs to adaptively react to a salient event. These objectives (called criteria in Table 7.1) concern relevance for the organism or its social group, implications for goals, coping potential, and significance for social norms and values. For example, the first column of table 7.1 predicts that an event will tend to generate enjoyment if it is relevant, predictable, very pleasant, intended, highly probable, and one to which the organism can adjust its behavior. However, Sander, Grandjean, & Scherer stress that emotional reactions cannot be reduced to a set of discrete rules, since they involve interactions among the various aspects of cognitive appraisal. Hence there is only a tendency to enjoy an event that has the characteristics just listed, not a rule that says that IF the characteristics apply THEN one has the experience of enjoyment. Thagard & Aubie (2008) describe a computational model that uses artificial neural networks to provide an overall coherent assessment of an event based on the objectives listed in Table 7.1. It uses 67 units (artificial neurons) to represent the 14 emotional states and many levels of objectives shown in Table 7.1. 262 links between the units represent

Table 7.1. Predicted appraisal patterns for some major emotions. ENJ/HAP, enjoyment/happiness; ELA/JOY, elation/joy; DISP/DISG, displeasure/disgust; CON/SCO, contempt/scorn; SAD/DEJ, sadness/dejection; ANX/WOR, anxiety/worry; IRR/COA, irritation/ cold anger; RAG/HOA, rage/hot anger; BOR/IND, boredom/indifference; Other/nat, Other or Natural (e.g. weather); Cha/int, Chance or Intent; Cha/neg, Chance or Negative Intent; Int/neg, Intent Negative; Vcon, Very Conducive. Criterion Relevance Novelty Suddenness Familiarity Predictability Intrinsic pleasantness Goal/need relevance Implication Cause: agent Cause: motive Outcome probability Discrepancy from expectation Conduciveness Urgency Coping potential Control Power Adjustment Normative significance Internal Standards External Standards








Low Open Medium High Medium

High/med Open Low Open High

Open Low Low Very low Low

Open Open Open Open Low

Low Low Open Open High

High Very low Low Open High

Low Open Open Open Medium

Open Intent Very high Consonant Conducive Very low

Open Cha/int Very high Open Vcon Low

Open Open Very high Open Open Medium

Other Intent High Open Open Low

Open Cha/neg Very high Open Obstruct Low

Other/nat Cha/neg Very high Dissonant Obstruct High

Other/nat Open Medium Open Obstruct Medium

Open Open High

Open Open Medium

Open Open Open

High Low High

Very low Very low Medium

Very low Very low Very low

Open Low Medium

Open Open

Open Open

Open Open

Very low Very low

Open Open

Open Open

Open Open

Criterion Relevance Novelty Suddenness Familiarity Predictability Intrinsic pleasantness Implication Cause: agent Cause: motive Outcome probability Discrepancy from expectation Conduciveness Urgency Coping potential Control Power Adjustment Normative significance Internal Standards External Standards








High Low Low Low

Low Open Medium Open

High Low Low Open

Very low High Very high Open

Low Open Open Open

Open Open Open Open

Open Open Open Open

Other/nat Open High Dissonant Obstruct Very high

Open Int/neg Very high Open Obstruct Medium

Other Intent Very high Dissonant Obstruct High

Open Open Very high Consonant Open Low

Self Int/neg Very high Open Open High

Self Intent Very high Open High Medium

Self Intent Very high Open High Low

Open Very low Low

High Medium High

High High High

Medium Medium High

Open Open Medium

Open Open Medium

Open Open High

Open Open

Open Low

Open Low

Open Open

Very low Open

Very low Very low

Very high High

Source: Sander, Grandjean, & Scherer (2005): 326.



positive constraints such as how an objective contributes to an emotion and negative constraints such as incompatibilities between emotions. A simple parallel algorithm spreads activation among the units starting with input that describes the characteristics of an event; emerging activation of some of the units and deactivation of others computes a maximally coherent reaction to the event. For example, when the units representing the criteria for enjoyment from Table 7.1 are activated, the unit for enjoyment gets high activation, and the unit for pride is also activated to a lesser extent. Some of the criteria in Table 7.1 are relevant to moral intuition as well as emotional appraisal in general: pleasantness, goal/need relevance, and standards have all been elements of different ethical theories. We will return to these topics in our discussion below of the normativity of the emotional consciousness theory of conscience. For now, it suffices to recognize that moral concerns are an important subset of the criteria expressed in Table 7.1 and incorporated in the EMOCON account of cognitive appraisal. In sum, conscience is a neural process that generates emotional intuitions combining somatic perception (the gut reaction) with cognitive appraisal concerning a special subset of goals.

7.2 Evidence for the Theory In order to evaluate our account with respect to older ones, we need to specify what should be expected of a theory of moral intuition from both descriptive and normative perspectives. The first requirement of such a theory is that it should explain how moral intuitions can be both cognitive and emotional. They seem to be cognitive in that they make claims about things in the world, for example that it was wrong for Jeffrey Dahmer to rape, kill, and eat his victims. Moral intuitions also have a large emotional component, as people react to repulsive actions such as Dahmer’s with anger and disgust. That moral judgments are generally emotional is evident not only from introspection, but also from studies in experimental psychology and neuroscience (Greene & Haidt (2002)). Second, a theory of moral intuition should be able to explain moral agreement: people often have the same moral intuitions in cases such as the Dahmer one. Thus conscience is not totally idiosyncratic, and seems to have less variation across individuals, cultures, and historical periods than, say, fashion sense. Third, a theory of moral intuition must explain why there is also a great deal of moral disagreement: people have very different moral intuitions concerning topics such as abortion, homosexuality, and capital punishment. Fourth, it is philosophically important for a theory of moral psychology to be normative as well as descriptive. In addition to explaining why people have the moral intuitions that they do, a theory should help to evaluate the ethical and epistemological validity of such intuitions. Are they indubitable as might be claimed by theories that take moral intuitions to be true a priori? Or have they no objective

C O N S C I E N C E : W H AT I S M O R A L I N T U I T I O N ?


force at all, as is insisted by theories that take ethical statements to be just expressions of emotional attitudes? Thus, a theory should provide insights into the normative status of moral intuitions as well as explaining their cognitive and emotional content and the existence of moral agreement and disagreement. We view it as a strong merit of our theory that it is not only consistent with, but helps to explain, a broad range of psychological and neurological evidence. This accomplishment will not impress mind–body dualists and anti-naturalists, who think that moral theory should operate in a logical space entirely separate from empirical matters. But even dualists should recognize that a moral psychological theory needs to be able to account for the aspects of moral intuitions listed above, including their cognitive/emotional nature, agreement/disagreement aspects, and normativity. If moral intuition is a kind of emotional consciousness as specified by the EMOCON model, it is easy to see how ethical reactions can be both cognitive and emotional. They are inherently emotional because they are carried out by the same neural processes that generate emotional reactions, including the perceptions of bodily states that give emotions—and moral intuitions—their visceral character. A purely somatic theory of emotions would have difficulty accounting for the cognitive content of ethical judgments, but this is not a problem for our emotional consciousness theory that allows a crucial component of appraisal with respect to goals. This appraisal, carried out by the full range of cognitive processes in the prefrontal cortex, can call on all the representational resources of the brain’s most intellectually sophisticated part. Through incorporation of the experimentally established neural interconnections of the prefrontal cortex with viscerally connected areas such as the insula and amygdala, the EMOCON model shows how moral intuitions can simultaneously and inextricably be both cognitive and emotional. To explain widespread agreement in moral intuitions among humans, for example in the Dahmer case, we can attend to great similarities in people with respect to both somatic perception and cognitive appraisal. Except in rare cases of damage to brain areas such as the amygdala or the insula, we have evidence from dissection and brain scanning experiments that all humans have the same brain structures and connections, including those shown in the EMOCON model in Figure 7.1. This similarity partly explains why emotions such as happiness, sadness, fear, anger, disgust, and surprise seem to be culturally universal (Ekman (2003)). People worldwide have the same basic emotions because they have the same physiology, including bodily processes such as blood circulation, respiration, and stress hormones, and also the same brain areas including the prefrontal cortex and amygdala. Physiological similarity alone, however, is not sufficient for a full explanation of moral agreement. Emotional consciousness, according to the EMOCON model, is also affected by cognitive appraisal with respect to goals. It seems, however, that there is a great deal of similarity in the goals that enter into people’s cognitive appraisal of scenes. People generally experience fear if their survival is threatened or if they are in danger of being hurt. People generally experience happiness when they can anticipate



pleasure or other rewards. Further, most people are concerned not only with happiness and freedom from pain for themselves, but also for other people, at least those with whom they have emotional connections such as family and friends (Batson (1991)). Hence, from the perspective of moral intuition as a kind of emotional consciousness, we can explain moral agreement as based on a combination of the similarities of people with respect to physiology and goals. Nevertheless, despite the general agreement in moral intuitions about cases such as Jeffrey Dahmer, there is also a great deal of disagreement. Here are some examples of moral judgments restricted to particular classes of people:     

People should not eat meat and milk together (Orthodox Judaism). Priests should be celibate (Catholicism). People should not get blood transfusions (Jehovah’s Witnesses). People should pray five times per day (Islam). People should not kill animals (Hinduism).

Many other examples of religious and cultural differences in ethical beliefs show that there is wide variation in moral intuitions (Prinz (2009)). Even within specific cultural groups, there is variation in moral intuitions. The most striking class of people who deviate from moral norms, estimated to be around 1% of the population, are psychopaths. Hare (1993: xi) describes them as follows: Psychopaths are social predators who charm, manipulate, and ruthlessly plow their way through life, leaving a broad trail of broken hearts, shattered expectations, and empty wallets. Completely lacking in conscience and in feelings for others, they selfishly take what they want and do as they please, violating social norms and expectations without the slightest sense of guilt or regret.

Whereas members of different religious groups have moral intuitions that others lack, psychopaths stand out in lacking moral intuitions that they should be concerned about the suffering of others. A theory of moral psychology should be able to explain both kinds of variation. There is mounting evidence for a neurophysiological explanation of psychopathy that is consistent with our emotional consciousness account of conscience. Blair, Mitchell, & Blair (2005) argue that psychopaths suffer from genetic anomalies that disrupt the functioning of the amygdala, a crucial part of the EMOCON model. These authors propose that atypical functioning of the amygdala in psychopaths produces impairments in emotional learning. Emotional dysfunction can lead to the learning of antisocial methods of achieving goals, because psychopaths do not receive negative reinforcement from the distress or disapproval of others. In psychopaths, moral socialization fails because they do not experience negative emotions when they see other people suffering, and they do not understand the efforts by caregivers to make them empathize with victims. Thus we can explain the great divergence between the moral intuitions of psychopaths and most members of society by physiological differences in brain structure, particularly the amygdala.

C O N S C I E N C E : W H AT I S M O R A L I N T U I T I O N ?


Damage to other brain areas can also lead to moral intuitions different from those among the general population. Koenigs et al. (2007) presented moral dilemmas to people with damage in the ventromedial prefrontal cortex (VMPFC), part of the EMOCON model known to be involved in emotional judgments. The patients were normal with respect to most kinds of moral dilemmas, but reacted differently to high-conflict personal moral dilemmas that pit aggregate moral welfare against direct harm. For example, most people have the intuition that it is right to save five lives at the expense of one, but would not physically push an individual in front of a train to accomplish this end. In contrast, patients with VMPFC damage react unemotionally and make the same utilitarian calculation in the personal case, lacking the highly emotional moral intuition that directly causing harm is especially wrong. We know of no studies of the effects of damage to the insula on moral judgments, but it is interesting that patients with insula lesions sometimes find it much easier to quit smoking, presumably because of a change in physical craving (Naqvi et al. (2007)). This finding does not concern moral intuitions, but illustrates the role of the insula in the somatic perception component of emotional consciousness. Most moral disagreements, however, are not caused by physiological differences, as the vast majority of people have the undamaged brain structures shown in the EMOCON model. Hence religious and cultural differences in moral intuition need to be explained by differences in cognitive appraisal. Cognitive appraisal is affected by at least the following factors: 1. Existence and priority of goals. 2. Beliefs about how to accomplish goals. 3. Other beliefs about people and the world. Religion and other cultural factors can generate radically different goals that people use to assess situations. For example, a religious upbringing can convince someone that the most important moral goal is compliance with the code of a particular religion as expressed in a text taken to be sacred, such as the Bible, Talmud, or Koran. People are brought up by parents with different religious backgrounds, so they acquire different moral codes that generate different cognitive appraisals and emotional reactions. Someone brought up to think that the Talmud is the best source of proper behavior will have very different reactions to various possible actions than someone who thinks that the source of moral authority is the Koran, New Testament Bible, or other text. Different religions and cultures also vary with respect to what they think constitutes moral goals, emphasizing to different degrees the importance of prayer, attendance at services, and obedience to moral authorities. Moreover, people can have different factual beliefs tied in with religious teachings, for example about the origins of persons. According to Catholics, people originate at conception, whereas for Hindus they are reincarnated. All of these differences can produce varying cognitive appraisals of situations, leading to different emotional reactions and moral intuitions.



Thus the emotional consciousness theory of conscience can explain individual and cultural differences in moral intuitions as arising sometimes from physiological differences, as in the case of psychopaths, but more often from goal-based differences in cognitive appraisal. This understanding of why there is much disagreement in people’s moral intuitions, as well as the earlier discussion of why there is also much agreement, provides the basis for a discussion of the normative significance of conscience.

7.3 Philosophical Implications: Normativity The central normative question about moral intuitions is: Does having a moral intuition that an act is right or wrong provide any degree of justification for the judgment that the act really is right or wrong? Extreme answers to this question include the affirmative one that moral intuitions can reflect judgments that are true a priori, and the negative one that moral intuitions are just emotional reflections of untutored prejudices and therefore have no evidential force. One of the advantages of the emotional consciousness theory of conscience is that it avoids both of these extremes and can provide some guidance about when moral intuitions deserve to be taken seriously. If emotions were just somatic perceptions, then emotional consciousness would be a poor contributor to the justification of moral judgments. The fact that you have a physical reaction to something provides no evidence at all that the reaction is morally appropriate. However, the EMOCON model contends that emotions are an integrated combination of somatic perceptions and cognitive appraisals, and the latter can be evaluated with respect to the extent to which they reflect considerations that are relevant to justifying moral judgments. The normative question is transformed into: Does the coherence-based assessment that generates different emotional reactions provide any reason to believe that the existence of a moral intuition can contribute to the justification of a judgment of right and wrong? To answer this question, we refer back to the model of cognitive appraisal described earlier. Some of the criteria listed in Table 7.1 are clearly very relevant to normative ethical judgments. For consequentialist ethics that judge the morality of an event based on its results, the most relevant criteria include pleasantness, goal relevance, outcome probability, and control. For deontological ethics that judge the morality of an event based on its compliance with ethical principles, the most relevant criteria are agency, motive, and internal and external standards. Hence the contribution of cognitive appraisal to emotional reactions can include many of the kinds of considerations that leading ethical theories advocate. Moreover, an emotional appraisal does not have to be limited to only consequentialist or deontological criteria, but can simultaneously take into account both consequences and principles. In medical ethics, for example, people commonly have to balance concerns about the well-being of patients and their families with principles such as autonomy and privacy. According to the emotional consciousness theory of moral intuition, this balancing is a largely unconscious process that takes place by

C O N S C I E N C E : W H AT I S M O R A L I N T U I T I O N ?


parallel computing in the brain’s neural networks, taking into account both goals and standards. Potentially, therefore, an emotional intuition can involve a highly sophisticated, multi-factorial appraisal of an event that incorporates elements of both consequencebased and principles-based ethical theories. An emotional intuition that performs such an incorporation is not just an unreflective bodily reaction, but rather an integration of legitimate contributors to ethical judgments with somatic perceptions. Hence, conscience can indeed sometimes be an indicator of the legitimacy of a moral judgment. Feeling bad about putting a personal item on a research grant can indeed be an indicator that this act would be wrong, if the feeling derives in part from an appraisal of the consequences of the act and its compliance with internal and external standards. The problem, of course, is that it is very difficult for people to identify the actual criteria that contribute to their cognitive appraisals. We have no conscious access to the neural processes that the EMOCON model describes as underlying our emotional reactions, and hence have no way of knowing what cognitive and somatic factors contributed to a feeling that an act is right or wrong. Although an emotional intuition may in fact be based in part on appraisals that incorporate legitimate ethical assessments about consequences and principles, we have no way of knowing in any particular case whether the appraisals we make are based on these or other considerations. Rousseau (1960: 61–2) described the difficulty of knowing the sources of our moral behavior: They [reflections] have taught me one great maxim of morality, the only one perhaps which is of practical use: to avoid situations which place our duties in opposition to our interests, and show us where another man’s loss spells profits to us. For I am sure that, in such situations, however sincere and virtuous the motives we start with, sooner or later and unconsciously we weaken, and become wicked and unjust in practice, although still remaining good and just in our hearts.

Thagard (2007) argues that understanding the unconscious neural processes underlying moral judgments makes it easy to see why conflicts of interest are such a large problem in government, business, and other enterprises. When your conscience tells you something, you simply have no way of knowing why it is telling you that. You may think that you are making the judgment that is best with respect to all relevant consequences and principles, but it may be that you are unconsciously responding to some aspect of self-interest that is particularly salient in the current situation. Although you cannot absolutely trust your conscience in a particular case, it may be possible to acquire inductive evidence that your moral intuitions have a good chance of being morally legitimate. For decision-making, Thagard (2006) recommends a procedure called informed intuition. This procedure recognizes that the formal methods of mathematical decision theory are rarely applicable in real life, because of lack of knowledge about relevant probabilities and utilities. But the procedure also acknowledges that intuitions about what to do may be based on prejudices and misinformation. Informed intuition operates as follows (Thagard (2006): 22):



1. Set up the decision problem carefully. This requires identifying the goals to be accomplished by your decision and specifying the broad range of possible actions that might accomplish those goals. 2. Reflect on the importance of the different goals. Such reflection will be more emotional and intuitive than just putting a numerical weight on them, but should help you to be more aware of what you care about in the current decision situation. Identify goals whose importance may be exaggerated because of emotional distortions. 3. Examine beliefs about the extent to which various actions would facilitate the different goals. Are these beliefs based on good evidence? If not, revise them. 4. Make your intuitive judgment about the best action to perform, monitoring your emotional reaction to different options. Run your decision past other people to see if it seems reasonable to them. Similarly, a moral intuition need not come out of the blue, but can be based on some conscious reflection on the relevant consequences or principles. Hence, if your current moral intuition is based on ethical information, there is an increased likelihood that your moral intuition yields a defensible moral judgment. Or, equally legitimately, a moral intuition for a particular act may be automatic because its context is analogous to previous situations in which you have exercised informed intuition. In cases like these, experience and analogy can enable your conscience to be a good guide to action. So you can sometimes trust your conscience, but what about the consciences of other people? How should you react to the moral intuitions of people who disagree with you? Moral disputes should not degenerate to exclamations that ‘my intuitions are stronger than your intuitions’. The EMOCON model suggests that there are various situations in which it can be legitimate to discount the moral intuitions of others. 1. Physiological defects. People with malfunctions in crucial brain areas, such as the amygdala, insula, or dorsolateral prefrontal cortex, are likely to have disrupted emotional consciousness and moral intuitions. To take an extreme example, we have no reason to respect the moral judgments of psychopaths who have defects in their amygdala-based social learning mechanisms. 2. Lack of moral education. Consider rare cases of people who have been raised without any training by parents or other caregivers in judgments of right and wrong. There is no reason to believe that the conscience of such people would be a reliable guide to morality. For example, people raised by amoral psychopaths would be unlikely to have acquired good moral intuitions, even if their emotional learning mechanisms are not defective. 3. Biased moral education. One of the contributors to cognitive appraisal in emotions is fit with internal and external standards. We have reason to discount the moral intuitions of people that we know have acquired illegitimate standards, for example racism and egoism, or the view that one should act only out of self-interest.

C O N S C I E N C E : W H AT I S M O R A L I N T U I T I O N ?


4. Situational distortions. As we saw in the case of conflicts of interest, it sometimes happens that our decisions and our moral intuitions are distorted or swamped by salient stimuli that lead to neglect of relevant ethical considerations. For example, someone may be experiencing temptation to have an illicit affair that contradicts the person’s own standards, but nevertheless may have the intuition that the affair is the right thing to do. In this case the intuition is governed more by lust than by the assessment of consequences and principles that mark a full cognitive appraisal. We are justified in rejecting the moral force of the intuitions of others if we can identify causes of the intuitions such as intense temptation. In such cases, people may be acting against their own consciences, as often happens in cases of weakness of will, which results when subcortical motivations such as food and sex dominate conscious cortical reflections about what ought to be desired (Thagard (2007, 2010)). We should, however, take seriously the moral intuitions of people who do not suffer from neural defects, defective moral education, or situational distortions. Their intuitions may be as well-based on experience as our own, especially if their backgrounds have ensured that the intuitions are highly informed. In cases of disagreement, people should attempt to move toward consensus by the usual means of verbal communication, such as exchanging facts and arguments. But if we are right that the moral intuitions are inherently emotional, the kind of communication needed to produce a moral consensus may need to be emotional as well as discursive. Thagard & Kroon (2006, reprinted in Thagard (2006)) propose a theory of emotional consensus that describes how agreement can arise from a combination of verbal and non-verbal communication, including empathy and emotional contagion. It would be interesting to explore how these mechanisms of emotional consensus might be used to increase agreement in moral intuitions. In sum, we have shown how our naturalistic, neural account of conscience can be used to address normative questions about the legitimacy of moral intuitions of ourselves and others. Conscience can inform us about what our moral goals are, as well as about good ways to meet these goals. But we must be wary of moral intuitions, our own as well as those of others, that may derive from defects in physiology, deficiencies in moral education, or situational distortions such as weakness of will.

7.4 Alternative Theories In philosophy as in science, a theory does not have to be perfect to be worthy of adoption, just better than the alternatives. We will now argue that our emotional consciousness theory of conscience is superior to alternative accounts based on theology, a priori reasoning, and moral grammar. Space does not permit detailed discussion of these alternatives, but we will outline how they compare with the emotional



consciousness theory with respect to the requirements discussed earlier: cognition/ emotion, agreement/disagreement, and normativity. 7.4.1 Theological theories of conscience According to theological theories, conscience is a God-given ability to distinguish right from wrong. Langston (2006) states that conscience was not explicitly discussed by Plato and Aristotle, but was an important topic for medieval thinkers such as Thomas Aquinas. Many religious people believe that right and wrong are determined by God’s prescriptions, and that conscience is a divine gift for telling whether an act is in line with these prescriptions. The most obvious problem with theological theories of conscience is that they make the strong metaphysical assumptions that there is a God who provided humans with a faculty of conscience, presumably operating in a non-material soul. Thagard (2000) argues that theism and dualism hang together in a highly coherent explanatory package, but that this package is less coherent with observed facts than a set of scientific alternatives provided by physics, biology, and psychology. If belief in God and the soul are not justified by inference to the best explanation, theological theories of conscience are dubious. Without good evidence for the existence of God and the soul, there is no reason to believe in them or in any theological theory of conscience. Even without this major metaphysical problem, theological theories face many difficulties. They provide no information about how conscience actually works, since the operations of the soul are inherently mysterious. Hence, they cannot explain why moral intuitions have both cognitive and emotional aspects. Introspection alone, without scientific experiments and theories based on them, tells us little about how cognition and emotion operate. In principle, theological theories of conscience could explain agreement in people’s moral intuitions by assuming that God provided us all with a similar faculty for identifying right and wrong. But disagreement is a much greater problem for such theories, as it is difficult to see from the theological perspective why people have ended up with such a wide variety of conflicting intuitions on topics such as abortion and capital punishment. People’s disagreements often correlate with disagreements about religious principles, as Catholics, Protestants, Muslims, Hindus, Buddhists, Jews, atheists, and others reach highly divergent views about the nature and commandments of God and about the validity of attendant moral judgments. All this is a big puzzle if conscience is a God-given faculty, but makes perfect sense if conscience is a kind of emotional consciousness shaped by the varied experiences of brains raised in different cultural circumstances. The issue of normativity could be addressed from a theological perspective by arguing that moral intuitions are a legitimate source of justified moral judgments when they duly convey information from God about his moral prescriptions. Philosophers have long argued, however, that theology cannot be the basis for morality, as it is always legitimate to ask whether the actions and will of God are

C O N S C I E N C E : W H AT I S M O R A L I N T U I T I O N ?


right or wrong. For example, one can challenge the ethics of the Old Testament God who told Abraham to kill his son Isaac. Given the dubiousness of the existence of God and the lack of evidence of a divine influence on people’s moral intuitions, we have little reason to expect theological theories of conscience to tell us much about the normative status of moral intuitions. 7.4.2 Moral intuitions as a priori Ethical intuitionism, the view that people can have immediate awareness of objective moral truths, is a distinguished position in ethical theory (e.g. Audi (2004); Huemer (2005); Moore (2003); Ross (1930); Stratton-Lake (2002)). We cannot provide a full discussion, but need to highlight how the emotional consciousness theory of moral intuition provides a major challenge to ethical intuitionism. We note first that ethical intuitionists have been largely silent about how human minds manage to have a priori knowledge of ethical truths. They seem to implicitly assume a dualistic view in which disembodied souls grasp non-natural objects, as in Plato’s theory that people can have objective knowledge of moral concepts by grasping the forms. But there is lack of evidence for souls and for the non-natural objects that they are supposed to grasp. Hence, ethical intuitionism does not provide much of an explanation for the cognitive and emotional aspects of moral intuitions. Ethical intuitionists who are not dualists are faced with the problem of explaining how a material brain can arrive at knowledge that is a priori. Intuitionism can explain moral agreement by hypothesizing that people can somehow grasp the same concepts or truths, but the common process by which they reach similar conclusions remains mysterious. Intuitionism is even less successful in explaining disagreement: if there are objective moral truths of which people can become intuitively aware, why do so few people manage to grasp the same truths? Stratton-Lake (2002) argues that moral propositions can be self-evident even if they are not obvious to every person, because of the difficulty of understanding them. But moral claims, unlike mathematical propositions, are usually not that difficult to understand; and many people who do understand claims proposed by some as self-evident have no difficulty rejecting them (e.g. abortion is wrong). Stratton-Lake (2002) further addresses the problem of disagreement in his discussion of Ross’s intuitionism. The charge with which he is concerned comes from Korsgaard (1996), that all an intuitionist can do in the case of disagreement is to assert dogmatically that a proposition is true. Stratton-Lake (2002) argues that Ross does not hold that ‘moral propositions are brute givens that cannot be questioned’ (2002: 114). Instead, they become self-evident after we have reflected on them, and have reached what Ross calls ‘sufficient mental maturity’. In other words, once we have given sufficient attention and reflection to a particular moral principle, we come to accept it as selfevident. Thus, disagreement can occur, and continue to occur until we have sufficiently reflected on the self-evident principles. As noted earlier, an intuitionist need not hold that self-evident principles are easily grasped or obviously true.



Nevertheless, there remains the possibility of unresolved moral disagreement. Ross seems to take it for granted that once one reflects on moral propositions, disagreement will dissolve, but dissolution may not occur. Moreover, the process by which one comes to disagree with another’s moral view remains unexplained on this account. Why does one agent think that a proposition is self-evident while another does not? An intuitionist can reply that perhaps one agent is tapping into certain self-evident principles, whereas another taps into different ones. This difference can account for how two agents may disagree about whether a moral proposition or set of principles is self-evident. However, this account does not give a full explanation of why disagreement continues to occur about the same issue. Consider again the issue of abortion. Often, agents have opposing, not just differing, views on the moral status of abortion. How could two agents discussing the same moral issue come to believe such drastically different viewpoints? Intuitionists may contend that one agent is unable to grasp the proposition, or to believe the truth of it, but why does this occur? If neither of the agents seems to hold a moral belief based on faulty information or biases, the cause of such disagreement is unknown. This sort of disagreement is difficult to explain if moral propositions are a priori, since there is no obvious reason why two agents cannot grasp the same propositions. Hence intuitionists cannot dodge the difficulty of explaining disagreement, whereas we showed that the emotional consciousness theory of moral intuition can robustly explain both agreement and disagreement. If ethical intuitionism were a viable theory, it could contribute substantially to the normativity problem of evaluating the validity of intuitions, marking at least some of them as yielding judgments that have a kind of objective, a priori truth. Unfortunately, intuitionism has no way of identifying which intuitions have the desired objective character, which is a major problem when intuitions so often differ. When one person is adamant that abortion is wrong and another person has an equally strong intuition that abortion is right, it is impossible to say which one has immediate awareness of some moral truth. In contrast, we argued that the emotional consciousness theory of conscience can help illuminate the conditions under which moral intuitions tend to yield objective judgments: when they are based on experience rather than on neural defects, poor education, or situational distortions. 7.4.3 Moral universal grammar A new, more psychological account of the origins of conscience is provided by the theory of moral universal grammar, according to which all humans are born with innate ethical principles (Hauser (2006); Mikhail (2007)). These rules are like the innate linguistic rules that Noam Chomsky claims are the basis for all human grammar. According to Chomsky, the diversity of human languages is built on a common set of innate grammatical principles that contain parameters that may be set in different ways to yield the grammars of different languages. Similarly, according to the theory of moral universal grammar, the diversity of moral views is built on a common set of

C O N S C I E N C E : W H AT I S M O R A L I N T U I T I O N ?


innate principles that contain parameters that allow for cultural diversity. Thus, moral universal grammar explains agreement and disagreement in ethical views in much the same way that linguistic theory explains commonality and divergence in human languages. As we remarked earlier, our emotional consciousness theory of conscience is in principle consistent with the view that moral judgments are partly innate, but we find this innateness hypothesis implausible. First, no one has identified anything like a brain module for moral reasoning. Whereas there are human brain areas that seem dedicated to linguistic processing (e.g. Broca’s and Wernicke’s areas), there are no areas that are known to play a special role in ethical judgments. The brain areas in the EMOCON model, such as the prefrontal cortex, the anterior cingulate, and the amygdala, are all involved in many kinds of human thinking. There is no apparent module for emotional consciousness, which arises from the interaction of multiple brain areas, all of which have diverse functions. Similarly, there is no apparent module for moral intuition, which also involves many brain areas known to be required for a wide range of cognitive and emotional processing. To paraphrase Bertrand Russell, postulating modules has all the advantages of theft over honest toil, and should not be done unless there is substantial neurological and psychological evidence for them. Second, it is not at all obvious that there are innate moral principles whose parametric variations can explain moral disagreements. Haidt ((2007); Haidt & Bjorklund (2008)) contends that cross-cultural studies suggest that humans have five major moral concerns: suffering, reciprocity, hierarchy, purity, and group boundaries. But he and the proponents of moral universal grammar have not shown how such concerns constitute the principles and parameters that are supposed to underlie ethical thinking. Moreover, these theorists have not accounted for the kind of individual and group moral diversity that we explained using the emotional consciousness theory of conscience. For example, some individuals (psychopaths) show little concern for human suffering, as do some cultures (Prinz (2009)): slavery and the suffering it causes were widely prevalent until the twentieth century. Hence, we are skeptical of the existence of a moral universal grammar that would provide an alternative theory of conscience to our emotional consciousness account, according to which moral intuitions largely derive from learning processes that are both cognitive and emotional. The analogy between moral intuitions and linguistic intuitions is too weak to support the hypothesis that conscience is innate. Our review of alternatives to the emotional consciousness theory of conscience has been unavoidably brief, but should serve to indicate ways in which alternative theories appear weaker. We have not discussed the process of reflective equilibrium, which views objective moral judgments as arising from an ongoing adjustment of intuitions and principles. Our account of moral intuition is broadly compatible with the reflective equilibrium metaphor, but we prefer an approach to ethical coherence that specifies in much greater detail how principles, judgments, actions, goals, evidence, hypotheses, and analogies can combine to maximize coherence (Thagard (2000): ch. 5; (2010):



ch. 10). Achieving reflective equilibrium only provides a justification of a system of principles and judgments if the system is more coherent than alternatives with respect to explanatory, deductive, conceptual, analogical, and deliberative constraints.

7.5 Conclusion We have contended that conscience is a kind of moral intuition, which is a kind of emotional consciousness. Moral psychology should be able to account for the cognitive/emotional character of moral intuitions, the existence of both ethical agreement and disagreement, and the normative status of ethical judgments. With respect to these requirements, the emotional consciousness theory appears to be superior to the main alternatives based on theology, a priori knowledge, and moral universal grammar. Of course, much remains to be done to explain in much more detail how the kinds of neural interactions in the EMOCON model generate diverse moral intuitions. We have not yet conducted simulations that show how the interactions between brain areas involved in various cognitive and emotional processes produce particular moral judgments. But we have outlined a descriptively adequate and normatively suggestive theory of conscience that should provide a basis for further naturalistic investigations of moral intuitions.1

References Audi, R. (2004) The Good in the Right: A Theory of Intuition and Intrinsic Value. Princeton, NJ Princeton University Press. Barrett, L.F. (2006) ‘Are emotions natural kinds?’, Perspectives on Psychological Science 1: 28–58. Batson, C.D. (1991) The Altruism Question. Hillsdale, NJ: Lawrence Erlbaum Associates. Blair, J., Mitchell, D.R., & Blair, K. (2005) The Psychopath: Emotion and the Brain. Malden, MA: Blackwell. Damasio, A.R. (1994). Descartes’ Error. New York: G.P. Putnam’s Sons. Ekman, P. (2003) Emotions Revealed: Recognizing Faces and Feelings to Improve Communication and Emotional Life. New York: Henry Holt. Greene, J. & Haidt, J. (2002) ‘How (and where) does moral judgment work?’, Trends in Cognitive Sciences 6: 517–23. Haidt, J. (2007) ‘The new synthesis in moral psychology’, Science 316(5827): 998–1002. —— & Bjorklund, F. (2008) ‘Social intuitionists answer six questions about moral psychology’. In: Moral Psychology, vol. 2, W. Sinnott-Armstrong (ed.), pp. 181–217. Cambridge, MA: MIT Press. Hare, R.D. (1993) Without Conscience: The Disturbing World of the Psychopaths Among Us. New York: Pocket Books.

1 We are grateful to Lorraine Besser-Jones, Patricia Marino, and two anonymous referees for comments on an earlier draft. This research is supported by the Natural Sciences and Engineering Research Council of Canada.

C O N S C I E N C E : W H AT I S M O R A L I N T U I T I O N ?


Hauser, M.D. (2006) Moral Minds: How Nature Designed Our Universal Sense of Right and Wrong (first edition). New York: Ecco. Huemer, M. (2005) Ethical Intuitionism. Houndmills, Basingstoke: Palgrave Macmillan. James, W. (1884) ‘What is an emotion?’, Mind 9: 188–205. Koenigs, M., Young, L., Adolphs, R., Tranel, D., Cushman, F., Hauser, M., et al. (2007) ‘Damage to the prefrontal cortex increases utilitarian moral judgements’, Nature 446(7138): 908–11. Korsgaard, Christine M. (1996) Creating the Kingdom of Ends. New York: Cambridge University Press. Langston, D. (2006) ‘Medieval theories of conscience’, The Stanford Encylopedia of Philosophy. Retrieved May 31, 2007, from Mikhail, J. (2007) ‘Universal moral grammar: Theory, evidence and the future’, Trends in Cognitive Sciences 11(4): 143–52. Moll, J., Zahn, R., de Oliveira-Souza, R., Krueger, F., & Grafman, J. (2005) ‘The neural basis of human moral cognition’, Nature Reviews Neuroscience 6(10): 799–809. Moore, G.E. (2003) Principia Ethica. Cambridge: Cambridge University Press. Naqvi, N.H., Rudrauf, D., Damasio, H., & Bechara, A. (2007) ‘Damage to the insula disrupts addiction to cigarette smoking’, Science 315(5811): 531–4. Nussbaum, M. (2001) Upheavals of Thought. Cambridge: Cambridge University Press. Prinz, J. (2004) Gut Reactions: A Perceptual Theory of Emotion. Oxford: Oxford University Press. ——(2009) ‘Against moral nativism’. In: Stephen Stich and his Critics, D. Murphy & M. Bishop (eds), pp. 167–89. Oxford: Blackwell. Raine, A. & Yang, Y. (2006) ‘Neural foundations to moral reasoning and antisocial behavior’, Social Cognitive and Affective Neuroscience 1: 203–13. Rolls, E.R. (2005) Emotion Explained. Oxford: Oxford University Press. Ross, W.D. (1930) The Right and the Good. Oxford: Clarendon Press. Rousseau, J. (1960) The Confessions. J.M. Cohen (trans.). London: Penguin. Sander, D., Grandjean, D., & Scherer, K.R. (2005) ‘A systems approach to appraisal mechanisms in emotion’, Neural Networks 18: 317–52. Scherer, K.R., Schorr, A., & Johnstone, T. (2001) Appraisal Processes in Emotion. New York: Oxford University Press. Stratton-Lake, P. (2002) (ed.) Ethical Intuitionism: Re-evaluations. Oxford: Oxford University Press. Thagard, P. (2000) Coherence in Thought and Action. Cambridge, MA: MIT Press. ——(2006) Hot Thought: Mechanisms and Applications of Emotional Cognition. Cambridge, MA: MIT Press. ——(2007) ‘The moral psychology of conflicts of interest: Insights from affective neuroscience’, Journal of Applied Philosophy 24: 367–80. ——(2010) The Brain and the Meaning of Life. Princeton, NJ: Princeton University Press. —— & Aubie, B. (2008) ‘Emotional consciousness: A neural model of how cognitive appraisal and somatic perception interact to produce qualitative experience’, Consciousness and Cognition 17: 811–34. —— & Kroon, F.W. (2006) ‘Emotional consensus in group decision making’, Mind & Society 5: 1–20.

8 Empathy and Empirical Psychology A Critique of Shaun Nichols’s Neo-Sentimentalism Lawrence Blum

Experimental philosophy often characterizes and belittles traditional philosophical inquiry as ‘armchair’ and counterposes it to going out in the world to see how things actually operate by running experiments or at least drawing on experiments run by others. Shaun Nichols is a leading proponent of the experimental philosophy method. His book Sentimental Rules: On the Natural Foundations of Moral Judgment (Nichols (2004)) makes extensive use of psychology experiments and findings in the course of building an argument about the nature of and close connection between emotion and morality. We can roughly distinguish two traditions in philosophical thinking that assert a close relation between emotion and morality, each rejecting a different form of rationalism. The first, which has come to be called ‘sentimentalism’, sees emotion, rather than reason, as the source of moral judgment. This tradition originated with the Scottish moralists, especially Hume and Hutcheson, and has had something of a revival in contemporary ethics (e.g. Blackburn (1998); Gibbard (1990); D’Arms & Jacobson (1994)). Nichols locates himself in this tradition, neo-sentimentalism, while staking out a distinctive position within it (discussed below). This view rejects a rationalism that takes the form of privileging reason as the foundation of (either as source, or as a mode of access to) moral judgment. The second tradition sees emotion as required for the human capacities necessary for leading a moral life, or being a moral person—that is, required for moral motivation or moral responsiveness more generally. Schopenhauer is an important exponent of this view. In his On the Basis of Morality Schopenhauer criticized Kant’s view that reason or rationality is fundamental to moral capacity—that reason alone (or predominantly) provides the motive for moral action—and suggested instead that the emotion-based motive of compassion was fundamental (Schopenhauer ([1841]/1965)). This tradition of concern with motive, action, and response is less concerned with moral judgment. Nichols does not explore or acknowledge this second tradition of thinking about the relation between emotions and morality; but parts of his argument bear on it.



Nichols agrees with Hume that moral judgment rests ultimately on sentiments or emotion rather than reason. He says that ‘core moral judgment’ involves an affective mechanism of response to distress, as well as a system of rules that he refers to as a ‘normative theory’. Core moral judgment involves distinguishing the conventional from the moral—using one’s fork to eat vs. not harming others. The ability to make this distinction shows up quite early in children. Nichols argues that the distinctive force and status of moral rules (among rules in general) is dependent on their eliciting strong affect, and this can happen only in persons who are able to respond emotionally to the suffering of others. For example, psychopaths are defective in their ability to make moral judgments, although they recognize the social existence of norms that forbid causing harm, and they also fail to respond emotionally to suffering in others. Nichols suggests that the emotional deficiency is the source of the failure to appreciate the moral character of the norms. I will argue that Nichols’s view suffers from several deficiencies: (1) It operates with an impoverished view of the altruistic emotions (empathy, sympathy, concern, compassion, etc.) as mere short-term, affective states of mind, lacking any essential connection to intentionality, perception, cognition, and expressiveness. (2) He fails to keep in focus the moral distinction between two very different kinds of emotional response to the distress and suffering of others—other-directed, altruistic emotions that have moral value, and self-directed emotional responses, such as personal distress, that do not. (3) Nichols is correct to see morality as requiring affectivity, the capability of emotional response to others; but his incorrect view of altruistic emotions (and of emotions in general) leads him to misstate the connection between morality and emotion. (4) Nichols fails to recognize Schopenhauer’s form of anti-rationalism as distinct from Humean sentimentalism; some of his arguments presented to support the latter instead lend support to the former. (5) Finally, while agreeing that moral philosophy is strengthened by knowledge of empirical psychology, I suggest that the foregoing failures of Nichols’s argument are partly due to his misuse of particular empirical results and findings, his being over-enamored of empirical psychology, and possibly in part to a weakened commitment to the distinctive contribution that the humanistic methods of philosophy make to our understanding of the moral enterprise.

8.1 Nichols on Empathy as Having the Same Feeling as the Other In the second chapter of Sentimental Rules, ‘Sparks of Beneficence: The Varied Emotional Responses to Suffering in Others’, Nichols turns his attention to the affective system of response to suffering for the emotion-based component of his sentimentalism. He also says he is interested in the underpinnings of altruistic motivation, presumably because altruistic motivation is a core dimension of morality. ‘My goal is restricted to the project of determining the cognitive and affective mechanisms underlying basic altruistic



motivation’ (2004: 33). These two foci pull Nichols in two different directions in the chapter. The focus on altruistic motivation as a distinctive type of motivation seems addressed to a Schopenhauer-like concern with moral motivation, on the assumption that altruistic motivation is a type of moral motivation. By contrast, the focus on emotional responses to distress blurs the distinction between moral and non-moral sentiments and emotions, since all Nichols needs for his sentimentalism is that the emotional response to distress be a strong one, not that it be distinctly altruistic, or otherwise moral, in character. Nichols begins by mentioning empathy, regarded by many psychologists as providing the motivation for pro-social behavior, psychologists’ term for behavior intentionally aimed at helping others (2004: 31). He defines empathy as ‘“a vicarious sharing of affect” or an emotional response in which the emotion is “congruent with the other’s emotional state or situation”’ (ibid.). Although it is not entirely clear whether Nichols regards this definition, which he attributes to a leading social psychologist in the empathy field (Eisenberg), as entirely satisfactory, he never proffers a different one, and appears to use this one in his subsequent argument.1 Nichols wants to make the point that we have a range of emotional responses to distress other than empathy, and these need to be recognized both for their own intrinsic interest, and also because of their role in providing what he sees as the emotional underpinning of moral norms against causing harm. The most important of these other responses are: ‘emotional contagion’ (being caused to have the same mental state as another by being in that individual’s presence), and ‘personal distress’ (being caused to have one of a number of mental states, e.g. alarm, anxiety, upset, distress by another’s distress). We ordinarily think of empathy as a moral phenomenon, in contrast to these other responses (which Nichols sometimes calls ‘self-directed’); but Nichols’s definition fails to capture that dimension. Having the same affect as another, and having one appropriate to another’s emotional state, are morally indifferent. If Don is sad and Juan is also sad, Juan’s sadness by itself is of no moral significance (in relation to Don), nor is it if we add the condition that Juan is caused to be sad by Don’s being sad. This is because without Juan’s recognition of or concern for Don’s sadness the connection to the other required for morality in this context is absent. Nor would merely having the same feeling, as described, normally be called ‘empathy’ in the usual use of that term. Empathy involves a feeling for another, that is, a feeling whose object is another’s state of mind or situation. When Nichols discusses ‘self-directed’ feelings, such as personal distress, he draws that terminology from the psychologist Batson, who introduces it to contrast with other-directed emotion. Nichols appears to follow Batson’s implication that this is a contrast with moral significance. A self-oriented feeling normally lacks moral significance, while feeling directed toward others may possess it.2

1 ‘In Eisenberg’s view, pure empathy is not other-oriented’ (Eisenberg (2000): 671). Eisenberg is referring to herself, and thus to her view, in the third person here. 2 Some self-oriented feelings do have moral significance, for example, a feeling of outrage at having been denied one’s rights or one’s dignity having been violated. But unless it involves a moral principle,



Empathy is an instance of a larger category of emotion that one might call (following Max Scheler) ‘fellow-feeling’ (Mitgefühl) (Scheler (1954)). Compassion, sympathy, and concern are other species of fellow-feeling. In fellow-feeling, we recognize that the other has a particular state of mind or situation, and we have a regard for her well-being with respect to that state of mind or situation. But we do not, as part of fellow-feeling itself, adopt or come to have that state of mind ourselves. If my friend is disappointed that she failed to attain a professional honor that she hoped for, I may be empathetic toward her with regard to this disappointment. But in being so I am not myself disappointed; I am not in a state of disappointment. I am disappointed for her; that is, I am acutely aware of her disappointment and perhaps I envision her disappointment firmly in my own mind. This is entirely compatible with my feeling quite satisfied myself, feeling that my own life (in contrast to my friend’s) is going very well. In sympathy, I also have some affect myself, but my state of mind is directed toward another in light of (what I take to be) something bearing on her well-being. The difference between sympathy and empathy is not entirely clear, and the common usage of these terms may not consistently mark a single distinction. Perhaps empathy implies a stronger and more acute imagining of what the other is going through than does sympathy; but neither involves actually having the same state of mind as the other, contrary to Nichols’s definition.

8.2 Fellow-Feeling as Directed Toward the Other’s Well-Being, and Some Faulty Accounts All the different forms of fellow-feeling have the following features: (1) they involve affectivity on the part of the observer; (2) they involve having as an intentional object another person in light of her state of mind or situation (or the state of mind or situation itself );3 and (3) the aspect of the person’s state of mind or situation that is conceptually tied to the fellow-feeling is the other person’s well-being. Fellow-feeling’s essential focus on the other’s well-being brings with it a fourth feature, a motivational dimension. Someone who has fellow-feeling for another has some concern for the other’s well-being, and thus an at least minimal ceteris paribus motivation to help the other. In any given situation there may be several motivational pulls, some of them running contrary to such help, so fellow-feeling does not always self-directed feeling is morally neutral. Batson’s terminology of ‘other-directed’ is not meant to include every feeling or emotion directed toward others rather than toward the self or the self’s situation. For negative emotions such as dislike, malice, or hatred are directed toward others, yet are not morally good. What he intends by ‘other-directed’ emotions must involve a positive regard for the other’s well-being in some way— a pleasure in the other’s good fortune, a worry about impending harm to her, or a concern for her distress. Schopenhauer captured these distinctions in his threefold differentiation of fundamental motives as they bear on morality—egoism (desire of one’s own weal), malice (desire of others’ woe), compassion (desire of others’ weal) (Schopenhauer ([1841]/1965): 145). Only compassion, according to Schopenhauer, is a moral motive, a motive with moral worth, and this seems to be Batson’s view, which Nichols implies that he adopts. 3 We might say ‘I empathize with Reginald in light of his sadness’ or ‘I empathize with Reginald’s sadness’. These seem very close in meaning, if not identical.



translate into helping. Nevertheless, if someone says that he empathizes with Rae with regard to a difficult emotional situation Rae is currently going through, yet is unwilling to proffer a minimal comfort that would cost him nothing and would seem an appropriate and helpful action for Rae, the idea of his having empathy for Rae starts to lose its grip.4 In this sense there seems to be a conceptual connection between having fellow-feeling for another and motivation to help that other.5 What is the affective state involved in fellow-feeling if it is not the same as or homologous to that of the person whose state of mind or situation is its object? More precisely, what is its felt quality? What is the phenomenological quality of being ‘sad for X’ when one is empathetic toward X in relation to her sadness, that differs from ‘being sad’ itself? I am not sure how to think about this. Some accounts suggest a ‘vicarious’ sadness; but I am not sure if this is not simply another way of saying one is ‘sad for’ the other and so does not serve to explain it. Others suggest the feeling is a less intense form of the feeling the other has. Perhaps. But what I want to emphasize here is that fellowfeeling, including empathy, is both affective and intentional.6 And yet we may not always be able to characterize the phenomenological feel of the fellow-feeling better than that. So the precise characterization of that phenomenal feel may not be central to our understanding of empathy, sympathy, and care. And that particular phenomenological feel may be relatively insignificant for the moral significance of fellow-feeling. Moreover, although fellow-feeling is necessarily affective, it does not necessarily involve an occurrent affective state at every moment of which it is true to say that one has fellow-feeling for X. This is most obviously true in the verbal forms of fellowfeeling: ‘I empathize with X’, ‘I sympathize with X’, ‘I am concerned about X’. Those characterizations can hold of me for an extended stretch of time during which I do not necessarily experience feeling-states of empathy, sympathy, or concern for the entire duration. But the same is true of ‘have empathy (sympathy, concern) for X’. Though these attribute the noun form of the emotions to the subject/observer, they do not entail that the subject is in a particular affective state for the entire period of which the attribution is true. As Peter Goldie notes, there is a difference between having an emotion and having an episode of emotional experience (2001: 2). We might use the single term emotion for either one, but we would more normally express the episodic form by saying ‘I am feeling sympathy (empathy, concern) for Midori at the moment’, while the former is expressed as ‘I have sympathy for Midori’ or ‘I sympathize with

4 In many situations, there may be no motivation to help because help is impossible or the observer is simply in no position to help and realizes this from the get-go. 5 I think ‘sympathy’ is sometimes used in a very minimal way to mean little more than ‘acknowledge’ but with the barest implication of caring or concern for the other. ‘Sure, I sympathize with Ashok. It’s too bad what’s happening to him’—but with the implication that I am not really very concerned and am not prepared to lift a finger to help him. But normally, I think we would question the accuracy or sincerity of someone’s saying that he had or felt sympathy for someone if he was unwilling or uninclined to take any steps to help when the cost to himself is negligible. 6 Peter Goldie (2001) emphasizes this point.



Midori’. In this latter use there may in many cases be no phenomenological feel in the state of mind of the one who has fellow-feeling.

8.3 Missing the Intentional, Other-Directed Character of Empathy and Fellow-Feeling As mentioned, in his definition of empathy, Nichols is reporting a widespread view of empathy within experimental psychology.7 His (and Eisenberg’s) view is not alone in failing to capture the distinctive, morally significant, features of fellow-feeling, and of empathy specifically. In a thoughtful analysis of empathy, the philosopher Nancy Snow says that it involves the observer (O) feeling the same feeling (S) as the victim (V), and feeling it because V feels it, knowing that she feels the same feeling as V.8 In support of the observer and victim’s having the same emotion, a feature I am challenging, Snow gives the following example: ‘When Barbara feels the same emotion as Bob, they both feel sad about the same fact, Bob’s father’s death’ (1991: 66). But while Barbara might feel sad about Bob’s father’s death, that is not part of her empathy for Bob. The latter requires only that she feel badly for Bob because of his sadness over his father’s death. If she knew nothing of Bob’s father, and thus had no basis for personal sadness at his death, she would or could still feel empathic sadness for Bob. Other accounts of empathy also fail to capture the intentional and other-directed structure of empathy that is a necessary element of its moral character. Sober and Wilson recognize the distinction between feeling X for someone, and feeling X full stop (Sober & Wilson (1998): 234). They express the difference by saying that Lakisha’s empathy for Victor entails that Victor’s feeling S causes Lakisha to feel S. (Snow’s account contains this feature also.) But this causal relationship does not capture the intentional one it is meant to. Lakisha might know that Victor is depressed, and his being depressed (or, Lakisha’s knowing that he is depressed) causes her to be depressed. Lakisha might say that his depression ‘brings her down’. (This is an example of ‘emotional contagion’, mentioned earlier.) Here Lakisha and Victor have the same emotion and his emotion has caused hers. But this is not her empathy for him. The two could be combined. Lakisha could genuinely have empathy for Victor in his depression, and could also come to be depressed herself because of his depression. It may not always be easy to distinguish these in practice. When I ‘absorb’ my friend’s anxiety or sadness, as we sometimes say, this can have a quality both of emotional contagion and of empathy. But this ‘absorbing’ is less likely to occur with regard to other objects of empathy, such as disappointment or physical pain. In any case, the contagion and the empathy remain two distinct psychic phenomena. 7 Here is a similar definition from a recent article in psychology: ‘empathy (i.e. a vicarious emotion that results in feeling the same as another)’ (Carlo, PytlikZillig, Roesch, & Dienstbier (2009): 277). 8 ‘Observer’ and ‘victim’ are Nichols’s terms for the person having the empathy and the person toward whom she has it and I follow his usage in this paper.



In his important discussion of various forms of fellow-feeling, Scheler notes an important moral difference between emotional contagion and fellow-feeling. If we come to be ‘infected’ by the other’s state of mind—say sadness or depression—this can well have the effect of crowding out genuine fellow-feeling for the other that is involved in empathy (Scheler (1954): 17). I may become so consumed by or at least stuck in my own sadness caused by the other’s sadness that I am unable to muster the other-directed attention and concern required for genuine empathy, or other fellow-feeling for her. Martin Hoffman, a leading psychologist of empathy, himself provides a different, but still faulty, account of empathy, failing to capture its intentional structure. He (rightly) rejects the common psychologists’ (and Nichols’s) view of empathy as ‘one’s feeling matching the other’s feeling’. But he puts in its place ‘the involvement of psychological processes that make a person have feelings that are more congruent with another person’s situation than with his own situation’ (2000: 30). He gives the example of feeling empathic anger on seeing someone verbally attacked when the attackee might be feeling sad or disappointed. The anger could be regarded as more appropriate to the victim’s situation than is the victim’s own sadness, and would not be appropriate to the subject’s situation which (let us assume) provides no basis for anger. But were the observer to feel anger and to believe that this is the emotional response the victim should appropriately have to her own situation, this would not be empathy. The observer’s empathy is directed toward the victim. It is not a self-standing feeling that is appropriate to the victim’s situation. Hoffman’s definition implies that the empathizer is in a way confused, having feelings inappropriate to his own situation but appropriate to those of another.9 This faulty definition does not vitiate Hoffman’s many important findings about and speculations regarding empathy, which presuppose that it is indeed essentially directed toward others.

8.4 Mental State or Life Situation as the Target of Fellow-Feeling So Nichols joins some other philosophers and psychologists in failing to articulate the intentional structure of empathy that helps to bring out its moral significance. Nichols’s account of empathy goes wrong in another way as well. This lies in his privileging the victim’s state of mind over her actual life situation in what it is that altruistic sentiments are directed towards. This comes out in his discussion of ‘mindreading’ necessary to activate ‘basic altruistic motivation’ (2004: 33). I must explain Nichols’s background discussion of mindreading before getting to this issue.

9 Schopenhauer recognizes this very error, in a criticism of the philosopher Cassina: ‘His view is that compassion arises from an instantaneous deception of the imagination, since we put ourselves in the position of the sufferer, and have the idea that we are suffering his pains in our person. This is by no means the case; on the contrary, at every moment we remain clearly conscious that he is the sufferer, not we; and it is precisely in his person, not in ours, that we feel the suffering, to our grief and sorrow’ (([1841]/1965): 147).



By ‘mindreading’ Nichols means the observer’s recognition or belief that the victim has a particular state of mind. Nichols sets out to defend the view that attribution of distress to the victim—that is, mindreading—is required for altruism. He defends this as a middle position between two others—(a) that no mindreading at all is necessary, and (b) that more sophisticated forms of mindreading (for example, the ability to imagine oneself in the other’s situation) are necessary for altruism. Against the ‘no mindreading’ view, Nichols cites experiments by Batson, an influential social psychologist researching altruism, in which a subject watches (what she takes to be) shocks administered to another. There are two different scenarios—one in which the subject can easily escape the situation and another in which she is compelled to stay for a more extended period of time. What Batson finds is that there are some people who will help the distressed person even when they are in the ‘easy escape’ situation, i.e. when they could relieve their own (i.e. ‘personal’) distress by escaping from the situation. Nichols reports this result by saying ‘so there is an important kind of altruistic motivation that cannot be satisfied by escaping the situation’ (2004: 38). Nichols implies that some people act altruistically to help when they could escape, while others act altruistically to help the victim as a way of relieving their own personal distress (and some will escape the situation as a way to relieve their own distress). But those who help the other in order to remove the cause of their own distress are not acting altruistically; they are simply engaged in helping behavior for self-interested reasons. What Batson’s experiment shows is that not all helping behavior is motivated by an attempt to relieve the observer’s own distress, not that the latter is actually a form of altruistic motivation different from a desire to help the other for her own sake.10 To put it another way, Batson’s experiment suggests that some persons are motivated by empathy to relieve the other’s distress even though they could escape the situation without doing so. There is by definition no kind of altruistic motivation that can be satisfied by escaping the situation, so an experiment cannot establish that there isn’t.11 10

Nichols at one point accepts that altruism refers to a motivational state and not simply an intentional helping of the other. He says, ‘Altruistic behavior is one form of prosocial behavior, but prosocial behavior also encompasses selfishly motivated behavior that helps others. For example, intentionally helping a stranger merely to impress onlookers counts as prosocial, though presumably not altruistic’ (2004: 31 n. 1). So this definition is out of line with his use of altruism in the quote about Batson’s experiment in the text. 11 In a gloss on Batson’s experiment, John Doris and Stephen Stich (whose work is associated with the experimental philosophy movement) say that Batson does not see his result as conclusively showing the existence of altruism. [T]here are also a variety of egoistic routes by which empathy might lead to helping behavior. Perhaps the most obvious of these is that empathy might simply be (or cause) an unpleasant experience, and that people are motivated to help because they believe this is the best way to stop the unpleasant experience that is caused by someone else’s distress. (Doris & Stich (2006): 23). If empathy is directed toward the other and carries with it some motivation to help the other, it cannot simply be ‘an unpleasant experience’ that could motivate the observer to help only as a means of relieving that unpleasantness. Doris and Stich are misled by the same faulty use of Batson’s experiments as Nichols is. Perhaps empathy can wane and be replaced by the mere unpleasantness of viewing the other’s distress (i.e. personal distress). But empathy itself cannot give rise to such personal distress. If the directedness of altruistic concern toward the other is or becomes absent, there is no longer any empathy.



Nichols wants to argue that some mindreading is necessary for altruistic motivation; but he does not need Batson’s or any experiment to establish this, nor would they be relevant. ‘Minimal mindreading’ is required for altruistic motivation by definition. Unless the observer knows that the other is in a state of distress (or need, or possible benefit, etc.), he cannot be motivated by altruism, as altruism requires recognition of the other’s state (and a direct motivation to improve it for its own sake). But that recognition precisely is mindreading (of at least the minimal form Nichols desires). Let me return to the issue of Nichols’s privileging the victim’s occurrent state of mind over her life situation as the object of fellow-feeling. He discusses another challenge to the minimal mindreading account. This is Sober and Wilson’s view that sympathy sometimes motivates altruistic behavior; that sympathy is different from empathy; yet that sympathy does not require mindreading. The first point is intuitively plausible, and Nichols does not dispute it. In support of the latter two points, Sober and Wilson construct an example of Walter discovering that, unbeknownst to his friend Wendy, her husband has been unfaithful to her. Walter could sympathize with Wendy even though he is not directing that sympathy to Wendy’s state of mind, since she is not aware of her husband’s infidelity (Nichols (2004): 38; Sober & Wilson (1998): 234–5). As Sober and Wilson put it, ‘You can sympathize with someone just by being moved by their objective situation; you need not consider their subjective state’ (Sober & Wilson (1998): 236). So being moved by an objective situation would be a genuine example of altruistic motivation seemingly without any mindreading. Nichols complains that Sober and Wilson provide no evidence for their ‘objective situation’ view (Nichols (2004): 39). But there seems to be a significant burden of proof issue that bears on Nichols’s privileging of social psychology experiments as the basis for his philosophical claims. For it seems clear that we are quite often moved by the plight of others just from knowing that plight and not from giving thought to, nor perhaps even knowing anything about, their state of mind, nor imagining what their state of mind is.12 If I receive a fund-raising letter from an international aid organization with a photo of a child suffering from AIDS, I might be moved to give money without thinking about what that child is feeling about her disease. Knowing she has AIDS might well be sufficient to move me to give. This is true in our more immediate environment as well. I might have sympathy for a colleague, with whom I am not intimate, who I learn has a child with mental health problems. I might express that sympathy if I run into the colleague, or even drop a note to the colleague—yet without giving thought to the colleague’s actual state of mind. It is enough that I know that her plight is a bad one for her.13

12 Nichols hypothesizes that the state-of-mind theory can be saved if Walter imagines what Wendy would (be likely to) feel were she to discover the infidelity. 13 This is not to say that it is typical to fail to give any thought to a person’s state of mind of whom one only knows her plight. Especially when we are emotionally close to someone, we are likely to do so.



So it certainly seems possible to have sympathy for someone’s plight without the ‘mindreading’. It would seem to me that the burden is not on Sober and Wilson to produce experimental evidence for this possibility, but for Nichols to show why our ordinary experience of this should not be credited. The argument here is not mere ‘armchair’ pumping of conceptual intuitions but simply reminding ourselves of a familiar experience (that we are sometimes moved by sympathy for someone’s objective situation without giving thought to the occurrent state of mind of our victim). This is a familiar and seemingly indispensable form of philosophical argument; and it seems on as sure a footing as a particular laboratory experiment in psychology. So Sober and Wilson seem correct to claim that we can be moved by sympathy (and perhaps other forms of fellow-feeling as well) by the ‘objective situation’ of others, not by their occurrent state of mind.

8.5 Young Children and ‘Core’ Cases of Altruism Nichols gives a second argument against the importance of sympathy in his project of understanding altruistic motivation that bears on his privileging of the occurrent state of mind. This depends on his use of examples of children’s altruism, an important part of his argument. Nichols says he is setting out ‘core’ cases of altruism (though he says new evidence might lead us to revise what we regard as core cases). He cites these examples: A twelve-month-old child retrieves a cup for a crying friend; a two-year-old accidentally harms a friend, looks concerned, and offers the hurt friend a toy (2004: 34). Referring to these examples (among others) Nichols says, ‘even if Sober and Wilson’s “sympathy” does exist, they provide no reason to think that it explains anything like the core cases of altruism with which we began’ (2004: 39). There are several problems with this argument. First, Nichols says that ‘children only begin exhibiting the characteristic signs of sympathy after the first birthday, and at this age, they probably have some rudimentary mindreading skills’ (ibid.). He takes this to show that the capacity for sympathy requires the capacity for mindreading. That is, Nichols treats the view that the general capacity for sympathy requires the general capacity for mindreading as equivalent to or at least as requiring the view (disputed by Sober and Wilson) that on every single occasion on which someone has sympathy for another, that person attributes a particular state of mind to the victim. We can think that a person who was completely unable to recognize what others are feeling is also incapable of sympathy. But this does not entail that on every occasion in which a person has sympathy he attributes a state of mind to the other. This point leads to a second, which concerns what it means to say that the childhood examples Nichols gives are ‘core’ examples of altruism that, Nichols says, Sober and Wilson fail to explain. At one point he takes it to mean cases that emerge early in children yet are pervasive among adults, and says that ‘comforting and helping others in distress’ are core in that sense (2004: 35). But the impression is also given that by ‘core’ Nichols also means uncontroversial cases of altruism. Yet Nichols’s examples might be typical of



childhood altruism in this latter sense, while lacking some central features of fully developed adult altruism. Indeed, it seems plausible that as children develop, they become more and more able to differentiate a victim’s plight from his current state of mind. So even if adults comfort and help others in distress, their doing so might have important features lacking in childhood cases of such, but also there might well be pervasive features of adult altruism that do not show up in very young children at all. Martin Hoffman, the psychologist of empathy mentioned earlier, puts the point this way: At some point in development, owing to the emerging conception of self and others as continuous persons with separate histories and identities, children become aware that others feel joy, anger, sadness, fear, and low esteem not only in the immediate situation but also in the context of a larger pattern of life experience. Consequently, although they continue to feel empathically distressed in response to another’s immediate pain or discomfort, they can also respond with empathic distress to what they imagine to be the other’s chronically sad or unpleasant life condition. (Hoffman (2000): 80)14

For adults, it would seem that the ‘core’ object of altruistic concern is not the state of mind as such but the victim’s plight. The state of mind is of altruistic concern insofar as it is a part of, or a symptom of, her actual or threatened ill-being. Normally the state of mind and the plight form a cohesive whole to which the observer responds with altruistic concern (when appropriate). Typically, our response to others, even when focused on an occurrent state of mind, takes the context of their life situation, as it takes place over time, into account, even if only implicitly. If my friend reports sadness about her boyfriend leaving her, but seems quite cheerful and behaves with a new sense of unburdenedness, I might infer that she is not really sad about her boyfriend leaving her, but only thinks she is; or that she is, but only to a small or superficial degree.15 But a person’s current state of mind can sometimes be out of line with her plight, as Sober and Wilson’s earlier example of the wife who does not know her husband is cheating shows. And when that is the case, it is clearly the plight that trumps the state of mind with regard to (appropriate) altruistic concern.16 It is plausible to adopt Hoffman’s developmental view cited above that part of moral maturity consists in coming to view the other’s life situation as the necessary context for a focus on her well-being, and that altruistic concern has this focus. Occurrent distress

14 Earlier I mentioned Hoffman’s potentially misleading use of ‘empathic distress’ in this passage. (See page 176.) 15 Eva Johansson (2008) argues that even young children are responding to the shared ‘life-world’ of other children—to their embodied selves, their socially situated selves, not merely to what one child takes to be another’s inner states. (The notion of ‘life world’ is taken from the phenomenological tradition, especially, in Johansson’s case, that of Merleau-Ponty.) The examples below (18) illustrate Johansson’s point, showing young children being concerned about another’s situation, not simply her occurrent state of mind. 16 Nevertheless, a state of mind can also be of moral interest independent of the victim’s objective situation. If my friend feels a sense of loss about something the loss of which I regard as in her long-term best interest (e.g. what I regard as a destructive relationship for her), it is appropriate for me to feel empathy for her sadness; yet also to have the appropriate fellow-feeling—e.g. relief, gladness—directed toward her overall situation. This is because her sense of loss is itself a genuine hurt she is feeling in the present. But it is still the state of mind as constitutive of well-being that is the appropriate object of fellow-feeling.



such as the children respond to altruistically in Nichols’s examples is expanded in later personal development to this larger canvas of the life situation. Indeed, it is worth noting that the examples from children that Nichols has chosen to present are biased in favor of his ‘state of mind’ account. The article from which he draws the examples contains other examples that have an ‘objective situation’ rather than a ‘state of mind’ focus and thus are closer in spirit to the life situation focus. For example, in one a 3½-year-old is playing with his 6-month-old sister on the floor. He sees a safety pin and takes it to his mother, saying that it would hurt his sister if she touched it. In another, a 3-year-old gives a friend her own Donald Duck hat because the friend has (recently, but not at the current moment) lost her own (Boston Celtics) cap. Both these cases involve a young child’s response to another’s objective situation rather than to a definite current state of mind.17 So Nichols has been selective in choosing his experimental examples. A fuller look at those examples does not support his privileging the victim’s state of mind over her plight as the ‘core’ object of fellow-feeling, nor, therefore, his argument for demoting the importance of sympathy in an understanding of altruistic concern (because of seeing it as not ‘core’). Following his presentation of the childhood cases to get to adult cases, Nichols says the following: ‘The clearest real-life examples of altruistic behavior in adults come from work on helping behavior in social psychology’ (2004: 34). To illustrate this, he describes in some detail a laboratory experiment of Clark and Word (1974) (the details are not relevant here) that finds a large percentage of people offering aid when no one is watching them. It is striking that Nichols finds such a laboratory experiment to be the ‘clearest real-life example’ of altruism. It would seem that the clearest real-life examples are those that come from our lives in our world. We can all produce examples in our own experience of people helping others from what every evidence available to us suggests is altruistic motivation. And some such real-life examples have been extensively studied, for example, rescuers of Jews during the Holocaust (Oliner & Oliner (1988); Monroe (1997)). As real-life examples of altruism they are likely to serve us better as material for reflection than social psychology experiments in laboratory settings. When we engage in that reflection, we are not simply ‘sitting in an armchair’. We are drawing on our moral sensibilities, our experience, and our reflective capacities to engage with human phenomena that we recognize simply because we are morally competent human beings living in a world of other human beings. It is noteworthy that, when Nichols talks about children, he does not cite experimental data but draws on observation of actual children as ‘core’ cases. It would seem that reflection on adult morality and adult altruism should similarly begin with cases that we can recognize from our world. In the scholarly community of experimental

17 Blum ([1986]/1994): 187). It is possible that in the second example given the observer infers a kind of distress in the victim; but even if so, the child’s life situation is part of the observer’s reason for responding the way she does, and Nichols’s account of childhood concern does not allow for this.



psychology, the Clark and Word experiments (and Batson’s as well) may well constitute important progress, because they challenge a reigning paradigm in that community that all human behavior is ultimately egoistically motivated.18 But the understanding yielded by these experiments cannot be equated with nor be paradigmatic of an understanding of the character of human altruism in all its complexity, such as philosophical reflection with its distinctive ways of surveying human capabilities and the human condition can provide. The nub of Nichols’s defense of minimal mindreading lies not in refuting the ‘no mindreading’ view, which I have argued is a non-starter, but in arguing that nothing more robust than minimal mindreading is necessary for altruistic motivation. I think his argument on this point is largely successful and I will present it briefly. At the same time, I think here too, Nichols operates with an impoverished view of fellow-feeling, missing its intentional structure, its focus on the other, its difference from an occurrent affective state, and, more generally, its organic connection to human relationality. The main form of robust mindreading Nichols considers is ‘perspective taking’, by which he means imagining oneself as having the mental states of another (2004: 8), or as taking the perspective of another person (2004: 9). These two characterizations are not really equivalent, but I understand Nichols to mean imaginatively projecting oneself into someone else’s circumstances as a way of envisioning what she feels. Nichols recognizes that perspective-taking in this sense can be a path to empathy. But his point is that empathy can be reached without it. In particular, it is implausible to think that the children whose empathy he has described have engaged or are capable of engaging in such a cognitively sophisticated process. If we are willing to accept these children as cases of fellow-feeling of some sort, then perspective-taking cannot be a necessary condition for empathy.19 This argument seems correct to me.

18 Eisenberg says that in the 1980s ‘there was considerable interest [among psychologists] in why people sometimes help others at a cost to themselves and whether truly selfless altruism exists’, contrasting this with an earlier period in which egoistic theories of human motivation were more dominant (2000: 672). See Scott & Seglow (2007): ch. 4 for a recent, brief account of the emergence of a recognition of genuine altruism (not merely subtle or disguised forms of pursuit of self-interest) in social psychology. Batson is reported to have done an about-face, originally holding to the egoistic hypothesis (Barasch 2009: 168). An earlier collection, Mansbridge (1990), lamented the dominance of egoistic paradigms in several different disciplines, including psychology, and tracked a then recent emergence of alternative views that allowed for genuine altruism. 19 Goldie makes the important point that imagining oneself in another’s place is, in contrast to empathy, by itself altruistically inert. It involves no necessary concern for the other, but can be engaged in from a variety of motives, including a malicious desire to savor the distress of the other (Goldie (2001): 215). Johansson provides a wonderful example of such imagining going disastrously wrong with young children. A 2½-yearold, Zakarias, is pulling the sweater of a 3-year-old, Sture. The teacher tries to get Zakarias to stop doing so by inviting him to think about how Sture is feeling through remembering when he himself had his sweater pulled by another child, Oskar, who is present. But the effort to remember does not lead Zakarias to empathy for Sture, but only to anger at Oskar (Johansson (2008): 42). None of this is to deny that from the perspective of moral education, encouraging people to see the world from another’s point of view is frequently an important way of encouraging fellow-feeling.



8.6 Appreciating the Other’s Difference from the Self Nichols also considers a form of mindreading more minimal than perspective-taking but still more robust than his own view. That is the ability to recognize that others might have different desires than one would have in the same situation. For example, Marisol might be able to ignore a public criticism of her work that would make Tamsin extremely upset were she its target instead of Marisol. Marisol’s fellow-feeling should be directed toward what Tamsin actually feels, not what Marisol would feel were she in Tamsin’s situation. And some people are quite poor at discerning the difference between what they themselves would feel in a given situation and what others in that situation do feel. Nichols calls failure to appreciate this sort of distinction between myself and others ‘mindreading egocentrism’. He distinguishes it from ‘ethical egocentrism’, which is a failure to be motivated by the well-being of others for their own sake. And he argues that the former does not entail the latter. That is, someone who keeps using his own needs and desires as a template for understanding others, and thus often misattributes their state of mind, could still be genuinely empathetic and concerned about their welfare. I think this argument is valid.20 But I also think that Nichols’s argument misses something important. He says, ‘Of course, if she is an egocentric mindreader, she may not be very effective in helping others, because she will not be sensitive to the variation in desires, feelings, and thoughts that actually exists among those she tries to help’ (2004: 51). This makes it seem that mindreading egocentrism is deficient only because of its consequences—it leads to inefficient or ineffective helping. But we often find something defective in the observer’s sentiment toward the other itself when it fails adequately to track the actual thoughts and feelings of the other—that is, when it fails to grasp the other’s actual situation. The victim may well have reason to feel that the observer is not really caring about her if he takes action that is apparently motivated by what he regards as her good, but is in fact beholden to a mistaken view of her and her good. Recognizing that mindreading egocentrism can be a deficiency of moral response, not only of outcome, is a strength of the ‘care ethics’ tradition, which emphasizes that caring about someone involves giving attention to her in her individuality and distinctness—seeing her as a distinct person with her own desires, needs, and interests that might well be different from that of the observer.21 This notion of ‘attention’ is derived from Iris Murdoch (who in turn got it from Simone Weil), who saw it as central to the moral attitude, fundamental to recognizing the ‘reality of other persons’, as she put it (Murdoch (1970); Weil (1962)). Mindreading egocentrism

20 Nichols attributes the ‘discrepant desire’ view (the view that altruistic motivation requires the ability to distinguish the desires of others from one’s own) to me (1994: 183–214, especially 193). 21 Noddings (1984) was the first in the recent care ethics tradition to emphasize this point, though the care tradition has subsequently been further developed, incorporating and building on her insight. Scheler (1954) was an earlier philosopher who strongly emphasized this point.



is morally problematic not only because of its likely suboptimal consequences but also because it betokens a defective emotional attitude toward the victim. Indeed, the notion of ‘altruism’ itself can fail us in this situation precisely because it suggests a purely motivational state severed from its embeddedness in broader sentiments and attitudes toward the other. Nichols is right to distinguish mindreading from ethical egoism, and to say that only the latter contravenes altruism. But mindreading egoism is itself contrary to the larger attitudes and sentiments of care and concern that provide the backdrop of altruistic motivation in an adequate account of moral responsiveness to the plight of others. That care and concern are attentive to the ways that the victim’s needs and desires may differ from the observer’s, as well as being aimed at what the observer takes to be the good of the other, is supported by the developmental story Hoffman tells, cited earlier (p. 180). Nichols emphasizes the way that toddlers sometimes offer their own comfort object to a distressed playmate, rather than the playmate’s comfort object, as an example of mindreading egocentrism (2004: 51). Yet Nichols’s account of his own examples is selective, omitting the child’s tracking of the other’s well-being that is involved in the example itself, in the source from which Nichols has taken it. He gives the example of Michael, a 15-month-old, struggling with his friend Paul over a toy, causing Paul to cry. Michael seems distressed and brings his own teddy bear and security blanket to Paul—thus, mindreading egocentrism. But what Nichols omits from Hoffman’s larger description of this example is that Paul continues to cry and Michael goes into the next room and brings Paul’s security blanket to him, implying that he has in some way recognized that his own comfort object will not do the trick for Paul.22 In his account of the development of empathy, Hoffman sees ‘self/other differentiation’ as the key element in that development from infancy to late childhood. For example, he sees ‘awareness of self and others as having independent internal states’ and ‘awareness of self and others as having their own personal histories, identities, and lives beyond the immediate situation’ as stages in that development.23 This is not a point against Nichols’s claim that mindreading egocentrism and ethical egocentrism are two different things. It is meant to support the view that we see mature moral responsiveness to others as involving an awareness of the ways that others can differ from the self, an attentiveness to the particularity of the other, and a desire for the other’s well-being, all wrapped up together. This complexity of the altruistic emotions cannot be captured by a view of them as simply occurrent affective states with no conceptual/intentional relationship to the human (or sentient) other.

22 The example, as mentioned, is drawn by Nichols from my essay ‘Moral Development and Conceptions of Morality’ ([1986]/1994), and I took it from Hoffman (1976). 23 Hoffman (1976): 64. This is a different point from the other drawn from Hoffman (1976: 80) that moral development involves the ability to differentiate a person’s state of mind from her ongoing situation. See p. 180, above.



8.7 Affect and Motivation: Nichols’s ‘Concern Mechanism’ After the defense of ‘minimal mindreading’ as necessary (but also sufficient) for the attribution of distress, Nichols says that he still needs to explain how that attribution leads to altruistic motivation: ‘I will try to characterize the affective response underlying altruistic motivation’ (2004: 52), which he calls the ‘Concern Mechanism’. It involves three components—(1) the attribution of distress to the victim; (2) an occurrent affective state produced by the attribution; and (3) such affect motivating the observer to take action to relieve the other’s distress. This analytical framework, I will argue, falls afoul of the problems I have noted so far in Nichols’s argument: his failure to sustain the morally significant distinction between self-directed and otherdirected emotions; his identification of complex altruistic emotions with mere occurrent affective states; and his omission of the intentional character of altruistic emotions. But in addition, this framework distorts both the perceptual and cognitive dimension of altruistic emotions, and the way that actions motivated by such emotions express them and thereby affect the character of the act itself. More generally, the ‘Concern Mechanism’ framework distorts the relationship between perception, emotion, and action. By using the word ‘concern’ for this mechanism in the context of saying that he is investigating altruistic motivation, Nichols implies that the mechanism provides a way of understanding concern for the other, that is, concern for the other’s well-being, a genuinely altruistic attitude. We would not ordinarily say that if I help someone out of a desire to relieve my own personal distress (caused by her suffering), I have acted out of ‘concern’. Yet when Nichols characterizes the Concern Mechanism, he includes within it ‘sympathy’ and ‘2nd-order contagious distress’. Second-order contagious distress is like the contagious distress mentioned earlier (p. 172), in that it is a distress in the observer caused by the distress in the victim. What makes it ‘second-order’ is that the observer first attributes distress to the victim and it is this attribution that produces the distress in the observer. This form of contagious distress can motivate the observer to help relieve the victim’s distress; but it does so only as a way to relieve the observer’s own distress, in the same way first-order personal distress can. So second-order order contagious distress provides a motive to help that is no more altruistic than that constituted by first-order contagious distress. Neither of these is a form of concern, because neither is a concern for the other. Second-order contagious distress thus does not belong in the same category as sympathy as a form of concern. A second problem with the Concern Mechanism is its rejection of the view that altruistic motivation can be produced without the intervention of an occurrent affect (step (2) of the Mechanism). I think that I have sometimes been motivated to help from the mere recognition that someone needs help, without the intervention of any



affect—either from the mere thought that the person needs my help or the thought that it is the right thing to do. This seems a fairly familiar phenomenon.24 I think Nichols dismisses this possibility because he conflates it with a much stronger one—that a human being would be capable of altruistic motivation without having any capacity for emotional responsiveness to the suffering of others. He attributes this latter view (though not distinguishing it from the former) to Sober and Wilson.25 I don’t think I would be capable of altruistic motivation were I to be a creature without any affective life at all. My ability to be moved by the mere recognition of plight in the absence of occurrent affect still depends on my having emotions, affectively based capacities that sometimes involve occurrent affect. So Nichols is incorrect to think that concern for others, when motivating helping action, always involves an occurrent affect. I can be concerned about Jill and do something to help her without feeling concern (or any other specific feeling toward Jill) at the moment. But more generally, if one appreciates the altruistic emotions as complex states of mind with intentional references to others’ well-being and concomitant forms of perception and cognition involved in them, the entire picture of the ‘Concern Mechanism’—not simply the claim that occurrent affective states are always required for altruistic motivation—is thrown into question. In the Mechanism, the observer first engages in an act of attribution of distress to the victim, temporally followed by and causing an affective state. But this is not the usual phenomenology of seeing someone in distress. Typically (if not in every instance, a qualification I will discuss in a moment) to see that someone is in distress already involves fellow-feeling toward her. ‘Distress’ is not an evaluatively and emotionally neutral attribution to another. Suppose I am at a meeting at work and a co-worker, Hector, is acting oddly. He seems distracted, as if his attention were elsewhere. Then I remember that a second co-worker has told me that Hector has been having some trouble with his teenage daughter that has been causing him some mental anguish. I now come to see his behavior as manifesting distress, which I had not done initially. And in doing so, I feel sympathy or empathy for him (for the example, it does not matter which). Viewing Hector sympathetically and seeing his distress are part of the same mental act for me. Seeing the other’s piece of behavior (his staring out the window, or fiddling with his smart phone) as manifesting distress and viewing the other


I am not relying here on the strong form of a Kantian thesis that a purely rational principle is capable of moving us to help others, no matter what we feel at the moment (although I believe that to be true). I think we can be motivated by the mere recognition of another’s need or suffering, not mediated by the thought that such an action is prescribed by a principle prescribing right action. 25 Sober and Wilson apparently make the evolutionist argument that we might have evolved dispositions to help one another in the absence of any emotional capacities (Nichols (2004): 52). I agree with Nichols in rejecting this view, although, again, Nichols equates this view with the view that on every occasion on which we are motivated to help others, we do so by way of an occurrent affective state.



with sympathy normally go together and are part of the same mental phenomenon. Fellow-feeling is an emotion that involves a way of viewing the other; one might say that it reveals aspects of the other’s well-being (and ill-being) that are invisible to the unsympathetic person. That is, to have empathy for someone is not simply to have a feeling state; it is to perceive the other in light of the other’s state of well-being (although the perception is not necessarily accurate).26 P.F. Strawson made a similar point in his influential essay, ‘Freedom and Resentment’ ([1962]/1974). He noted a set of attitudes toward other persons—for example, resentment, gratitude, forgiveness, anger, love—that arise naturally in interaction and relationships with them, involving responses to the good will or ill will they show us. He calls these ‘participant, reactive attitudes’. Strawson contrasts with them a set of attitudes we can take up toward others that psychically remove us from interpersonal involvement with them and view them as something like objects, to be manipulated or controlled (sometimes for their own good). From the participant’s point of view, these reactive attitudes are part and parcel of our relationships with others and we cannot envision having such relationships without those attitudes. But we are capable of adopting an objective point of view that inhibits these attitudes and calls forth other ones (bafflement that someone could act that way, concern to make sure the person does not do that same thing again, scientific curiosity about what makes such a person ‘tick’). Similarly, I am arguing that our normal mode of engagement with others means that when we experience someone as in distress or suffering, we are seeing her in a compassionate light; we have some sort of fellow-feeling toward her. We can pull back from that kind of engagement with others, enabling us to see their distress from a distance, as it were, for a number of reasons (compassion overload, a desire to avoid the moral pull of the other’s distress that we know will accompany our fellow-feeling). But the normal case of interacting with others involves perceiving someone as in distress and having fellow-feeling toward them as bound up with each other, just as experiencing them as being hateful toward us and resenting them for doing so are. In a sense, then, Nichols’s notion of ‘attribution’, which might seem like a perfectly neutral description carrying no philosophical baggage, is misleading in the normal situation of seeing another as in distress. It implies that there is never an intrinsic or internal connection between this viewing the other as in distress and having fellow-feeling for her. Nichols implicitly replies to this argument by saying that psychopaths and perhaps other evil persons are perfectly capable of seeing someone as in distress without experiencing any fellow-feeling toward that person. So the attribution and the fellow-

26 On the idea that emotion affects the reality perceived by the observer, see Goldie (2008): 233, and, more generally, Murdoch, Nussbaum (e.g. (2004): 24–31), and others emphasizing the perceptual dimension of emotional experience. I am also influenced by Vetlesen (1994: 154–63); e.g. he says, ‘I argue that to “see” someone as suffering is already to have established an emotional bond between myself and the person I “see” suffering’ (1994: 159).



feeling must be two different mental events only contingently related. But this does not follow. Just because a psychopath can attribute distress without having any accompanying fellow-feeling does not mean that the attribution and the fellow-feeling are not tied together in a holistic way in the case of non-psychopathic human beings. In this regard it is significant that we regard psychopaths as humanly as well as morally defective beings. Nichols and others in the neo-sentimentalist tradition have pointed out that psychopaths have three deficiencies. They are generally unresponsive to others’ distress; they lack motivation to help others (unless they see doing so as in their interest); and they do not make moral judgments. (They make normative judgments, but Nichols makes a compelling case that they are unable to appreciate the force of the ‘moral’, shown in part in their failure to be able to distinguish between (what we regard as) judgments of moral wrongness and judgments of conventional wrongness.) These deficiencies suggest that psychopaths’ responses to distress cannot tell us what the connections between fellowfeeling and seeing distress are among normal persons. A further consideration calling into question a purely contingent link between seeing someone as suffering and responding to it is the set of expressions we have that suggest different degrees of understanding of another’s suffering or distress. We often say someone ‘has not taken in’, ‘has not really understood’, or ‘understands only superficially’ someone else’s suffering, if he verbally attributes suffering to the person but fails to respond to it. That is, these expressions imply some sort of failure of the understanding—then a cognitive failure—not only a failure of appropriate emotional response, thereby implying some sort of internal link between cognizing of distress and responding emotionally to it in an appropriate manner. Nichols may reply to this as he does to a different but related case (2004: 113) that the person understands perfectly well that the other is suffering but just doesn’t care. But I think our notion of ‘understand’ in application to the plight of others does not draw this distinction so sharply; it allows that there are undoubted kinds of understanding that are absent in those who fail to respond with fellow-feeling of some kind (Starkey (2008): 425–54). This is not to deny that there can be a minimal threshold for Nichols’s notion of ‘attribution’ of distress or suffering, according to which one can say the attributor understands but does not care. My point is that this phenomenon can also be expressed in the epistemic language of failing to understand, to really understand, or to have a full understanding, and that Nichols’s conception cannot capture this epistemic usage.

8.8 Action Expressive of Emotion The ‘Concern Mechanism’ idea runs into other, not unrelated, problems regarding the relationship posited between affect and helping ((2) and (3), on p. 185). That posited relationship is that the affect supplies a motive for the act, but the character of the act itself is not affected by this motive. That implication is also carried by the somewhat antiseptic



terminology of ‘prosocial behavior’ or ‘helping behavior’ favored by psychologists. The problem here is that the character of action that expresses particular emotions is (not always but often) very much affected by the particular emotion in question. The action of helping someone out of compassion or concern has quite a different significance from helping someone as a way to relieve a cause of the observer’s own distress.27 It feels like, and is, a different action, both from the observer/agent’s and from the victim/recipient’s point of view.28 That is, emotions such as fellow-feeling do not only motivate action; they express themselves in action. The image of a ‘mechanism’ (the ‘Concern Mechanism’) tends to mask this important point about the relation of emotion and action by implying a too-sharp divide between the motive/emotion as a purely inner psychological event and the external behavior it is taken to prompt. This problem with the ‘mechanism’ imagery is comparable to the previously discussed one of making a too-sharp divide between seeing distress and emotionally responding to it. In both cases the oversimplified image of the emotion of fellow-feeling as an occurrent, short-lived, affective event is at fault. Once we recognize the perceptual, cognitive, intentional, and expressive dimensions of altruistic emotions, we will be less attracted to the metaphor of a ‘mechanism’.

8.9 Nichols’s ‘Sentimental Rules’ Recall that Nichols’s discussion of responses to distress is in service of a theory of moral judgment. ‘Core moral judgment depends on two mechanisms, then, a normative theory prohibiting harming others, and some affective mechanism that is activated by suffering in others’ (2004: 18). In contrast to Hume, for whom sympathy (the emotion on which moral judgment rested) is a moral emotion, for Nichols the emotions that are implicated in moral judgment are not moral emotions. ‘None of the affective mechanisms explored in this chapter counts as a moral sense’ (2004: 62), by which he means a sense that tracks vice and virtue in persons and acts. Nichols takes the argument of his chapter 2 to show that non-psychopathic persons have strong emotional reactions to others’ distress, and this powerful emotional reaction attaches to the norms prohibiting causing distress, providing them with a distinctive status (in contrast, for example, to mere conventional norms) that constitutes what we understand as ‘moral’. They ‘supply the sentiment to moral judgment’ (2004: 63).

27 This is not always so, because in some situations we may care only that a certain result ensues from an action, not what sentiment or emotion the action expresses. For example, if I am being held up at gunpoint and a stranger rescues me by disarming the gunman, I may not care whether the stranger is motivated by compassion or to win a bet with his friend that he could accomplish this rescue effort. (Even in this case, however, I think the motive would make some difference to me. I would appreciate the rescuer in a different way if I thought his action expressed compassion or helpfulness than if I thought he was motivated by gain.) 28 This point was made by Bernard Williams in (Williams [1966]/1973), in saying that the recipient sometimes desires the ‘human gesture’ from the agent; but the contrast there was with action done out of duty. Blum elaborated Williams’s point in Blum (1980): ch. 7. Goldie makes a similar point in Goldie (2001): 41.



This view seems to me to have contributed to Nichols’s failure to sustain an appreciation of the moral difference between altruistic emotions and other, egoistic, responses to distress and suffering. He is indeed correct to say that altruistic emotions do not constitute a moral sense as he understands it. Compassion or empathy for another is a response to (or, as I have argued, reveals) suffering; but it does not, or at least not necessarily, involve a moral judgment about the rightness or wrongness of an action. (Nichols rightly cites in support of this view that children younger than two are capable of altruistic behavior but not of moral judgment.) Nevertheless, there is a vast moral difference between compassion/empathy and the other egoistic responses to suffering that Nichols discusses—contagious distress, personal distress, second-order contagious distress, and enduring representation.29 One is a morally worthy response, the other a morally neutral one, everything else being equal. Nichols sometimes recognizes this morally significant difference, as we have seen, but only fitfully. The Schopenhauerian tradition would have helped him out, with its insistent focus on the morality of motives, not of judgments. But the focus on the strength of the emotional response, and not its moral character, contributes to the problem. (But that focus is what is distinctive of Nichols’s view among neo-sentimentalists.)30 The failure to see the distinction between egoistic and altruistic responses to distress vitiates Nichols’s claim to have provided an explanation of moral judgment. That a norm prohibits a type of action that tends to cause strong feelings in observers does not seem to account for its being regarded as a moral norm, which would seem to depend not only on the strength of the feelings but also on their character. Suppose all altruistic emotions, emotions tied to concern about the well-being of others, disappeared in human beings, but emotions (such as contagious and personal distress) remained. That is, individuals’ distress caused other people distress but no one had any concern for any one else’s distress for its own sake. In such a situation, how would norms arise prohibiting the causing of distress? It would seem that the same result—discouraging the causing of personal distress—could be achieved by norms that prevented or discouraged people in distress from contact with others not in distress. Beyond that, even if the sorts of norms Nichols is envisioning arose, it is not clear that persons so constituted would be capable of regarding the norms in question as moral ones. Nichols recognizes that it is possible to recognize the validity of a norm that others think of as moral without recognizing the moral character of that norm. He thinks that psychopaths recognize that it is wrong to harm others, but think that the reason for its wrongness is social convention rather than morality ((2004: 13 n. 4) and 29 ‘Enduring representation’ is when the observer withdraws from the scene that contains the cause of his distress but doing so does not rid him of a mental representation of that distress, which thus continues to distress him. 30 ‘Norms prohibiting actions that are likely to elicit strong negative affect will be treated as distinctively wrong’ (Nichols (2004): 63f.). In a later article, Nichols summarizes the position: ‘[O]ur natural emotional reactions to harming others confer a special force on the rule against harming others, and this connection with emotion likely played a critical role in the cultural success of the rule’ (Gill & Nichols (2008)).



elsewhere).31 In that spirit, I would question whether someone who was completely incapable of altruistic emotion, that is, whose only emotional reactions to the distress of others are egoistic in character, would be able to see the causing of distress to others as morally wrong. If people’s emotions were confined to self-oriented ones, it is not clear that these could underpin or sustain moral judgment. So Nichols’s sentimental rules view fails to give an adequate account of moral judgment.32 Elements of Nichols’s argument point more toward Schopenhauer’s than to the Humean one that Nichols favors. That is, compassion is (for Schopenhauer) the fundamental moral motive not because it issues from a moral judgment, but because it is directed toward the weal and woe of others for their own sake. And the evidence that Nichols cites from children provides part of a plausible developmental story as to how human beings develop the capacity for compassion. Children are able to respond in a caring way toward others before they make moral judgments.33 And adults, while unlike very young children in being able to make moral judgments, are capable of acting morally without invoking, even implicitly, moral norms in doing so. That is, to follow Schopenhauer, they can act with compassion, thereby doing the moral thing, without invoking moral norms. And this is how emotion comes into morality; humans will not be capable of that moral responsiveness without emotional capacities. So emotion is necessary for morality, but not in the way that Nichols postulates. Nichols and so many other contemporary moral theorists assume that the ultimate object of a theory of morality is an explanation of moral judgment. But why not see as an equally worthy object a theory of moral responsiveness—something that includes moral emotions and moral motives, and actions animated by them? With such a perspective, one would be much less likely to be misled into short-changing the moral distinction between personal or contagious distress and genuine concern for others. More generally, I have argued that Nichols operates with an impoverished view of altruistic emotions (empathy, compassion, concern) as mere short-term, affective states of mind. He fails to recognize their essentially intentional structure, that they are directed toward others in light of their well-being; their essential connection to perception and cognition; and their expressive dimension that is central to their moral significance. Although I have not attempted to demonstrate this, I suspect that


From Nichols’s discussion of the basis on which psychopaths say that a type of action is wrong, I am not sure whether the psychopaths he discusses actually regarded the action as wrong (though for reasons of convention) or that they merely recognized that others thought the actions were to be discouraged and expressed this by saying they were ‘wrong’. That is, I am not sure that psychopaths understood, or applied, a concept of ‘wrong’ for any actions that were not in their own perceived self-interest. I am not sure that psychopaths did not take ‘wrong’ to mean ‘socially disapproved of ’. 32 This argument draws on Jones (2006). 33 Nichols sometimes implies that the child’s capacity for emotional contagion is a developmentally necessary stage for empathy and concern; but the psychologist Zahn-Waxler argues that children are able to exhibit concern for others apart from emotional contagion, and do not need to have gone through a stage of contagion (Zahn-Waxler et al. (1992)).



Nichols’s love affair with experimental social psychology has led him to retreat from the rich tradition of philosophical moral psychology.34

References Barasch, Marc Ian (2009) The Compassionate Life: Walking the Path of Kindness. San Francisco: Berrett-Koehler. Blackburn, Simon (1998) Ruling Passions. Oxford: Oxford University Press. Blum, Lawrence (1980) Friendship, Altruism, and Morality. London: Routledge & Kegan Paul. —— ([1986]/1994) ‘Moral Development and Conceptions of Morality’. In Blum (1994): pp. 30–62. —— (1994) Moral Perception and Particularity. Cambridge: Cambridge University Press. Carlo, G., PytlikZillig, L., Roesch, S., & Dienstbier, R. (2009) ‘The Elusive Altruist: The Psychological Study of the Altruistic Personality’. In: Personality, Identity, and Character, Darcia Narvaez and Daniel Lapsley (eds), pp. 271–94. Cambridge: Cambridge University Press. Clark, R. & Word, L. (1974) ‘Where is the apathetic bystander? Situational characteristics of the emergency’, Journal of Personality and Social Psychology 29: 279–87. D’Arms, Justin & Jacobson, Daniel (1994) ‘Expressivism, Morality, and the Emotions’, Ethics 104: 739–63. Doris, John & Stich, Stephen (2006) ‘Moral Psychology: Empirical Approaches’, Stanford Encyclopedia of Philosophy, Edward Zalta (ed.) Eisenberg, Nancy (2000) ‘Emotion, Regulation, and Moral Development’, Annual Review of Psychology 51: 665–97. Gibbard, Allan (1990) Wise Choices, Apt Feelings. Cambridge, MA: Harvard University Press. Gill, Michael & Nichols, Shaun (2008) ‘Sentimental Pluralism: Moral Psychology and Philosophical Ethics’, Philosophical Issues 18(1): 143–63. Goldie, Peter (2001) The Emotions. Oxford: Oxford University Press. —— (2008) ‘Getting Feelings into Emotional Experience in the Right Way’, Emotion Review 1 (3): 232–9. Hoffman, Martin (1976) ‘Empathy, Role-taking, Guilt, and Development of Altruistic Motives’. In: Moral Development and Behavior, T. Lickona (ed.), pp. 124–43. New York: Holt, Rinehart, and Winston. —— (2000) Empathy and Moral Development: Implications for Caring and Justice. Cambridge: Cambridge University Press. Johansson, Eva (2008) ‘Empathy or Intersubjectivity? Understanding the Origins of Morality in Young Children’, Studies in the Philosophy of Education 27: 33–47. Jones, Karen (2006) ‘Metaethics and Emotions Research: A Response to Prinz’, Philosophical Explorations 9(1): 45–53. Mansbridge, Jane (1990) (ed.) Beyond Self-Interest. Chicago: Chicago University Press. Monroe, Kristen Renwick (1997) The Heart of Altruism. Princeton, NJ: Princeton University Press. Murdoch, Iris (1970) The Sovereignty of Good. London: Routledge. 34 I wish to thank Paul Harris, Steven Levine, Kelso Cratsley, Jennifer Radden, and Peter Goldie for comments on earlier drafts or presentations of this article.



Nichols, Shaun (2004) Sentimental Rules: On the Natural Foundations of Moral Judgment. Oxford: Oxford University Press. Noddings, Nel (1984) Caring: A Feminine Approach to Ethics and Moral Education. Berkeley: University of California Press. Nussbaum, Martha C. (2004) Hiding From Humanity. Princeton, NJ: Princeton University Press. Oliner, Samuel & Oliner, Pearl (1988) The Altruistic Personality. New York: The Free Press. Scheler, Max (1954) The Nature of Sympathy. New Haven: Yale University Press. Schopenhauer, Arthur ([1841]/1965) On the Basis of Morality. New York: Bobbs-Merrill. Scott, Niall & Seglow, Jonathan (2007) Altruism. Maidenhead, England: Open University Press. Snow, Nancy (1991) ‘Compassion’, American Philosophical Quarterly 28: 195–205. Sober, Elliott & Wilson, David Sloan (1998) Unto Others: The Evolution and Psychology of Unselfish Behavior. Cambridge, MA: Harvard University Press. Starkey, Charles (2008) ‘Emotion and Full Understanding’, Ethical Theory and Moral Practice 11 (4): 425–54. Strawson, Peter F. ([1962]/1974) ‘Freedom and Resentment’, Freedom and Resentment and Other Essays. London: Methuen. Vetlesen, Arne Johan (1994) Perception, Empathy, and Judgment: An Inquiry into the Preconditions of Moral Performance. Pennsylvania: Pennsylvania State University Press. Weil, Simone (1962) ‘Human Personality’. In her Selected Essays 1934–1943, pp. 9–34. New York: Oxford University Press. Williams, Bernard ([1966]/1973) ‘Morality and the Emotions’. In his Problems of the Self: Philosophical Papers 1956–1972, pp. 207–29. Cambridge: Cambridge University Press. Zahn-Waxler, Carolyn, Radke-Yarrow, Marian, Wagner, Elizabeth, & Chapman, Michael (1992) ‘Development of Concern for Others’, Developmental Psychology, 28: 126–36.

This page intentionally left blank


This page intentionally left blank

9 Reactive Attitudes Revisited John Deigh

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 contemporary analytic 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 attitudes 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 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.

1 Proceedings of the British Academy 48 (1962): 1–25. Reprinted in Free Will, Gary Watson (ed.), 2nd edition (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 Straaten (ed.) (Oxford: Clarendon Press, 1980), pp. 14–47.



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, in section 9.1, 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. Once this conception is appreciated, it will be apparent that Strawson saw the practice of attributing moral responsibility as emerging from these states and the interpersonal relationships to which they are integral and not as necessarily embedded in those relationships. In section 9.2 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 supersede. In section 9.3, 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.


R. Jay Wallace, Responsibility and the Moral Sentiments (Cambridge, MA: Harvard University Press, 1996). Stephen Darwall, The Second-Person Standpoint: Morality, Respect, and Accountability (Cambridge, MA: Harvard University Press, 2006). 4



9.1 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 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 forward, 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 morally wrong ones. For though such facts would sit well with the pessimists, the optimists would surely need to be convinced that a deterministic universe 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 behavior that appear to be completely consistent with determinism. Specifically, he looks to what appear to him to be plain, natural facts about men and women 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. The plain, natural facts about men and women as social beings that Strawson adduces—what he calls commonplaces—are these.5 A central concern of people who live together in communities is what their intentions and attitudes are toward 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 toward them. The goodwill each expects of others consists in the others’ keeping, as best they can, from harming him in their conduct toward 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


Strawson, pp. 75–7.



with whether others are well- or ill-disposed toward 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 toward him. Thus, while one naturally takes pleasure in benefits that come one’s way on account of another’s behavior, 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 toward 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 toward 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 toward one. Other examples are love, forgiveness, and hurt feelings. He calls these personal reactive attitudes. One cannot overstate the importance to Strawson’s argument of the distinction between being benefited or harmed inadvertently or accidentally and being benefited or harmed intentionally out of evident good or ill will toward 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 toward 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 an enormous range of transactions and encounters’ and to think, ‘in turn . . . of the kind of importance we attach to the attitudes and intentions toward us of those who stand in these relationships to us, and of the kinds of reactive attitudes to which we ourselves are prone’.6 A decade later, John Rawls, in A Theory of Justice, makes it explicitly. Concerning the three psychological laws that, on his theory of moral development, explain the transitions between the stages of moral growth in a child’s acquisition 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


Ibid.: p. 76.



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.7

The depth of this fact (and so the ability to discern the dispositions and attitudes of others toward one) was observed two centuries earlier by Rousseau, whom Rawls acknowledges as his predecessor.8 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 wellbeing 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 toward 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.9

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 toward 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.10 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 person reason to abandon his belief that the offender acted with indifference or malice toward him. The offender in making such pleas is in effect saying that he did not, 7 8 9 10

John Rawls, A Theory of Justice (Cambridge, MA: Belknap Press, 1971), pp. 494–5. John Rawls, ‘The Sense of Justice’, Philosophical Review 72 (1963): 281–305. J. J. Rousseau, Emile or On Education, Allan Bloom, trans. (New York: Basic Books, 1979), pp. 213. Strawson, p. 76.



despite his having committed the offense, lack the goodwill toward 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 toward 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 toward him. One may of course still be liable to some feelings toward 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 that 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 toward him creates. In each of these cases, we maintain an objective view only temporarily or only toward 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 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’.11 To make this explanation Strawson adds to the class of reactive attitudes vicarious analogues of the personal reactive attitudes. These are attitudes one takes toward the goodwill or ill will of another, not in reaction to the manifestation of that will in the person’s actions toward one, but in reaction to its manifestation in his actions toward 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 ‘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.12 Strawson, in other words, takes indignation to be a moral attitude in virtue of its being impersonal or disinterested. He does not, then, in conceiving of it as a thirdparty 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 toward 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 11 12

Ibid.: p. 93. Ibid.: p. 84.



regard on the part of others toward 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 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 toward 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.13 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 toward one and thus coming to expect and demand such goodwill of others toward 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 toward others that 13 David Hume, Enquiries Concerning Human Understanding and Concerning the Principles of Morals, P.H. Nidditch (ed.), 3rd edition (Oxford: Clarendon Press, 1975), pp. 215–18.



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.14 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 toward him that otherwise would be forthcoming.15 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 toward others generally in his or her dealings with them, and this generalized, impersonal expectation and demand amounts to an understanding of everyone in the community as lying under moral obligations to act with some degree of goodwill toward 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.16 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 toward 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 14 15 16

Strawson, p. 84. Ibid.: p. 86. Ibid.: p. 85.



they act with a degree of goodwill or regard toward him as general requirements to act with that degree of goodwill toward 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 toward 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.17 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 disinterested or impersonal 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 him in supposing that we get our ideas of moral obligation and moral responsibility directly from generalizing the expectations of and demands for goodwill toward ourselves that we make on others, which is to say, by vacating our personal view and taking up an impersonal 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 impersonal 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 toward others that we come to demand of everyone including ourselves when we take up the impersonal or disinterested view.18 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 toward him and at the same time expecting and demanding on his behalf that we so act, we reproach ourselves for our lack

17 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. 18 See David Hume, A Treatise of Human Nature, P.H. Nidditch (ed.), 2nd edition (Oxford: Clarendon Press, 1978), p. 479.



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.

9.2 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 toward 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 small children or social animals other than humans from being susceptible to them. His account fits well within Rousseau’s and Rawls’s accounts of moral development, on which discernment of the good and ill will that others manifest in their behavior toward one occurs at a young age, 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.19 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 to which rational beings alone are susceptible and as appropriate only when taken in response to the actions of rational 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 to resentment he must believe that others with whom he is engaged in interpersonal relationships are required to act in certain ways toward 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 toward him.20 What provokes one person P to feel resentment toward another Q, then, is Q’s violating a requirement imposed on him to act in certain ways toward 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 toward 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 toward each other. Thus, when one feels indignation toward 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 toward others. This belief is implicit in one’s readiness to blame oneself for failing to meet expectations of how to act toward 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, 19 20

Oliver Wendell Holmes, Jr., The Common Law (Boston: Little, Brown and Co., 1881), p. 3. Wallace, pp. 21–2; Darwall, p. 68.



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 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.21 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.22 Consequently, both Wallace and Darwall conceive of the agents toward 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 toward 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 toward the person who feels resentment, as Strawson does, he takes them to be choices.23 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 toward 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 attitude. On Wallace’s conception, that is, resentment toward someone does not include any sense of that person’s being ill-disposed toward 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


21 22 23

Wallace, pp. 35–6. Darwall, pp. 74–9. Wallace, p. 128.



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.24 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.25 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 toward 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 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 24 Wallace, pp. 36–7, 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. 25 Darwall, pp. 57–61 and 67–8.



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 toward 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 toward one. It does not in its original form contain any thought of the actor’s being morally responsible for harming one, though after one 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 disinterested or impersonal 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.26 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 toward one. That

26 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 toward 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.



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

9.3 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, siblings, and playmates. As I noted earlier, his understanding is consistent with Rousseau’s and Rawls’s accounts of moral development, and 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 toward 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 toward whom resentment is felt are rational agents as essential to the attitude. 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 toward 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.27 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 toward 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 toward 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 toward 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 toward the offender because he will cease to regard the latter as having acted with either ill will toward 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 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 27 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.



toward the offender. Accordingly, one could say that pleas of the second type tell us that S views the actions of those toward 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 toward whom one is liable to feel those attitudes as being 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 toward 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 toward 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 toward 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.28 Yet being unfortunate in one’s formative circumstances hardly implies that one cannot bear ill will toward 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 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 toward one and the more one’s dealings with him are


Strawson, p. 79.



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 toward 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.

10 Responsibility and Dignity Strawsonian Themes Bennett W. Helm

10.1 Strawsonian Themes Peter Strawson’s landmark ‘Freedom and Resentment’ was a largely successful attempt to reshape the debate about the nature of freedom and responsibility. Since its publication, many philosophers have approached the question of responsibility not by attempting to identify the properties or features in agents necessary for responsibility but rather by examining our practice of holding agents responsible. Strawson argued that we ought to understand what it is for an agent to be responsible in terms of how we appropriately hold her responsible, and so an account of freedom must look at least in part to our ‘reactive attitudes’. As he describes it, the ‘reactive attitudes are essentially natural human reactions to the good or ill will or indifference’ people have towards each other.1 Consequently, Strawson includes in his list of reactive attitudes such emotions as gratitude, resentment, forgiveness, love, hurt feelings, esteem, indignation, contempt, self-respect, affection, disapprobation, guilt, remorse, and shame. There does seem to be something very right about Strawson’s general strategy. Being a responsible agent is a matter of being answerable or accountable for what you do, where that answerability is often to a fellow person, who can justly complain when we fail to act properly; in many cases2 such complaints are more or less successful attempts to hold each other to account. So our attempts to hold each other accountable do seem to be relevant to our being responsible. Nonetheless, Strawson’s discussion leaves many unanswered questions, including questions about what exactly the reactive attitudes are and how they

1 Peter F. Strawson, ‘Freedom and Resentment’, Proceedings of the British Academy 48 (1962): 195. The quote is taken from Strawson’s account of the personal reactive attitudes (like gratitude and resentment), which he distinguishes from their ‘vicarious’ analogues (like indignation) and from the self-reactive attitudes (like guilt). With my gloss at the end, I have tried to generalize Strawson’s remark to include these other types of reactive attitudes. 2 But surely not all: we can grumble and swear as a way of venting our anger or frustration, without thereby holding anyone to account.


B E N N E T T W. H E L M

involve holding others responsible. Moreover, what is at stake here is not just that we hold others responsible but that we do so appropriately; we therefore need to understand what the conditions of such appropriateness are—and, if Strawson’s strategy is right, do so in a way that does not simply presuppose that persons are responsible. Following this general Strawsonian strategy, Jay Wallace has presented an intriguing account of responsible agency in terms of the reactive attitudes that tries to answer these questions. For present purposes, Wallace’s account differs from Strawson’s in two key respects. First, Wallace criticizes Strawson’s understanding of what the reactive attitudes are as too broad a class of attitudes to be informative: This inclusive interpretation of the reactive attitudes has generally been followed in discussions of Strawson’s lecture, which display a tendency to construe the reactive attitudes as any emotions that involve or point us toward, interpersonal relations. If this is right, then such attitudes as embarrassment, friendly affection, and sympathy would have to be counted as reactive attitudes as well.3

Wallace’s complaint is that including emotions like embarrassment, affection, and sympathy in the class of reactive attitudes will leave us unable to make sense of two claims that Wallace thinks Strawson rightly makes: first that the reactive attitudes contribute to our understanding of what it is to hold each other responsible and thereby to be responsible; and second that the reactive attitudes are incompatible with the objective attitude, a way of responding to someone as a mere object of treatment or policy or manipulation rather than as a responsible agent. For these two claims point to an essential connection between the reactive attitudes and human agency, a connection which (Wallace thinks) the inclusive interpretation of the reactive attitudes would simply give up. By contrast, Wallace proposes that we understand the reactive attitudes more narrowly, as emotions responsive to the violations of ‘expectations’ or ‘demands’—that is, to violations of ‘practical requirement[s] or prohibition[s] in a particular situation of action’.4 Consequently, he holds the paradigm reactive attitudes to be just resentment, indignation, and guilt—emotions that involve holding persons responsible for violations of such expectations; other reactive attitudes are simply refinements of these three.5 Although I think Wallace is right to think of the reactive attitudes partly in terms of their connection to expectations or demands, he is clearly missing something in understanding


R. Jay Wallace, Responsibility and the Moral Sentiments (Cambridge, Mass.: Harvard University Press, 1994), 27. Wallace’s target in this criticism really is Jonathan Bennett, ‘Accountability (II)’, in Free Will and Reactive Attitudes: Perspectives on P. F. Strawson’s ‘Freedom and Resentment’, ed. Michael McKenna and Paul Russell (Burlington, VT: Ashgate, 2008), 47–68. Strawson himself, in understanding the reactive attitudes to be responsive to the good or ill will of others, would have a principled reason to reject at least embarrassment and sympathy as reactive attitudes. Nonetheless, the underlying criticism applies equally well to Strawson: we need a clear, explicit delineation of the class of reactive attitudes, and Strawsonians—including Strawson himself—have generally failed to provide one. 4 Wallace, Responsibility and the Moral Sentiments, 22. 5 Indeed, Wallace thinks the relevant expectations and the reactive attitudes are mutually dependent, with neither intelligible apart from the other.



them as responses only to violations of these expectations. Why not understand gratitude or approbation, for example, as the proper response to someone who notably exceeds these expectations in their interactions with us or with others?6 Moreover, why shouldn’t we include among the reactive attitudes emotions that are responsive not simply to what has happened but also to what may well happen, such as trust or distrust? One might reject this latter suggestion by pointing to Strawson’s claim that what makes something a reactive attitude is the fact that it is a reaction to events in the world that reveal the ‘good or ill will or indifference’ of others; so construed, of course reactive attitudes are not reactions to future possibilities. Yet we might wonder whether this construal of the reactive attitudes is too narrow. After all, that a response involves an evaluation of future possibilities does not show that it is not reactive to persons in the way relevant to their being reactive attitudes. Finally, we need to know more about the sorts of expectations or demands relevant to the reactive attitudes. Are they interpersonally valid, so that what I can reasonably expect of others is what they can reasonably expect of me? And what precisely is the connection between such expectations and the reactive attitudes such that they are, as Wallace says, mutually dependent on each other? The second key difference between Wallace and Strawson is this. Although Wallace follows Strawson in understanding what it is for someone to be responsible in terms of how others treat her as responsible, he offers an explicitly normative understanding of the connection between the two: someone is responsible just in case it is fair for us to treat her as responsible, where such fairness depends on her having the ability ‘to grasp moral reasons and to control [her] behavior by their light’.7 Thus, her having such powers of rational self-control justifies our treating her as responsible, which in turn makes her actually be responsible. However, one might wonder at this point whether this abandons Strawson’s central insight: why can’t we simply cut out the middleman—the appeal to the reactive attitudes, which might seem to be only a rhetorically significant sideshow—and instead understand responsibility simply in terms of these powers of self-control? It is not clear that an appeal to the reactive attitudes’ role in holding someone responsible—though surely interesting in its own right—really contributes anything essential to an account of responsibility. Indeed, something like this criticism is offered by Manuel Vargas: on the face of it, what it is for someone to be responsible depends solely on facts about the agent; appealing to some further thing— the social practices of holding people responsible—seems unnecessary.8 One reply to this criticism, albeit not one developed by Wallace, is to deny that having the relevant powers of rational self-control is conceptually prior to being fairly treated as responsible and instead to understand each to be a necessary condition of the 6 A similar worry about Wallace’s restricted understanding can be found in Bennett’s discussion of the reactive attitudes as sometimes blame-related and sometimes praise-related. See Jonathan Bennett, ‘Accountability (II)’. 7 Wallace, Responsibility and the Moral Sentiments, 16. 8 Manuel Vargas, ‘Responsibility and the Aims of Theory: Strawson and Revisionism’, Pacific Philosophical Quarterly 85, no. 2 (2004): 225.


B E N N E T T W. H E L M

other. The thought here is that being responsible is a matter of being answerable or accountable—that is, capable of responding or answering to others, accounting for what you have done; such a capacity to answer to others, to take responsibility for what one does, might plausibly be thought a necessary condition of responsible agency. If appropriately holding someone responsible depends on his being capable of answering such a demand, and if his being capable of answering in this way requires that he be a part of a practice of holding others responsible, then we get the required conceptual circle. Moreover, it does seem that in holding others responsible and providing an answering response each centrally involves the reactive attitudes: we hold others responsible partly through making them the target of resentment and indignation, and they themselves take responsibility partly by feeling guilt and expressing these feelings in making apologies and reparations. My aim in this paper is to pursue—somewhat tentatively and hesitantly—roughly this line of thought, though it is a line of thought that raises many questions. 1. Why should a capacity to ‘take responsibility’ be a necessary condition of responsible agency? Can’t someone be incapable of ‘consideration’ for others and yet still be responsible for her actions? 2. What exactly is the connection between holding someone responsible and her taking responsibility for what she does? 3. Why are the reactive attitudes necessary for taking responsibility or holding someone responsible? My tentative claim is that we can make some progress here by thinking carefully about Strawson’s off-hand remark concerning the way the reactive attitudes are ‘humanly connected’. For, I shall argue, the reactive attitudes are partly defined through their interpersonal, rational interconnections, such that they together constitute our respect for persons.9 Being one of us—a person who respects and is respected by others and so who can both hold others responsible and take responsibility—is a necessary condition of the possibility of responsible agency. Arguing for this conclusion will require first (in section 10.2) a detour through my account of emotions quite generally and their relation to import—to caring about something. I shall then (in section 10.3) use this as a way of coming to understand more clearly what the reactive attitudes are and how they can constitute respect for persons, before turning (in section 10.4) to think about what all this might tell us about the nature of responsible agency.

9 Although Strawson, Wallace, and others commonly understand the reactive attitudes to be responses called for by someone’s status as a person (or the ‘human community’, as Strawson describes it (Strawson, Freedom and Resentment, 201)), I suspect that it is too narrow. Rather, I suspect that someone’s status within communities smaller than the community of all persons can properly call for reactive attitudes. However, I cannot argue for this here, and so I shall follow them here in this narrow understanding.



10.2 Background: Emotions and Import For something to have import to someone, for him to care about it in one way or another, is for it to have a kind of worth to him. Of course, there are multiple kinds of import things can have for us, multiple types of evaluative attitude we can take towards things: we can care about something, or, more deeply, value it as a part of the kind of life we find worth living;10 we can love a particular person by identifying with him by treating his identity as a person in ways analogous to our own;11 and, as I shall argue later in this paper, we can respect someone as a person. In each of these cases, the worth involved is worthiness of attention and action; after all, it is hard to make sense of something as having import to someone if he never or rarely notices when it is affected favorably or adversely, and the differences among these various kinds of import amount to differences in the kind of attention and action that import demands from us. How is such import, such worthiness of attention and action, to be understood? In my earlier work,12 I argued that there is a two-way conceptual connection between the emotions and import. First, emotions are responsive to the imports things have for us; such responsiveness is evident in part in the kinds of objects the emotions have. Thus, it is commonly recognized that each emotion has a target: that at which the emotion is intuitively directed. When I am angry at the squirrels for eating my tomatoes, the squirrels are the target of my anger. Each emotion also has a formal object: the evaluation of the target that is characteristic of that emotion type. In being angry at the squirrels, I evaluate them as somehow offensive; were I afraid of them instead, I would evaluate them as dangerous. Here, offensiveness and dangerousness are the formal objects of anger and fear respectively. Finally, in order to understand why in having an emotion one evaluates the target in light of the formal object, we need to appeal to a third, not commonly recognized object of emotions: their focus. Why are the squirrels dangerous and so appropriate objects of fear for me? The answer depends on their relation to my tomatoes, which they tend to eat. Indeed, that I care about the tomato harvest is what makes intelligible not only my finding the squirrels to be dangerous (in fearing them) but also my finding them to be offensive (in getting angry at them). The focus of an emotion, then, is the background object having import to which the target is related in such a way as to make intelligible the target’s having the evaluative property defined by the formal object. That an emotion’s focus has import to the subject is a necessary condition of the emotion’s warrant, and it is this fact about emotions that makes them intelligible as responsive to import—as intentional feelings of import. The second conceptual connection is that projectible, rational patterns of emotions constitute import. The idea is that particular emotions are rationally connected to other 10 Bennett W. Helm, Emotional Reason: Deliberation, Motivation, and the Nature of Value (Cambridge: Cambridge University Press, 2001). 11 Bennett W. Helm, Love, Friendship, and the Self: Intimacy, Identification, and the Social Nature of Persons (Oxford: Oxford University Press, 2010). 12 See, for example, Helm, Emotional Reason, especially chapters 3–4.


B E N N E T T W. H E L M

emotions having the same focus. Thus, if I am afraid that the squirrels will eat my tomatoes, I rationally ought to be relieved if they don’t, or angry or disappointed when they do, frustrated that the fence I built isn’t keeping them out (or pleased that it is), and so on. (How do you suppose I ought to feel when a hail storm skirts to the north or when I harvest my first tomato or when I discover my neighbor has been poaching them?) These rational connections arise because emotions are not simply responses to antecedently existing import but rather are commitments to that import and so to having further emotions with a common focus in the appropriate circumstances. (See Figure 10.1.) Moreover, the pattern defined by these rational connections includes not just emotions but also desires and evaluative judgments: given my fear of the squirrels, I ought to want to keep them out, judge that they are a nuisance, and so on. Indeed, understanding desire as a part of this pattern enables us to make sense of desire as finding its object to be worth pursuing—as having import—rather than simply as a matter of goal-directedness. In general, such a pattern of emotions and desires focused on tomatoes is a pattern of responding to the tomatoes as worthy of attention and action, and it is precisely such a pattern that constitutes my caring about the tomatoes— that constitutes the import they have for me. To have import just is to be the focus of a projectible, rational pattern of emotions and desires. Consequently, we might say that caring and import are the subjective and objective sides of the same phenomenon. This two-way conceptual connection between our emotions, desires, and judgments on the one hand and import on the other reveals that talk of the evaluative attitudes of caring, valuing, and so on and talk of import are alternative ways of describing the same phenomenon—the latter objectively in terms of that to which we respond and the former subjectively in terms of the response we have to it. Of course, something has import to us, we care about it, not simply de re but rather under a particular description. Thus, I might care about my car merely as an efficient, reliable means of transportation. When its fuel efficiency drops or when it starts to suffer regular breakdowns, I may get upset, but I am unconcerned by scratches in its paint job or minor dings in the body that don’t otherwise affect its performance. On the other hand, I may care about my car as a collector’s item, finding fuel efficiency and reliability relatively









tomatoes fear

sadness worry


Figure 10.1 Pattern of emotions focused on tomatoes, constituting import.



unimportant, but get very upset by blemishes in its appearance. In each case, the pattern of emotions and desires partly delimits what is worth attending to and acting on behalf of, thereby defining a description under which their common focus has import to me.

10.3 Respect and Dignity Given this sketch of an account of emotions and import, it is time to return to the question of how to understand the reactive attitudes and, in particular, the ‘human connection’ Strawson saw among them. In initially describing the reactive attitudes, Strawson contrasts the human connection among them with what he calls a ‘logical’ connection, the latter of which he describes as follows: The personal reactive attitudes rest on, and reflect, an expectation of, and demand for, the manifestation of a certain degree of goodwill or regard on the part of other human beings towards ourselves. . . . The generalized or vicarious analogues of the personal reactive attitudes rest on, and reflect, exactly the same expectation or demand for the manifestation of a reasonable degree of goodwill or regard, on the part of others, not simply towards oneself, but towards all those on whose behalf moral indignation may be felt, i. e. as we now think, towards all men. The generalized and non-generalized forms of demand, and the vicarious and personal reactive attitudes which rest upon, and reflect, them are connected . . . logically.13

The ‘logical connection’ Strawson describes involves these emotions’ formal objects: resentment (a ‘personal reactive attitude’), indignation (the ‘vicarious analogue’ of resentment), and guilt (a ‘self-reactive attitude’), for example, are all negative evaluations of someone for violating an expectation or demand of goodwill towards someone else (though the persons involved and their relation to the subject of the emotion may vary from one reactive attitude to the other). It is this logical connection that Wallace focuses on in arguing that we should understand the reactive attitudes rather narrowly, as encompassing as paradigms only these three emotions. By contrast, Strawson describes the human connection as follows: All these three types of attitude [i. e. the personal, vicarious, and self-reactive attitudes] are humanly connected. One who manifested the personal reactive attitudes in a high degree but showed no inclination at all to their vicarious analogues would appear as an abnormal case of moral egocentricity, as a kind of moral solipsist. . . . But it is barely more than a conceptual possibility; if it is that. In general, though within varying limits, we demand of others for others, as well of ourselves for others, something of the regard which we demand of others for ourselves. . . . For all these types of attitude alike have common roots in our human nature and our membership of human communities.14

In this passage, Strawson is pointing not just to certain similarities in the semantic content of the formal objects of various reactive attitudes (the logical connection


Strawson, Freedom and Resentment, 200.


Ibid.: 201.


B E N N E T T W. H E L M

described above); rather, he is pointing to connections that hold between feeling one reactive attitude and feeling others. There would be something odd about feeling resentment towards someone who did something to you and yet to fail to feel indignation towards her were she to do it to someone else or even to fail to feel guilty were you to do that same thing to another person. Similarly, there would be something odd about feeling gratitude towards someone for intervening on your behalf and yet to fail to feel approbation towards her were she to do so on behalf of someone else or even to fail to feel pride were you to do so for another. The oddity, as Strawson indicates, has its source in the kinds of demands or expectations we have of people’s treatment of other people quite generally, but it is an oddity that points to the way these demands or expectations can be unified through the sort of human connection among reactive attitudes he is after. Thus it begins to seem both that the human connection, by unifying such expectations, is more fundamental to understanding the reactive attitudes than the logical connection, which simply takes them for granted, and consequently that Wallace is wrong to focus simply on the logical connection in trying to understand what the reactive attitudes are. I shall try to refine and justify this suspicion in what follows. It should be clear that this human connection, and the demands or expectations on which it rests, are normative, involving standards for how we ought to behave and respond to our circumstances. What exactly is the scope of these norms? As just indicated, the normative expectations, as I shall call them, apply not simply to particular individuals: what I expect of you in your behavior towards me I should also expect of myself or, indeed, of anyone’s behavior towards anyone.15 As Strawson’s appeal to our ‘membership in human communities’ indicates, these normative expectations are not simply my expectations for myself and others, but are rather our expectations for ourselves, for how we are to behave and respond. In short these normative expectations must be interpersonally valid. Moreover, these normative expectations are about how we treat each other as persons, and they arise out of the concern we have for persons as such: we ought to behave and respond to each other in ways that conform to these normative expectations because being a person merits such treatment. And persons merit this treatment because of the kind of import they have: dignity; indeed, it is the dignity of persons that explains the interpersonal validity of these normative expectations. This connection between the dignity of persons and the relevant normative expectations suggests the following hypothesis: whereas the logical connection, articulated in terms of similarities among these emotions’ formal objects, involves an understanding of the reactive attitudes as essentially reactions to the upholding or violation of these normative expectations, the human connection should be articulated in terms of the

15 The normative expectations relevant to the reactive attitudes extend more broadly than this: we might feel indignation or contempt towards someone who tortures animals or vandalizes works of art or destroys natural beauty. Moreover (as I indicated in note 9) in other cases the community of individuals to whom these normative expectations apply can be smaller than the community of all persons. To simplify discussion, however, I shall set both these types of case aside.



reactive attitudes sharing a common focus, namely persons as such, so that the reactive attitudes taken together form a projectible, rational pattern. Given the account of emotions and import offered in section 10.2, this would mean that there is a two-way conceptual connection between such projectible, rational patterns of reactive attitudes and dignity, and the import of persons as such. The idea is that the reactive attitudes are not just responsive to dignity; in addition, projectible, rational patterns of reactive attitudes constitute that dignity. Recall how this works in the case of caring and import (cf. section 10.2). For something to have import is a matter of its being worthy of attention and action; emotions, desires, and evaluative judgments constitute the relevant modes of attention and action. When these responses are tied together in the appropriate sort of rational pattern, we can say that such attention and action are worthwhile in that they are rationally called for by the rest of the pattern. It is in this way that our caring about something—our exhibiting such patterns of emotions—can simply be another way of describing (from a subjective point of view) the very same phenomenon we describe (from an objective point of view) in terms of import. The same applies to dignity: such objective import of persons can be redescribed subjectively in terms of a distinctive mode of caring about persons as such, exhibited in projectible, rational patterns of the reactive attitudes. Indeed, such caring about persons as persons just is respect. So just as particular modes of caring are appropriate responses to import that simultaneously constitute that import, so too particular modes of respect—the reactive attitudes—are appropriate responses to dignity, simultaneously constituting that very dignity.16 This hypothesis, of course, needs to be fleshed out in more detail. But it is worth noting first that this understanding of the human connection in terms of these emotions’ focus largely preserves Strawson’s inclusive interpretation of the reactive attitudes as involving not just guilt, resentment, and indignation but also (contra Wallace) positive emotions responsive to the notable upholding of the normative expectations, like gratitude, approbation, and what Jonathan Bennett calls ‘self-congratulation’.17 Moreover, such an understanding addresses Wallace’s complaint that the inclusive interpretation makes it hard to understand how the reactive attitudes are connected in such a way as both to contribute to our understanding of responsible agency and to make sense of their incompatibility with the objective attitude—that is, with an understanding of persons as mere objects of manipulation or treatment. For the reactive attitudes hang together by virtue of their common focus on persons as such, involving

16 For remarks concerning respect as the appropriate response to dignity, see J. David Velleman, ‘Love as a Moral Emotion’, Ethics 109, no. 2 ( January 1999): 338–74. However, my view differs from Velleman’s in at least two respects. First, whereas Velleman follows Kant in thinking of respect as motivating only by forbidding violations of others’ dignity (on which note the similarity to Wallace’s conception of the reactive attitudes), I understand respect to motivate positive actions, partly through my more inclusive understanding of the reactive attitudes, as will become clear below. Second, Velleman seems to understand dignity to be ontologically and conceptually prior to respect, a priority which I of course am rejecting, instead finding respect and dignity to be ontologically and conceptually on a par. 17 Bennett, ‘Accountability (II)’.


B E N N E T T W. H E L M

an at least implicit understanding of persons that essentially contrasts with that of the objective attitude. In what follows I shall undertake an initial clarification and defense of this hypothesis. In particular, I shall be concerned in the remainder of this section with two issues. First, ‘respect’ can mean many different things, and the connection between respect (in two of these meanings) and the reactive attitudes needs to be clarified; that will be my aim in section 10.3.1. Second and more troubling is a potential disanalogy I have glossed over in comparing the relation between import and caring to that between dignity and respect. Caring is, it might seem, something an individual does, and so the import that is thereby constituted must be relative to that individual; however, although respect, like caring, is something an individual does (or fails to do), how could it constitute the dignity of persons, which surely exists irrespective of whether I respect others or not? Indeed, such a thought is behind the objection to Wallace posed in section 10.1 that the role of the reactive attitudes in an account of responsibility is simply as a middleman that might well be simply eliminated insofar as the question concerning whether someone is free and responsible might seem to be settled prior to anyone’s recognizing this in responding with the reactive attitudes. The solution, as I shall suggest in section 10.3.2, is to understand respect as something we call upon each other to do, whether or not each and every one of us does so individually, so that dignity is the import of persons not simply to me but rather to all persons. 10.3.1 Recognition respect Stephen Darwall usefully distinguishes two kinds of respect, appraisal and recognition respect, as follows: Appraisal respect is esteem that is merited or earned by conduct or character. By contrast, the respect we can demand as persons regardless of our merit [i.e. recognition respect] is no form of esteem at all. Recognition respect concerns not how something is to be evaluated or appraised, but how our relations to it are to be regulated or governed. Broadly speaking, we respect something in the recognition sense when we give it standing (authority) in our relations to it.18

Given this contrast, it is clear that the kind of respect I have suggested is constituted by patterns of reactive attitudes, both positive and negative, would be recognition respect.19 Recognition respect (henceforth just ‘respect’) is a form of caring: caring about someone as being a person. Such caring needs to be distinguished from caring about someone as this particular person, in which your concern is for him as having a 18

Stephen L. Darwall, The Second-Person Standpoint: Morality, Respect, and Accountability (Cambridge, Mass.: Harvard University Press, 2006), 122–3. 19 Darwall worries that the distinction between recognition and appraisal respect cannot hold up, for we regulate our appraisal respect by a proper appreciation of someone as a person, which makes appraisal respect look like an instance of recognition respect (ibid.: 124 n. 9). As I shall suggest below, we can resolve this worry by understanding recognition respect to be an evaluative attitude that is constituted by patterns of reactive attitudes focused on persons as such; among the reactive attitudes making up such a pattern is appraisal respect.



particular identity. Such caring about particular persons, as I have argued elsewhere, is love, constituted by patterns of what I have called person-focused emotions, like pride and shame, which are responsive to the identities of particular persons.20 By contrast, respect is a matter of caring about someone simply as being a person, regardless of her identity; it is an evaluative attitude responsive to her dignity. What, then, is it to be a person, and how should we understand what the dignity of persons consists in? As a first pass, persons are autonomous and responsible agents. This includes not only (a) autonomy and responsibility for action but also (b) autonomy and responsibility for one’s identity, and (c) autonomy and responsibility for belief. In each case, such autonomy and responsibility involve the exercise of a capacity to be responsive to reasons, whether practical or theoretical. Indeed, the normative expectations—what we expect of each other as persons—are defined partly in terms of our autonomous agency: that we persons not (unduly) interfere with each other’s autonomy, that we be, at least to some minimal degree, kind to each other, and so on, for the dignity of persons is such that we merit such non-interference, kindness, and so on. I hypothesized that a projectible, rational pattern of reactive attitudes constitutes respect, where this pattern is defined by the common focus of these emotions. What exactly is that focus? On first glance it may seem that if I respect you, you must be the focus of these reactive attitudes. Thus, I ought to feel gratitude when you do me a good turn (or approbation when you treat someone else well), come to trust you when, having displayed goodwill, you make me a promise, feel betrayed when you violate that trust and subsequently come to feel distrust or apprehensiveness towards you; and of course I ought to feel resentment when you harm me (or indignation when you harm someone else), and guilt or remorse when I harm you. Indeed, appraisal respect and appraisal disrespect would seem themselves to be reactive attitudes: emotions that are elements of this overall pattern. Here we seem to have a pattern of emotions each of which is a commitment to the import of a common focus, namely you as a person, where it is the feeling of the import of this focus that explains the evaluation of the target of each of these emotions and explains the rational interconnections among these emotions. However, such an understanding of the focus of the pattern of reactive attitudes constituting respect is too simple. When I feel indignation for what you did to Alice, is my indignation focused on (and so a mode of respect for) you or Alice? Such indignation does seem to be a part of a pattern of emotions ‘clustered around’ Alice, including gratitude, trust, approbation, and so on. Yet it also seems to be a part of the pattern of emotions ‘clustered around’ you.21 Indeed, that

Bennett W. Helm, ‘Love, Identification, and the Emotions’, American Philosophical Quarterly 46, no. 1 ( January 2009): 39–59; Helm, Love, Friendship, and the Self. 21 It might be thought that indignation which targets you cannot be a part of an attitude of respect for you insofar as it is not a way of giving you some ‘standing’ or ‘authority’ in my relations to you (Darwall, The Second-Person Standpoint: Morality, Respect, and Accountability, 123). However, this would be short-sighted, for in feeling indignation or resentment towards you, I am thereby properly recognizing and responding to your status and import as a person whose behavior is to be governed by the relevant normative expectations. (We do not, after all, feel resentment towards animals, for example.) 20


B E N N E T T W. H E L M



Past-oriented gratitude




Being a person apprehensiveness

Alice resentment indignation

Figure 10.2 Pattern of reactive attitudes focused on being a person (solid arrows) and subfocused on Alice (broken arrows), constituting respect for her as a part of a more general respect for persons generally.

the rational interconnections among the reactive attitudes are not confined to those ‘clustered around’ particular persons is part of Strawson’s point in identifying the human connection among the reactive attitudes. Consequently, if indignation would be appropriate were you to harm Alice in some way, it would also be appropriate were Barry to harm you in the same way; likewise resentment would be appropriate were Charlie thus to harm me, and guilt were I thus to harm Dawn. Moreover, the appropriateness of these emotions does not depend on whether or not I know the parties involved: other things being equal, I ought to feel indignation at seeing a stranger screaming at and hitting her child. If we are to explain these rational interconnections in terms of the way these emotions are commitments to the import of a common focus, we must understand that focus not simply to be a particular person but rather to be all persons generally. In this way, respect for persons is something like a dog lover’s concern for dogs. That concern is general, applying to any and all dogs as dogs, such that it expresses itself in a concern for particular dogs as a part of this general concern. Of course, the dog lover will exhibit patterns of emotions directed at particular dogs that are sub-patterns of the broader pattern constitutive of the general concern. I have argued elsewhere22 that we should understand the emotions making up these sub-patterns to be subfocused on the particular dog and focused on dogs generally, where this talk of a ‘subfocus’ is intended to capture the hierarchical nature of the rational structure of caring involved in such a case. Likewise, my gratitude towards you should be understood as focused on persons as persons and subfocused on you, thus manifesting recognition respect for you as a part of my respect for persons generally. (See Figure 10.2.)23 Yet unlike the concern for dogs, the reactive attitudes generally do not have a single subfocus. As indicated above, my indignation for Helm, ‘Love, Identification, and the Emotions’; Helm, Love, Friendship, and the Self. Indeed, as suggested in note 9, this structure of focus and subfocus should be understood as a feature of recognition respect in all its forms. For, although I have been focused on recognition respect for persons, we can also make sense of finer-grained varieties of recognition respect: for tennis players, for philosophers, for Americans, and so on. In each case, recognition respect for a particular X must be a part of respect for Xs quite generally. 22 23



what you did to Alice is equally a part of a sub-pattern of reactive attitudes subfocused on you and one subfocused on Alice. The reactive attitudes thus generally24 have a dual subfocus, thereby committing one to the import each has as being a person as a part of a more general commitment to the import of persons more broadly. If this is right, then recognition respect of a particular person is constituted by projectible, rational patterns of reactive attitudes subfocused in part on that person and focused on persons generally. 10.3.2 Addressability and the objectivity of dignity I indicated above that dignity, as the import of persons as such, is interpersonally valid, something to which each person ought to respond with respect, regardless of whether each does or not. It seems, therefore, that dignity must be conceptually and ontologically prior to respect, which is best understood as merely a response to that dignity. This may seem to run contrary to my claim that respect, while being responsive to dignity, simultaneously constitutes that dignity. As I indicated above, my response to this objection involves understanding respect to be something we call on each other to manifest towards each other.25 Notice that the rational interconnections among the reactive attitudes are not simply intrapersonal; they are interpersonal as well. If my resentment of you is appropriate, then so is your guilt and another’s indignation. If my approbation of her is appropriate, then so is your gratitude and her self-congratulation. Now in one sense this is exactly what we should expect given the interpersonal validity of dignity. Thus, an affront to someone’s dignity ought, other things being equal, to be met with resentment, guilt, and indignation, and a notable kindness upholding someone’s dignity ought, other things being equal, to be met with gratitude, self-congratulation, and approbation from the ‘victim’, the ‘perpetrator’, and onlookers, respectively. And this is, of course, correct as I have acknowledged: emotions in general are responses to the import of the circumstances, and the reactive attitudes are no exception. Yet it should also be clear that this is far from the whole story. Our reactive attitudes, Strawson tells us, ‘rest on, and reflect, an expectation of, and demand for, the manifestation of a reasonable degree of goodwill or regard’ that people have for each other.26 What we expect here is not simply that people act (or refrain from acting) towards each other in certain ways. For our reactive attitudes themselves are manifestations of our regard—our respect—for each other, and so the reactive


But perhaps not always: Jonathan Bennett describes a case of indignation in response to someone’s heedless disregard for natural beauty; presumably such indignation only has a single subfocus. (See Bennett, ‘Accountability (II)’, }16.) 25 This response, I shall argue, makes dignity intelligible as interpersonally valid, though it must be admitted that it does not on its own provide a full account of the objectivity of dignity; after all, we collectively might be mistaken about what has dignity and what does not. I shall have a little more to say about this in section 10.4, at least pointing in the direction of a more complete account. 26 Strawson, Freedom and Resentment, 200.


B E N N E T T W. H E L M

attitudes themselves are among the responses we normatively expect from each other. Thus, my resentment calls on you to feel guilt and others to feel indignation; my gratitude calls on you to feel self-congratulation and others to feel approbation. In the face of the call of my resentment, your failure to respond with guilt can itself be felt as an affront to my dignity and so be the ground of further resentment. Our reactive attitudes thus call on others to respond with corresponding reactive attitudes of their own. Indeed, our forward-looking reactive attitudes can be calls for each other to behave in certain ways as well: trust, as Karen Jones tells us, involves not just ‘an attitude of optimism that the goodwill and competence of another will extend to the domain of our interaction with her’ but also ‘the expectation that the one trusted will be directly and favorably moved by the thought that someone is counting on her’.27 Being worthy of trust—one aspect of being respectable, of laudably manifesting the import of persons—means, at least in part, being susceptible to this call of trust and to subsequent calls of gratitude or betrayal. In short, the reactive attitudes involve a kind of address to persons, calling for an answer in the context of certain normative expectations; such addressability and consequent answerability make intelligible how ‘the reactive attitudes are incipient forms of communication’.28 In presenting the account of respect in section 10.3.1, I argued that the reactive attitudes generally involve a dual subfocus, committing one to the import of both perpetrator and victim as a part of a more general commitment to the import of persons. We might now put this claim more strongly: in feeling gratitude or resentment or trust or apprehension, it is not simply that I am recognizing that you and I each have import as persons but rather that I am recognizing us as fellow persons, as each rationally bound to the call of the other and so each answerable to the other, at least within certain limits (on which more below). This perhaps can explain Christine Korsgaard’s remarks about how we can obligate each other: If I call out your name, I make you stop in your tracks. . . . Now you cannot proceed as you did before. Or, you can proceed, all right, but not just as you did before. For now if you walk on, you will be ignoring me and slighting me. . . . By calling out your name, I have obligated you. I have given you a reason to stop.29 Karen Jones, ‘Trust as an Affective Attitude’, Ethics 107, no. 1 (October 1996): 8. Nonetheless, I think Jones fails properly to acknowledge the way the competence of others on which we rely in trusting them essentially involves their competence at meeting the normative expectations of the relevant community and so involves their properly respecting others. Here I think Robert Solomon is right to think that trust is part of the background responsiveness we have to each other that is ‘an essential ingredient in’ our general social environment (Robert C. Solomon, ‘Trusting’, in Heidegger, Coping, and Cognitive Science: Essays in Honor of Hubert L. Dreyfus, Vol. 2, ed. Mark A. Wrathall and J.E. Malpas (Cambridge, Mass.: MIT Press, 2000), 231). Indeed, Solomon is right to think that this is a feature that many accounts of trust ignore at their peril, and it is a feature that I think can best be illuminated by understanding trust (and distrust) to be a reactive attitude and so a part of our respect for persons as such. 28 Gary Watson, ‘Responsibility and the Limits of Evil: Variations on a Strawsonian Theme’, in McKenna and Russell, Free Will and Reactive Attitudes, 122. 29 Christine M. Korsgaard, The Sources of Normativity (Cambridge: Cambridge University Press, 1996), 140. 27



In this way, Korsgaard says, ‘you make yourself an end for others; you make yourself a law to them’, and you do this by ‘see[ing] your self, identify[ing] yourself, as just someone, a person, one person among others who are equally real’.30 Where I differ from Korsgaard is in thinking that the origin of this obligation is not to be found primarily in our shared language but rather in our shared emotional commitment to the import we each have as fellow persons.31 Of course, we sometimes overstep these limits of our answerability to each other or disagree about where they lie. If I trust that my junior colleague will mow my lawn this weekend, rather than being susceptible to this call, she will rightly resent my presumptuousness in a way that (we should hope!) is backed up by others’ indignation towards me, calling on me to feel guilty and to apologize. Here my trust presumes a kind of expectation that not only is not shared but also manifests a lack of respect for my colleague as a person. More interesting cases are those of disagreement, where what is at stake is partly the concept of a person. Suppose that, as I struggle trying to do something, you try to offer me a hand, a gesture which I reject resentfully, finding it an affront to my autonomy. Such resentment is consistent with the broader pattern of reactive attitudes (including desires and evaluative judgments) that I have, such that I feel scorn towards those who cannot do things on their own, righteously refuse to help others when they struggle, and so on—a pattern of reactive attitudes that involves an understanding of persons as strongly autonomous, with expectations of non-interference. Now you may well feel differently, resenting my resentment (and, on other occasions, my lack of helpfulness), distrusting me, and so on such that you display a pattern of reactive attitudes that involves an understanding of persons in terms of strong expectations of kindness and compassion. The dispute in this case concerns the concept of a person and the normative expectations it involves, and its resolution must involve something like philosophical arguments concerning that concept. Yet an elucidated concept of a person must be capable of informing our reactive attitudes: a proposed articulation that failed in significant ways to be taken up into our reactive attitudes would need to be reconsidered and potentially revised.32 Indeed, it is partly in this way that the rational patterns of emotions serve to delimit the type of focus to the import of which they are committed. Consequently, the concept of a person, understood not merely as a biological concept but as having normative implications—as the concept of something that has import to us—is not intelligible apart from these patterns of reactive attitudes.


Ibid.: 143. In saying this, I do not mean to deny that the shared nature of language is important. As I shall acknowledge below, to be intelligible as such our reactive attitudes must be informed by the linguistic concept of a person, and the shared nature of linguistic concepts is also important to understanding the unity of human connection that can be found among the reactive attitudes. 32 For detailed arguments supporting this claim, see Helm, Emotional Reason, especially chapter 7. 31


B E N N E T T W. H E L M

There is, of course, no guarantee that the people within a given community will exhibit a wide degree of coherence in their reactive attitudes such that we can say that these reactive attitudes form a single pattern informed by a shared concept of a person. Of course, actual communities of persons will to some degree exhibit failures of coherence in their responses to each other, though clearly minimal such failures will not be problematic and can simply be written off as ‘noise’ in the overall pattern. However, if such failure is significant—if community members are not intelligible as imposing on each other (and responding to) the call I have characterized as essential to the reactive attitudes and if, therefore, they no longer recognize each other as fellow persons—then such failure can undermine the idea that members of the community even have the capacity for the reactive attitudes at all. Yet when that call is in place, it and its communicative function, together with its essentially shared nature of linguistic concepts, impose considerable pressure for the formation of such a single pattern, informed by the concept of a person. When such a pattern does form—when, as Strawson says, the reactive attitudes of different people ‘have common roots in our human nature and our membership of human communities’33—the pattern is that of the whole community, constituting the shared import persons have to that community; such import is therefore the interpersonally valid import of fellow persons as such. Such import, I submit, is dignity.

10.4 Holding Responsible and Taking Responsibility I have hypothesized that the reactive attitudes are emotions that are essentially focused on persons generally, such that projectible, rational patterns of reactive attitudes constitute caring about persons as such—they constitute respect. Such rational patterns, moreover, are interpersonal by virtue of the way the reactive attitudes involve addressability—a call for the other to respond—and answerability, so that particular reactive attitudes involve the sense of ourselves as fellow persons bound together in a normative community. Consequently, the import these rational patterns constitute, and to which we respond with respect, is interpersonally valid: it is dignity. So what does all this tell us about the nature of responsible agency? Recall Wallace’s claim that being responsible requires being fairly held responsible, which in turn requires having powers of rational self-control. My criticism was that this makes it seem like we can cut out the middleman of our holding each other responsible and instead understand responsibility directly in terms of these powers of rational selfcontrol unless rational self-control itself is intelligible only in terms of one’s being held responsible. We are now in a position to see how this might be true. Although in understanding the reactive attitudes to be responses to the import of persons as such I have focused on these emotions as modes of attending to that import,


Strawson, Freedom and Resentment, 201.



we must not forget that for something to have import it must not only be worthy of attention but also worthy of action as well. Consequently, emotions are a source of motivation for action, and the reactive attitudes are no exception; indeed, in the case of the reactive attitudes our being motivated to act in various ways is a part of our answerability to the call of others. Of particular relevance to understanding responsible agency are what Strawson calls the ‘self-reactive attitudes’: reactive attitudes that are, as we can now say, subfocused on oneself as the ‘perpetrator’, such as guilt, remorse, and self-congratulation. For in this case our answerability to others, as the capacity to be appropriately responsive to their dignity in their role as ‘victims’ of our actions and responses, just is a capacity to take responsibility for those actions and responses, as when guilt motivates us to apologize and make amends. Although I have thus far focused on various emotions, we can press further this idea that import involves worthiness of action by recognizing that desires and evaluative judgments are a part of the same pattern of reactive attitudes constituting respect.34 For if I am to be properly responsive to your dignity, I ought, other things being equal, to be motivated by an awareness of the way the relevant normative expectations apply in a particular circumstance; such motivation can come either directly by having the relevant desires or indirectly by making the relevant practical judgments and exercising my will to act accordingly.35 Thus, I do not push ahead of you in line at the store to get the last loaf of bread, even though I am hungry, because as a part of respecting others I recognize that taking turns is what is (normatively) expected. This is a kind of rational self-control, but it differs from the kind of rational self-control a dog exhibits in delaying snatching the treat balanced on its nose in that it is responsive to the dignity of others; as such, we might say in a Kantian vein, it is motivated out of duty. It is, I submit, only the later sort of rational self-control that is relevant to responsibility. The upshot is that whereas Wallace claims that being responsible requires being appropriately held responsible which in turn requires that one can exercise rational selfcontrol (and, we might add, take responsibility for what one does), we can now see that exercising the relevant sort of rational self-control and taking responsibility themselves require that one be a fellow person with others in a community of respect in which each can hold the others responsible for their actions. Hence being responsible requires not only being properly held responsible, but also, what is more fundamental and affords greater insight, requires being appropriately both a subject and an object of recognition respect—requires being a fellow person in a community of persons. The obvious objection to this line of thought concerns the status of the relevant communities. Surely such communities cannot be defined by the de facto rational interconnections we happen to find among their reactive attitudes. After all, members of certain groups (slaves, women, foreigners, and so on) might fail to be fully a part of 34

I have argued for this claim in Helm, Emotional Reason, chapter 5. For details on how I think we can exercise the will by making evaluative judgments, see ibid., especially chapter 6. 35


B E N N E T T W. H E L M

the relevant community of fellow persons as a matter of fact, but that does not mean they are not fellow persons who are fully responsible. In short, the community can go wrong about who belongs to it and so who the fellow persons are, and this suggests that being held responsible is not necessary for being responsible, contrary to what I just argued. Moreover, on the account just presented it looks as though a particular person could simply opt out of this community of persons by refusing to participate in the reactive attitudes, thereby failing to be responsible. And that too seems wrong: misanthropes can be responsible—and can properly be held responsible—too. In reply, it is surely right that more needs to be said about who ought to belong to such a community of fellow persons—about who is owed respect and so who really has dignity, whether we members of this community recognize it or not. However, I am not able to say much more about this here: doing so would amount to presenting substantial chunks of a completed meta-ethics. Indeed, it is largely for this reason that I have described this account as a hypothesis, to be advanced only tentatively. Nonetheless, my claim is that however we ought to understand such rational standards for community membership, however we ought to understand the concept of a person and its extension, we cannot do so except from within our evaluative attitude of respect, suitably modified and enriched. For, as I have argued, respect, dignity, the concept of a person, and the capacity for the reactive attitudes are all conceptually interconnected, with none of them intelligible apart from the rest. And that provides a kind of research program into meta-ethics that is worth taking seriously. All of this suggests that Strawson was right: being responsible is something we can only understand in a broader context of holding each other responsible and so, I would add, being able to take responsibility yourself; all of this requires understanding the reactive attitudes as involving a distinctive human connection. And it suggests that Wallace is right, too: if we are to understand the connection between being responsible and holding responsible, we must understand it to be a normative connection. In each case, we can understand what is right about the view only through the notion of the focus of a reactive attitude and thereby to the relation between the reactive attitudes and both respect and dignity. In sum, being responsible is a matter of being susceptible, out of respect, to the call of another’s dignity.36

36 Thanks to the audience at the conference on Emotions and Agency (Centre for the Study of Mind in Nature, Oslo), Alexa Forrester, and especially Kathryn Kutz for helpful discussions on the subject matter of this paper.

11 Guilty Thoughts Angela M. Smith

11.1 Introduction Most of us, at some time or another, have felt guilty over an unexpressed thought, attitude, or emotion. Perhaps we have caught ourselves taking secret pleasure in a close friend’s uncharacteristic failure, or feeling distrustful of a loved one’s fidelity, or viewing a stranger through the lens of an odious stereotype. Perhaps we have received the confidences of another with contempt, or have felt resentment rather than gratitude toward someone who has done us a kindness. It is quite common, in such cases, for people to say that they feel ‘guilty’ about these thoughts and attitudes, even when they are quite sure that they have not been, and will not be, expressed or acted upon in any way. The feeling of guilt, it seems, attaches to the mere having of these thoughts and attitudes, and is not inhibited by the knowledge that these mental states are and will remain wholly private. There is something rather puzzling about these self-attributions of guilty feelings, at least if one accepts a commonly held view about the conditions under which it is rational or appropriate to feel guilt.1 According to this common view, it is rational to feel guilty only if one is in fact guilty of some moral transgression; and one is in fact guilty of a moral transgression only if one has freely and knowingly performed some action that violates a moral duty we owe to others. Private, unexpressed attitudes, it might seem, are not ‘actions’ in the relevant sense, nor are they the sorts of things that we can reasonably be said to ‘owe’ to others. So it is unclear what we could be feeling guilty about in cases of the sort mentioned above. It is possible, of course, that those of us who feel guilty about unexpressed attitudes are simply feeling guilt irrationally. There are other cases in which we are inclined to say that a person’s guilt is irrational, such as the continuing guilt a lapsed Catholic may feel about sexual practices she no longer believes to be immoral. Perhaps what we have

1 Morris (1987) draws attention to this puzzle and considers various ways of responding to it. Though he ultimately defends a position very different from the one I defend in this chapter, I am much indebted to his discussion.



here is just another instance of the same phenomenon—in this case, perhaps, a residue of our early childhood inability to see the difference between thought and deed. But I’m interested in exploring another possibility, namely, that our guilt feelings in such cases may in fact be a justified response to a genuine moral transgression. If we take this possibility seriously, we will need to ask what kind of moral transgression could be involved in an unexpressed attitude, and to consider whether any plausible account of morality could allow for the existence of such moral transgressions. My aim in this chapter is to argue that we can make sense of the idea of ‘wronging’ another through an unexpressed attitude, and that a contractualist theory of morality can provide us with a compelling account of the normative basis of such attitudinal wrongs. Before turning to the positive part of this chapter, however, I need to spend some time looking at some objections that have been raised to the very idea that we could morally owe attitudes to others. Some have argued that we lack the kind of control over our attitudes that is necessary in order for them to be subject to moral demands. Others have argued that since unexpressed attitudes cannot cause harm to others, we cannot have a moral obligation to have any particular attitudes toward them. Still others have argued that it would be psychologically dangerous to allow moral obligations to extend into the sphere of private thought and feeling. I will go through these objections fairly quickly, in the hope that I can say enough in response to at least leave open the possibility that we could be morally obligated to have certain attitudes toward others. In the second part of this chapter, I will consider two proposals about how to make sense of our guilt over unexpressed attitudes that do not depend on the claim that we have a moral obligation to others to think or feel about them in certain ways. A virtue ethicist, for example, might argue that some thoughts, emotions, and attitudes reveal vices of character, and our guilty feelings are a justified response to the recognition of such vices. A consequentialist, on the other hand, might argue that bad attitudes have a tendency to lead to wrongful actions, and our guilty feelings serve as a useful warning sign to deter us from behaving in morally impermissible ways. Though I think there is much to be said for these proposals, I will argue that they cannot in fact make sense of our guilt feelings in the cases under consideration. We must look elsewhere, I believe, for the moral underpinnings of these feelings. In the third part of the chapter, I will argue that a contractualist moral theory, of the sort that T.M. Scanlon defends, can provide the requisite moral underpinnings. According to Scanlon (1998: 162), the distinctive importance and authority of moral requirements is grounded in a certain ideal of relations with others, which he describes as a relation of mutual recognition.2 But in order to genuinely stand in this relation to

2 See also Scanlon (2008). I suspect Scanlon himself would not accept my claim that it is possible to ‘wrong’ another through an unexpressed attitude (since he denies that questions of permissibility apply to attitudes). But he certainly accepts that certain attitudes can be ‘blameworthy’ in virtue of falling short of the standards required for standing in certain relationships with others. Guilt, in turn, is a reasonable response to the judgment that one’s attitudes are blameworthy in this sense.



others, I will argue, we must have certain attitudes toward them. If this is correct, then our feelings of guilt over unexpressed attitudes may reflect a judgment that we have failed to accord to others the basic respect and recognition they are due as fellow members of the moral community. Far from being irrational, then, our guilty feelings can be seen as a profound expression of our aspiration to live with others on mutually acceptable terms.

11.2 Three Objections Let me begin, then, by looking briefly at three objections to the very idea that we can be morally obligated to have certain attitudes toward others. The first objection challenges the claim that our thoughts, emotions, and attitudes can be subject to moral demands, on the grounds that we do not have the kind of control over these states that is necessary in order for it to be reasonable to expect us to comply with such demands. Call this the Control Objection. The second challenges the claim that we can wrong another if we do not harm her in any way. Since unexpressed attitudes (by stipulation) do not cause any material harm to others, we cannot be said to ‘wrong’ another simply by having an unexpressed attitude toward her. Call this the No-Harm Objection. The third objection challenges the claim that moral obligations apply to attitudes on the grounds that this would be psychologically damaging to us. It might seem that morality is demanding enough when it comes to our actions; to allow it to extend into the mental realm would be downright oppressive. Call this the Psychological Health Objection. Let’s examine each, in turn. 11.2.1 The Control Objection The central intuition underlying the Control Objection is that we cannot have a moral obligation to do something it is not within our power to do. So, for example, I cannot have a moral obligation to lift a car off an injured person because it is not within my physical power to do so. (Superman, on the other hand, might well have such an obligation.) This intuition needs some refinement, of course, since in some cases our inability to do something is a direct result of earlier acts of negligence on our part. So, for example, the fact that I do not have control over my car while driving in a selfinduced drunken stupor does not relieve me of my moral obligation not to hit pedestrians. But, with suitable refinements, it seems plausible to hold that a moral demand is unreasonable if it is addressed to an agent who simply does not have it within her power to comply with it.3 Now, if we think about particular states of desire, emotion, and belief, it seems clear that they are not under the control of the will in the way that our ordinary physical actions are. I cannot stop feeling angry at will or summon up a desire to help others at 3 The principle underlying this intuition seems to be Kant’s well known ‘ought implies can’ principle. Sher (2005) raises some compelling doubts about the justification of this principle.



will, for example. Nor can I believe things at will or stop caring about things at will. Though there are things we can do to try to influence our attitudes in certain ways, it is also notoriously true that these efforts are not always effective. Reflecting on the apparent unruliness of these states, it is natural to ask the following question: if we do not have direct control over our attitudes, then how can we be morally obligated to have particular attitudes? Morality is supposed to be ‘action-guiding’; it cannot tell us to do (or fault us for not doing) something that is not within our power to do. R. Jay Wallace (1996) gives a clear statement of what I have called the Control Objection. In his view, moral obligations can apply only to states that are susceptible to being influenced directly by reasons: That is, what one is obligated to do must be the sort of thing that could be motivated by one’s grasp of the reasons expressed in moral principles; otherwise the commitment to justification that is inherent in the stance of holding people responsible cannot be sustained. But neither emotions nor feelings nor mere bodily movements appear to be phenomena of this sort. Emotions and feelings have their reasons of course, in the sense that they often have propositional objects, which are explained by their privileged connection to certain sorts of beliefs . . . Particular states of emotion or feeling, however, are not the sorts of states that can directly be controlled by the reasons expressed in moral principles: such states as love, esteem, and goodwill are generally not states that could be produced simply by the belief that there are moral considerations that make them obligatory. This is why we cannot plausibly interpret moral obligations as governing the quality of people’s will, where such qualities are construed broadly, to encompass emotions and feelings quite generally. (Wallace (1996): 131–2)4

According to Wallace, ‘only if an action expresses a choice of some sort can we say that a moral obligation has either been violated or complied with’ (1996: 128). We reveal our commitment (or lack of commitment) to moral principles through the choices expressed in our actions, not through the attitudes, emotions, or feelings we happen to have toward others. On this view, we cannot wrong another simply by having an attitude of contempt, ill will, or distrust toward her; we can only wrong her by impermissibly choosing to express these attitudes in some way through our actions. I think Wallace is right to emphasize the connection between our moral obligations and our capacities to recognize, assess, and respond to reasons. It seems intuitively correct to say that ‘what one is obligated to do must be the sort of thing that could be motivated by one’s grasp of the reasons expressed in moral principles’. And I think he is right to say that this close connection is essentially tied to the commitment to justification implicit in attributions of responsibility. To say that a person is responsible for some thing is to say that she is a legitimate target of demands for justification for that

4 For other explicit statements of this view, see Sidgwick (1981) and Taylor (1970). Taylor is particularly explicit on this point: ‘One can do certain things . . . but emotions can only . . . be “suffered” . . . One can . . . move his limbs, voluntarily, or even pursue a certain train of thought, voluntarily; but one can in no similar way feel elation or dejection . . . [O]ne could be commanded . . . to swing his arms (an action) but could not intelligibly be commanded to love or hate (passions)’ (1970: 241; cited in Thalberg (1978): 381).



thing. It would be inappropriate to demand a justification for something that does not reflect an agent’s judgments about reasons, such as her athletic ability or her shoe size. ‘Why are your feet so big?’ may be a demand for explanation, but it cannot be a demand for rational justification. ‘Why did you punch Bobby in the nose?’, by contrast, is a demand for justification: that is, a demand that the agent either justify her action with reasons, or acknowledge and respond appropriately to the fact that it cannot be so justified. The question, then, is whether this sort of demand is appropriate only with regard to phenomena that can be ‘directly produced’ by our judgments about reasons. Wallace is concerned about the fact that our emotions and feelings cannot always be summoned ‘at will’ when we judge that they are morally obligatory; and he concludes from this fact that they are not the sorts of states that can be subject to moral demands. But the relevant question, it seems, should not be whether a state can be produced ‘at will’, but whether it is the sort of state that is ‘susceptible to being influenced directly by reasons’, for it is the connection to reasons that makes demands for justification appropriate. Perhaps an analogy would be helpful at this point. We generally hold people responsible for their factual beliefs, and we think those beliefs are appropriately subject to certain rational ‘demands’. A person who believed something on the basis of obviously insufficient evidence, or who reasoned to her conclusion in an obviously fallacious manner, is open to rational criticism for failing to meet the epistemic requirements that apply to her. Insofar as beliefs are judgment-sensitive states, they are the sort of thing that it is reasonable to expect people to be able to justify. But few would say that we have voluntary control over our beliefs, nor would they say that these epistemic demands are unreasonable because we do not have it within our power to produce a belief ‘at will’. A similar thing should be said, I think, when it comes to desires, emotions, and other attitudes. These, too, are judgment-sensitive states, and it makes sense to expect people to be able to justify them.5 When we react in a prejudiced way toward another, for example, that usually reflects an implicit judgment based on reasons. A feeling of distrust, for instance, may reflect the judgment that a person is unlikely to perform a certain task well because of the color of her skin or her ethnicity. A feeling of guilt in response to such a reaction, in turn, may reflect the judgment that distrusting someone on the basis of such reasons is morally unjustified. In both cases, our emotional reactions are sensitive to our (in some cases mistaken) judgments about reasons, even if these states cannot be produced ‘at will’. So if what matters in determining the reasonableness of a moral demand is that it be addressed to a state that is susceptible to being influenced directly by reasons, I do not see why our desires, emotions, and other attitudes should be excluded from the scope of moral requirements. As the case of


I defend this claim more fully in Smith (2005).



guilty feelings nicely illustrates, in many cases our emotions are ‘directly controlled’ by the reasons expressed in moral principles. I am not convinced, therefore, by the Control Objection. There may be other reasons to think that moral obligations should be confined to actions, but the fact that we lack direct volitional control over our attitudes does not seem sufficient to show that they cannot be subject to moral demands. 11.2.2 The No-Harm Objection Even if I am correct that there is no reason, in principle, that moral obligations cannot apply to attitudes, it might still be argued that such obligations would be unreasonable on other grounds. For example, one might follow John Stuart Mill (1978: ch. 4) in claiming that ‘The only purpose for which power can be rightly exercised over any member of a civilized community against his will, is to prevent harm to others.’ Mill clearly intended his Harm Principle to place limits on the scope of both moral and legal obligations, as he included ‘the moral coercion of public opinion’ among the forms of ‘power’ that society can exercise over an individual. But if preventing harm to others is the only legitimate basis upon which a moral obligation can be justified, then it would seem that our attitudes cannot be legitimately subject to moral demands. For it does not seem possible to harm another simply by holding an attitude toward her. We may have moral obligations not to express certain attitudes, because of the harm such expression might cause. But we cannot be obligated simply to have (or not have) certain attitudes, because we cannot harm someone simply by thinking or feeling about her in a certain way. If it is not possible to harm another through an unexpressed attitude, in other words, it is not possible to ‘wrong’ her through one. In order to think clearly about this issue, however, we need to look more closely at the notion of a ‘harm’ and its relation to the notion of a ‘wrong’. As Joel Feinberg points out, the word ‘harm’ ‘is both vague and ambiguous, and entangled with other concepts, like “wrong”, in ordinary usage’ (1990: xxviii). Sometimes when we speak of a ‘harm’, what we mean is a setback to the agent’s interests. Thus a person can be harmed by the deliberate actions of others (when those actions thwart or in some other way adversely affect her interests), but one can also be harmed in this way by accidents of nature, malfunctioning machines, bad luck, etc. It is natural to assume that a person who is harmed in this sense must be aware of the harm done to her, if it is genuinely to count as a setback to her interests. But this will depend upon how widely or narrowly we specify both the notion of an ‘interest’ and the notion of what legitimately counts as a setback to it. As Feinberg points out, however, one might also speak of a ‘harm’ in a different sense, as ‘the concept of a suffered wrong or violated right’ (1990: xxviii). It seems one could suffer a harm in this moralized sense without suffering a harm in the nonmoralized sense of a ‘set-back to interest’. For example, I might ‘harm’ you in this moralized sense by wrongfully reading your diary without your permission, but in doing so I might gain valuable information about your likes and dislikes that actually helps me to advance your interests more effectively. In such a case, it seems I have



harmed you in the moralized sense, but I have not harmed you in the non-moralized sense of adversely affecting your interests.6 What we need to know, then, is what sense of ‘harm’ is at work in the argument sketched above against the claim that we can be morally obligated to have certain attitudes toward others. If the claim is that the only justification for a moral obligation is to protect others from suffering a setback to their interests (that is, from suffering a harm in the non-moralized sense), then we will have to admit that we have no moral obligation to refrain from secretly reading another’s diary (assuming we can do so without causing any setback to her interests). But this seems implausible. There are just too many cases in which it seems clear that we have a moral obligation to refrain from certain activities, even when we are convinced that the activities in question will have no or even a positive effect on another’s interests. The concept of what it would be wrong to do is not so closely tied to the concept of a setback to interests. This leaves us, then, with the moralized sense of ‘harm’—the concept of a suffered wrong, or violated right. But if we interpret the No-Harm Objection using this sense of ‘harm’, it simply begs the question. For now the argument would be that we cannot have a moral obligation to have certain attitudes toward others, because we can only be morally obligated to respect others’ rights, and no one has a right to expect certain attitudes from us. But the question at issue is precisely whether people are entitled to expect certain attitudes from us. The No-Harm Objection, understood in this sense, gives us no independent reason for thinking the answer to this question is ‘no’, for it simply assumes that which it is intended to prove. It seems to me, then, that we cannot dismiss the idea that we might morally owe certain attitudes to others on grounds that it is not possible to harm another through an unexpressed attitude. And, for the record, I’m also not entirely convinced that we cannot do harm to another, in the non-moralized sense of ‘causing a set-back to her interests’, simply in virtue of holding an unexpressed attitude toward her.7 But I will not rely on this controversial claim in anything that follows. 11.2.3 The Psychological Health Objection Let me turn, then, to a final worry one might have about the claim that we can be morally obliged to have certain attitudes toward others. It has sometimes been suggested that the mental sphere must be kept free from the oppressive constraints of morality, and that seeking to impose moral order on what Thomas Nagel has called ‘the sheer chaotic tropical luxuriance of the inner life’ would be psychologically 6 For a helpful discussion of some of the difficulties Mill’s Harm Principle faces in dealing with cases of ‘harmless wrongdoing’, see Ripstein (2006). 7 Dan-Cohen (1999) argues that thoughts by themselves can have negative ‘extrapersonal effects’, and therefore one can cause what I have called ‘non-moralized’ harm to others simply by holding unexpressed attitudes toward them. Though I find Dan-Cohen’s argument intriguing and suggestive, I think it is also possible to wrong another through an unexpressed attitude even if it does not cause any ‘relational’ harms of the sort he discusses.



disastrous for us (1998: 4). In his article ‘Concealment and Exposure’, for example, Nagel expresses concern about the psychological effects of allowing socialization to penetrate the private sphere of thought and feeling: ‘To internalize too much of one’s social being and regard inner feelings and thoughts that conflict with it as unworthy or impure is disastrous. Everyone is entitled to commit murder in the imagination once in a while, not to mention lesser infractions’ (1998: 7). In this he echoes Freud, who criticizes the ‘cultural super-ego’ for failing to limit its demands to what is psychologically attainable by us: [The cultural super-ego] . . . does not trouble itself enough about the facts of the mental constitution of human beings. It issues a command and does not ask whether it is possible for people to obey it. On the contrary, it assumes that a man’s ego is psychologically capable of anything that is required of it, that his ego has unlimited mastery over his id. This is a mistake; and even in what are known as normal people the id cannot be controlled beyond certain limits. If more is demanded of a man, a revolt will be produced in him or a neurosis, or he will be made unhappy. The commandment, ‘Love thy neighbour as thyself ’, is the strongest defence against human aggressiveness and an excellent example of the unpsychological proceedings of the cultural super-ego. The commandment is impossible to fulfil; such an enormous inflation of love can only lower its value, not get rid of the difficulty. Civilization pays no attention to all this; it merely admonishes us that the harder it is to obey the precept the more meritorious it is to do so. (1961: 90)

Two distinct claims can be gleaned from these passages by Nagel and Freud, both of which seem eminently reasonable but fail to show that moral obligations do not apply to attitudes. The first claim is that moral norms regulating the appropriate public expression of attitudes should not themselves be thought to regulate the content of those attitudes. It is ‘pernicious’, Nagel claims, to think ‘that socialization should penetrate to the innermost reaches of the soul, so that one should feel guilty or ashamed of any thoughts or feelings that one would be unwilling to express publicly’ (1998: 9). This is quite right, and it would be a serious mistake to think that one should feel guilty about having an attitude simply because one would be unwilling to express it publicly. Our concern for other people’s feelings and for the maintenance of certain relationships should make us reticent to express many attitudes that are not morally objectionable in themselves. For example, one might be quite justified in feeling annoyed by the endless boasting of a conceited colleague, but judge, correctly, that it would be damaging to departmental collegiality to express this annoyance in any way. Here the attitude might be entirely warranted, even if it would be morally objectionable to give voice to it. In some cases, moreover, it is simply not our place to tell other people what we think. It may be inappropriate for me to express to an acquaintance my disapproval of the way she treats her husband, for example, even if my attitude itself is perfectly warranted and even if I think it would be a good thing if she came to see her own behavior in this light. My unwillingness to express such an attitude may reflect, among other things, an appreciation of the fact that I do not stand in an appropriate



relation to her to do so, and that I do not have a relevant stake in the matter. As Nagel rightly points out, public conventions of reticence and discretion exist in part to make possible the smooth functioning of social life, and part of what this requires is a fairly well-developed sense of what is, and what is not, ‘one’s business’. But it would be a mistake for me to conclude that, because I am unwilling to publicly express my negative attitude in this situation, I should feel ashamed or guilty for having it. As long as I believe the attitude in question is morally justified, there is no reason for me to feel guilty for having it. In some cases, however, it seems we are unwilling to publicly express an attitude precisely because of its objectionable content (and not merely because it would violate these social conventions of reticence and discretion). Consider, for example, a person who considers herself very progressive and liberal, but who discovers she is harboring serious prejudices against homosexuals. She might be unwilling to publicly express her attitudes precisely because she believes that they cannot be morally justified. In cases of this sort, I would argue, there might well be grounds for feeling guilty. But the guilt in such a case does not reflect a mistaken application of public norms of reticence to the private sphere of judgment; it reflects an independent assessment that the attitudes themselves are morally unjustified, and that they ought not to be held. Nagel is not always clear about this distinction. He seems to conflate cases in which we mistakenly feel guilty about an attitude simply because it does not conform to an acceptable public persona (which can lead to repression), and those in which we may be justified in feeling guilty about an attitude because in itself it reflects a disrespectful, belittling, or prejudiced attitude toward another. He may conflate these cases because he assumes that morality is exhausted by these norms of social interaction (and hence that there never are moral grounds for feeling guilty for an unexpressed attitude). But that is a substantive moral claim that would require independent defense. The second reasonable claim implicit in the passages by Nagel and Freud is that we should not feel guilty about failing to meet morally unreasonable psychological demands. A moral theory that demanded that we love everyone equally, or that we never feel anger or resentment toward those who have wronged us, would be psychologically unrealistic because the demands it makes are rationally and morally indefensible: love (or at least most forms of love) requires a kind of intimacy with others that it is neither possible nor desirable to try to achieve with humanity at large; resentment expresses a demand for reasonable regard, which is a perfectly appropriate response to have toward those who have flouted that demand in their treatment of us. We do not have unlimited ‘control’ over our attitudes because we do not have unlimited control over the reasons that count in their favor. But there is no reason to think that the moral demands applying to attitudes would have to be unreasonable in this way. Morality may demand that we not harbor malicious, disrespectful, or prejudiced attitudes toward others, but this demand seems morally justifiable and therefore psychologically unexceptionable. Essentially, it is a demand that our attitudes toward others be based on morally justifiable grounds. It is far from obvious why the



attempt to meet such a demand should be thought to lead to the kind of repression and neuroses that Nagel and Freud fear. If this is correct, then the sorts of concerns Nagel and Freud raise about the psychological dangers of applying moral obligations to the mental sphere rest upon a misunderstanding of what those moral obligations would have to be like. There is no reason to think that these obligations would require that we become moral saints, never entertaining a negative thought toward others, or that we feel guilty for any attitude we would be unwilling to express publicly. For the sad fact of the matter is that in many cases negative attitudes toward others are perfectly warranted (even if it would be wrong to express them in some cases), and an acceptable morality would acknowledge that fact.

11.3 Character and Consequences I hope at this point to have at least raised doubts about some common arguments that have been given for why our attitudes are not subject to moral demands. If the very idea of such attitudinal demands is not conceptually or practically untenable, then there is room to say that feelings of guilt over unexpressed attitudes might well be rational in some cases. Such feelings might be a justified reaction to a violation of a moral obligation we have to others. It may be, however, that we do not need to appeal to the existence of such moral obligations in order to make sense of our guilt feelings over unexpressed attitudes. There are various ways in which we might argue that our attitudes are a legitimate subject of moral concern without committing ourselves to the strong claim that we morally owe it to others to think or feel about them in any particular way. In this section, I want to look at two proposals of this sort, one from a virtue-ethical perspective and the other from a broadly consequentialist perspective. Though I think there is much to be said for these proposals, I will argue that they cannot really account for the nature of our moral experience in cases where we feel guilty about an unexpressed attitude. In the next section, I will suggest that a contractualist moral theory does a better job of explaining the moral basis of our guilty feelings in such cases. 11.3.1 The virtue-ethical response A virtue ethicist reading this chapter might well be perplexed by the suggestion that there is any puzzle at all to be solved regarding our guilty feelings over unexpressed attitudes. The sense that there is a puzzle, she might claim, is due to a misguided picture about the nature of morality—a picture that is, unfortunately, still widely accepted. According to that misguided picture, morality is fundamentally about ‘what we owe to others’, and can be understood primarily in terms of overlapping sets of rights and duties. The recent movement in moral philosophy toward more virtue-based approaches to ethics attests to the dissatisfaction many philosophers now feel with the standard rights- and duties-based theories of Kantianism, contractualism, and utilitarianism.



Rosalind Hursthouse, for example, gives representative voice to this dissatisfaction in her recent book On Virtue Ethics when she writes, [I]t has become all too common to allow a vague concept of justice and rights to encompass large areas of morality that virtue ethicists believe are better dealt with in terms of other, more concrete, virtues. According to virtue ethics . . . what is wrong with lying, when it is wrong, is not that it is unjust (because it violates someone’s ‘right to the truth’ or their ‘right to be treated with respect’) but that it is dishonest, and dishonesty is a vice. What is wrong with killing, when it is wrong, may be not so much that it is unjust, violating the right to life, but, frequently, that it is callous and contrary to the virtue of charity. (1999: 6)

Once we begin thinking of morality more broadly in terms of the cultivation of virtues of character, however, there should be nothing puzzling in the idea that attitudes should be seen as a central focus of moral concern. To have a virtue, after all, involves not only acting well, but also having appropriate attitudes, emotions, and patterns of awareness. Attitudes can be assessed as morally good or bad (or virtuous or vicious) from a virtue-ethical perspective, therefore, without our having to establish that we ‘owe’ such attitudes to others, or that others can reasonably ‘demand’ them of us. I have a great deal of sympathy for this proposal, and more generally for the idea that morality itself should be understood as concerning more than just ‘what we owe to others’. And I certainly agree that people harboring malicious, belittling, or prejudiced attitudes toward others can be said to be lacking in virtue. Virtues are states of character that dispose us to respond well in both our actions and our passions.8 A person who is somehow able to perform the right actions but whose attitudes are deficient has not yet achieved full virtue. Even if these attitudes are never expressed, their presence indicates a defect of character, a defect, moreover, that reflects poorly on the person who possesses it. My concern about this proposal, however, is that it does not seem to me to adequately capture the nature of our moral experience in cases where we feel guilty about an objectionable attitude toward another. The virtue-ethical approach directs our attention to the character of the agent who possesses the objectionable attitude, and (correctly) points out that such an attitude indicates a moral deficiency—a vice—in that person’s character. An agent who accepts this judgment and who cares about being a good person will likely be pained by this recognition, and will see this as an occasion for moral self-improvement. But what this account does not seem to capture—and what Hursthouse, at least, thinks it need not and should not try to capture—is the further idea that in holding such an attitude toward another, we have done a wrong to her. The sense that we have been unfair, unjust, or uncharitable to someone else, and that she is

8 As Aristotle puts it, ‘Virtue, then, is a state of character concerned with choice, lying in a mean, i.e. the mean relative to us, this being determined by a rational principle, and by that principle by which the man of practical wisdom would determine it. Now it is a mean between two vices, that which depends on excess and that which depends on defect; and again it is a mean because the vices respectively fall short of or exceed what is right in both passions and actions, while virtue both finds and chooses that which is intermediate’ (1980: 1106b35–1107a5).



owed better, is not captured by the idea that we are morally deficient in holding such an attitude.9 This is a significant omission, I think, because many of us feel that we have not only perfectionistic reasons for modifying objectionable attitudes (i.e. reasons stemming from our desires to become better people), but also reasons of justice for doing so (i.e. reasons stemming from the legitimate moral claims of others). And this corresponds to the fact that we generally feel guilt rather than shame upon the recognition that we are harboring a morally unjustified attitude toward another: our attention in such cases is usually directed toward the unmet claims of others, rather than toward the gap between who we are and who we would like to be.10 Let me pause for a moment to say a bit more about the significance of this distinction. Much has been written over the last thirty years or so about the difference between feelings of guilt and feelings of shame. Rawls was one of the first to highlight this distinction in A Theory of Justice (1971), and many now endorse his basic analysis of the normative bases of these moral emotions and how they differ. Guilt, on Rawls’s view, tends to direct us toward ‘the infringement of the just claims of others and the injury we have done them’, while shame tends to direct us toward ‘our disappointment with ourselves for failing to live up to our ideals’ (1999: 391; cf. 420–5). Though the very same action can often occasion both guilt and shame, their normative focus is quite different. As Rawls puts it, ‘In general, guilt, resentment, and indignation invoke the concept of right, whereas shame, contempt, and derision appeal to the concept of goodness. And these remarks extend in an obvious way to feelings of duty and obligation (if there are such), and to proper pride and a sense of one’s own worth’ (1999: 423). The question, then, is how we should understand the negative feeling we have when we recognize that we are holding an objectionable attitude toward another. In my experience, that feeling does have elements of shame (‘I’m a bad person for thinking that’), but the predominant feeling is one of guilt (‘She deserves better’). Of course, it is possible that people’s experiences vary in this regard, and for some the feeling of shame might fully capture the emotion’s moral content. But for those of us who feel elements of guilt as well, the virtue-ethical approach will seem to leave something important out of the picture. 9 I believe something like this point underlies Watson’s distinction between what he calls ‘aretaic appraisal’ and ‘accountability appraisal’ (Watson 1996). He points out that we can morally appraise a person’s excellences and faults—or virtues and vices—without thinking that that person is accountable to us to have or to cultivate these character traits. As he puts it, ‘Unless we think she is responsible to us or to others to live the best life she can—and that is a moral question—we do not think she is accountable here’ (1996: 231). In the case of many moral faults, such as vanity, greediness, and pusillanimity, this sort of aretaic appraisal seems perfectly appropriate. We can think poorly of someone because of her vanity, for example, without thinking that she owes it to us to rid herself of this vice. It seems to me that the notion of accountability is appropriate, however, when it comes to objectionable attitudes that are directed toward others. 10 Shame, rather than guilt, seems an appropriate response to attitudes of the sort mentioned in the previous footnote, e.g. vanity, excessive pride, greediness, slavishness, self-pity, etc., and I think the virtueethical approach does a good job of explaining our moral reactions in these cases. My suggestion is that it has a harder time explaining our moral reactions in cases involving morally unjustified attitudes toward others.



11.3.2 The consequentialist response This brings me, then, to the second proposal about how we might make sense of our guilt feelings over unexpressed attitudes without having to say that we ‘owe’ attitudes to others. I have found, in talking with people about this issue, that many of those who are sympathetic to the idea that attitudes are a legitimate subject of moral concern are sympathetic precisely because they believe that such attitudes will (eventually) lead to harmful conduct toward others. It is not really possible, they insist, to keep racist, sexist, and otherwise objectionable attitudes fully private, and it is their tendency to lead to wrong actions that explains why they are appropriately subject to moral scrutiny. On consequentialist grounds, then, it can be argued that we have a moral obligation to eliminate bad attitudes because of the potential harm they pose to others. The obligation here is not strictly to have certain attitudes toward others, but rather to ensure that we do not put others at risk by allowing potentially harmful mental conditions to persist. Here again, I do not disagree that the idea of a ‘fully private’ attitude, an attitude that has no effect at all on a person’s actions or behavior, is less plausible than it might initially seem. Our attitudes have a way of manifesting themselves in our outward conduct in all sorts of subtle and not-so-subtle ways, and a person who thinks she can completely cordon off a pernicious attitude and keep it from influencing her actions is probably deceiving herself. Given the close ties between our attitudes and our actions, then, why not just say that we have a moral obligation to rid ourselves of attitudes that have a tendency to lead to wrong actions? Though our attitudes themselves might not be subject to moral demands, we have an important obligation regarding our attitudes, namely, to try to eliminate the ones that might lead us to act wrongly. And when we feel guilty about a bad attitude, this may be because we recognize and feel bad about the potential harm such attitudes pose to others. As with the virtue-ethical response sketched above, I think there is a lot to be said for this proposal. But I have two general concerns about this line of argument. My first concern is similar to the concern I had with the virtue-ethical approach: namely, that this account seems to miss something important about the nature of our moral experience when we feel guilty about an unexpressed attitude. The moral concern about unjustified attitudes toward others, on this account, seems to be located entirely in the fact that they are likely to lead to bad consequences, and not in anything having to do with the intrinsic nature of these states. This is to treat the wrong involved in having an objectionable attitude as akin to the wrong involved in leaving broken glass on one’s walkway or failing to have one’s brakes checked regularly. In each of these cases, the wrong involved is the wrong of negligence: culpably permitting a dangerous condition to obtain that could lead to harm or injury to others. But it seems exceedingly odd to try to account for the wrongfulness of objectionable attitudes on a negligence model, treating pernicious attitudes like unmonitored broken glass or faulty brakes. Why? Because our attitudes are not merely ‘conditions’ we find in ourselves



that we have to monitor to make sure they do not pose a risk to others; they are, rather, judgment-sensitive states that, by their very nature, reflect our judgments about the value or status of other people. The moral significance of these states seems to attach in the first instance to their rational content, not to the likelihood of their leading us to perform harmful actions. While the likelihood of leading to harmful actions might be an additional reason to be concerned about objectionable attitudes, it does not seem to me to be what primarily establishes their moral significance. This is related to the second concern I have about this proposal, which is that it does not really seem to explain why we feel guilty upon finding that we have an objectionable attitude toward another. Strictly speaking, the moral demand on such a view does not apply to our attitudes themselves, but rather to our self-monitoring activities. The demand is that we not allow an attitude to persist that could lead us to engage in wrongful actions toward others, and we will violate this demand if we knowingly or negligently fail to rid ourselves of a potentially harmful attitude. But if this is the only moral demand that applies to us with respect to our attitudes, then it would make no sense for us to feel guilty simply upon finding that we hold a potentially harmful attitude (any more than we should feel guilty upon finding that there is broken glass upon our walkway, or that our brakes require fixing); in each of these cases, as long as we take the appropriate steps to eliminate the dangerous condition, there would be no grounds for feelings of guilt. So this line of response cannot really explain the moral phenomenology involved in the cases under consideration. In response, a defender of this view might say that guilt serves the useful function of alerting us to those attitudes that might lead us to act wrongly. While it may be true that we have not done anything wrong simply by having an objectionable attitude toward another, our guilt draws our attention to these conditions and encourages us to fix them before they can cause harm to others. But this sort of instrumentalist justification for our feelings of guilt seems strained: we are to understand our guilt in such cases not as a rational response to our failure to accord others proper attitudes of respect or regard, but as a practically useful warning sign that we may be in danger of acting immorally.11 Perhaps this is the best we can do by way of offering a ‘justification’ for our feelings of guilt in these cases; but ‘from the inside’, at least, there certainly seem to be more plausible internal justifications for these moral responses. If we really mean to take seriously the possibility that we can sometimes justifiably feel guilty over an unexpressed attitude, therefore, then we cannot rest content with either the virtue-ethical account or the consequentialist account. We will need a moral theory that can justify feelings of guilt rather than shame, and that can explain how our

11 D’Arms & Jacobson (2000a, 2000b) might see this as an instance of what they call ‘the conflation problem’: in this case, the problem would be failing to distinguish prudential reasons for feeling guilt—such as that it will lead to good consequences—and reasons bearing upon whether guilt is in fact rationally warranted in the context (because, e.g., one has wronged another).



attitudes themselves can violate a duty owed to others. In the next section, I will argue that Scanlon’s contractualist view can meet these theoretical demands.

11.4 Mutual Recognition In his 1998 book What We Owe to Each Other, and in his more recent book Moral Dimensions: Permissibility, Meaning, Blame (2008), T.M. Scanlon defends a contractualist account of morality according to which we can understand the content of morality by considering what principles each of us could reasonably be asked ‘to employ as a basis for deliberation and accept as a basis of criticism’ (1998: 298). Morality, in Scanlon’s view, should be understood as a ‘system of co-deliberation’ (ibid.) structured by the aim of finding a set of principles for the regulation of behavior that no one could reasonably reject. The notion of ‘justifiability to others’ is normatively basic in Scanlon’s account in two senses: it both characterizes the basic content of morality, and explains its distinctive reason-giving force (1998: 5, 189). It characterizes the basic content of morality in the sense that thinking about questions of right and wrong is a matter of considering what principles no one, suitably motivated, could reasonably reject. And it explains its distinctive reason-giving force in the sense that it is the value of standing in relations of justifiability with others that explains why we are motivated to comply with moral demands. Scanlon expresses this latter point in the following way: The contractualist ideal of acting in accord with principles that others (similarly motivated) could not reasonably reject is meant to characterize the relation with others the value and appeal of which underlies our reasons to do what morality requires. This relation, much less personal than friendship, might be called a relation of mutual recognition. Standing in this relation to others is appealing in itself—worth seeking for its own sake. A moral person will refrain from lying to others, cheating, harming, or exploiting them, ‘because these things are wrong’. But for such a person these requirements are not just formal imperatives; they are aspects of the positive value of a way of living with others. (1998: 162)

He goes on to explain how he interprets the content of our feelings of guilt when we violate moral demands: Duty is most familiar in its negative form, in the feeling of unwelcome constraint and the experience of moral guilt. According to the account I am offering, the pain of guilt involves, at base, a feeling of estrangement, of having violated the requirements of a valuable relation with others. So understood, this familiar negative aspect of morality corresponds to a positive ‘pull’: the positive value of living with others on terms that they could not reasonably reject. (1998: 162)

I think we have the materials here to construct a plausible account of how it is possible to wrong another through an unexpressed attitude. That account will make use of two ideas articulated in the passages above: first, that what fundamentally grounds our interest in morality is the value of standing in a certain relation with others, which Scanlon calls a relation of mutual recognition; and second, that feelings of guilt are a



reaction to perceived violations of the requirements governing this basic moral relationship. If it can be shown that this basic moral relationship, like other relationships, requires the presence of certain attitudes on the part of the parties involved, then failing to have such attitudes can reasonably be said to constitute a ‘wrong’ to the other party in the relationship. Taking our cue from Scanlon, then, we might begin by considering some of the duties we have toward others not simply qua persons, but qua friends, lovers, or members of some other special relationship. While it is true that the moral relationship is likely to be ‘much less personal’ than these other relationships, Scanlon suggests that we might understand the sorts of duties governing our moral relationships on the model of those governing these more personal relationships.12 In particular, if it is appropriate to speak of attitudinal obligations in the case of personal relationships, it might also make sense to speak of attitudinal obligations in the case of impersonal moral relationships. But does it make sense to speak of attitudinal obligations in the case of personal relationships? We might consider what these relationships would be like if we denied the existence of such obligations. Imagine, then, a friend or lover holding a view of friendship or fidelity according to which all that is required in order to be a good friend or partner is that one’s outward conduct conform to certain standards. Your ‘friend’, for instance, thinks that it is sufficient for friendship that she spends time with you, is available for overt support and encouragement, displays interest in your activities, etc. She has seen a lot of movies and has read a lot of books, so she has a pretty good idea of how friends are supposed to act toward one another. Indeed, her outward behavior is irreproachable. Let’s now assume, however, that her attitudes do not match her behavior. In fact, she actually despises you and thinks you are a conceited twit. She secretly hopes that you will fail in your various endeavors, and is pleased when you do. She finds your conversation boring and your insecurities pathetic. But she doesn’t see how any of these attitudes are incompatible with being a good friend: after all, she doesn’t express them in her actions in any way, and, we may suppose, she never will. We can tell various stories about why such a person would want to maintain such a ‘friendship’, but to make things simple let’s just assume that she thinks friendships, in general, are a good thing and that you’re the only person around with whom she has been able to develop what she regards as such a relationship. As discussed earlier, one might, of course, question whether it would really be possible to keep one’s true attitudes private in the way this example supposes. Wouldn’t they show through in one’s actions in some way or other? But for the sake of argument, let’s assume this is possible. A more interesting question is whether such a person has violated any duties of friendship. Keep in mind that this person has no ulterior motives in concealing her true attitudes from you. She really does, sincerely, want to be a good friend, and she really does, sincerely, believe that what this requires is


Scanlon pursues this analogy more fully in (2008): ch. 4.



simply that she act toward you in certain ways. And keep in mind, as well, that you don’t know what her true attitudes are and that you never will, for she behaves perfectly toward you. There is no chance that, down the road, you will find that she has been merely using you, or that she is secretly laughing at you with her ‘real’ friends behind your back. Do you, nevertheless, have grounds for complaint? I would argue that you do. The relation of being a friend, I would contend, is a relation with certain normative demands, expectations, and responsibilities built into it. These demands, expectations, and responsibilities pertain not only to one’s outward behavior, but to one’s attitudes, as well. We reasonably expect our friends to have attitudes of care and concern for us, to respect us, and to take pleasure in our accomplishments and feel sadness in our losses. Indeed, when we speak of ‘duties of friendship’, we have in mind this whole complex of behavioral and attitudinal demands and responsibilities. I think it would be reasonable for us to feel let down by a friend who failed to have some or all of these attitudes toward us, to feel that they have failed to relate to us in the ways called for by the special relationship in which we stand to them. The point of this example is to bring out the fact that there is a distinction between what we might call merely behavioral friendship—friendship that involves merely conforming our outward behavior to the norms of friendship—and a deeper attitudinal friendship—friendship that involves, in addition, the presence of attitudes of sincere care and concern.13 It seems clear that in order to be a genuine friend (and thereby comply with the requirements constitutive of friendship), one must meet the conditions of attitudinal friendship, not just behavioral friendship. I now want to suggest that we can draw a similar distinction in the case of what Scanlon calls the moral relationship. In order to genuinely stand in a relation of mutual recognition with others (and thereby comply with the requirements constitutive of this relation), we must meet the conditions of attitudinal recognition, not just behavioral recognition. Now, one might be tempted to think that friendship is a special case, and that it does nothing to show that we morally owe any attitudes toward people in general.14 Indeed, one might think that part of what distinguishes friendships from the sorts of relations we have with strangers is precisely that, in friendships, the relevant norms, expectations, and responsibilities extend beyond the sphere of action and into the sphere of attitudes. The fact that friends may reasonably demand certain attitudes from us, however, does nothing to support the claim that strangers may do so as well. And, it might be further argued, the attitudinal demands that friends make of us are not moral in nature, anyway: to fail to meet such demands is a failure of friendship, not of morality.


I am grateful to Laurence BonJour for suggesting this way of formulating the distinction. Scanlon (2008: 140) argues that ‘good moral relations’ with others requires not only that we have certain standing intentions to regulate our behavior toward them in various ways, but also that we be disposed to have certain other attitudes toward them, such as being disposed to be pleased when things go well for them. 14



Before addressing these objections, it should be pointed out that the case of friendship does show two very important things: first, that we do in fact care about the attitudes people have toward us in certain contexts; and, second, that we do think these attitudes are properly subject to normative demands in those contexts (whether we regard those demands as moral or not—I need not take a stand on that issue here). This in itself seems to show that the notion that we might ‘owe’ certain attitudes to others is not a complete non-starter. But I think we can say more than this. It is true that part of what distinguishes friendships and other personal relationships from the sorts of relations we have with strangers is that we (reasonably) hold our intimates to higher demands and expectations than we do strangers. We certainly do not expect complete strangers to care about our lives and well-being in the same way or to the same degree that we expect our intimates to do so (indeed, it would be somewhat creepy if they did). But I think this point about the relative stringency of the demands that apply to us in the case of personal versus impersonal relationships actually cuts across the behavioral/attitudinal distinction described in the friendship example above. That is, while it is true that we place more stringent attitudinal demands on friends than we do on strangers, it is also true that we place more stringent behavioral demands on friends than we do on strangers. What distinguishes friendships from more impersonal relationships, it seems to me, is not that the former are subject to attitudinal demands while the latter are not; it is that the general normative demands—both attitudinal and behavioral—that we make of friends are more stringent than the general normative demands—both attitudinal and behavioral—that we make of strangers. Or, at least, an argument needs to be given for why our attitudes are appropriately subject to normative requirements in some relationships (personal ones) but not in others (impersonal ones). For my part, I find it just as bizarre to say that a person could stand in a relationship of mutual recognition with me whilst holding attitudes of contempt, disregard, or ill will toward me as it is to say that a person could stand in a relationship of friendship with me whilst holding attitudes of dislike, distrust, or indifference toward me. In both cases, standing in the appropriate relationship with others seems to require not just that we act in certain ways, but also that we think and feel in certain ways. If this is correct, then the important question becomes what sorts of attitudes it is reasonable for us to expect others in general to hold toward us. And it is here that I think Scanlon’s contractualist account is perhaps uniquely well suited to give us guidance. On Scanlon’s view, we answer this question—about the reasonableness of principles requiring or forbidding us to have certain attitudes toward others—by considering what reasons people might have for either wanting or not wanting to be bound by the principles in question (1998: 195). On the one hand, we must consider the burdens such principles would impose upon those who must comply with them. On the other hand, we must consider the burdens the absence of such principles would impose upon those who may have reason to favor them. Finally, we should also consider ‘the other



implications (for both agents and others) of having agents be licensed and directed to think in the way that the principle requires’ (Scanlon (1998): 203). To illustrate this process, we might begin by considering whether anyone could reasonably reject a principle requiring that we never have negative attitudes toward others. Given how much hatred, malice, and distrust there is in the world today, some people might find this a very attractive moral principle. Upon reflection, however, it seems clear that there are decisive reasons for rejecting such a demanding moral requirement. As discussed earlier in this chapter, the fact of the matter is that negative attitudes toward others are often warranted. If someone treats me or other people with moral disrespect or indifference, for example, it seems perfectly appropriate to respond to such moral wrongs with negative attitudes of disapproval, resentment, or indignation. A principle forbidding such attitudes would, in effect, prevent us from acknowledging the fact that the victims of wrongdoing have moral standing and are deserving of better treatment. For similar reasons, a principle requiring that we have attitudes of love and deep concern for all others could reasonably be rejected. These attitudes both define and make possible our most personal and intimate relationships with others. A principle requiring that we hold such attitudes toward everyone would rule out the possibility of forming these distinctive relationships, which for many of us are essential components of a good life. This alone provides sufficient grounds for reasonably rejecting such a principle. We have good reasons, then, for rejecting very demanding attitudinal principles. Do we also have good reasons for rejecting less stringent attitudinal principles? For example, are there reasonable grounds for rejecting a principle requiring us to have basic attitudes of respect and goodwill toward others? Such a principle, which I will call the Attitudinal Recognition Principle, would require us to regard others as fellow members of the moral community with a basic moral dignity, and would make it morally impermissible to think or feel about them in ways that are inconsistent with this basic moral regard. So, for example, it would rule out attitudes of contempt toward others on the basis of morally irrelevant features (such as their race, sex, or sexual orientation); it would forbid us from regarding others in demeaning or belittling ways (e.g. merely as sex objects, or under certain pejorative stereotypes); and it would require us to have at least minimal feelings of goodwill toward others, such as dispositions to hope that things go well for them, dispositions to feel compassion when they suffer, and dispositions to feel indignation about violations of their moral status on the part of others. Are there reasonable grounds for rejecting even this more limited attitudinal principle? The first thing that should be said is that such a principle, as limited as it is, would still have substantial costs for some. Some people’s sense of their own self-worth is premised on attitudes of contempt and ill will toward others. White supremacists, male chauvinists, and homophobics, for example, often derive a sense of their own moral worth and value only by comparing themselves favorably to others whom they deem inferior on the basis of their race, sex, or sexual orientation. To demand that they give up such



attitudes might therefore be deeply damaging to their own self-conceptions. Given this cost, it could be argued that what is reasonable is a principle requiring that people not express or act upon attitudes of disrespect or ill will (which we might call the Behavioral Recognition Principle), but that it is unreasonable to demand that they not have such attitudes at all. In order to assess the force of this objection to the Attitudinal Recognition Principle, we must consider whether there are reasons counting in favor of this principle that are sufficient to overcome it. And here I must acknowledge that there is room for reasonable disagreement over this question. But I do want to suggest that there are some very strong reasons for preferring the Attitudinal Recognition Principle over the Behavioral Recognition Principle. These reasons have to do with the sort of moral relationship we can reasonably aspire to have with others under each of the proposed principles. As Scanlon rightly points out, ‘because principles constrain the reasons we may, or must, take into account, they can affect our relations with others and our view of ourselves in both positive and negative ways’ (1998: 203–4). Simply knowing that others recognize moral reasons to hold or not to hold certain attitudes toward me, for example, can be important to ‘defining my standing as an independent person who can enter into relations with others as an equal’ (1998: 204). And knowing that others do not recognize the authority of such reasons can adversely impact the sorts of relations I can enter into with them. This point about how the principles we recognize can affect the sorts of relationships we can aspire to have with others is a quite general one. For example, simply knowing that others did not accept that friendship required the presence of certain attitudes would significantly affect the sort of relationship I could aspire to have with them. I could not take for granted, for example, that my friends would genuinely care about me and be concerned about my welfare, even if I knew that they would always behave in a supportive way toward me. In a society in which merely behavioral friendship were the norm, I could not hope to achieve the kind of intimacy and mutual concern with others that is possible in a society where friendship is understood as requiring the presence of certain attitudes. Similarly, knowing that others did not accept that the relation of mutual recognition required the presence of certain attitudes would significantly affect the sort of moral relationship I could aspire to have with them. I could not take for granted, for example, that they would not wish ill upon me or regard me with contempt, even if I knew that they would behave in morally permissible ways in their interactions with me. In a society in which merely behavioral recognition were the norm, I could not hope to achieve the kind of moral fellowship and mutual goodwill with others that is possible in a society in which mutual recognition is understood as requiring the presence of certain attitudes. In both of these cases, what would impair my ability to enter into certain valuable relationships with others is just the fact that those relationships were understood as requiring only certain forms of behavior but not certain kinds of attitudes. For those of us who aspire to stand in deeper forms of moral fellowship with others, then, there are strong reasons to insist upon the Attitudinal



Recognition Principle over the Behavioral Recognition Principle. It is only if such a principle is widely accepted and adhered to that it will be possible for us to enter into such valuable moral relationships with others. As I said above, I think reasonable people can disagree about the respective merits of the Behavioral Recognition Principle and the Attitudinal Recognition Principle, and my aim here is not to try to convince anyone that they must accept the latter. My aim, rather, has been to suggest that a plausible argument can be given, within a contractualist framework, for why we should accept a principle requiring that we hold certain basic attitudes of respect and goodwill toward others. And it seems to me that this also provides us with a reasonable account of why we sometimes feel guilty over unexpressed attitudes toward others. If one aspires to stand in a relation of mutual recognition with others, and one thinks, as I do, that standing in such a relation requires the existence of certain attitudes of respect and goodwill, then failing to have such attitudes toward others can seem to be a violation of what we owe to them as fellow members of the moral community. This seems to me to make much better sense of our guilt feelings over unexpressed attitudes than either the virtue-ethical account or the consequentialist account. And, more importantly, it gives us a way of understanding how our guilt feelings in such cases, far from being irrational or neurotic, might in fact be a reflection of our positive commitment to living with others on mutually acceptable terms.

11.5 Conclusion This chapter began with an apparent puzzle about the sense of guilt many of us feel over some of our unexpressed thoughts, attitudes, and emotions. Feelings of guilt, it seems, are appropriate only in response to our own moral transgressions. But, it is commonly thought, merely having an unexpressed attitude toward another cannot constitute a moral transgression against them, because moral norms do not extend into the private, mental sphere. I have tried to show that a contractualist account of morality can allow for the existence of attitudinal obligations to others, and hence for the possibility of wronging another through an unexpressed attitude. I have not tried to argue that we must accept the existence of such obligations, but only that there is a plausible and appealing way of thinking about our moral relation to others within which such obligations make sense.15 15 I am grateful to Ann Baker, Jason Benchimol, Larry BonJour, Andrew Fyfe, James Mahon, James Mish’alani, and William Talbott for extremely helpful written comments on earlier drafts of this chapter (which prompted a good bit of rethinking on my part about its overall argumentative structure). Paul Glezen, Sandra Reiter, Jean Roberts, and Cathy Yu have been invaluable conversational partners about this issue over many years. I also received very helpful comments on earlier versions of this material from the following audiences: the University of Washington’s Workshop on Moral and Political Philosophy, participants at a mini-conference on Free Will jointly sponsored by the Philosophy Departments at the University of British Columbia and Simon Fraser University, members of my graduate seminar on Moral Agency at the University of Washington, and the Philosophy Departments at Washington and Lee University and Florida State University. I am grateful to the participants in all of these sessions for their probing questions and suggestions.



References Aristotle (1980) Aristotle’s Nicomachean Ethics. Translated by W.D. Ross, revised by J.D. Ackrill and J.O. Urmson. Oxford: Oxford University Press. D’Arms, Justin & Jacobson, Daniel (2000a) ‘The Moralistic Fallacy: On the “Appropriateness” of Emotion’, Philosophy and Phenomenological Research 61(1): 65–90. —— (2000b) ‘Sentiment and Value’, Ethics 110(4): 722–48. Dan-Cohen, Meir (1999) ‘Harmful Thoughts’, Journal of Law and Philosophy 18: 379–405. Feinberg, Joel (1990) Harmless Wrongdoing. Oxford: Oxford University Press. Freud, Sigmund (1961) Civilization and its Discontents. Translated and edited by James Stratchey. New York: W.W. Norton & Company. Hursthouse, Rosalind (1999) On Virtue Ethics. Oxford: Oxford University Press. Mill, John Stuart (1978) On Liberty. Edited by Elizabeth Rapaport. Indianapolis: Hackett Publishing Company. Morris, Herbert (1987) ‘Nonmoral Guilt’. In: Responsibility, Character, and the Emotions, F. Schoeman (ed.), pp. 220–40. Cambridge: Cambridge University Press. Nagel, Thomas (1998) ‘Concealment and Exposure’, Philosophy and Public Affairs 27(1): 3–30. Rawls, John (1971) A Theory of Justice. Cambridge, MA: Harvard University Press. —— (1999) A Theory of Justice. Revised edition. Cambridge, MA: Harvard University Press. Ripstein, Arthur (2006) ‘Beyond the Harm Principle’, Philosophy and Public Affairs 34(3): 216–46. Scanlon, Thomas (1998) What We Owe to Each Other. Cambridge, MA: Harvard University Press. —— (2008) Moral Dimensions: Permissibility, Meaning, Blame. Cambridge, MA: Harvard University Press. Sher, George (2005) ‘Kantian Fairness’, Philosophical Issues 15(1): 179–192. Sidgwick, Henry (1981) The Methods of Ethics. Indianapolis: Hackett Publishing Company. Smith, Angela (2005) ‘Responsibility for Attitudes: Activity and Passivity in Mental Life’, Ethics 2: 236–71. Taylor, Richard (1970) Good and Evil. New York: Macmillan. Thalberg, Irvin (1978) ‘Mental Activity and Passivity’, Mind 87: 376–95. Wallace, R. Jay (1996) Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Watson, Gary (1996) ‘Two Faces of Responsibility’, Philosophical Topics 24(2): 227–48.

12 Moral Sentiment and the Sources of Moral Identity Jacqueline Taylor

When we turn to the contemporary revival of the moral sentiments in moral philosophy, some of the most influential accounts focus exclusively on the negative moral sentiments, typically on blame, anger, and guilt. While good reasons are offered for this narrower focus, we might still ask whether we lose sight of an important aspect of our ethical life by excluding the positive sentiments of praise, admiration, pride, and selfesteem. My aim in this essay is to retrieve the insights of the eighteenth-century accounts of moral sentiment, particularly those of Hume, that point to the importance of moral approval, admiration, and pride as crucial sources of moral identity and agency. I shall suggest that the contemporary sentiment-based accounts that emphasize the negative sentiments neglect the broader canvas of our ethical experience and the importance of the varied attitudinal relations between the characters who populate it. Our capacities to admire and value one another for our moral dispositions and moral effectiveness, and to develop pride in such dispositions and effectiveness, play a key role in helping to create empathic and caring individuals. I will begin by looking at the approaches of the contemporary sentiment-based accounts and at what prompts the narrower focus on negative sentiments, such as blame, anger, and guilt. Next, I show that the Humean sentiment-based account of ethics provides us with a range of attitudes or sentiments, both positive and negative, directed towards character and conduct, and I describe how the principle of sympathy, when suitably corrected, allows for interpersonal communication of our sentiments and moral attitudes. My reconstructed Humean view allows us to examine more deeply the relation between positive moral sentiment, such as praise, and the agent’s pride or self-esteem. I will consider possible objections to privileging praise and pride, as well as the role of shame, pride’s opposite, in moral censure and improvement. I argue that the Humean account is compatible with important developments in contemporary psychology that stress the importance of meeting basic human needs, including needs for maintaining and enhancing self-esteem and for connection with



others, that are crucial for the development of empathy and moral effectiveness. I argue that moral competence, sustained by pride and moral approval, goes beyond not violating moral obligations, and also concerns having a well-developed sense of self, such that one can appropriately empathize with and focus on the needs of others.

12.1 In his classic essay, ‘Freedom and Resentment’, P.F. Strawson noted that ‘there is a whole continuum of reactive attitude and feeling’, and that different theoretical frameworks might start at different places on that continuum (1962: 6). R. Jay Wallace rejects the notion that the entire continuum of reactive attitudes has relevance for our holding people morally responsible. Instead, Wallace focuses on certain negative attitudes because holding people responsible ‘is essentially a disposition to respond in certain ways to the moral wrongs people commit’ (1994: 12). Moral wrongs are violations of moral expectations. What distinguishes moral expectations from nonmoral ones is that the former are justifiable in terms of distinctive moral reasons, and in particular, these moral reasons undergird those expectations that we regard as moral obligations (1994: 34–7). The moral attitudes expressed in response to wrongdoing, considered as the neglect or violation of some moral obligation, include blame, anger, and indignation. Guilt is the first-person counterpart, the reactive attitude expressed by the wrongdoer, to the blame or anger of others. While Wallace identifies shame as a moral sentiment, it does not count for him among the reactive attitudes. One might feel moral shame, he suggests, for not living up to standards of excellence one has set for oneself. The reactive attitudes that Wallace focuses on are, he says, ‘explained exclusively by beliefs about the violation of moral obligations (construed as strict prohibitions or requirement), whereas other moral sentiments are explained by beliefs about the other various modalities of moral value’ (1994: 38). Blame, anger, and guilt are thus the moral sentiments central to morality, as Wallace defines it. Alan Gibbard, whose account of the moral sentiments has a basis in evolutionary theory, also focuses on the negative moral attitudes. Borrowing J.S. Mill’s narrow sense of morality, used with respect to wrong acts, Gibbard wants to examine the claim that wrong acts ought to have a sanction, such as law, public opinion, or conscience. He is particularly interested in the relation between public opinion and conscience. In the case of our affective responses to wrongdoing, this concerns the relation between anger and blame on the one side, and the guilt of the wrongdoer on the other (1990: 41). Like Wallace, Gibbard sees this narrow sense of morality as one that assigns a special role to obligation. And like Wallace, he thinks that guilt, rather than shame, is the characteristic first-person reaction of the wrongdoer. Drawing on a traditional distinction between guilt and shame, Gibbard points out that shame on the part of an agent consists of an awareness of some personal inadequacy. Shame might motivate one to withdraw from others, or to cultivate the qualities or abilities that would give one more value in social cooperative schemes. In contrast, guilt, as something experienced



by the agent (as opposed to something attributed to her), consists of an awareness of having committed a wrong. Guilt results not so much from inadequacy as from the agent lacking sufficient motivation to act rightly or not to act wrongly on some particular occasion. Guilt ‘meshes’ with anger, since anger is an appropriate response to someone lacking the motivation to act as she ought. Those affected by or aware of the wrongdoing respond with anger, which, as a negative sanction against her conduct, in turn motivates the agent to repair the damage she has done or otherwise make amends. Gibbard argues that anger is assimilated to moral judgment because it makes sense to be angry when someone acts in a way we typically find objectionable. Thus anger and guilt, rather than disdain and shame, comprise the characteristic affective reactions of the narrower morality system. Gibbard does acknowledge that there are other morally important emotions, including moral admiration, that are part of morality ‘construed more broadly’, but he does not develop an account of how these other emotions are morally significant or look in depth at how a broader morality relates to the narrower morality system (1990: 51).

12.2 The last time the moral sentiments were a central part of the philosopher’s moral vocabulary, before falling out of favor when Reid and Kant turned moral attention to the will and intention, was in the eighteenth century, particularly in British philosophy, in the works of Anthony Ashley Cooper, the Third Earl of Shaftesbury, Francis Hutcheson, David Hume, and Adam Smith. These philosophers appealed to the notion of a moral sense or moral sentiments. They focused on our sentiments of praise and blame, and tended to give greater attention to moral praise than to blame or disapprobation. While their approaches were united by the need to challenge Hobbes’s and Mandeville’s characterization of human nature as fundamentally self-interested, each gave a different account of the relation between our moral sentiments and the balancing of our self- and other-oriented passions and affections. For Shaftesbury, the superior charm of moral approval, when it reflexively surveys the agent’s own kind affections, makes it a master-pleasure, one that thereby brings self-interest into agreement with virtue. The agent attentive to the needs of others, with the proper balance of self-regarding and social affections, has in addition the pleasures of self-approbation, which spur him on to cultivate and sustain virtue. Hutcheson’s philosophical task lies in showing both the fundamental value of the various forms of benevolence of which we are capable, and how the perceptions of an internal moral sense make that value salient for us. Adam Smith gives a central place to the agent’s self-command combined with a sympathetic attunement to the emotions and situations of others. Hume eschews the notion of an innate moral sense and focuses on the cultivation of moral sentiments grounded in our natural capacity for sympathy. Since sympathy gives us a felt awareness of the sentiments of others, including their sentiments towards us, Hume can forge an important link between our practices of praise and approval and the



pride or self-esteem of praiseworthy agents. Pride, and particularly pride in virtue, serves several ethical functions. Pride has a special effect on the mind: it gives the proud person an affective awareness of her own advantageous qualities and her merit. Such awareness in turn provides her with a sense of competence and confidence: being praised for and thus proud of her virtuous character reveals that through her virtuous activity she makes a difference in the community in ways that she and others value. And having pride in one’s character, and in being the person one is, provides important protection of one’s dignity against the threats of oppression or calumny by others. As we shall see in more detail later, psychologists agree with Hume on the value of pride or self-esteem. Self-esteem is essential for having a sense of self-efficacy, a sense of one’s competence, and of what one can accomplish. Self-esteem is also crucial both in cultivating pro-social helping behavior, and in overcoming moral indifference in the face of human evil. For the Humean, our responses, at least in part, convey our evaluation of one another’s enduring dispositions and character traits. A moral concern with character helps us to distinguish, for example, between unintentional harm and a disposition to cruelty or callousness. Similarly, someone who from kindness does something to further our well-being is more likely to earn our gratitude than the same action done from self-serving motives. Focusing on character, not simply on the intention, action, or its consequences provides a guide for the attitudes it is appropriate to take towards one another. To orient our attention towards an agent’s character and conduct also moves us away from Wallace’s and Gibbard’s narrower focus on the sentiments of blame, anger, and guilt that are appropriate in response to the violation or neglect of obligation or principle. Looking more holistically at character accommodates our interest in an agent’s broader pattern of conduct, and in the dispositions we find praiseworthy or admirable as well as what we find blameworthy. For the Humean, as for other moral philosophers interested in character, we care about the cultivation of virtue, and about what is admirable for its own sake and not simply about those actions that fall in line with obligations or moral principle. Character, and the particular virtues and vices that comprise character, are among the central concepts of a virtue ethics, and Hume is often regarded as a virtue theorist (Baier (1985); Swanton (2007)). It is worth noting, however, that the Humean focus on character differs from the contemporary Aristotelian versions of virtue ethics. Contemporary virtue theories in the Aristotelian tradition, for example, tend to focus on the fully virtuous agent’s deliberative perspective, on what makes for accurate perception or sensitivity to what has ethical salience, and how his action aims at what is fine or the best, where no strict principles may serve to guide him (e.g. Nussbaum (1988); Hursthouse (1999)). Aristotelian versions of virtue ethics focus on character because it is the unity of the virtues, the virtues of character and the virtues of intelligence, of the fully virtuous agent that gives authority to his deliberative perspective. The fully virtuous agent has achieved the proper mean state of various dispositions (the virtues of character) that afford him an emotional attunement to what in a given



situation has most salience (e.g. his sense of courage is a stance reflecting the proper degree of both fear and confidence that enables him to discern things that really are dangerous), while practical intelligence guides his assessment and response (e.g. this is a danger that requires him to stand firm). The Humean steps back from this deliberative perspective to look at the relation between an agent’s character and our evaluative, sentiment-based response to her character. It is this emphasis on the evaluative perspective, where the focus falls on character, which motivates my interest in giving a place to the role of moral praise and pride in good character. The relation between moral sentiment and character shifts our concentration from the agent’s deliberative perspective to the evaluative perspectives we take towards each other and our appraisals of character. Attending to this relation between character and sentiment-based evaluation suggests that, first, apart from knowing what is required to deliberate well, we also want to know what kinds of persons we are dealing with in our everyday encounters. Second, while we have a special interest in knowing the characters of those around us, we also take a general interest in the range of human character and conduct that we find when we turn to survey human history, accounts of people distant from us, as well as the characters in drama and fiction. Finally, focusing on the relation between character and sentiment draws attention to how we evaluate the characters of others. It is important both that we get an accurate assessment of someone’s character, and that we have an appropriate sentiment-based response to it. This interest we take in human character and the lives of others derives from our capacity for empathy, or what Hume calls sympathy. While we tend to associate sympathy with pity, Hume intends it as a technical term, referring to our capacity to communicate and respond to one another’s emotions. Hume’s understanding of how sympathy works anticipates one contemporary, social psychological model of empathy. According to Stephanie Preston and Frans de Waal, empathy comprises a set of capacities, ranging from simpler subcortical processes such as empathic contagion, to more sophisticated forms of cognitive empathy and helping behaviors that depend on the more highly developed prefrontal cortex in the brain (Preston and de Waal (2002)). Hume’s own account of sympathy reflects this range of capacities. Sympathy explains our feeling with another, for example, when we feel pain in response to his pain. And it explains our having an empathic response that is not the same as the person’s own feeling, but is nonetheless about the other rather than ourselves, as, for example, when we feel compassion in response to her sorrow, or resentment in response to the injustice done to her. Sympathy also keeps us attuned to situations that tend to elicit particular emotional responses so that, for example, we sympathize with someone facing difficulties even if he does not show the emotions that the situation typically arouses in people. Sympathetic communication also forms the basis of our moral appraisal of one another’s character. If we lacked the capacity for sympathy, we would be indifferent to the pains and pleasures of others, and to morality itself.



Sympathy explains not only our interest in one another, but also the fact that we naturally make appraisals of the features of persons that advantage or disadvantage them in various ways. These features include physical attributes (such as beauty, strength, and capabilities, and their opposites), talents, possessions, wealth, and power (or the lack of these), as well as mental qualities, such as character or personality, and the attitudes and beliefs that reflect our commitments and values. Our appraisals of one another in response to these features take various forms: we admire, esteem, respect, or love people for good qualities (the useful and agreeable), and blame, contemn, disdain, or hate, for harmful or disagreeable qualities. We recognize both moral and non-moral versions of these attitudes. Nevertheless, while our admiration of someone’s musical talent or physical strength is a kind of aesthetic admiration, and our admiration of her courage is typically a moral admiration, there can sometimes be important overlap between moral and non-moral sentiments, and both kinds of sentiment have their basis in our capacity for sympathy. In both kinds of case, we sympathize with the advantage or disadvantage that accrues to the person possessing the valuable or disvaluable feature, or to others who are affected by her having this feature. For example, we know from experience that wealth enables the rich man to enjoy the conveniences of life; we sympathize with his enjoyment and admire him for having such enjoyable possessions. The Humean focuses on our appraisals of and responses to character and the various dispositions that typically provide some pattern to our deliberations, choices, and conduct. Of course, attention to individual intention outside the more holistic context of someone’s character can also matter to us. A person’s intention in acting provides some evidence of her good or ill will, her love, respect, or contempt. At times, knowledge of her intention may have more relevance to us and thus override the information we have (or thought we had) about her character, with respect to our natural or moral response to her. But typically knowledge of an agent’s character gives us a fuller sense of the kind of person she is, since her character will be tied to her reputation, and will inform our sense of her trustworthiness or untrustworthiness, or our expectations of her for how she treats the various people with whom she has some relation, or how she will conduct herself generally. This emphasis on the evaluative perspective made possible by our capacity for sympathy also explains the distinctive taxonomy of Humean virtues and vices. Hume defines a virtue as a quality of mind or character that is useful or immediately agreeable, either to the person possessing the quality or to others affected by her character or conduct. Conversely, a vice is a quality of mind or character that is harmful or immediately disagreeable to the agent or those affected by her character or conduct. A useful trait is one that we identify as having a tendency to contribute to the wellbeing of the agent with the trait or of members of her moral community. An immediately agreeable trait might be one such as wit or charm that we simply find pleasing, although such traits might not contribute to the good of society. Together, the qualities of utility and immediate agreeableness unify all the traits that are virtues, just as harmfulness and disagreeableness unify the traits that are vices. Some traits have



both useful and agreeable aspects. Compassion, for example, contributes to the wellbeing of others, and we approve of it on those grounds or because someone has benefited from another’s compassion. At the same time, we may also take delight in the actions and character of the compassionate person herself. When we sympathize with the effects of some trait of character, for example, with the relief produced in someone who is the object of another’s compassion, we are pleased that she is relieved, and we direct our pleasure towards the character of the compassionate agent. In recognizing that compassion tends to contribute to others’ well-being and in finding the compassionate action pleasing, our attitude towards the agent is one of moral approval. The taxonomy of useful and agreeable, harmful and disagreeable, has some explanatory benefits, and we can make sense of them by once again drawing a contrast with the Aristotelian view of virtuous character. As we noted earlier, the Aristotelian focuses on the deliberative perspective of the fully virtuous person: his possession of all the virtues affords him moral knowledge. While such knowledge does not come automatically in any given case, he has what it takes to comprehend and appreciate what has ethical salience in the situations he encounters. He thus stands as the exemplar for the rest of the community, and especially for those who are striving to achieve full virtue. In the figure of the fully virtuous agent we find a unity of all the essentially valuable qualities of character and practical intelligence. Rather than embracing this ideal, the Humean accepts that most people will have mixed characters: some traits that yield benefit or enjoyment for himself or others, and some traits that may bring harm to himself or others, or make himself or others uneasy. The Humean emphasis on moral sentiment and making accurate appraisals of character points to our awareness of this imperfect nature of human character, and our need to acquire sufficient experience of the mixtures of traits, as well as of situations that might encourage certain kinds of conduct, and of institutions and policies that help to shape certain kinds of character rather than others. Our practices of moral evaluation, which I review below, require that we communicate our appraisals to one another, subjecting them to mutual scrutiny and correction, and that we adhere to accepted forms of moral discourse that make our appraisals intelligible to others. In contrast to the Aristotelian view, then, the Humean account of moral evaluations shows that moral knowledge is a collective resource, arrived at through our conversations and debates about our understanding of the characters of individuals and about the value of particular traits. Drawing attention to the useful and agreeable as the categories that unify the virtues also accommodates the historical and cultural reality of different ideals of character and different rankings of the virtues. On the Aristotelian view, the set of virtues unified by practical wisdom comprises the dispositions needed for human flourishing. According to Martha Nussbaum, these virtues are precisely the dispositions that equip us to deliberate well and make good choices in domains of human activity in which we must all participate (Nussbaum (1988)). In contrast, Hume includes a long list of qualities that count as useful or agreeable, and it is unlikely that any individual could



cultivate and possess all such qualities. Moreover, when we consider all the good things that can count as ends for us, things that we recognize as part of a good life, there are more ends than can be realized in a single life or even be made available to the members of a given society. In deliberately pursuing certain good ends over others, individuals and communities may need to privilege some traits over others. Attention to the varieties of ways of living, both throughout history and across cultures, shows us that different societies value different ends, which leads to different rankings of traits of character. One community may privilege refinement, wit, and the arts, while another places greater value on industriousness, curiosity, and innovation. Placing the focus on the usefulness and agreeableness of traits, as those features that unify a set of traits as virtues, accommodates this variation in the conception of social well-being.

12.3 I said earlier that we have a particular stake in accurately assessing character, and that our ability to do so makes it possible for us to create and acquire moral knowledge as a collective or shareable resource. Such knowledge concerns the traits we admire or deplore, how to cultivate the virtues or stem the development of vicious traits, and includes attention to how education, policies, laws, and institutions such as government, the family, and religion, encourage or block the development of virtuous characters. Creating the circumstances in which such knowledge is produced and acquired also reflects good practices of moral evaluation that allow for open-ended conversations and debates about characters, education, policy, and about the practices of evaluation themselves. A climate of trust that allows for moral inclusion, enabling others to voice their views, and for mutual scrutiny and correction of one another’s judgment, makes it possible to have healthy moral debate. Inclusion, mutual scrutiny, and correction may require ongoing negotiation of our moral understandings, and so may not lead to moral consensus, even though that may be the ideal at which we aim. Humean moral knowledge, while not dynamic and progressive in the way scientific knowledge characteristically is, nevertheless has an organic character insofar as we may change our views of virtue and vice in light of a greater understanding of human nature or appreciation of the views of others. One reason that the Humean emphasizes the accurate assessment of and response to character is because the moral sentiments are grounded in our capacity for sympathy, and sympathy naturally develops and operates in ways that can render us partial or biased. Typically, when we sympathize with another person, we focus on the other, on her feelings, opinion, situation. Sympathy thus takes us out of ourselves, allowing us to feel as the other does, and to be responsive to her feelings or situation. Sympathy gives us an interest in others, distinct from our own more selfish interests. But some sympathetic processes are automatic and devoid of reflection. Recall that I said Hume’s own account of sympathy anticipates in some key ways the model of empathy devised by Preston and de Waal. While some researchers distinguish between emotional



contagion or imitation and empathy, Preston and de Waal include contagion and imitation in their model. These processes emerge earlier than more cognitive forms of empathy (the distress of infants in response to an already distressed infant is a wellknown example), and facilitate emotional development. As the notion of emotional contagion suggests, sympathetic communication is partly constitutive of our emotions and sentiments. Contagion is an important part of human (and other animal) development, through which infants and children ‘try on’ emotions and attitudes. While both imitation and contagion decrease as the individual develops and learns to regulate her emotions, both tendencies persist even in mature people (Preston and De Waal (2002): 5). Cases of emotional contagion may have an ethical or an aesthetic dimension, as with, for example, the mass behavior of crowds, the zeal of those persuaded by rhetoric, and an audience’s enhanced experience of theater through shared emotional responses. Other factors, including similarity of some aspect of another to oneself, familiarity with another, past experience, and conditioning, can also generate more automatic sympathy-based responses. Familiarity facilitates sympathy so that we naturally sympathize more easily with those to whom we are close, or with friends, colleagues, and acquaintances. We also sympathize more easily with those who resemble us in ways we find significant, for example, a shared language or shared values; such resemblances will vary according to context. Our habits of sympathy and how we sympathize—who we empathize with and under what circumstances—typically also reflect what we might think of as an emotional education, where we might be taught to pay attention or defer to certain kinds of people, and perhaps to take less interest in certain others. Given certain cultural contexts, habits of sympathizing may reflect learned biases (including implicit bias) on the basis of, for example, gender, race, religion, or socio-economic status. Sympathy thus tracks and reinforces both our natural partiality and learned biases. Some of our more partial sympathetic habits are important for us, operating to keep us attuned and responsive to the emotional lives of those around us and those about whom we care most. Nevertheless, we need an impartial sympathy to serve as a foundation for the moral sentiments. Both tendencies, towards partiality and bias, require correction when we make moral assessments of others. While we should value the virtues wherever we find them, not simply in those we care about or those who impact our lives in some way, we tend to respond more passionately to the situations or characters of those whom we know or care about. Our sympathy may be blunted if the agent and those whose situations or lives she affects lie at some remove from us, such that our sympathetic sentiments are not what they would be had we a better empathic understanding of her case. More troubling are the biases we form so that we either fail to sympathize with some people altogether or we rely on stereotypes and automatically attribute certain character traits, needs, or beliefs to those whom we perceive to belong to a particular social group. Research in social psychology shows that relying on negative stereotypes leads people to attribute negative dispositional qualities to those with membership in the stereotyped group. Making such dispositional attributions will often reflect false



beliefs, either for the perceived group as a whole or for a given individual member. A critical aspect of Humean moral knowledge concerns having an accurate assessment and appreciation of what someone is really like. The person who remains in the grip of prejudice and so has an inaccurate view of certain others lacks knowledge of what those others are really like. Importantly, inaccurate evaluations affect those to whom they are directed, not only distorting others’ views of their character or reputation, but also affecting their own self-conception when they sympathize with those evaluations. Our sympathy is especially attuned to others’ opinions of evaluations concerning ourselves. If others have inaccurate beliefs about our character and thus make inappropriate evaluations, this may still affect our self-conception even if the other’s evaluation does not confirm what we think about ourselves. Especially with those whose opinion and evaluation we have special reason to be concerned about (those we care about, or those with some authority over how aspects of our lives go, for example), what we regard as their undeserved blame or praise for us may have more effect on our sense of self-worth. Undeserved contempt still stings, and undeserved praise should make us uneasy. Moreover, what we may at first regard as an undeserved blame or praise may alter our own selfconception, as we sympathize with others’ judgment and thus see ourselves as they see us. The Humean focus on evaluation emphasizes the importance of moral discourse and conversation in our evaluations of one another. What Hume calls the ‘common point of view’ is a perspective we adopt in order to make ourselves mutually intelligible to one another. This requires, first, that we use the discourse of morality, referring to character and conduct, rather than that of interest or bias. Clearly there are complex matters here, both with respect to how moral vocabulary may nevertheless be employed to disguise interest, and how ideology may pervade the descriptive and evaluative content of our moral discourse (Taylor (2002a, 2002b)). Second, we need to consider the effects of particular character traits, and whether they tend to promote the well-being or the suffering or uneasiness of both the agent and those affected by her character or actions. This consideration requires us to extend our sympathy to all those affected and consider their responses and concerns. Third, in offering a moral evaluation we should make our judgment in such a way that we can expect that others will agree with us, or if they do not, we should be open to being corrected or learning more, or to correcting and teaching others. Most important is the requirement to cultivate virtues of good judgment (Falk (1986); Taylor (2002b, 2009)). These include relying on our capacities for good reasoning, including fact-gathering, making appropriate distinctions or comparisons, and drawing the appropriate conclusion. We also need to cultivate our capacity for sympathy or empathy, to put aside our own concerns or biases so that we can properly take up the perspectives of those whose concerns are relevant to the case at hand. As with any set of virtues required for a particular social role or task, experience clearly matters, and even with the same amount of experience, some people will do better at cultivating the virtues needed to evaluate well. In some cases, for example, in making certain legal decisions about competence or responsibility, we may need to rely on experts, those with a greater understanding of psychology or the



law than most of us have. We then have to deal with weighing the evaluations of some against those of others, and finding ways to do so without the unfair exclusion of views that nevertheless may have less relevance for the case at hand.

12.4 We need to cultivate the virtues of good judgment to evaluate appropriately one another’s character and conduct. Our evaluations are not, however, affectively neutral. We are ineluctably, and in various ways, pleased or pained by both our own and others’ character and conduct. The account above shows that our evaluations of one another’s character, and also of our own character, are in effect sympathetically engendered moral sentiments. These sentiments comprise various forms of approbation and disapprobation. With respect to virtue, we may, for example, feel esteem for the hardworking, the kind, and the just, admiration for the heroic, or charm in response to the graceful or witty person. There is likewise a broad range of vices to which we have different sentimental responses. We abhor cruelty and injustice, are contemptuous of the small-minded or cowardly, and made uneasy by boorish or buffoonish behavior (Williams (1973)). We have various ways of expressing our sentiments, both straightforward and subtle, using verbal expressions of praise, congratulation, anger, or blame, for example, as well as through facial expression, body language, and our actions. As we saw above, those towards whom we direct our sentiments naturally sympathize with our sentiment, which in turn tends to elicit in them a self-regarding emotion, pride or shame, for example, in response to their sense of themselves given how we regard them. Hume refers to this interactive process as a kind of sympathetic reverberation or mirroring. In this section, I want to look at how sympathetic mirroring works in the moral context, and then discuss the moral function or role of the positive sentiments of praise, pride, and self-esteem. The concept of mirroring has been much discussed in the psychoanalytic literature, and the process there described has some relevance to the Humean process of sympathetic mirroring. For psychoanalysts such as Heinz Kohut or D.W. Winnicott, mirroring between a parent and infant is critical to the infant’s emotional development and sense of self-worth. In empathizing with the infant’s emotional exhibitionism, a parent feels an emotional response of approval of or pride in the infant. In feeling the regard or joy of his parent, the infant over time acquires a confidence about expressing himself. Kohut’s further theoretical development of the parent as the all-powerful ‘selfobject’ urges that in merging with the idealized parent the child also achieves a sense of security and power (Kohut (1971, 1977); Miller (1996)). Children who do not receive this empathic mirroring, who are instead disapproved of or neglected, have what is perhaps their most basic human psychological need, for security, frustrated, and are deprived of the foundation for developing self-esteem. While psychoanalytic theorists deploy this metaphorical use of the term mirroring to explain healthy psychic development, their insights are supported by contemporary neuroscience. Mutual emotional expressiveness



between an infant and parent is crucial to the ability of the infant to organize emotion, and to give the infant a sense of security and effectiveness, which in turn helps with development of empathic capacities (Preston and de Waal (2002): 8). We should note Hume’s attention to the more complicated emotional interaction of typically more mature persons, who have already developed an emotional repertoire as well as a sense of self. Humean mirroring includes the empathic process, a kind of sympathetic concern, with the recognition and appreciation of the feelings and situation of the person with whom one empathizes. But it also includes censure and disapproval, and can be complex as when we empathically respond to the situation in which one person has received an injury, while disapproving of the one responsible for the injury, and sympathizing with the emotional responses of both the injured and the perpetrator. Hume’s own description of emotional mirroring shows a profound insight into the social dimension of our psychological nature, so I will quote a long passage in full from his Treatise of Human Nature, and then move on to an analysis of the process more generally. In general we may remark, that the minds of men are mirrors to one another, not only because they reflect each other’s emotions, but also because those rays of passions, sentiments and opinions may be often reverberated, and may decay away by insensible degrees. Thus the pleasure, which a rich man receives from his possessions, being thrown upon the beholder, causes a pleasure and esteem; which sentiments again, being perceiv’d and sympathiz’d with, encrease the pleasure of the possessor; and being once more reflected, become a new foundation for pleasure and esteem in the beholder. There is certainly an original satisfaction in riches deriv’d from that power, which they bestow, of enjoying all the pleasures of life; and as this is their very nature and essence, it must be the first source of all the passions, which arise from them. One of the most considerable of these passions is that of love or esteem in others, which therefore proceeds from a sympathy with the pleasure of the possessor. But the possessor has also a secondary satisfaction in riches arising from the love and esteem he acquires by them, and this satisfaction is nothing but a second reflection of that original pleasure, which proceeded from himself. This secondary satisfaction or vanity becomes one of the principal recommendations of riches, and is the chief reason, why we either desire them for ourselves, or esteem them in others. Here then is a third rebound of the original pleasure; after which ’tis difficult to distinguish the images and reflections, by reason of their faintness and confusion. (Hume (2007):

This passage points to the complexity of the dynamic interplay and the phenomenology of emotional communication. Hume’s example begins with a rich man simply taking pleasure in his riches. Observers sympathize with his joy and, recognizing that wealth affords a power ‘of enjoying all the pleasures of life’, they feel a sympathetically engendered pleasure, which takes the form of esteem for him. In turn, the rich man sympathizes with the observers’ esteem for him, and feels a ‘secondary satisfaction’ in his wealth, which is ‘a second reflection’ of his original pleasure. That is, the rich man now feels proud, given that he is the possessor of this wealth. His pride or selfsatisfaction adds an additional value to the wealth, becoming ‘one of the principal recommendations of riches’, and giving the observers an additional reason to esteem



the rich man. The observers sympathize now with the rich man’s pride, as well as his pleasure in his wealth, and in ‘a third rebound of the original pleasure’, esteem him for his pride as well as the pleasure afforded him by his wealth. The passage also points to the circumstances that make pride an appropriate emotional response. To feel proud appropriately one must have some positive quality that is durably connected to oneself. This helps to explain pride as a disposition as well as an emotion. In having a valuable feature of personality, a sense of wit, for example, others’ admiration over time can help one develop a sense of pride in one’s wit. Experiencing pride as an occurrent emotion typically includes a positive valuing of oneself, experienced as a form of pleasure, along with a belief that one has the valuable quality. Pride as a disposition does not entail thinking well of oneself all the time (although some vicious forms of pride such as arrogance or conceit do tend to make someone selfabsorbed or self-centered). Rather, the proud person has a sense of his own merit or a sense of confidence, at least with respect to that in virtue of which he is proud. What one takes pride in must also receive public acknowledgement as something valued. The public acknowledgement may be from a small group (some types of wit are too refined or dark for everyone to enjoy them), or it may be a community-wide valuing, as is the case with many of the traits held to be virtues. Shifting this analysis to the moral context, we can add some points about the importance of self-conception and shared values for one’s sense of moral identity. Imagine someone taking pleasure in an act of kindness she performs for another. She is pleased that the other’s well-being is improved, and her kindness gives her a pleasing sense of connection to the person she is helping. Observers sympathize both with the recipient of the kind act and with the kind agent. They sympathize with the recipient’s gratitude and improved sense of well-being, and with the agent’s pleasure in performing the kind action. As in the case of the rich man, the observers’ sympathy elicits new emotions and sentiments in them. They may feel relief or joy for the person benefited, and perhaps moral approval for gratitude, an approval with which she will in turn sympathize. The observers feel a moral approval or esteem for the kind agent, and the agent in turn sympathizes with their approval or esteem. She sympathetically experiences their pleasure, directed towards her in virtue of her kindness, and moreover, recognizing that she has done something that others find valuable, she feels a pride in her kind disposition. Sympathizing with her pride in her kindness, the observers feel a new esteem for her: they esteem her for taking pride in her virtue, in addition to approving of her for her kindness. Finally, following Hume in the passage we looked at above: the agent’s pleasurable sense of pride in her kindness and others’ esteem of her for her pride in her kindness becomes one of the principal recommendations of kindness and virtue. The psychic shifts to different forms of pleasure show that for both the virtuous agent and the observers, the trajectory from pleasure in kindness and esteem for kindness, to pride and the observers’ esteem arising from their sympathy with pride, creates an increasingly intimate and meaningful exchange between the agent and the observers.



Moreover, to say that to be proud of one’s own virtue or to be esteemed by others for the pride one takes in one’s virtue comprises one of the principal recommendations of virtue and attests to the importance of this emotional connectedness between the agent and the observers. Pride is fundamentally an interpersonal emotion, an emotional response that arises or is sustained by the positive appraisal of others. In the rest of this section, I want first to look in more depth at the interpersonal nature of pride. I then turn to the importance of pride in meeting and sustaining some basic human needs, for self-esteem and connectedness to others, and its role in sustaining a sense of moral effectiveness and in developing empathic capacities. Pride is not often discussed in contemporary moral philosophy, and its status as a virtue has had a checkered history since ancient times. It is all but neglected in contemporary ethics. Consider the ancient concept of megalopsychia or magnanimity: the magnanimous man is one in possession of all the virtues and who thus recognizes himself as worthy of great things; Julia Annas describes Aristotelian magnanimity as a ‘super-virtue’ (1993: 117). Magnanimity contrasts with an overweening form of pride, viz. hubris or insolence (towards the gods). Nevertheless, as Kristja´n Kristja´nsson points out, few if any contemporary Aristotelians admit magnanimity into the modern catalogue of virtues. While this might be because we may think ourselves more realistic than Aristotle and other ancients, such that magnanimity is an ideal rather than a virtue whose cultivation we typically encourage, Kristja´nsson, following Bernard Williams, suggests another reason. For both the Greeks and Hume, in giving a central role to pride, as well as to shame, human interdependence will also have an important role. To be susceptible to pride and shame makes us sensitive to and concerned with the esteem and recognition or disapproval of others. As Kristja´nsson notes, Aristotle’s view of the virtuous agent as ‘a prideful person is more other-entwined and other-identified than most modern moral theorists are willing to accept’ (2001: 176). Williams, in his discussion of shame and the ancient Greeks, points out that on the Kantian (and also Plato’s) conception of morality, giving ethical significance to shame, with the internalization of the other and of social expectations, makes the Greeks socially heteronomous, and ‘represents a compromise of genuine autonomy’ (1993: 98). The neo-Kantian, Christine Korsgaard, takes just this critical stance towards Hume and other eighteenth-century moral sentimentalists, writing that for them, ‘the approval and disapproval of others is the fundamental moral phenomenon, from which all our ideas spring. There is something obviously unattractive about taking the assessment of others as the starting point in moral philosophy’ (1996: 189). Yet as Hume’s mirroring passage suggests, the virtuous person is not proud of her virtue simply because others approve of her, nor does she act virtuously to gain others’ approval. In possessing the traits of kindness and pride in virtue, she has traits that others value, and which make her praiseworthy in their eyes. The approval of others is a recognition that she is valuable, and her response of pride is an affirmation of her selfworth. The approvers’ further approbation of her pride in her virtue is in turn apt to sustain both her sense of pride and her disposition to kindness. The kind agent’s pride



also affirms the value of kindness, and the mirroring between her pride and the observers’ esteem reflects a shared affirmation of the value of both pride in virtue and virtue. Pride is an emotion of self-assessment, which like shame or guilt makes one aware of oneself in a certain way. Pride makes us aware of ourselves as the possessor of valuable qualities, the value of which is shared by others who admire or esteem us for these qualities. Pride thus contributes to a sense of confidence and of competence. The person who takes pride in her kindness has confidence that she can effectively help others. According to June Tangney, one of the leading theorists in the psychology of what are called self-conscious emotions, pride has a self-regulatory function. Tangney cites Mascolo and Fischer, who regard pride as a self-regulating emotion ‘generated by appraisals that one is responsible for a socially valued outcome or for being a socially valued person’ (Mascolo & Fischer (1995): 66). With respect to pride’s self-regulatory function, pride enhances self-worth and encourages conduct ‘that conforms to social standards of worth or merit’ (Tangney (2003): 395). In contrast to Korsgaard’s depiction of it, I have characterized the Humean view about how those social standards of merit are set as one that requires public conversation and debate, mutual intelligibility, scrutiny, and correction. Each of us participates in the practice of moral evaluation, and given our social connectedness and need to cooperate with one another, most of us consider it important that we arrive at correct assessments of what is praiseworthy and blameworthy. Pride has a natural developmental trajectory. Psychologists tend to distinguish between pride and self-esteem, regarding pride as an emotion that contributes to the enhancing and maintaining of self-esteem. Self-esteem is an evaluation of oneself, combining both beliefs about oneself and a positive or negative affective component. Experiences of pride and of being praised by others are crucial to the development of positive self-esteem. Mascolo and Fischer have helpfully charted the developmental trajectory of pride. Very young children (18 to 24 months) experience pride in relation to some outcome they caused, and by age 4 or 5, children take pride in a valued trait. By age 10–12, children can take pride in their character, for example, if they possess traits that make them competent persons (Mascolo & Fischer (1995)). As noted above, pride contributes to the child’s sense of self-worth, and given that her pride is a response to praise, or imagined praise, it also serves to forge a positive connection with others. Pride has a sequential development of skills for appraising oneself, for having a just measure of one’s strengths and merit. The developmental account also shows how pride develops from an episodic emotion (pride in result) to a disposition, as children learn to take pride in their enduring character traits. Moreover, it reinforces a Humean point, namely, that it is important and meaningful for us to see ourselves as others see us, and for our sense of ourselves to be informed by others’ evaluative responses to us. Bernard Williams has pointed out that moral philosophers who emphasize, as Wallace and Gibbard do, guilt rather than shame as the relevant first-person moral emotion, associate guilt with autonomy, and regard guilt as the appropriate response to



voluntary wrongdoing. But guilt is ‘isolated’ from other elements of one’s self-conception, including those elements in which we can take pride or of which we can be ashamed. Williams argues for the importance of shame precisely because ‘it embodies conceptions of what one is and of how one is related to others’ (1993: 94). Shame can motivate self-improvement and move us to earn the respect or esteem of others in ways that guilt cannot. Pride, shame’s opposite, likewise embodies one’s self-image, and is an important source of one’s moral identity, an identity informed by one’s connection to others. Like shame, pride also functions to protect one’s sense of dignity or self-respect, and to preserve one’s personal standards. In his Enquiry Concerning the Principles of Morals, Hume draws attention to the importance of a sense of pride as protection of one’s dignity (1998: section 7). The person with a due sense of pride is less likely to behave in ways that are servile or degrading. A healthy pride provides self-assurance about what one thinks right, even in the face of injustice and oppression. I conclude by showing how Hume’s insight is borne out by research in psychology. Psychologist Ervin Staub, a child survivor of the Holocaust, has devoted his life to understanding altruism and helping behavior, aggression, and harm-doing, as well as the bystander phenomenon. According to Staub, self-esteem, the maintaining and enhancing of self-esteem, a sense of competence or effectiveness, the esteem of others, and a positive connectedness with others, comprise the most basic human needs (2003: ch. 5). Meeting these needs constructively contributes to the continued growth of the person. These needs co-exist. Satisfying the need for the esteem of others both contributes to self-esteem and helps one achieve a positive connection with others. Meeting the need for self-esteem allows one to have a more purposeful life. It also leads to a feeling of effectiveness, to confidence in fulfilling one’s own goals, as well as to being open to others, with the possibility of empathy. Having a positive connection with others, through their esteem, as well as having selfesteem, encourages one to see others as valuable, and promotes empathy and caring for others. These needs can come into conflict; for example, it may appear safer to maintain self-esteem rather than take a risk that would enhance self-esteem. The need to maintain self-esteem may thus be met at the cost of not meeting the need to enhance self-esteem, or vice versa. Destructive satisfaction of needs, such as being punished or having to have too strong an identification with a group, interferes with personal growth, and Staub associates it with hostility and aggression. In relation to our connectedness with others, Staub distinguishes between what he calls the connected self and the embedded self. The connected self has positive connections with others, but can also stand apart from others, and disagree with them or with their actions. The embedded self, in contrast, characterizes someone whose sense of self is grounded in a dependence on others. Embedded selves are less likely to voice opposition to the group to which they belong, and have little sense of an identity distinct from group identity. As we saw above, children who are encouraged to develop an appropriate pride in their abilities and traits learn to value themselves and acquire a healthy self-esteem. The valuable traits or abilities that make them unique



give them a sense of themselves as unique individuals, and distinct from others. At the same time, they stand in a positive relation to those who praise them for these traits and abilities, and whose values they share. Moreover, Hume’s point that pride protects one’s sense of dignity or self-esteem has relevance here. Gabriele Taylor has written that ‘shame is the emotion of self-protection’, because fear of shame can prevent one from betraying one’s values, thus preserving one’s sense of self-respect (1985: 81). Similarly, to have a well-grounded sense of pride can help one resist others’ attempts to belittle, contemn, degrade, or dehumanize one, thereby preserving one’s sense of dignity. To have self-esteem is, as we noted above, to be well positioned to value others, to be sensitive to their needs, and to help them. To be proud of one’s capacities to help others in need or to resist oppression or injustice is to have a confidence in one’s moral competence, and one’s effectiveness as an agent.

References Allport, G. (1954) The Nature of Prejudice. Reading, MA: Addison-Wesley. Annas, Julia (1993) The Morality of Happiness. Oxford: Oxford University Press. Baier, Annette C. (1985) ‘Theory and Reflective Practices’. In her Postures of the Mind: Essays on Mind and Morals, pp. 207–27. Minneapolis: University of Minnesota Press. Blackburn, Simon (1998) Ruling Passions: A Theory of Practical Reasoning. Oxford: Oxford University Press. Casey, John (1990) Pagan Virtue: An Essay in Ethics. Oxford: Oxford University Press. Falk, W.D. (1986) ‘Hume on Practical Reason’. In his Ought, Reasons, and Morality, pp. 143–59. Ithaca, NY: Cornell University Press. Gibbard, Allan (1990) Wise Choices, Apt Feelings: A Theory of Normative Judgment. Cambridge, MA: Harvard University Press. Goffman, Erving (1963) Stigma: Notes on the Management of Spoiled Identity. New York: Simon & Schuster. Hume, David (1998) Enquiry concerning the Principles of Morals, Tom Beauchamp (ed.). Oxford: Clarendon Press. —— (2007) A Treatise of Human Nature, David Fate Norton & Mary J. Norton (eds). Oxford: Clarendon Press. Hursthouse, Rosalind (1999) On Virtue Ethics. Oxford: Oxford University Press. Kohut, Heinz (1971) The Analysis of the Self. NY: International Universities Press. —— (1977) The Restoration of the Self. NY: International Universities Press. Korsgaard, Christine M. (1996) ‘Creating the Kingdom of Ends: Reciprocity and responsibility in personal relations’. In her Creating the Kingdom of Ends, pp. 188–223. Cambridge: Cambridge University Press. Kristja´nsson, Kristja´n (2001) ‘Pridefulness’, The Journal of Value Inquiry 35: 165–78. MacIntyre, Alasdair (1982) ‘How moral agents became ghosts or why the history of ethics diverged from that of the philosophy of mind’, Synthese 53: 295–312. Mascolo, M.F. & Fischer, K.W. (1995) ‘Developmental transformations in appraisals for pride, shame, and guilt’. In Self-Conscious Emotions: The Psychology of Shame, Guilt, Embarrassment, and Pride, J.P. Tangney & K.W. Fischer (eds), pp. 64–113. New York: Guilford Press.



Miller, Alice (1983) For Your Own Good: Hidden Cruelty in Child-Rearing and the Roots of Violence. New York: Farrar, Straus & Giroux. Miller, Susan B. (1996) Shame in Context. Hillsdale, NJ: The Analytic Press. Nussbaum, Martha C. (1988) ‘Non-Relative Virtues: An Aristotelian Approach’. In: Midwest Studies in Philosophy, vol. XIII, Ethical Theory: Character and Virtue, Peter A. French, Theodore Uehling, Jr., & Howard Wettstein (eds), pp. 32–53. Notre Dame: University of Notre Dame Press. Preston, Stephanie D. & de Waal, Frans B.M. (2002) ‘Empathy: Its Ultimate and proximate bases’, Behavioral and Brain Sciences 25: 1–20. Staub, Ervin (1989) The Roots of Evil: The Origin of Genocide and Other Group Violence. Cambridge: Cambridge University Press. —— (2003) The Psychology of Good and Evil: Why Children, Adults, and Groups Help and Harm Others. Cambridge: Cambridge University Press. Strawson, P.F. (1962) ‘Freedom and Resentment’, Proceedings of the British Academy 48: 1–25. Swanton, Christine (2007) ‘Can Hume Be Read as a Virtue Ethicist?’, Hume Studies 33 (April): 91–113. Tangney, June Price (2003) ‘Self-Relevant Emotions’. In: Handbook of Self and Identity, Mark R. Leary & June Price Tangney (eds), pp. 384–400. New York: The Guilford Press. Taylor, Gabriele (1985) Pride, Shame, and Guilt. Oxford: Oxford University Press. Taylor, Jacqueline (2002a) ‘Humean Ethics and the Politics of Sentiment’, Topoi 21: 175–86. —— (2002b) ‘Hume on the Standard of Virtue’, The Journal of Ethics 6: 43–62. —— (2009) ‘Hume’s Later Moral Philosophy’. In: The Cambridge Companion to Hume, second edition, David Fate Norton & Jacqueline Taylor (eds), pp. 311–40. Cambridge: Cambridge University Press. Wallace, R. Jay (1994) Responsibility and the Moral Sentiments. Cambridge, MA: Harvard University Press. Williams, Bernard (1973) ‘Morality and the emotions’. In his Problems of the Self, pp. 207–29. Cambridge: Cambridge University Press. —— (1993) Shame and Necessity. Berkeley and Los Angeles: University of California Press.

13 On Alienated Emotions1 Talbot Brewer

Emotions provide the kind of materials we need if we are to fashion an evaluative stance of our own. While our emotions sometimes speak our stance with precision and eloquence, more often they surge forth in half-articulate form, and in such cases we are left with the task of making our inchoate stance both more definite and more transparent to ourselves. Emotions have, then, a dual relation to the self. They give expression to the self ’s pre-reflective evaluative posture towards the world, and they provide crucial raw material for the lifelong task of working up this initial posture into a mature and discerning outlook upon the values in play in our changing circumstances. My aim is to cast light on the nature and stakes of a divorce that occurs between the self and its emotions when these emotions protrude before it as something alien. Emotions might be alien either in an objective sense or in a subjective sense. Objectively, emotions are alien when they express the evaluative outlook not of the person in whom they arise but rather of some other agent. Subjectively, emotions are alien when the person in whom they arise refuses to acknowledge the evaluative outlook embedded in them as his own. While alienated emotions can arise in many different contexts, I will be focusing in particular on those caused by service economy work. I choose this focus because the rise of the service economy has greatly increased the number of workers who are called upon daily to conform their displays of feeling, if not their feelings themselves, to ends and interests that are not their own. I hope to shed some light on what it might mean for this most personal form of labor— emotional labor, as it has come to be called—to be sold for a wage. What I will suggest is that this sort of alienation tends either to interrupt or to disfigure a lifelong task that is an indispensable part of a well-lived human life—the task, as I call it, of self-elaboration.

1 An earlier draft was presented at the New School for Social Research, and I benefited greatly from the comments of Jay Bernstein and others in the audience. I owe thanks also to Charles Mathewes and Matthew Crawford for their illuminating comments on an earlier version.



13.1 Alienated Labor in Industrial Production Before we can make headway in our exploration of alienated emotional labor, it will help to get a bit clearer on the more general notion of alienated labor, as it arises in the early work of Karl Marx. It is a humdrum observation that in a capitalist economy, workers alienate their labor in the sense of selling use of it to their employer for certain stretches of time. Marx’s insight was that in the context of industrial capitalist labor, this straightforward kind of alienation entails a number of other, more troubling kinds. It entails the alienation of the worker not only from his daily laboring activity, and from whatever goods are produced by that activity, but also from what Marx called his species-being.2 For Marx, humans fully realize their species-being only when they freely and selfconsciously contribute to the creation of a fit setting for human life. Human beings are not the only animals that work upon inorganic nature to provide themselves with a suitable habitat. Bees, beavers, and ants all do this. However, as Marx points out, other animals perform such work only in accordance with an instinctually determined blueprint, while human beings are capable of revising their mode of reshaping their world so as to permit new experiments in living.3 A dramatic example of this process is the rise of the Internet. The Internet comes into existence from a vast, coordinated effort to reshape our environment in such a way as to bring human beings into more continuous and productive conversation with each other. It provides its users with a kind of psychological prosthetic device, one that places within their immediate reach an unprecedented accumulation of information, misinformation, rantings, musings, images, reflections, etc. In sculpting the Internet, human beings collectively imprint the natural world with their intersecting pictures of what sorts of conversations and information human beings ought to have. It is worth emphasizing that for Marx, human labor ought to be driven by, and ought in turn to refine, evolving ideas of the human good. Our laboring activity ought to move in a cycle between formulating a picture of how the world might be altered so as to accommodate more valuable activities, making over the world in accordance with our picture, and arriving through experience with the world so refashioned at another, yet more adequate picture of how the human world might be remade. Call this the Marxist, or materialist, conception of human self-elaboration. We will soon consider another, closely related conception of self-elaboration—one whose proximate aim is to remake the psyche rather than the material environment. In capitalist work conditions, Marx thinks, the proper cycle between practical thinking and labor is broken, and labor ceases to be a mode of self-elaboration. The laborer is called upon to put into action the plans of others. There is no opportunity for him to reshape his work in accordance with his evolving sense of what the product might contribute to human lives, or of what it is for a human life to go better rather than worse. The task of

2 Marx works out these ideas in the section entitled ‘Estranged Labour’ of the Economic and Philosophical Manuscripts of 1844. The relevant portions can be found in Marx (1975): 322–34. 3 Marx (1975): 329.



product design is assigned to specialists, as is the determination of production techniques. In order to minimize production costs, the production process is decomposed into simple constituent operations, each performed repetitively by a single worker along an assembly line. This tends to increase production speed while also permitting the employer to save money by replacing skilled craftsmen with unskilled workers.4 The customer enjoys a lower price. But the worker comes to be employed only in his capacity as an unusually dexterous if not entirely reliable machine, and not in his capacity as a thinking and evaluating being. The worker is thereby alienated from his own activity. The activity is directed by another, whose interest lies in making it repetitive and mindless.

13.2 The Rise of the Service Economy In the generation or two after Marx, some thinkers began to hope that technological advances in production would soon make it possible to defang the problem of industrial labor by greatly reducing the fraction of their time on earth that human beings would devote to work. Edward Bellamy expressed this hope for a nearly workless human world in Looking Backward—a book that inspired a nationwide utopian-socialist political movement when it appeared in 1888.5 Since Bellamy placed his vision in the year 2000, we are now in a position to conclude that it has not arrived on schedule. Recent history suggests, rather surprisingly, that increased productivity does not generally lead to less work. As Juliet Schor notes in The Overworked American, the material consumption rates achieved in 1948 could have been sustained in the 1990s even if every single worker took every other year off. Yet the average American family contributed about sixteen more weeks of full-time work to the formal economy in 1988 than in 1967.6 Writing twenty years after Bellamy, the eminent German sociologist Georg Simmel offered a less utopian and more believable prospect. Simmel recognized that bluecollar labor would not disappear entirely, and that it would probably continue to be performed in hierarchical institutions under the direction of a managerial class. Nonetheless, he saw reason to hope that it would soon be done in such a way that each worker would be able to enjoy ‘the all-decisive feeling of dignity’ and of ‘a life which is its own master’.7 The route Simmel envisioned to the restoration of human dignity lay, interestingly, not in mitigating but in deepening those qualities of factory labor that Marx had viewed as most objectionable. If work was bound to become increasingly mindless and mechanical, Simmel reasoned, this might at least permit the worker to go through the motions while holding in reserve whatever is most worthy about his humanity, thereby leaving himself essentially unsullied.

4 For an excellent discussion of alienated labor and its relation to the fragmentation of laboring activity into simple mindless steps, see Braverman (1974): chs 3 and 4. 5 See Erich Fromm’s Foreword to Bellamy (1960): vi. 6 Bluestone & Rose (1997). 7 Simmel (1971): 341.



Whatever one may think of the appeal of Simmel’s vision, we are now in a good position to assess its prescience, and it appears to have been just as removed from the actual currents of history as Bellamy’s. As it turns out, the same historic shift in work has dashed the hopes of both thinkers. The mechanization of production has resulted not in the withering away of work but rather in a radical expansion of the service sector of the economy, and it is typically not possible for service economy workers to reserve their most valuable and distinctively human capacities from their work. Is this a blessing or a curse? I think it would be a mistake to affirm a sweeping verdict of any kind. Service economy work can provide individuals with ennobling vocations and inspiring social roles. It can also be stultifying and degrading. Yet if this sort of work can be alienated, it generally can’t be alienated in quite the same way as factory labor. It is not normally possible for managers to break service work down into simple mechanical motions that require no thought. Mindless and mechanical service is subpar service. Nor is the product of service work removed entirely from the worker’s possession and transferred to a far-away purchaser. Such work often requires the manipulation of one’s personal demeanor and the gradual reshaping of one’s emotional register. For this reason, service work opens up the possibility of a distinctive and particularly insidious sort of alienation—alienation from those portions of one’s own psyche and self-presentation that have been made over in conformity with the employer’s profit interests.

13.3 Emotional Labor in the Contemporary Service Economy This fundamental shift in the nature of work provides the backdrop for Arlie Hochschild’s influential sociological investigation into what she calls ‘emotional labor’.8 Service economy workers typically have to represent the employer to the customer not only in speech and action but even in those semi-conscious projections of emotions that are achieved by subtle facial expressions and body language. It is true, of course, that service workers may serve up some concrete deliverable—a glass of wine, say, or a boarding pass. But there are typically other, less tangible goods that the service economy worker is supposed to provide: an atmosphere of conviviality and warmth, perhaps, or a credible show of solicitous and personal concern. In the long run, these intangible goods may have a greater impact on the bottom line of the employer than the tangible ones. These intangible goods cannot conceivably be delivered impersonally and mechanically, in the manner of the assembly line riveter. As Hochschild puts the point, ‘Seeming to “love the job” becomes part of the job; and actually trying to love it, and to enjoy the customers, helps the worker in this effort.’9 This is what leads Hochschild to claim that service economy work requires emotional labor, by which

8 9

Hochschild (1983). Ibid.: 6.



she means engaging for a wage in ‘the management of feeling to create a publicly observable facial and bodily display’.10 Hochschild’s most widely cited field research focused on the emotional labor performed by flight attendants. Hochschild conducted a first-hand study of the extensive training programs that airlines have devised for their flight attendants. The aim of these programs is to hone the on-the-job emotional displays of flight attendants so as to enhance customer contentment with the travel experience. It is no surprise that these training programs place great emphasis upon the smile. ‘Your smile is your biggest asset’, flight attendant trainees were informed during a Delta Airlines training session. ‘I want you to go out there and use it. Smile. Really smile.’11 As Hochschild notes, these words are more than a bit misleading. The wage agreement between the airline and the flight attendant converts the smile—that peculiar facial contortion through which human beings consummate their joy and invite others to share it—into a corporate asset. It is only insofar as the smile becomes a corporate asset that it is an asset of the employee. If the play of the smile deviates from the interests of the corporation—if it routinely flashes, for instance, in such a way as to betray the silliness or implausibility of a prescribed sales pitch—then the smile ceases to be a corporate asset, and it thereby ceases to be a saleable asset of the employee. This means that the smile and other displays of emotion can gain value as saleable assets only by losing value as inchoate expressions of the employee’s sense of the evaluative contours of the world. The way to excel in the job of flight attendant is to make over one’s emotional register so that one’s emotional display, and perhaps one’s emotions themselves, are in an important sense not one’s own—that is, so that they manifest the potentially alien evaluative standpoint of the corporation. This requires the decoupling of one’s emotional display from changes in the social environs—e.g. to the coldness, lust, or condescension of clients. Put another way, it requires a social isolation that wears the mask of extreme sociability.12 This petrified friendliness, this unsocial sociability, has begun to take an increasingly extreme form. A Japanese software company named Omron recently made the news for releasing a computer program that monitors the emotional labor of service employees, providing a continuous ranking of their smiles on a scale of 0–100. When employees’ grins fall below their ‘personal best’, the program sends messages exhorting them to try just a bit harder, while keeping a running record of their labial shortcomings that can be reviewed by managers. The first sale of this new program occurred in 2009, when more than 500 customer service workers in Japan’s Keihin Electric Express Railway were placed under its unblinking eye.13


Ibid.: 7. Ibid.: 4. 12 See Hochschild (1983): 25–6 on the training of flight attendants to remain calm and friendly when confronted with ‘irates’—that is, angry or abusive customers. 13 See: 11



Not surprisingly, sociologists have found that emotional labor is associated with high levels of stress and job dissatisfaction, a sense of inauthenticity in one’s work life, and emotional numbness and other symptoms of depression.14 What is perhaps more interesting is that they have also found that the longer workers stay on the job, the more likely they are to make over their emotional registers so that they actually feel job-appropriate emotions, and that those who remake their emotional registers in this way find their work less burdensome than those who engage in mere surface acting.15 What these findings suggest is that permitting one’s emotions to be made over so that they express the interests of another person (i.e. the corporation) is a common coping mechanism for those who engage in corporate forms of emotional labor.16

13.4 The Commodification of Servility Hochschild’s work on emotional labor in the service economy clearly struck a nerve. It ignited a veritable cottage industry of studies of emotional labor and its effects. This enterprise would make little sense if there were no good reason to worry about the effects of emotional labor, and it is clear from the tone of these studies that emotional labor does strike many researchers as worrisome. Yet on reflection it is not easy to articulate a sensible ground for such concern. There are, after all, many useful or agreeable things that we cannot get unless someone serves us: food in a restaurant, product information in a store, health care at a clinic, a haircut at a barber shop. We presumably want these goods to be served up politely rather than brusquely, cheerfully rather than coldly. A smile wouldn’t hurt. Leave it to the dour academic to rain on this parade of good cheer. The rest of us, it might be said, will go on smiling and will hope that others reciprocate. Well, not quite all the rest of us. When McDonald’s opened its first franchise in Russia, the newly hired staff was trained in the delivery of the requisite smile of American service work. Muscovite customers seem to have found these smiles neither welcoming nor friendly. On the contrary, customers reportedly felt they were being mocked.17 With a little reflection I think we can begin to appreciate what lies behind this reaction. All parties


Grandey et al. (2005): 900; Pugliesi (1999): 130 and 147; Tracy (2000): 95–6. Tracy (2000): 115–16; Gross (1998); Pugliesi (1999): 147; Brotheridge & Grandey (2002). 16 Within some corporations, professionals in ‘human resources’ (a slightly ominous phrase) seek to promote this preferred outcome by exercising ‘cultural control’ (a more decidedly ominous phrase) over the workplace. The idea is to use a combination of rituals, training programs, official dispensations of criticism and praise, internal publications, moderated discussions at employee retreats, financial rewards and promotions to establish a local currency of social esteem that tracks corporate profit interests. (Ashforth & Humphrey (1993): 102.) The brilliant television show The Office gets a good deal of its satirical punch from its exploration of this process. It sets a rather pitiful attempt to breed a corporate culture against the backdrop of a work life whose aims and daily tasks are very nearly meaningless. Dwight is the superficial but efficient sort of toady that emerges when corporate acculturation efforts succeed; Michael represents the crumbling of integrity visited upon semi-intelligent and vaguely good-willed managers who must represent corporate aspirations with an outward pretense of conviction; the rest are casting about for some way to keep a trace of integrity while paying lip service to corporate-prescribed aspirations and convictions. 17 Ashforth & Humphrey (1993): 102. 15



know that the smiles in question would not have been produced if not for the confluence of the employer’s interest in securing money from the customer and the employee’s interest in securing money from the employer. Given this, it would take a rather thick customer to regard the professional smile as a genuine and personal expression of delight to see them—rather than, say, their wallets—coming through the door. But if this is right, why exactly are these smiles so widely believed to be good for business? Is there some common human need or desire—beyond the mere desire for politeness or civility—that the professional smile could satiate even if all parties knew its origins to be fundamentally mercenary? I think there is. To see what this desire might be, it will help to consider a distinction Rousseau made between two basic forms of human self-love. The first, amour-de-soimeˆme, is a pre-social preference for whatever it is that satisfies such self-limiting natural desires as hunger, the desire for sleep, and the appetite for sex. The second, amourpropre, is an essentially social desire for signs in the words and actions of others that they recognize one’s standing or worth. Amour-propre can take the benign and perhaps even laudable form of a concern that others recognize one’s equal dignity or worth. But in Rousseau’s view it often takes the inflamed form of a desire that others acknowledge one’s superior importance.18 When amour-propre takes this inflamed form, its satiation requires the frustration of some other human being’s reasonable desire for merely equal recognition. My suggestion is that much of the emotional labor of the service economy is calculated to indulge inflamed amour-propre. The plausibility of this hypothesis stems partly from its ability to explain why the transparently pecuniary motives that drive this emotional labor do not cancel its perceived value. A spontaneous and heartfelt smile does not indicate a deferential attitude or a subservient posture towards its object. A forced smile often does. Such a smile can make it public between two people that one recognizes a dependence upon the other that requires him to ingratiate himself to the other without any expectation that the other will return the favor. We can gain some insight into the psychodynamics of emotional labor from the writings of Sarah Tracy, a sociologist who worked on a luxury cruise ship for eight months in an undercover participatory study of this work environment. What is most astounding about Tracy’s work is the nearly Orwellian effort made by management to sculpt the emotional displays of the cruise ship service staff. From the point of view of management, the single most important part of the job is attitudinal: the continuous ingratiating demeanor, the permanent smile.19 On Tracy’s ship, these attitudinal standards were driven home by a corporate service credo posted in all crew bathrooms and on the walls of all crew cabins, just above the headboard of the bed, hence in an

18 While the whole of Rousseau’s Second Discourse and much of Emile can be read as a meditation on amour-propre, the pithiest and most pointed account of it is found in endnote 15 of the Second Discourse. For an outstanding discussion of this key element of Rousseau’s moral psychology, see Neuhouser (2008). 19 Tracy’s study, cited above, is one of the more interesting of the large number of sociological inquiries into emotional labor that were inspired by Hochschild’s The Managed Heart.



ideal position to be the last thing seen before dropping off to sleep and the first thing seen upon awakening. Employees were required to carry copies of the credo on their person at all times, and to wear pins designed to remind them of it. The credo included items like ‘We never say no’, ‘We smile, we are on stage’, and ‘We use proper telephone etiquette . . . and answer with a smile in our voice’.20 (Everywhere one looks in these service economy expectations, one sees the smile, staring back with disturbing predictability, disturbing because it closes off the possibility of a response fitted especially to the occasion—a response that reveals a particular human being or that conveys a particularized goodwill.) It is worth noting the use in this service credo of the first-person plural pronoun and the assertoric voice, which conspire to present these items as the self-consciously affirmed practices, or the self-generated identity, of a distinctive social group. In practice, of course, the items of the credo function not as shared beliefs but as work requirements. They are enforced not only by managers but by all passengers, via the comment cards that they are encouraged to submit after all cruise experiences and that function as the primary arbiters of retention and promotion decisions. This practice, Tracy notes, has the effect of putting the customer in the position of a second boss, one whose gaze takes in (or, perhaps more importantly, cannot be known not to be taking in) nearly every nook of the ship.21As one manager put it to his crew, service workers must regard themselves as ‘public property’ whenever they are in public areas of the ship—which is to say, almost continuously, from early morning until evening every day, during travel shifts of eight to ten months.22 The effect of this concerted emotional labor is to deliver up, at a price that is within reach for upper-middle-class Westerners, a level of continuous, detailed, and deferential personal service that has historically been a mere fantasy for anyone not born an aristocrat or appointed a colonial governor. The enactment of this fantasy is affordable partly because cruise ship lines are floating nations unto themselves, able to broker an exchange between clients from the world’s wealthiest countries and workers from some of the world’s deeply impoverished countries. (Two of the most common countries of origin for cruise ship workers are Indonesia and the Philippines, both with per capita GDPs of less than 4% of that of the United States and about 2% of Norway’s.) The result is a dramatic shrinking of the distance that ordinarily separates the world’s ‘haves’ and ‘have-nots’, and a consequent exposure of the normally veiled relations of economic interdependence that make it a routine matter for moderately well-to-do Western consumers to purchase several days of manual labor, congealed in an imported consumer product, with the net earnings of a single hour of work time. While this everyday, entirely routine arrangement is in its basic terms quite regal, these basic terms are rarely noticed by any of the parties engaged in it. Indeed, these 20 21 22

Tracy (2000): 107. Ibid.: 105–6 and 108–9. Ibid.: 106.



parties are often quite literally oceans apart, and are rarely aware of each other’s existence, much less of their roles in each other’s lives. The anonymity of their relationship ordinarily prevents the wealthy consumer from seeing first-hand that he continuously enjoys the ministrations of what is in effect a small army of personal servants. He usually gives little thought to the remarkable fact that many other human beings are effectively obliged to place inordinate value on his smallest conveniences and comforts, and to impose great inconveniences upon themselves in order to supply these comforts. The cruise ship environment removes the cover of impersonality from these deeply inegalitarian arrangements. One might hope here for a sudden moment of politicizing clarity. Yet luxury cruises are not known as seedbeds of radicalism. What apparently materializes, instead, is a new and distinct saleable commodity, one that cannot be delivered at a distance: inequality itself, served up in personalized and sometimes downright obsequious form. Cruise ships are hardly the only loci of the face-to-face consumption of personalized deference. This same intangible product is part of the package provided by many service workers who ply their trade on dry land. Consider, for instance, that growing cadre of migrant workers who have left their homes in impoverished nations to tend the homes of wealthy Westerners. Barbara Ehrenreich, who worked as a cleaner while preparing to write a book on the working poor, offers this description of the interpersonal dynamics of paid housecleaning: ‘To make a mess that another person will have to deal with—the dropped socks, the toothpaste sprayed on the bathroom mirror, the dirty dishes left from a late-night snack—is to exert domination in one of its more silent and intimate forms. One person’s arrogance—or indifference, or hurry—becomes another person’s occasion for toil.’23 It seems likely that an important function of a great deal of service work is to provide the customer with an encoded assurance of superior social standing. Still, luxury cruise lines deliver up commodified servility of an especially pure sort, much of it virtually bereft of any tangible benefit. Consider in this vein David Foster Wallace’s observations, during a luxury cruise he took as a correspondent for Harper’s magazine, on the labor of ‘the towel guy’. This is the name Wallace devised for the crew member whose task it was to place cleanly laundered towels upon any deck chair just before a guest sits down in it, and to shuttle these same towels to the laundry bin as soon as any guest stands up. The ‘towel guy’ on Wallace’s ship was apparently so efficient that it was impossible to use the bathroom without having one’s towel removed, only to be replaced by a fresh one upon return to the chair. One wants to know exactly why the cruise ship line sees fit to impose this small inconvenience upon its passengers. Similar questions can be raised concerning the waiter hovering by the table who, upon serving a lobster, will ‘incline over you with gleaming claw-cracker and surgical fork and dismantle the lobster for you, saving you the green goopy work that’s the only remotely rigorous thing about lobster’; and the


Ehrenreich (2002a): 88.



‘waiter who will materialize as you peel away from the buffet and will carry your tray’ to your table; and the ‘crew of maıˆtre d’s and supervisors watching the waiters and sommeliers and tall-hatted buffet-servers to make sure they’re not even thinking of letting you do something for yourself that they could be doing for you’; and the cabin cleaner who enters the cabin several times a day, when one is at a meal or wandering the deck, to clean and straighten and put yet another mint-chocolate on the pillow.24 A great deal of this labor must be placed in the same category as having one’s napkin refolded when one goes to the bathroom in a fancy restaurant—it seems to confer no tangible benefit. But if, following our current hypothesis, we admit as an intangible benefit the placation of the inegalitarian amour-propre of the clientele, then a straightforward explanation presents itself. The small inconvenience of being the target of hyper-attentive service might be compensated by tangible signs that another human being must cater to one’s smallest ‘needs’ and must do so under the obligation of maintaining a perfect smile. This serves up something that no machine, however deft its ministrations, could possibly serve up. It serves up a continuous confirmation, in the words and acts and bodily comportment of another human being, of one’s superior importance. One might test the hypothesis at hand by considering how one would respond if one’s lobster were being cut, one’s tray solicitously carried, one’s room cleaned, or one’s towel neurotically kept fresh by a close friend who happened to be employed by a luxury cruise firm. I presume that this would cause discomfort for many people who would not be excessively disturbed by the receipt of comparable service from a stranger, or by transactions involving the purchase of tangible goods or skilled tradework from an employed friend. It counts in favor of the present hypothesis that it handily explains this disparity. It traces the disparity to a cognizance that part of what one is consuming when one’s lobster is cut, one’s tray carried, one’s towel freshened, etc., is precisely another human being’s servility—something no one could comfortably consume from a genuine friend. Yet if this diagnosis is correct, it raises sharp questions about whether one should consume this putative good from a fellow citizen, or a fellow human being. And indeed, given the impact of Hochschild’s work on emotional labor, it seems that many of us do feel a gnawing discomfort with the personal relationships in which we are enmeshed as service economy clients. Yet it would take the stubborn devotion of a hermit to avoid them entirely, and a gnawing discomfort might well be insufficient to neutralize their character-distorting effects on providers and recipients. As noted above, sociological studies have found that emotional labor is associated with high levels of stress and job dissatisfaction, a sense of inauthenticity in one’s work life, and an array of symptoms of depression including especially emotional 24 See Wallace (1997): 293–4 and 297–9. I owe to Wallace the phrase ‘the professional smile’, and I have benefited from and been influenced (and greatly entertained) by his general description of the human significance of service work.



numbness.25 While I do not doubt these aggregate correlations, I suspect that many of these studies have cast too wide a net, and that researchers might have generated more illuminating and perhaps also more dramatic results by focusing on that special sort of emotional labor that commodifies servility. This sort of work has only a superficial similarity to the ennobling emotional labor required of a nurse or a medical doctor if she is to muster a decent bedside manner, yet sociological studies of emotional labor often include doctors and nurses alongside house cleaners and waiters.26 Nor does this sort of work have much in common with the labor required of a school teacher or professor if her critique of a student’s performance is to avoid tearing the fabric of hope that permits the student to throw himself into his studies, yet sociologists often count school teachers and college faculty members too among the ranks of the emotional laborers.27 While it is true that all of these jobs involve the self-management of emotions, doctors and nurses and teachers engage in this sort of self-management in order to promote a result that is conceptually independent of the emotional display itself, and that can be affirmed as vitally important without the least belittlement of the emotional laborer’s standing vis-a`-vis other human beings. Still, I do not think that the problems with service economy work can be traced entirely to the commodification of servility. This sort of work is particularly bad because it encourages workers to inhabit an undignified picture of their relative worth, while indulging clients in an unmerited picture of theirs. But the problem does not lie solely in the objectionable content of the picture of value that such work encourages workers to internalize. The more basic problem is that it encourages the internalization of an outlook that one values only as a means to securing a wage and not as a faithful register of what is independently valuable. Given this, such labor threatens to disfigure the lifelong activity through which we human beings give shape to our understanding of the human good. This more basic problem is endemic in the contemporary service economy, and extends to many jobs that are highly remunerated and widely coveted.

13.5 Emotions and the Task of Self-Elaboration We cannot hope to shed light on the potential harms of emotional labor without attending to the question of what exactly emotions are and what role they play in human life. I believe, and have argued elsewhere, that emotions are vivid ‘seemings’ of


Grandey et al. (2005): 900; Pugliesi (1999): 130 and 147; Tracy (2000): 95–6. Brotheridge & Lee (2002): 61 include all health professionals and office workers who have frequent contact with customers in their study. Glomb, Kammeyer-Mueller, & Rotundo (2004) include nurses, medical doctors, and psychiatrists. In a review of the literature on emotional labor, Zapf (2002): 238 cites a number of studies that have included nurses. Brotheridge & Grandey (2002) include social workers, who are more commonly perceived as paternalistic than as servile, in their study of emotional labor. 27 Pugliesi (1999) includes professors within her study of the deleterious consequences of emotional labor. Zapf (2002): 238 cites other studies that include teachers. 26



the evaluative contours of one’s changing circumstances, and that they serve to set the stage for evaluative understanding and action.28 Fear, for example, involves a vivid sense of some element of one’s circumstances, or some possibly impending event, as a threat to something one values, and it sets the stage for acting so as to defuse or escape the threat. Indignation involves a vivid sense of some person or act as offensive or disrespectful, and it sets the stage for active response to the perceived affront. This view of emotions locates them between conceptually denuded feelings such as the pain associated with a headache, and self-consciously affirmed stances on value such as those expressed in our considered evaluative beliefs. Emotions differ from headaches in that they are intentional states—they are about the world, and they present us with a picture of the value or importance of various features of our circumstances. It is not possible to have a delusional or unwarranted headache, but it is an all-too-familiar human possibility to have a delusional and unwarranted jealousy or fear. These points have sometimes been taken to support the Stoic thesis that emotions are a species of beliefs or judgments, but this is too quick.29 We can be fearful even when we believe that we are in no danger, and resentful even when our considered judgment is that no offense has really been given. When we are gripped by an emotion, we construe the world in a particular way, but we need not believe that it actually is that way.30 It is worth making two brief points about the complex distinction between emotions and evaluative judgments. The first point is that emotions organize our patterns of awareness so as to make it tempting to affirm that things really are the way the emotion portrays them—to believe, that is, that the emotion really is warranted. Fear, for instance, tends to make us see even the most innocuous events as signs of impending danger, while those in the throes of love often ‘see the best’ in all other people. By contrast, one can affirm an evaluative belief yet find that one’s pre-deliberative experience makes it seem as if the belief is false. The second point is that emotions sometimes inflect one’s conscious awareness well before one can say which emotion one is feeling, or what exactly has provoked an emotional response. By contrast it does not make sense to ask oneself whether one really is in pain, and it does not seem possible to have a belief or to make a judgment while not yet knowing what it is about.31 It is a truism that we are entangled with our own emotions in a much more intimate way than we can possibly be entangled with the emotions of others. But how exactly


Brewer (2009): chs 1, 2, and 4. Influential contemporary versions of this cognitivist conception of emotions can be found in Solomon (1976) and Nussbaum (2003). 30 A compelling construal-based account of emotions can be found in Roberts (2003). 31 This is a delicate point, and it may well need qualification. One difficulty concerns the thought that when we act we implicitly make the practical judgment that certain features of our circumstances count as sufficient reason to perform the act under some description of it. This is the thesis that underlies Kant’s confidence that a subjectively adopted principle or maxim can be assigned to every human action. If we accept this Kantian notion of practical thought and action, then we might well have reason to accept that we can and do make practical judgments without conscious awareness of their content. On the other hand, this implication might seem odd enough to ground a reductio of the view. On this question, see Brewer (2002). 29



should we characterize this intimate entanglement? One part of the answer is epistemic. While our knowledge of our own emotions is fallible, and while others sometimes know what we are feeling better than we do, still we have a vantage point on what we feel that is different in kind from the vantage point of others. This is a conceptual point. No one else could have the sort of access to my feelings that I have. If someone else did feel exactly what I feel in the same way that I feel it, the feelings felt by that person would simply be the other’s feelings—qualitatively identical (as it happens) to mine, but not mine.32 Here then is an intimate connection between people and their feelings, one that is grounded in the fundamental nature of feeling. But as long as we think of the connection in terms of having a special and direct vantage point from which to identify what one feels—the internal equivalent of a front row seat—we have not yet fully grasped the especially intimate connection between the self and its feelings. The connection is not at heart passive or spectatorial. It is active and creative. What do I do when a new but still quite indefinite feeling wells up in me? Well, sometimes nothing at all. But quite often I make it my project to determine what exactly is bothering me. If we reflect on exactly how we do this, we see that this task is essentially first-personal. It is not a task that could in principle be handed over to another, even if the other had more accurate insight into one’s psychological states than one has oneself. The task is not to identify one’s mental state but to try to get clear on what one sees when one looks through it. I do not determine what is bothering me by turning my gaze inward, in the spirit of someone seeking to describe the precise phenomenological contours of some alien thing that has lodged itself in my stream of awareness. The proper categorization of a feeling depends not upon how the emotion strikes the person feeling it but how the emotion’s object strikes the person feeling it. If I am bothered by something my lover has done and want to determine what my incipient feeling is, I do not direct my attention inwards at the feeling itself (whatever that would mean); I look outwards at my lover’s doings with an eye to seeing more clearly what is troubling about them. As I do this, my feeling might congeal into a vivid sense of having been wronged (i.e. resentment) or a vivid concern that the attention of a lover is straying (i.e. jealousy), or it might coalesce into some other emotion—e.g. fear of abandonment, or forlornness, or distress at not being fully understood. What gives more determinate shape to the emotion is not my arrival at a judgment concerning how things are. I might not end up making any such judgment. What gives more determinate shape to my emotion is the reshaping of how things appear to me that occurs when I attend carefully to the question how things are. It is through this sort of active and concerted attention to what is not myself, then, that I perform the work of self-elaboration. It might seem puzzling that we can determine what emotion we are feeling by getting clear on the emotion’s object. Isn’t it one question what emotion we are feeling


Cavell (1969) makes this point, though with reference to pain. See pp. 247–53.



and quite another question what has prompted us to feel the emotion? I don’t think so. There is a basic conceptual link between kinds of values and kinds of evaluative responses. We do not have one array of concepts for characterizing kinds of value and another, independently intelligible array of concepts for characterizing kinds of evaluative responses. Many of our most discerning evaluative responses cannot be characterized without ineliminable reference to the values they bring into view, and these values cannot be characterized without ineliminable reference to the evaluative response that brings them into view. For instance, we cannot know what it is to be amused without having some inkling of what it is to find something amusing or funny, but we cannot grasp what it is to be funny without having some inkling of what it is to be a fitting object of amusement.33 This conceptual interdependence obtains between many emotions and the values they bring into view: fear and the fearful, sadness and the sad, grief and the grievous, admiration and the admirable, respect and dignity, and so on. Even when we move beyond the domain of the nameable emotions, we see the same pattern of conceptual interdependence between specific evaluative properties and our sensitivity to them. To take but a few examples, we sometimes experience people or things as gracious, courageous, beautiful, majestic, humane, elegant, wise, amusing, or deserving of gratitude. To say that these are all ways of finding something good or valuable is to bring out a family resemblance among them, not to suggest that they involve a uniform attitude of approval. It is one thing to appreciate the elegance of a mathematical proof and quite another to appreciate someone’s humanity. These specialized ways of finding things valuable cannot be characterized without reference to the values they bring into view, and vice versa. When we try to get straight on what we feel, then, we do not face two separate tasks: to make our evaluative stance articulate and to clarify whatever values are in view. These are at heart the same task—the task, as I am calling it, of self-elaboration. It is well established that if we are asked to report our emotions, our answer will depend upon our interpretation of the practical or evaluative contours of our situation. That is, if a subject is caused to have an emotion-like psychological perturbation and asked what emotion he is feeling, he will tend to report whatever emotion it would make most sense to feel given his understanding of his circumstances.34 On the view of the emotions that I have offered, this experimental finding comes as no surprise. It does not reveal us to be fabulists who invent emotions in order to salvage the intelligibility of brute psychological processes. It simply manifests the conceptual connection between having an emotion and having a construal of one’s circumstances that makes that emotion appropriate. Since the construal is an essential constituent of the emotion itself, the emotion cannot float free from the achieved construal.


For a highly influential elaboration of this view, see chapters 7 and 8 of McDowell (1998). For a summary discussion of some of the more important experiments in this area of research, see Velleman (2009): 37–9 and (2006): 224–52. 34



Our emotions, then, manifest our way of seeing the evaluative contours of our circumstances. We cannot back away from the view of the world that they serve up to us the way that we can simply refuse a view put forward by a stranger. If we believe that our emotions provide a badly distorted picture of value, still we must admit that our running apprehension of the world is impaled upon these distortions, and that the distortions pose a serious threat to the possibility of acting spontaneously in ways that we will not later regret. Our emotions provide raw materials for the task of selfelaboration, but they are raw materials of a very special sort. One cannot simply refuse the material the way that a potter could return defective clay to the supplier and demand some more suitable material. Emotional materials are not entirely raw; they are works in progress. They do not come from a source external to the activity of selfelaboration. They are the form taken by our inchoate conceptualization of our circumstances. They manifest the way in which evaluative concepts have taken up residence in our sensibility. To be dissatisfied with these materials is to be dissatisfied with some element of the self we are continually elaborating. To approach the point from a slightly different angle, if we regard our emotions as unwarranted—that is, if they serve up a picture of the evaluative contours of our circumstances that we reject on due reflection—we must admit that something has gone wrong with some element of our practical outlook on the world. I believe it would overstate the case to claim, with Richard Moran, that an attitude that cannot be known by reflection on its object must be regarded as rationally defective, and indeed as ‘a kind of fixation’.35 To mention one frequently cited example, this standard would dismiss as a mere fixation the sense of loyalty that makes it impossible for Huck Finn to turn in the runaway slave Jim, since Huck’s considered view is that he is doing something terribly wrong by refusing to go to the authorities. Still, Moran’s comment is very nearly right: there must be a defect somewhere in a person’s evaluative outlook if the person has an emotion that cannot be known by reflection on its object. If our practical outlook is free of such defects, then we can refashion our emotions by peering through the evaluative lens that they offer and trying to arrive at a considered assessment of their objects. A central part of the work of self-elaboration, then, is the effort to bring one’s halfformed evaluative sensibility into words. As Charles Taylor has stressed, one’s own inchoate evaluations are not the sorts of things that can be expected to hold still while one attempts to capture them in words.36 They take on a more fine-grained (or more completely articulated) shape as one brings them into words. This, however, does not mean that these inchoate evaluative outlooks are capable of taking any shape whatever. It is always possible that my words will fail to capture the glimmer of value contained in my own incipient emotions, and hence that my considered account of my feelings will not be inflated with the vividness or the urgency of the feeling I am trying to articulate. 35 36

Moran (2001): 107–8. Taylor (1976): 281–300.



There is, however, also the possibility that my articulation of my inchoate feeling will ‘take’—embedding itself, so to speak, in the emotion itself, and giving a more determinate and perhaps more discerning character to the evaluative outlook in which the emotion partly consists. The work of self-articulation extends not only to inchoate feelings but also to incipient thoughts. In this latter form, it is a familiar part of philosophy. We philosophers engage in this sort of work whenever we find ourselves searching for the right words to stretch our unfolding sentences towards an intimation of insight lying at the horizon of our understanding. In such cases, one has the sense of trying to wrest the thought that has inspired one’s philosophical interest from a partially veiling obscurity and to bring it into words that vindicate one’s incipient excitement about it. One has been faithful to the insight if the incipient excitement vests itself in the words one finds, and faithless if one’s more definite articulation breaks away from that incipient insight without any sense of having found it to be illusory. We are also engaged in the task of self-articulation when we stand before a painting in a museum and try to say what exactly we like about it. This sort of quest for aesthetic articulacy does not leave our experience of the painting unchanged. Ideally our experience gains depth and becomes more discerning as we bring its evaluative contours into words. When I engage in the elaboration of incipient thoughts or feelings, I must be recognizing myself in those thoughts and feelings—acknowledging them, that is, as structuring the way things are appearing just now to me. The sense of ‘appearing’ here needs clarification. If I close my eyes and imagine a unicorn, I experience what might perhaps be called an appearance, but this is not yet a case of the sort of ‘appearing’ at issue here. I am aware that I am merely imagining the unicorn, hence the conjured image does not alter how things appear to me to be. The thoughts and feelings that are proper raw materials for self-articulation are different. I am involved in them because they condition how things appear to me to be (though, as I have stressed, perhaps not how I judge them to be). This is what makes it possible to shape myself—that is, to clarify or sharpen my posture towards the world—by working towards a clearer apprehension of the objects of my thoughts and feelings. There is a crucial normative aspect to the attitude I must take towards my emotions and desires when I take them up as raw materials for the work of self-elaboration. I must regard them as worthy of elaboration precisely in virtue of the fact that I rely upon them as the lens through which I discern the good. If I do not see them as valuable in this particular way, then I will not have reason to elaborate them by looking through them and trying to discern what is there to be seen. I might take myself to have reason to shape them in accordance with some other standard, dissociated from the faithful apprehension of values that are there to be seen, but I will not recognize a reason to shape them by the sole standard of faithfulness to an independent order of value. Since I can hardly think it important for my emotions to track the good unless I think it important for me myself to track the good and to think and act in its light, it follows that when I engage in the work of self-elaboration I implicitly recognize myself



as someone who is striving for a deeper appreciation of the good, and whose progress in that endeavor genuinely matters. In his illuminating discussion of reification, Axel Honneth argues convincingly that one fails to recognize one’s fundamental value as a human being if one treats one’s own incipient thoughts and feelings either as mere means for bringing about some conceptually separate end, or as fixed entities before which one can assume only a passive or contemplative stance.37 Either of these self-relations counts, by Honneth’s lights, as a form of self-reification—that is, as a case of regarding oneself as a mere thing rather than as a human being. Curiously, Honneth does not extend this view of reification and recognition to the interpersonal case. Yet he takes himself to be discussing the sort of recognition that grounds a proper moral concern, and presumably the basis of a proper moral concern remains the same across the first-person/second-person divide. I think that Honneth’s picture of self-recognition is quite apt, and that it ought to be extended to the interpersonal case. To recognize a human being is to recognize a being that is always already engaged in the work of self-elaboration whenever it exists at all. What is to be recognized is not the momentary status or product of this work, but the work itself—that is, the ongoing activity of self-elaboration in which a distinctively human life consists. Such recognition has different practical implications in the firstperson and second-person cases, since we cannot elaborate the feelings of others merely by straining to see through them and articulate their objects, nor are we generally called to bring the inchoate sensibilities of others into words. But we can recognize this practical asymmetry while insisting that first-personal and second-personal recognition both have the same ongoing life activity as their proper object. It might seem that there are two quite separate problems with the sort of service economy work that we’ve been discussing.38 First, it interferes with the integral selfelaboration of the worker. Second, it invites other human beings (clients) to withhold recognition of the worker’s equal standing. On the account of recognition just sketched, there is an intimate connection between these two problems. A failure to recognize another human being’s equal worth just is a failure to acknowledge the value of the activity of self-elaboration through which that other human being is actualized. If we view emotional labor through this lens, its badness does not depend entirely upon the subordination of the service provider to the client. A wide array of service economy positions, including many that are quite prestigious, foster a purely instrumental stance towards one’s own incipient thoughts and feelings, while enmeshing their occupants in an array of social relations that reinforce that stance.39 On the view under consideration, these positions all encourage intrapersonal and interpersonal relations of


Honneth (2008): 63–74. I thank Jay Bernstein for pressing me on this issue. 39 The sociologist Richard Sennett has argued that ‘the culture of the new capitalism’ is distinguished by its tendency to foster an extremely malleable personality type—one capable of continuously reinventing itself to meet changing workplace demands. See Sennett (2006): ch. 1. 38



misrecognition and therefore impede the fundamentally important life activity through which human beings strive to hone their evaluative outlook so that it tracks genuine goods. As we saw above, Marx (or, at least, the Marx of the 1844 Manuscripts) found the ergon or characteristic activity of the human species in a cyclical form of labor through which we remake the world so as to accommodate our evolving sense of how human beings ought to live, then refine our sense of how human beings ought to live through sustained interaction with the world that we have fashioned for ourselves. Selfelaboration shares the continuous cyclical structure of the sort of labor that interested Marx, but the cycle is both tighter and more personal—moving not between considered evaluative views and material objects, but only between considered and spontaneous elements of one’s evaluative outlook. While I do not think that self-elaboration is an exhaustive account of the human ergon, it seems to me that it is an essential element of the kind of life to which human beings are called, and in which their best and highest potentialities are actualized. It is through the work of self-elaboration that we are able to take responsibility for our characters by engaging in the daily task of giving determinacy to our conception of the good life. The most fundamental problem with emotional labor is that it tends to impart an alien evaluative outlook—one whose contours are shaped by the interests of the employer and not by the worker’s own lifelong effort to see what is genuinely good. While this sort of labor works on the psyche of the laborer, still it is not an asocial activity, much less a solipsistic one. Our emotional comportment has an active and social face. Many of our emotions concern the evaluative contours of our social relations. Anger, for instance, involves a vivid impression as of having been offended, while indignation involves a sense of having been done a serious moral wrong. Outward displays of these emotions can have a dramatic effect on the social relations upon which they comment. Suppose a driver has just rudely cut in front of me. If I angrily extend my middle finger and thrust it upwards towards the driver, and if the other driver clearly sees what I have done, I simultaneously give expression to my anger and somehow also return my social circumstances to a state that no longer calls for anger.40 Indeed a well-executed finger can leave its perpetrator with a sense of triumph. By contrast, if one feels compelled to swallow one’s anger or indignation, the containment in itself can count as acquiescence in the offending party’s contempt for one’s standing, and hence can constitute a further indignity. It is partly through outward displays of emotion that we work out the normative contours of our social relations. This too is the sort of work that one conducts over time, in dialectical interaction with the social world whose contours are being shaped. Here the dialectical circle extends outwards from one’s own subjectivity to an intersubjective social world of which one is at most a co-author.


For a marvelous disquisition on ‘the finger’ and its mysterious powers, see Katz (1999): 60–6.



13.6 What’s So Bad about Emotional Labor? In Kazuo Ishiguro’s The Remains of the Day, the narrator, Stevens, is looking back on his long service as head butler of Darlington Hall, a grand English manor that has recently been purchased by a wealthy American businessman.41 The role of butler is not something that can sensibly be scripted unilaterally. It is a form of service to another, and as such it arises at the meeting point of the expectations of the employer and the aspirations of the butler. Stevens and the late Lord Darlington had a shared understanding of the role, informed by long-standing English traditions to which they both were heir (Stevens’ father was a butler; Darlington’s, of course, a peer). While serving dinner to Lord Darlington, for instance, Stevens would seek ‘to achieve that balance between attentiveness and the illusion of absence that is essential to good waiting’.42 But with Darlington dead and gone, Stevens is seeking to alter his aspirations to fit the unfamiliar expectations of an American employer. This alteration will require more than a nip here and a tuck there. It will require Stevens to learn the unfamiliar art of ‘bantering’ with another human being, speaking of personal matters, offering opinions on the events of the day, laughing at jokes, and occasionally venturing witticisms of his own. To master these new job requirements, Stevens is pressed to see himself through the eyes of his new employer, and from that perspective he shows up as a living curio of a vanishing past that can be shown off to guests as they admire the other fine period details of Darlington Hall. Presumably it is the strain of having to imagine himself through his new employer’s eyes, and to conform his speech and his emotional display to new and unfamiliar expectations, that has prompted Stevens to take stock of his life and ponder what to do with what remains of his days. Stevens is especially preoccupied with the question of just how good a butler he has managed to be. He has settled upon a phrase for the rare quality that distinguishes great butlers from merely good ones. A great butler must have ‘a dignity in keeping with his position’.43 Here is Stevens’ attempt to explain this imagined species of dignity: Lesser butlers will abandon their professional being for the private one at the least provocation. For such persons, being a butler is like playing some pantomime role; a small push, a slight stumble, and the fac¸ade will drop off to reveal the actor underneath. The great butlers are great by virtue of their ability to inhabit their professional role and inhabit it to the utmost; they will not be shaken out of it by personal events, however surprising, alarming or vexing.44

Stevens recounts with particular pride the unbroken and outwardly perfect service he offered to an international assemblage of aristocrats and statesmen who, as it happened, were gathered at Darlington Hall on the night that Stevens’ own father—who had also 41 42 43 44

Ishiguro (1988). Ibid.: 72. Ibid.: 33. Ibid.: 42–3.



been a butler—was dying upstairs. When the head housekeeper, Miss Kenton, breaks the news to Stevens that his father has succumbed, he sends her upstairs to close his father’s eyes so that he can continue to tend to the guests. ‘You see’, he explains to Miss Kenton, ‘I know my father would have wished me to carry on just now . . . To do otherwise, I feel, would be to let him down.’45 It soon becomes obvious to the reader that Miss Kenton has fallen in love with Stevens, and that Stevens has strong feelings for her that he will not permit himself to acknowledge. When these feelings well up, Stevens seems not to understand entirely what they are. Nor does he engage in the work of self-elaboration through which these feelings might be brought to articulacy. Rather than dwelling on Miss Kenton with an eye to determining what significance she has for him, he turns back to his work with redoubled attention to detail. Put another way, he enacts his favored conception of ‘dignity in keeping with his position’—disciplining his feelings and thoughts so as to prevent any breach in his occupancy of his professional role. We see this pattern when Miss Kenton receives word of the death of the aunt who raised her and who is her closest remaining family member. The thought of Miss Kenton crying alone in grief behind the closed door of her room, just a few feet away from him, causes what the ordinarily fastidious and highly articulate Stevens manages to describe only as ‘a strange feeling’.46 Yet when, under the influence of this ‘strange feeling’, Stevens approaches Miss Kenton with the vague idea of offering some comfort, he finds himself instead offering a pointed critique of her recent job performance.47 That is, he presses his felt perturbations into the alien mold of his professional preoccupations, bypassing the task of bringing them faithfully to articulacy, and he simultaneously invites Miss Kenton to respond to her own grief with the numbing practice of ‘dignified’ self-denial that he has perfected in his own case—a possible basis for sharing a fate, no doubt, but not a very promising recipe for happiness. In the end Miss Kenton declines this invitation. Though she continues to harbor some hope that Stevens will relent, she begins to spend time with another man who works at a nearby manor. When she tells Stevens that this man has proposed to her, and that she might possibly accept his proposal and leave Darlington Hall, Stevens’ ‘dignity’ intercedes again. He assures Miss Kenton that she has ‘his warmest congratulations’, then insists with rather excessive vehemence that he really must return to his work.48 Later that evening Stevens finds himself once again just outside the closed door of Miss Kenton’s room, certain for a second time that she is crying alone just a few feet away from him. He pauses a long moment, tray in hand, uncertain what to do. Then he hurries upstairs to wait unobtrusively in the vicinity of the drawing room, in case the gentlemen gathered there might need his service. At first he is ‘somewhat downcast’,

45 46 47 48

Ibid.: 106. Ibid.: 176. Ibid.: 177–8. Ibid.: 219.



but within an hour ‘a deep feeling of triumph started to well up within [him]’—a feeling he attributes to his awareness of having ‘just come through an extremely trying evening, throughout which [he] had managed to preserve “a dignity in keeping with [his] position” ’.49 What is Stevens’ plight to us? Isn’t it of primarily antiquarian interest, conditioned as it is by a long tradition of class privilege and class deference? I don’t think so. In Stevens we have a revealing if somewhat exaggerated portrait of the hidden costs (the ‘inner externalities’, as it were) of certain kinds of emotional labor. The cost is that Stevens’ inchoate feelings are left in crude form, deprived of that extra degree of intelligibility and determinacy that might have been conferred upon them by a sustained effort to articulate the potential goods they intimated. It is only in retrospect that Stevens begins to make sense of the perturbations that have marked his relations with Miss Kenton, and by that time what remains of the day is not sufficient to permit any true compensation for the cost exacted by his notion of dignity. The distortion we’ve been considering in the self-elaboration of certain service workers is, I think, essentially the same as the distortion that Socrates identifies in the rhetorician’s misuse of logos. The rhetorician makes use of this distinctively human capacity for thought and speech as a mere means of persuasion, which in turn is bent to the task of satisfying desires for pleasure or wealth or power. This alters the proper relation between the rhetorician and his own tongue, since he views his words as calculated instruments for remaking the world, and thereby cannot make use of them as the medium of the continuous effort of reflection through which humans can lend clarity and articulacy to their evolving idea of how best to live. The practice of rhetoric effectively subverts the activity of self-elaboration through careful reflection, with the effect that one is continuously shaped by one’s sense of what others would find flattering.50 The emotional laborer is interrupting this same active and self-formative thought. One’s emotions cease to play their proper role as embodiments of one’s past evaluative reflections and starting points for the next stretch of reflection. They are taken up, instead, as instruments to be reshaped in service to desires for pleasure or wealth or professional success. This interrupts the ongoing task of clarifying one’s evolving sense of how it would be good to live, by pressing a crucial constituent of that activity into the service of one’s pecuniary interests. This, I think, is the central harm of emotional labor, and its deleterious effects might well be quite widespread in today’s economy.

13.7 Concluding Methodological Remarks There was a time when social and political theorists were deeply concerned with the effects of capitalist forms of commerce on the character of citizens. Some sided with Montesquieu, Condorcet, Adam Smith, and David Hume in asserting that commerce 49 50

Ibid.: 227. Plato, Gorgias, 481c–e.



exercised a civilizing influence on those engaged in it, fostering virtue and/or making manners soft and gentle. Others sided with Rousseau in insisting that capitalism deafened humans to the voice of pity and fed egocentric longings, thereby making it possible for humans to revel in dominating their fellow human beings and in benefiting from the misfortune of others.51 Within the ranks of Anglo-American political philosophers, these debates have for the most part been left behind. This is due in part to the dependence of these debates on empirical claims that are now regarded as the provenance of sociology, psychology, economics, and political science. (It is an implicit supposition of the current essay that there is valuable work to be done in the interstices between philosophy and these empirical disciplines.) But it is also due to the enormous influence of John Rawls, whose work has focused the attention of Anglo-American political philosophers on questions of justice in the distribution of rights, income, wealth, and opportunity. In my view, there are interesting social and political pathologies that cannot be illuminated adequately in the language of justice and injustice, at least in the austere usages to which that language is put by Rawls. An example is the sort of self-alienation that arises in the contemporary service economy. I do not mean to suggest that the work patterns in question are not unjust, nor that their injustice is not visible in a Rawlsian conceptual framework. Rawls counted ‘the social bases of self-respect’ among the primary goods that must be distributed fairly in an ideally just society, and some service economy labor does seem to yield unjust distributions of this important good.52 Rawls was also deeply concerned with the conditions under which a lively concern for justice could be passed down from one generation to the next, so as to stabilize just institutions.53 This might provide another opening for considering the normative stakes of patterns of servility in the service economy, since these might impede the rise of a lively sense of the equal standing of all citizens as fellow participants in the project of political self-rule. Either of these Rawlsian concerns might imply that justice requires not only a redistribution of opportunities to work and of the financial rewards of work but also a more basic reshaping of the way work is organized. The considerations put forward in this essay, then, might bring out a largely untapped radicalism in Rawls’ conception of political justice. Yet there are serious obstacles to such a proposal. In the first instance, it would require an expansionist notion of the subject matter of a philosophical conception of justice, one that brings the souls of citizens within the ambit of justice. Second, it would require the invocation of a thick and controversial conception of human flourishing, one that would deviate from the basic liberal commitment to neutrality among competing conceptions of the human good. I have not taken on these challenges. Instead, I have turned directly to 51 52 53

An informative history of these debates can be found in Hirschman (1982). See Rawls (1971): section 67. Ibid.: ch. VIII.



the question whether certain kinds of work in the service economy impede the flourishing of those who spend their waking hours engaged in them. Whether or not such reflection yields insights into the injustice of our socio-economic order, it might illuminate the distortions of character that are endemic to it. It is not only among professional philosophers that such questions have fallen into disfavor. With a few notable exceptions, these questions get scant attention in popular political deliberation and debate. Capitalism is frequently subjected to radical criticism for its effects on the natural environment, but there is far less discussion of its effects on the human environment.54 I think there would be considerable value in sustained philosophical reflection on this latter topic. This essay represents one small contribution to that project.

References Ashforth, Blake & Humphrey, Ronald (1993) ‘Emotional Labor in Service Roles: The Influence of Identity’, The Academy of Management Review 18(1): 88–115. Bellamy, Edward (1960) Looking Backward, 2000–1887. New York: New American Library. Bluestone, Barry & Rose, Stephen (1997) ‘Overworked and Underemployed: Unraveling the Economic Enigmas’, The American Prospect (March/April): 58–69. Braverman, Harry (1974) Labor and Monopoly Capital: The Degradation of Work in the 20th Century. New York and London: Monthly Review Press. Brewer, Talbot (2002) ‘Maxims and Virtues’, The Philosophical Review 3(4): 539–72. —— (2009) The Retrieval of Ethics. Oxford: Oxford University Press. Brotheridge, Celeste & Grandey, Alicia (2002) ‘Emotional Labor and Burnout: Comparing Two Perspectives of “People Work” ’, Journal of Vocational Behavior 60: 17–39. —— & Lee, Raymond (2002) ‘Testing a Conservation of Resources Model of the Dynamics of Emotional Labor’, Journal of Occupational Health Psychology 7(1): 57–67. Cavell, Stanley (1969) Must We Mean What We Say? Cambridge: Cambridge University Press. Crawford, Matthew B. (2009) Shop Class as Soulcraft: An Inquiry into the Value of Work. New York: Penguin. Ehrenreich, Barbara (2002a) ‘Maid to Order’. In Ehrenreich & Hochschild (2002b): 85–103. —— & Hochschild, Arlie (2002b) (eds) Global Woman: Nannies, Maids and Sex Workers in the New Economy. New York: Henry Holt and Company. Glomb, Theresa M., Kammeyer-Mueller, John D., & Rotundo, Maria (2004) ‘Emotional Labor Demands and Compensating Wage Differentials’, Journal of Applied Psychology 89(4): 700–14. Grandey, Alicia, Fisk, Glenda, & Steiner, Dirk (2005) ‘Must “Service With a Smile” Be Stressful? The Moderating Role of Personal Control for American and French Employees’, Journal of Applied Psychology 90(5): 893–904.

54 There are some notable exceptions to this sweeping generalization. One excellent new work in this genre that found a very large audience (it was a New York Times bestseller) is Crawford (2009). Another notable exception is the hit television show The Office, which focuses in large part on the psychodynamics of alienated white collar work.



Gross, James J. (1998) ‘Antecedent- and Reponse-Focused Emotion Regulation: Divergent Consequences for Experience, Expression and Physiology’, Journal of Personality and Social Psychology 74(1): 224–37. Harre´, Rom (1986) (ed.) The Social Construction of Emotions. Oxford: Basil Blackwell. Hirschman, Albert O. (1982) ‘Rival Interpretations of Market Society’, Journal of Economic Literature 20(4): 1463–84. Hochschild, Arlie Russell (1983) The Managed Heart: Commercialization of Human Feeling. Berkeley and Los Angeles: University of California Press. Honneth, Axel (2008) Reification: A New Look at an Old Idea. Oxford: Oxford University Press. Ishiguro, Kazuo (1988) The Remains of the Day. New York: Vintage Books. Katz, Jack (1999) How Emotions Work. Chicago and London: The University of Chicago Press. McDowell, John (1998) Mind, Value and Reality. Cambridge, MA: Harvard University Press. Marx, Karl (1975) Early Writings. New York: Penguin Books. Moran, Richard (2001) Authority and Estrangement: An Essay on Self-Knowledge. Princeton, NJ: Princeton University Press. Neuhouser, Frederick (2008) Rousseau’s Theodicy of Self-Love: Evil, Rationality and the Drive for Recognition. New York: Oxford University Press. Nussbaum, Martha (2003) Upheavals of Thought: The Intelligence of Emotions. Cambridge: Cambridge University Press. Plato (1986) Gorgias. Indianapolis: Haektt Puplishing Co. Pugliesi, Karen (1999) ‘The Consequences of Emotional Labor: Effects on Work Stress, Job Satisfaction and Well-Being’, Motivation and Emotion 23(2): 125–54. Rawls, John (1971) A Theory of Justice. Cambridge, MA: Harvard University Press. Roberts, Robert C. (2003) Emotions: An Essay in Aid of Moral Psychology. Cambridge: Cambridge University Press. Rousseau, Jean-Jacques (1997) The Discourses and Other Early Political Writings. Cambridge: Cambridge University Press. ——(1979) Emile: Or, On Education. New york: Basic Books. Sennett, Richard (2006) The Culture of the New Capitalism. New Haven: Yale University Press. Simmel, Georg (1971) On Individuality and Social Forms. Chicago: The University of Chicago Press. Solomon, Robert C. (1976) The Passions. New York: Doubleday. Taylor, Charles (1976) ‘Responsibility for Self ’. In: The Identities of Persons, Ame´lie Rorty (ed.), pp. 281–300. Berkeley and Los Angeles: University of California Press. Tracy, Sarah (2000) ‘Becoming a Character for Commerce: Self-Subordination, and Discursive Construction of Identity in a Total Institution’, Management of Communication Quarterly 14: 90–128. Velleman, J. David (2006) Self to Self. New York: Cambridge University Press. —— (2009) How We Get Along. Cambridge: Cambridge University Press. Wallace, David Foster (1997) A Supposedly Fun Thing I’ll Never Do Again. New York and Boston: Back Bay Books/Little, Brown and Company. Zapf, Dieter (2002) ‘Emotion Work and Psychological Well-Being: A Review of the Literature and Some Conceptual Considerations’, Human Resource Management Review 12: 237–68.

Index Abramson, Kate 8 Action-guidingness 118, 122–3, 238 Adams, Robert M. 87 Affect appraisal 6 Affective attitude 184 Affective awareness 260 Affective concepts 6, 119, 128, 130 affective components 7, 10, 43, 45, 55–6, 110, 259, 271 affective components integrated 57, 174, 184 thick affective concepts 6 Affective neutrality 267 Affective perception of value 8, 23, 260 affects and salience 45, 51 Affect program 17, 141 Agential authority 20–1, 260. See also Autonomy Ainslie, George 48 n. 5 Alienation 27, 275, 276, 278 Altruism 171, 189, 190, 202 mutual affection 202 Altruistic attitude 185. See also Altruism Ambivalence 92 Anderson, Elizabeth 10, 117 n. 1, n. 2, 119 n. 7 Anderson, Scott 78 n. 36 Anger 8, 23, 26, 40, 44, 65 n. 9, 67–9, 141, 151, 155–6, 180, 187, 217, 221, 243, 258–9, 267, 292 and danger, 155–7, 221, 267 empathic 176 Annas, Julia 6 n. 4, 270 Anscombe, Elizabeth Gertrude M. 3, 12, 13 Antonaccio, Maria 4 Aquinas 164 Aristotle 4–7, 15, 164, 245 on practical syllogism 44 on pride 270 on self-love 83–7 Attachment 22, 200 secure attachment 89 theory 82 to persons and institutions 202, 204 Attitudes constitutive of rational agency, 63, 75–8 Attitudes, participant vs. objective 218 Aubie, B. 151, 153 Audi, Robert 165 Authenticity and lack thereof 280, 284 Autonomy 1, 8

agential autonomy as requirement of rational agency 18, 21 and guilt 271 as responsiveness to reasons 227 cognitive autonomy 18, 227, 231 moral autonomy and respect 72–8 personal autonomy 21, 71, 82, 104, 153, 160 Bacin, Stefano 72 n. 24 Bagnoli, Carla 4, 9, 10, 16, 21, 26, 66, 74–6, 148 n. 19 Baier, Annette 6, 260 Baker, Ann 255 n. 15 Barasch, Marc Ian 182 n. 18 Barcan, Marcus Ruth 66 n. 11 Baron, Marcia 10, 66 n. 12, 70 n. 19 Batson, C.D. 158, 172, 173 n. 2, 177–8, 182 Bauman, Zygmut 103, 108 Bedford, Edward 25 Behaviorism 9 Bellamy, Edward 277, 278 Ben-Ze’ev, Aaron 16, 22, 23, 100, 102, 113, 140 Benchimol, Jason 255 n. 15 Benjamin, Martin 99 Bennett, Jonathan 26, 192 n. 2, 214 n. 2, 218 n. 3, 219 n. 6, 225, 229 n. 24 Berker, Selim 15, 27 n. 11 Bernstein, Jay 273 n. 1, 291 n. 38 Biddle, Francis 99 Bjorklund, F. 166 Blackburn, Simon 7, 65, 66, 118, 125, 128, 144, 170 Blame 2, 19, 25, 65, 205, 210, 219 n. 6, 236 n. 2, 249, 257–60, 262, 266–7, 271 See also Punishment, Responsibility Bluestone, Barry 277 n. 6 Blum, Lawrence 4 n. 2, 22, 24, 181 n. 17, 189 n. 28 BonJour, Lawrence 251 n. 13, 255 n. 15 Borges, Maria 10, 70 n. 18 Brady, Michael 24 Bratman, Michael 53 n. 9 Braverman, Harry 277 n. 4 Brentano, Franz 8 n. 8, 117 n. 1 Brewer, Bill 123, 137 n. 7 Brewer, Talbot 27, 286 n. 28, 286 n. 31 Brotheridge, Celeste 280 n. 15 Butler, Samuel 87 Bykvist, Krister 119 n. 5



Carlo, G. 175 Carr, David 6 Carter, Ian 27 n. 11, 78 n. 36 Cavell, Stanley 287 n. 32 Chang, Ruth 47 Change in attitude 67–8, 98 Character 2, 5, 11, 26, 45, 82–3, 88, 236, 244, 258–60 and moral knowledge 264–6 Aristotelian vs. Humean accounts 4–5, 260–4 distortion 284, 296–7. See also Authenticity Chazan, Pauline 83 Chisholm, Roderick 119 n. 5, 122 Churchland, Paul 16 Clark, Maudemarie 78 n. 36 Clark, R. 181–2 Clarke, Samuel 62, 71, 117 n. 1 Cognitive appraisal theory 151. See also Cognitivism Cognitive sciences 12–19, 20 Cognitivism 5–7 critique of 7 Cohen, David 16 Compassion 1, 19, 24, 26, 170–3, 205, 231, 254, 263 critique of Nichols’ account of 189–92 Kantian accounts of 68–9, 70 n. 19 Concern mechanism 185–9. See also Nichols Condorcet 295 Conflicting attitudes 92 Conscience 1, 10, 24, 64, 150, 152, 156, 158, 160–4, 258 Corbı`, Joseph 78 n. 36 Cosmides, Leeda 26 n. 10 Cratsley, Kelso 193 n. 34 Crawford, Matthew 273 n. 1, 298 n. 54 Crisp, Roger 18, 19, 20, 65 n. 12, 124 n. 19 Cuypers, Stefaan 78 n. 36 D’Arms, Justin 18, 45, 62, 118–23, 126, 144, 248 n. 11 Damasio, Antonio 6, 16, 17, 151 Dan-Cohen, Meir 241 n. 7 Dancy, Jonathan 124 Danger, 140, 141, 143–4, 146, 157, 221, 236, 247–8, 261, 267, 286 Danielsson, Swen 117, 119 n. 9, 124, 127 n. 21 Darwall, Stephen 11, 25, 26, 41 n. 21, 66, 75 n. 30, 75 n. 32, 118, 121, 198 n. 4, 208–13, 226 Day, T.J. 97 Deigh, John 6, 20–2, 25, 65 Deonna, Julien 117 De Sousa, Ronald 5, 6 n. 7, 7, 17, 22, 25, 42, 45, 65 n. 5, 67 n. 14, 120 n. 10, 135 n. 1, 141 n. 11 Detached intellect 16, 20

De Waal, Frans 261, 264–5, 268 Diamond, Cora 4 Dienstbier, R. 175 Dillon, Robin 27 Disaffection 106–10 Distress 174, 187 helping other people’s 189–90 Dokic, Jerome 117 Doring, Sabine 120 n. 10, 135 n. 1, 136 n. 6 Doris, John 15, 177 n. 11 Eisenberg, Nancy 172 n. 1, 173, 175, 182 n. 18 Elgin, Catherine Z. 135 n. 1, 137, 142, 145, 146 Elster, John 16 Emotional attitudes 157 Emotions analogy with sensory perception 127, 137–8 critique of the analogy with sensory perception 139–45 Kant’s critique of the perceptual model 70 as appraisal 24, 108, 136, 139–60 as appraisal of character 261 as blind sensations 5, 70, 136 n. 2 as perceptions 2, 7, 8, 10, 23–4, 45, 136, 185–6, 191, 205, 207, 260 as perspectival 22, 136 as reason-trackers and reason-providers 45, 189 as somatic perceptions 151 education of 6, 13, 19, 65, 72 n. 23, 163, 264–5 involuntary 5, 9, 10. See also Inclinations, Cognitivism misdirected 42, 52 of others 270 of responsibility 271 passive 10, 16, 17, 48, 59, 69 n. 17, 70 n. 18, 97, 136, 287, 291 recalcitrant 18, 129, 136 n. 3, 139 Empathy 2, 4, 163, 171, 261–7, 272 and sympathy 173 as “fellow feeling” 173–5 as intentional 174–5. See also Nichols, Scheler Enforcement view of moral emotions 65 Engstrom, Stephen 14 Evans, Dylan 26 n. 10 Ewing, A.C. 117 n. 1, 119 n. 9, 124, 127 n. 21 Faucher, Luc 117 Feinberg, Joel 240 Fine, Cordelia 15 Fischer, K.W. 271 Fitting Attitude Analysis 117–22 Flanagan, Owen 4 Foot, Philippa 5 n. 4, 13, 25, 65 Forrester, Alexa 234 n. 36 Frank, Robert 17, 26 n. 10


Frankfurt, Harry 22, 89, 90, 91, 92, 93 Freud, Sigmund 242, 243, 244 Frijda, Nico 46 Fromm, Erich 277 n. 5 Fyfe, Andrews 255 n. 15 Gallagher, Shawn 16 Gibbard, Allan 8, 9, 26 n. 10, 65–7, 118, 121, 125, 144, 258, 259, 260, 271 Gill, Michael 190 n. 30 Glezen, Paul 255 n. 15 Glomb, Theresa M. 285 n. 26 Gluer, Kathrin 136 n. 4 Goldie, Peter 143, 174, 182 n. 19, 187 n. 26, 189, 190, 193 n. 34 Goldin, M.P. 99 Gordon, Robert 18 Graham, Pauline 112 Grandey, Alicia 280 n. 14, 280 n. 15 Green, Joshua 14, 15, 16, 153, 156 Greenspan, Patricia 7, 16, 18, 20, 39, 42, 44, 45, 46, 48, 55, 58, 66 Griffith, Paul 2 n. 1, 129 n. 26 Guilt 1, 19, 25, 26, 44, 56, 65–6, 139–40, 153, 155, 158, 206, 209, 220, 213, 217, 219, 220, 232–5, 240–3, 257–9, 271 Haidt, John 15, 16, 153, 156 Hanfling, Oswald 85–7 Harcourt, Edward 21 Hare, Richard M. 3 Harris, Paul 193 n. 34 Hauser, Mark 166 Hedonic attitude 91 Helm, Bennett 6, 11, 17, 20, 22, 25, 222 n. 10, 222 n. 11, 222 n. 13, 261, 263, 286 n. 29 Hirschman, Albert O. 296 n. 51 Hochschild, Arlie 278–81, 284 Hoffman, Martin 176, 180, 184 Holmes, Jeremy 89 Holmes, Oliver Wendell 209 n. 19 Homiak, Macia 88 Honneth, Axel 291 Hume, David 7–8, 11, 20, 170, 257 and Strawson 205–8 as protection of dignity 271–2 on moral evaluation 264–7 on pride 257–9 on self-love 87 on sympathy 260–2, 267–70 Humean sentimentalism 117 n. 1, 42, 127, 150, 172, 189, 191 Hursthouse, Rosalind 245, 260 n. 13 Hurtig, Kent 117 Husserl, Edmund 8 n. 8, 117 n. 1 Hutcheson, Francis 70, 118, 120


Ibsen, Henrik 92 Inclinations 9 Kant’s account of 70, 75 Internalism 118, 123 Intuitions (moral) 14, 15, 24, 54, 66 n. 10, 118, 179 as emotional consciousness 150–61. See also Conscience Ishiguro, Kazuo 293 Jacobson, Daniel 18, 45, 62, 118–23, 126, 144, 248 n. 11 Johansson, Eva 180 n. 5, 182 n. 19 Johnston, Marc 8, 62, 118 n. 4, 120 n. 10, 123 n. 17, 128 n. 23, 129 n. 26, 135, 151 Johnstone, T. 151 Jones, Karen 11, 15, 45, 120 n. 11, 191 n. 32, 230 Kammeyer-Mueller, John D. 285 n. 26 Kant, Immanuel 2 argument against emotions 69–72 Humean critique of 259–66 legacy 9–14 on dignity 233 on self-government, self-discipline, and self-esteem 75 Schopenhauer’s critique of Kant’s rationalism 170–3, 190–1 Katz, Jack 292 n. 40 Kayser, K. 106, 107, 109 Kennett, Jeanette 15 Kenny, Anthony 2, 5, 120 Kerstein, Samuel 39 n. 1 Klein, Melanie 68 n. 13 Kohut, Heinz 267 Korsgaard, Christine M. 5, 9, 11, 13, 14, 20, 66, 70 n. 18, 70 n. 21, 71 n. 22, 75 n. 31, 76, 77, 78 n. 35, 165 on humanity and dignity 230–1 on identity 271 Kristja´nsson, Kristja´n 270 Kutz, Kathryn 234 n. 36 Lacewing, Michael 16 Langton, Rae 117 Lear, Jonathan 4, 21, 83 Leighton, Stephen 39 n. 1 Lemaire, Stephan 117 Lemos, Noah 117 n. 1, 119 n. 5 Levine, Steven 193 n. 34 Levy, Neil 12 Loewald, Hans 82 Louden, Robert 10 Love 1, 10, 22, 23, 41, 58, 59, 64, 68–9, 187, 200, 201, 213, 268, 278, 286, 287, 294 and compromise 101–2



Love (cont.) as a person-focused emotion 221 for humanity 238, 242, 243, 250, 262 of life 73, 82 Romantic 95–114. See also Altruism, Compassion, Respect Ludwig, Pascal 117 Mabbott, J.D. 87 Mackie, John L. 8, 66, 131 MacLeod, Colin 117 Macpherson, Fiona 148 n. 19 Mahon, James 255 n. 15 Malatesti, Luca 27 n. 11 Marx, Karl 276–7, 292 Mascolo, M.F. 271 Massin, Olivier 117 Mathewes, Charles 27 n. 1 McDowell, John 4, 5, 7, 14, 62, 117 n. 1, 120 n. 10, 125, 137, 144 McKenna, Michael 218 n. 3, 230 n. 28 Meinong, Alexius 8 n. 8, 117 n. 1, 120 n. 10 Menzies, Peter 128 Merleau-Ponty, Maurice 180 n. 5 Mikhail, John 166 Mill, John Stuart 64–6, 240, 241 n. 6, 258 Miller, Susan 267 Millgram, Elijah 27 n. 11 Mish’alani, James 255 n. 15 Moll, J. 153 Monroe, Kristen R. 181 Montesquieu 295 Moral assessment 2, 9, 162, 163, 161, 271, 290 self-assessment 26, 261, 271, 290 Moral attitudes of approval/disapproval 158, 183, 204–5, 210–11, 262–3. See also Moral assessment disapproval and censure 268 disapproval and recognition 270. See also Blame, Shame, Guilt, Pride Moral authority 21, 25, 62–78, 159, 198, 236, 254, 260, 266 categorical 63–77 moral authority as felt 77 moral authority to address demands 198, 209, 210–13, 226, 236 subjective vs. normative 63–5, 72–5 Moral motivation 8, 10, 13, 39, 40–2, 63, 72, 170–2 non-moral motivation 56–60 Moral obligation 39, 42, 45–8, 56, 65, 70–2, 97,198, 206–13, 236–8, 240, 242, 244, 246, 250, 255 moral competence beyond moral obligation 258, 259, 260, 284 Moral psychology 3, 4, 19, 28 Aristotelian approach 4–7, 270

empirical psychology 12–17, 157, 170–92 evolutionary psychology 9, 18, 26 n. 10, 65 n. 8, 140, 186, 258 Humean approach 7–9, 263–6 Kantian approach 10–12, 65, 83, 93 social psychology 265, 296 Moral Sense School 70, 259 Moran, Richard 3, 4, 289 Morris, Christopher 39 n. 1 Morris, Harry 235 n. 1 Mulligan, Kevin 8, 118 n. 3, 120 n. 8 Murdoch, Iris 3, 4–13, 183, 187 n. 26 Nagel, Thomas 12–13, 66, 241–3, 264 Negative vs. positive attitudes 91, 96 Nichols, Shawn 14, 16, 24, 171–92 Nietzsche, Friedrich 25 Nodding, Nel 183 n. 21 Normative force vs. motivational force 64, 71, 122 Normativity as reason-givingness 62, 65 normativity requirement for evaluative judgments 122–3, 128, 131 of emotional consciousness 156–7, 160–4, 166 Nussbaum, Martha C. 4, 5, 20, 23, 27, 151, 188, 261, 263, 286 n. 29 O’Neill, Onora 9, 67, 75 n. 31, 76 n. 32, 77 n. 33 Oakley, Justin 9,16 Oddie, Graham 118 n. 4, 130, 138 Ogien, Ruwen 117 Oliner, Pearl 181 Oliner, Samuel 181 Olson, Jonas 117, 119 n. 9, 124, 127 n. 21 Panksepp, Jaak 2 n. 1 Paradigm scenario 7 n. 7 Peacocke, Christopher 6, 130 n. 27, 136 Pedrini, Patrizia 27 n. 11 Pettit, Philip 55, 127, 128, 129, 130 Piller, Christopher 117 Plato 4, 150, 164, 165, 270, 295 n. 50 Pleasure feelings 10, 43, 56, 107, 113, 154, 158, 173 n. 2, 201, 235, 251, 259, 261, 263, 268–9, 295 and the self 86–93 Plutchik, Robert 26 n. 10 Practical Reason 3, 17, 19, 21, 39–44, 47, 49–52, 55–6 Aristotelian accounts 4–7, 86 efficacy of practical reason 62, 77–8 Humean Accounts 264 Kantian Accounts 9–11, 70–8. See also pride, Self-love, Moral authority Preston, Stephanie 261, 264–5, 268


Price, Huw 117 Prichard, Henry 62 Pride 26, 65 n. 7, 155, 156, 224, 227, 246, 247 n. 10, 257–73 Prinz, Jesse 15, 16, 18, 26 n. 10, 118, 119, 120 n. 10, 129 n. 26, 135, 151, 158, 167 Private attitudes 26, 160, 235, 242, 247, 255 Propositional attitudes 55 Pugliesi, Karen 280 n. 14, 280 n. 15, 285 n. 25, 285 n. 27 Pugmire, David 57, 76 Putnam, Hilary 4, 13 PytlikZillig, L. 175 Rabinowicz, Wlodek 117, 118 n. 2, 122, 125 Radden, Jennifer 193 n. 34 Railton, Peter 65 n. 7, 117, 121 Raine, A. 153 Rannow-Rasmussen, Toni 117, 118 n. 2, 122, 125 Rational assessment 5 coherence-based assessment 160–1 Rational authority 39, 41–2, 84 Rational choice 9, 17 and compromise 100–4 Rationalism 9, 10, 62–4, 68, 70, 77 anti-rationalism 170, 171 Rawls, John 75 n. 29, 201–2, 208, 213, 246, 296 Rawls on self-respect 296–7 Reactive attitudes 25, 187–203. See also Strawson natural vs. moral 208, 211–14 personal 201–10 self-reactive 206 vicarious vs. personal 204, 206, 208 Reath, Andrews 9, 10 Reciprocity 75, 97, 167, 202. See also Respect as mutual recognition Reflexive or self-directed attitudes 85, 89, 90 Regret 66, 107,158, 202 Reid, Thomas 254 Reisner, Andrew 117 Reiter, Sandra 255 n. 15 Resentment 23, 26, 67–8, 96, 201–15, 243, 246, 255, 261, 287 just resentment 218, 220–34. See also Respect Respect 10, 19, 21, 25, 26 and responsibility 76–8, 223–38 as mutual recognition 249–55 as respect for persons 220, 223 as respect for the lawgiving capacity 73, 75–82 as the subjective experience of autonomy 73 mutual respect 74 recognitional vs. appraisal respect 256 respect, dignity and addressability 229–32 self-respect 272, 288, 296 Responsibility 9, 17, 25–7, 48


and reactive attitudes 197–217, 232–4 and respect 76–8, 223–38, 249–55 for character 292 for emotions 9, 17, 25–7. See also Education Ripstein, Arthur 241 n. 6 Roberts, Jean 255 n. 15 Roberts, Robert 136 n. 1, 286 n. 30 Robinson, Jenefer 6 Roesch, S. 175 Rorty, Ame´lie O. 4, 6, 17, 22, 27, 45 Rose, Stephen 277 n. 6 Rosen, Gideon 117 Ross, D.W. 165, 166 Rotundo, Maria 285 n. 26 Rousseau, Jean-Jacques 75 n. 29, 161, 202, 208, 213, 281, 296 Russell, Paul 218 n. 3, 230 n. 28 Sabini, John 9, 70 n. 18 Sankowski, Edward 18 Scanlon, Thomas 66, 75 n. 30, 117 n. 1, 119 n. 7, 124, 236, 249, 250–4 Scheler, Max 8 n. 8, 117 n. 1, 173, 176, 183 n. 21 Scherer, K.R. 151 Schopenhauer, Arthur 170, 171–2, 173 n. 2, 176 n. 9, 190–1 Schor, Juliet 277 Schorr, A. 151 Schroeder, Timothy 117 Scottish Moralists 170 Self-alienation 296 Self-love 1, 21, 73 n. 26, 74, 83–93, 281 Sennett, Richard 291 n. 39 Sensen, Oliver 78 n. 36 Sentimentalism 7, 11 critique of the Humean or empiricist variety 24, 25, 64, 66, 71, 72, 170–2 neo-sentimentalism 8, 16, 23, 24, 25 normative and descriptive varieties 119–25 objection of circularity 128–30 varieties of neo-sentimentalism 8, 117 Shaftesbury, A.A. Cooper, Earl of 7, 120 n. 10, 259, 260 Shame 1, 26, 118, 121–6, 136, 141–3, 153, 217, 227, 242–7, 258–9, 267, 270–3 vs. guilt 258 vs. pride 257. See also Pride Sher, George 237 n. 3 Sherman, Nancy 10, 14, 42, 66 n. 12, 70 n. 19 Sidgwick, Henry 238 n. 4 Silver, Maury 9, 70 n. 18 Simmel, Georg 277 Sinnott-Armstrong, Walter 15 Skorupski, John 7, 25, 64, 75 Slote, Michael 8 Smith, Adam 7, 259, 296



Smith, Angela 18, 26, 239 n. 5 Smith, Michael 55 Sober, Elliott 175, 179, 180, 186 Solomon, Robert 5, 6, 16, 18, 230 n. 27, 286 n. 29 Sreenivasan, Gopan 117 Stark, Susan 5 Starkey, Charles 27, 188 Staub, Ervin 272 Stich, Stephen 15, 177 n. 11 Stocker, Michael 6, 18, 23 Stoicism 151, 286 Stratton-Lake, Peter 165 Strawson, Sir Peter (Frederick) 25, 75 n. 30, 187, 197–216, 223, 225, 232–4, 254, 258 Wallace’s and Darwall’s interpretation 217–21. See also Kant, Hume Stroud, Sarah 117, 128 n. 24 Stuart, Hampshire John 3 Swanton, Christine 260 Sympathy 10, 41, 59, 171, 173–9, 181, 185, 203, 205, 218, 245, 257, 259 Hume’s account 261–76 Talbott, William 255 n. 15 Tangney, June 271 Tappolet, Christine 18, 21, 23, 117 n. 1, 117 n. 2, 120 n. 10, 129 n. 26, 131 n. 28 Taylor, Charles 289 Taylor, Gabriele 26, 238 n. 4, 266 Taylor, Jacqueline 26, 67 n. 14, 266 Teroni, Fabrice 117, 136 n. 6 Thagard, Paul 19, 20, 24, 151, 161, 168 Thalberg, Irwin 16, 238 n. 4 Thompson, Michael 3, 5, 11, 13, 14, 41 Todd, Cain 148 n. 19 Tooby, John 26 n. 10 Tracy, Sarah 280 n. 14, 280 n. 15, 281–2 Van Willigenburg, Theodor 99 Velleman, David 11, 22, 65 n. 7, 225 n. 16

Virtue 1, 14, 41, 42, 56, 59, 205, 296 Hume’s account vs. Aristotelian account 263–70 virtue-ethics 4 n. 3, 42, 236, 244–7, 255, 261 Wallace, Jay R. 6, 13, 25, 87, 198, 208–18, 219, 223–39, 258–60, 271. See also Strawson Wasserman, David 39 n. 1 Watson, Gary 197 n. 1, 230 n. 28, 246 n. 9 Wedgwood, Ralph 120 n. 10, 121 n. 14 Weil, Simone 92–3, 183 Well-being 43, 59, 153, 160, 173 of others 173, 180, 183–7, 190 Wiggins, David 7, 8 n. 8, 62, 117 n. 1, 120, 121, 126, 128, 129, 144 Williams, Bernard A.O. 2–13, 24–5, 63–7, 72, 75, 77, 119, 189, 267, 270–2 Wilson, Catherine 117 Wilson, David Sloan 175, 179, 180, 186 Winch, Peter 92–3 Winkelman, Stephen 78 n. 36 Winnicott, Walter 267 Withing, Jennifer 14 Wittgenstein, Ludwig 14, 198 Wollheim, Richard 8 Wood, Allan 10, 70 n. 20, 71 n. 23 Word, L. 181–2 Wright, Crispin 118 n. 4, 130 Wrong kind of reason objection 118, 125–31 Wuerth, Julian 9 Yang, Y. 153 Yu, Cathy 255 n. 15 Zagzebski, Linda 6, 135 n. 1 Zahn, R. 153 Zahn-Waxler, Carolyn 191 n. 33 Zapf, Dieter 285 n. 25, 85 n. 27 Zemach, Eddy 8 Zhu, Jing 19, 20